Land Reclam. 37.indb - Annals of Warsaw University of Life

Transkrypt

Annals
of Warsaw
Agricultural
University
Land Reclamation No 37
Warsaw 2006
Contents
ROMANOWICZ R.J. Data Based Mechanistic rainfall-flow models for climate change
simulations
3
RAVAZZANI G., MANCINI M., MERONI
C. Design hydrological event and routing
scheme for flood mapping in urban area 15
DYSARZ T., WICHER-DYSARZ J. Assessment of hydrologic regime changes induced
by the Jeziorsko dam performance and morphodynamic processes in the Warta river
43
BANASIK K., BYCZKOWSKI A. Estimation of T-year flood discharge for a small
lowland river using statistical method
27
MAJEWSKI G., PRZEWOŹNICZUK W.
Characteristics of the particulate matter
PM10 concentration field and an attempt to
determine the sources of air pollution in the
living district of Ursynów
55
WOODWARD D.E., SCHEER C.C.,
HAWKINS R.H. Curve number update used
for Runoff Calculation
33
BARYŁA A. Runoff volume and slope gradient relationship – laboratory investigations
69
PAŃKA D., ROLBIECKI R., RZEKANOWSKI CZ. Influence of sprinkling
irrigation and nitrogen fertilization on health
status of potato grown on a sandy soil
75
MUSIAŁ E., BUBNOWSKA J., GĄSIOREK
E. Variation of climatic water balance and
heat balance for various ecosystems in
Wrocław in the years 1964–2000
101
MUSIAŁ E.,, BUBNOWSKA J., GĄSIOREK
E., ŁABĘDZKI L. Heat balance and climatic
water balance in vegetation period of spring
wheat
83
VABOLIENÉ G. Investigation for biological
nitrogen removal from wastewater using simultaneous nitrification/denitrification technology
111
KASPERSKA-WOŁOWICZ W., ŁABĘDZKI
L. Climatic and agricultural water balance
for grasslands in Poland using the penman-monteith method
93
LITWIN U., JANUS J., ZYGMUNT M.
Development of technologies used in agricultural engineering work on an example of
selected stages of land consolidation process
123
Series Editorial Board
Elżbieta Biernacka, Warsaw Agricultural
University, Chair
Wojciech Bartnik, Cracow Agricultural
University
Marian Granops, Warsaw Agricultural
University,
Waldemar Mioduszewski, Institute for
Land Reclamation and Grassland Farming,
Poland
Marek Lechowicz, Warsaw Agricultural
University,
Józef Mosiej, Warsaw Agricultural
University
Gunno Renman, Royal Institute of Technology, Stockholm, Sweden
Michael Hirschi, University of Illinois at
Urbana-Champaign, USA
Lubos Jurik, Slovak Agriculture University,
Nitra, Slovakia
SERIES EDITORS
Józef Mosiej – Chairman
Gunno Renman
Janusz Kubrak
EDITORIAL STAFF
Jadwiga Rydzewska
Krystyna Piotrowska
WARSAW AGRICULTURAL UNIVERSITY PRESS
e-mail: [email protected]
ISSN 0208-5771
Edition 500 copies
PRINT: Agencja Reklamowo-Wydawnicza A. Grzegorczyk
Annals of Warsaw Agricultural University – SGGW
Land Reclamation No 37, 2006: 3–13
(Ann. Warsaw Agricult. Univ. – SGGW, Land Reclam. 37, 2006)
Data Based Mechanistic rainfall-flow models for climate change
simulations
RENATA J. ROMANOWICZ
Lancaster University, United Kingdom.
Abstract: Data Based Mechanistic rainfall-flow models for climate change simulations.
Analysis of climate change indicates an increased
probability of high rainfalls and related flooding.
Climate change may also lead to an increase in
rainfall variability and the probability of extreme
events. This forces researchers to focus on both
high and low flow models. This paper presents the
application of a Data Based Mechanistic (DBM)
approach and Stochastic Transfer Function (STF)
methods to rainfall-flow modelling, with special
emphasis on low flows. We present two different
models, one with nonlinearity applied to the input,
and the second with nonlinearity on the output. In
both cases, the application of stochastic methods of
identification and estimation of model parameters
allows for the evaluation of predictive uncertainty
of the estimated flow. The first model introduces
input nonlinearity in the form of effective rainfall.
The second model applies the STF approach to
log – transformed flow acting as a surrogate of
water storage in the catchment. Each of the
models has two modules: a groundwater storage
module and a surface water module. Logarithmic
transformation of the output introduces a bias
towards low flows. In order to model also high
flows, the fast flow component is transformed
using the linear STF model. Both models are
applied to a karstic catchment, the River Thet in
the United Kingdom.
Key words: DBM and STF models, low flow,
karstic catchment.
INTRODUCTION
The increased variability of rainfall
magnitude and the resulting increased
variability of river flows observed in
the past two decades may have been
caused by climate changes related to
NAO intensification, NSO cycles and/
or global warming. This time period is
not long enough to provide the statistical
evidence supporting any particular thesis
(Robson et al. 1998, 2002). However,
globalisation of the world economy
requires global thinking about the
possible water crisis in order to avoid
global disaster. Ongoing developments
in modelling future climate and its
influence on water resources requires
incorporating models able to simulate
rainfall-flow processes under changing
climatic conditions. From the point of
view of water resources management
and planning the prediction of extreme
(both high and low) flows is the most
critical.
Low-flow hydrology deals with river
flows during the annual dry season. River
flow is a result of complex processes of
rainfall-water transport operating on
a catchment scale. Usually we distinguish
fast and slow components of river flow.
The slow component, with a large time
constant, is attributed to base flow,
because the majority of the streamflow
during the low flow period originates
from groundwater storage. Groundwater
storage may be related to the drainage
of the saturated top soil zone or
a groundwater aquifer. Another source of
4
R.J. Romanowicz
low flow may be relatively slow moving
groundwater drainage in fracture zones
with significant lateral components. Yet
another source may be permanently
wetted channel bank soils, the bottoms
of alluvial valleys or wetlands. The
existence of these different sources
of low water depends on catchment
geology. Smakhtin (2001) lists several
different processes which he attributes
to low flows, such as karst formations,
lakes and glaciers. Additionally there
are also anthropogenic impacts on low
flows, among others, deforestation,
afforestation, industrial and agricultural
direct water abstraction, irrigation, return
flows and dams and river regulation.
The first statistical approach towards
the separation of quick and slow flow
components was introduced by Young and
Beven (1994), who developed a bilinear
model for the rainfall-runoff process
with input nonlinearity representing the
effective rainfall. The main underlying
concept was the development of a Data
Based Mechanistic (DBM) approach to
modelling, combining the data-based
model with a mechanistic interpretation
of its components (Young 1998). This
approach applies Stochastic Transfer
Function (STF) methods and uses flow
as a surrogate of information on the soil
water content in the catchment. Young
(2003) extended this methodology to
model rainfall-runoff processes “offline”, i.e. in a simulation mode applying
a simple first order linear model to
describe the soil moisture storage in
the catchment used for the nonlinear
transformation of rainfall. In this paper
we propose an extension of this approach,
consisting of using a Transfer Function
model representing the flow which is
subsequently used as a surrogate of
the catchment soil moisture storage.
We then compare this model to a novel
methodology for modelling low flows. It
applies the STF approach to a logarithm
of flow, representing water storage in
the catchment (Romanowicz 2006). The
logarithmic transformation enhances
the model performance for low flow
values. It introduces a bias to the model
solution, as the mean of the logarithm
is not equal to the logarithm of its mean
value. We apply a separation of the fast
and slow components of the estimated
STF model and attribute the base-flow
to the slow component only. The fast
component, representing catchment
quick response related to surface or
near surface waters, may be modelled
using the ordinary STF model without
a logarithmic transformation of flow. The
model structure has to vary depending on
the flow values. We introduce potential
evaporation as additional information on
flow levels in the catchment to distinguish
between high and low flow periods. It
is clear that the base-flow module may
be calibrated on information-rich data,
i.e. in the catchment with a noticeable
base flow component in the catchment
outflow. The model is suitable for climate
change simulations as it uses rainfall and
evaporation as input variables, without
the necessity of explicit modelling of
soil moisture content in the catchment.
Moreover, the application of stochastic
methods of identification and estimation
of model parameters allows for the
evaluation of predictive uncertainty
of the estimated flow. It is interesting
to note that water resource engineers
use the logarithm of flow to test model
performance for low flow, but this
Data Based Mechanistic rainfall-flow models for climate change simulations
transformation has never been explicitly
applied in a rainfall-runoff model with
the aim of low flow modelling.
Section 2 presents the structure of
both DBM STF models. In section 3
we present the application of a standard
DBM rainfall-flow model developed for
climate scenario simulations to the River
Thet in south-east England. In section 4
we describe the application of the logtransformed low flow model to the same
catchment. Section 5 is a discussion of
the results.
DATA BASED MECHANISTIC
(DBM) RAINFALL-FLOW
MODEL
In this section we present two DBM STF
based models developed for the purpose
of climate change simulations. The
first model follows the DBM approach
to rainfall-flow modelling introduced
by Young (1998, 2003) and Young
and Beven (1994). This model applies
the nonlinear transformation of the
rainfall. The second model, described in
Romanowicz (2006), applies a nonlinear
transformation of the flow. Thus the first
model has an input nonlinearity, while the
second uses the output nonlinearity. This
nonlinear transformation is introduced
in order to represent process nonlinear
behaviour of the process thus assuming
that its dynamics is linear.
DBM model with nonlinear
transformation of the rainfall
The methodology follows the
approach described in detail by Young
(2003) with the difference of applying
a Multiple Input Single Output (MISO)
5
Stochastic Transfer Function (STF)
model to obtain an estimate of the flow
variable, which issubsequently used as a
surrogate of soil moisture content in the
catchment. It is assumed that the model
structure has the first order:
b1
st =
rt −δ1 +
1 − a1 z −1
+
b2
1 − a2 z
−1
evt −δ2 + ζ t
(1)
where:
st – is the observed flow used as a
surrogate for soil moisture at the end of
sample time t;
rt, evt – denote, respectively, rainfall and
evaporation at the same sample time t;
δ1, δ2 – denote any pure, ‘advective’ time
delay for each input;
ζt – represents the noise (which in some
cases can be considered as zero mean,
serially uncorrelated Gaussian white
noise);
a1, a2, b1, b2 – model parameters and
the operator denotes a backward shift in
time, i.e. z–1ut = ut–1
The parameters are estimated using
the Simplified Refined Instrumental
Variable (SRIV) algorithm in the
CAPTAIN Matlab toolbox (e.g. Young
1989, 2000, Young et al. 1999).
Later on, the rainfall-runoff process is
described by a second order model, which
is decomposed into the fast and slow
components (Young and Beven, 1994).
Looking for a physical interpretation,
the slow model component represents
ground water storage and the fast
component is related to quick response of
the catchment. In the case of daily time
periods the third component, representing
6
R.J. Romanowicz
direct runoff, may also appear. The model
has the following form:
Fast component:
Z1,t =
β1,1
1 + α1 z −1
uefft −δ
Slow component:
Z 2,t =
β2
1 + α 2 z −1
uefft −δ
Direct component:
Z3,t = β3uefft
uefft = rt ⋅ stp
yt = Z1,t + Z 2,t + Z3,t + ξt
where:
α1, α2, β1, β2 – parameters derived from
the identified second order STF model;
uefft – denotes the effective rainfall
obtained from the rainfall multiplied by
a nonlinear gain expressed by the power
of the flow estimate obtained from (1).
All the parameters of both equations
and the power p are estimated
simultaneously
using
optimisation
methods from MATLAB® optimisation
toolbox. Due to the application of
stochastic methods of estimation, this
methodology gives estimates of the
variance of the predictions, as well as the
variance and the correlation structure of
the model parameters.
DBM model with logarithmic
transformation of flow
In a second approach we apply STF
method to the logarithm of flow. The
physical explanation of this procedure
is described in detail in Romanowicz
(2006). In brief, we look at the rate of
change of the flow rather than the water
balance in the catchment. As mentioned
in the introduction, the resulting flow
estimates are based towards low flows.
The model structure is similar to that
shown in equation (2) with second order
model dynamics, thus allowing for the
decomposition of the model into fast and
slow components:
Fast component:
Z1,t =
+
β1,1
1 + α1 z −1
β1,2
1 + α1 z −1
rt −δ1 +
evt −δ2
Slow component:
Z 2,t =
+
β2
1 + α 2 z −1
β2,2
1 + α 2 z −1
rt −δ1 +
evt −δ2
(3)
yt = exp(Z1,t + Z2,t + ηt)
where: α1, α2, β1,1, β1,2, β2,1, β2,2 –
parameters derived from the identified
second order STF model.
Due to the introduced logarithmic
transformation of flow, the variance of
model predictions is heteroscedastic and
has the form:
μ f = exp[(2μ s + σ2s ) / 2]
σ2f = exp(2μ s + 2σ2s ) − exp(2μ s + σ 2s )
(4)
where: μs – denotes the mean value of
the state variable Zt = Z1,t + Z2,t,
σ2s – denotes the estimate of its prediction
variance.
In order to model also high flows we
should include a high flow module in our
model. The high flow module may consist
of a linear STF model based on either the
errors between the low flow component
and the observed flow or directly on
the observations (without logarithmic
transformation) for the periods with
high flows. Yet another option is a
nonlinear transformation of a fast flow
component. This module would require
the use of additional information on the
conditions in the catchment provided by
temperature or potential evaporation.
The low flow module can have the form
of an STF model for log transformed
flows or we can use the slow component
as a representation of a base flow in the
catchment. In other words, the difference
lies in the way we construct the base flow
module, either as a full module obtained
from log-transformed flows or its slow
component only. The choice will depend
on the catchment characteristics. For dry
catchments, dominated by base flow, the
first approach may be suitable. However,
for catchments with predominantly high
flows, the second approach will be more
suitable.
APPLICATION TO RAINFALL-FLOW PROCESS IN THE RIVER
THET, UK: DBM MODEL WITH
EFFECTIVE RAINFALL
The Thet catchment, situated in
the SE of England, is characterised by
predominant base flows, due to its chalk-
7
based geology. The catchment has an area
about 307 km2 with mainly arable land and
very mixed geology. There are reservoirs
in the catchment affecting runoff as well
as industrial and agricultural abstractions
and effluent returns. Additionally, runoff
is affected by groundwater abstraction
and recharge.
In this section we apply the nonlinear
rainfall-flow model (1–2) for climate
change simulations with the nonlinearity
on the input to the Thet catchment. In
order to represent the nonlinear relation
between flow and rainfall we apply
a State Dependent Parameter (SDP)
approach (Young 2000) to the daily
rainfall and flow data. Figure 1 shows the
resulting nonlinear gain for the rainfall
as a function of flow. This nonlinearity
is subsequently approximated using the
power law, as shown in eq. 2. However,
instead of the observed flow we use the
estimate of flow obtained from the first
order MISO model (1) identified from
the data.
FIGURE 1. SDP identified nonparametric gain for
Thet
The first order model for soil moisture
has the form:
8
R.J. Romanowicz
st =
0.0278
rt +
1 − 0.9844 z −1
0.0131
−
evt −1 + ξt
1 − 0.9844 z −1
(5)
The second order rainfall-flow model
with nonlinearly transformed rainfall has
the form:
Fast component:
Z1,t =
0.0532
1 + 0.7953 z −1
uefft −δ
FIGURE 2. DBM model with effective rainfall,
validation, Thet; 55% of data variation explained
Slow component:
Direct component:
purpose of comparison with the DBM
model with effective rainfall presented
in the previous section, we use the same
as before datasets for calibration and
validation stages.
Z3,t = 0.0017 ueft
Low flow model
Z 2,t =
0.0107
1 + 0.979 z −1
uefft −δ
(6)
uefft = rt ⋅ st0.21
yt = Z1,t + Z 2,t + Z3,t + ξt
This model has an additional direct
component representing catchment
response shorter than a day. The years
1970–1972 were used for the calibration
and the model explained 88% of the
observed flow variance. Validation was
performed on the years 1990–1993
and only 55% of the flow variance was
explained (Figure 2).
The best identified MISO STF model has
a structure described by [2 2 2 1 2 0],
which can be decomposed into the fast
and slow components for both rainfall
and evaporation. The model has the
following form:
log(Q1 ) =
0.011rt −1
+
1 − 0.9931z −1
0.027 rt −1
0.0061evt −2
+
+
+
−1
1 − 0.7350 z
1 − 0.9931z −1
0.0078evt −2
−
+ ξt
1 − 0.7350 z −1
(7)
DBM MODEL WITH LOG-TRANSFORMED FLOW
In this section we present an application
of the DBM model with log-transformed
flow to the Thet catchment. For the
where rt denotes rainfall and evt denotes
evaporation.
The variance of the transformed model
estimates (one step-ahead prediction) is
0.038. The results of the calibration give
a 92% explanation of the data variation
for water storage and 89% of flow after
back transformation. The decrease of
efficiency is related to the bias towards
lower flows, which is introduced by the
logarithmic transformation.
The decomposition of the model
into slow and fast component is shown
in Figure 3. The slow component has a
time constant of 144 days and describes
the base flow in the catchment, while
the fast component has a time constant
of about 3 days. Due to the logarithmic
transformation of flow, after back
transformation both components are
multiplied by each other.
( )
The efficiency RT2 of the model
during the validation period (all 27 years)
is 77.6%. The results of the validation on
the years 1990–1993 are shown in Figure
9
4. These three years were characterised
by a very low flow, and the model gives
worst fit to the observations out of the
whole validation period.
Full DBM model with
log-transformed flow
For high flows the DBM low flow
model output has to be augmented by the
surface flow from saturated areas in the
catchment. We use evaporation levels
to distinguish between the low and high
flow periods. For evaporation higher
than a certain level, the model output is
equal to the exponent of groundwater
storage in the catchment. When the
evaporation is lower than this threshold
value, the flow is increased by the value
obtained from the STF model identified
from the errors between the base flow
and observed flow. The threshold value
of evaporation is optimised together
with the parameters of the linear in flow
FIGURE 3. Decomposition of the model into fast (lower panel) and slow component (upper panel); the
observations are shown by a dotted line; calibration stage, years 1970–1973
10
R.J. Romanowicz
–1972). The validation was performed
on the years (1990–1993) and is shown
together with 95% confidence bands in
Figure 5.
FIGURE 4. The DBM log-flow model: validation
stage, years 1990–1993; 57.6% of the data variation explained; observed daily flows are shown by
dots, continuous line denotes model simulations,
shaded area represent 95% confidence bands
model. This model gives nonzero output
only for the evaporation smaller than a
threshold value. The form of the STF
model, identified using Captain toolbox
IV (instrumental variable) methods is
[1 1 1 2 5], it is a first order model as
follows:
FIGURE 5. Full DBM model – validation on the
years 1990–1993, 65.1% of data variation explained; observed daily flows are presented by
dots, continuous line denotes model simulations,
shaded area represent 95% confidence bands
⎧
⎫
0.08
0.1233
⎪
⎪
Qt ,sim − Qt ,obs = ⎨
r −
ev
−1 t − 2
−1 t −5 evt < 0.4 ⎬
1 − 0.6886 z
⎪1 − 0.6886 z
⎪
evt > 0.4 ⎭
⎩ 0
where Qt,sim denotes the flow
estimated using (7) and Qt,obs denotes
observations.
The full DBM rainfall-flow model
is the sum of (7) and (8). The threshold
value for the evaporation was estimated
as 0.4, using an optimisation routine
with simultaneous estimation of the
model parameters. This model gave
a 93% explanation of the flow variation
for the calibration period (years 1970–
(8)
Figure 6 shows the log-log plot of
the full DBM model for all the available
years 1970–1997. As might be expected,
it differs from the groundwater storage
model only for high flows. Compared to
that model, the gain is 2% over the whole
validation period and it considerably
improves the high flow predictions.
Further work should be done on the
use two different time constants for
evaporation and rainfall and on the use
11
FIGURE 6. Log-log results for the
full DBM model, River Thet, validation on all the data from the years
1970–1997.
of temperature as a threshold value. In
the model presented here the choice
of the threshold, although based on
data, is site specific. It is related to the
seasonal temperature variation through
evaporation and thus explains high flows
which usually occur in winter in the
UK. It might not be general enough for
other countries with different climatic
conditions.
CONCLUSIONS
We have presented two DBM models
suitable for off-line climate change
simulations. The first rainfall-flow
model applies nonlinearity on the input,
following the methodology of Young
(2003). The second DBM model for low
flows applies log-transformed flows in
order to obtain better estimates of the
base flow. The model uses evaporation as
additional information on soil moisture
conditions in the catchment.
The full DBM model consists of baseflow and high-flow components. Baseflow is determined by the separation of
a low flow module into slow and quick
catchment responses. The logarithm of
flow is treated here as a substitute for
groundwater storage in the catchment.
In hydrology the assumption of lognormal distribution for the flow is well
justified (Yevjevich 1972). In the result
of logarithmic transformation of flow
variable, the variance of the predictions
of flow is heteroscedastic, depending
on the predicted estimates of flow. The
models have been tested on the River
Thet catchment, UK. The low flow model
performs better for this karstic catchment
than the model with effective rainfall. In
this application the simulated base flow
matches well the observed low flow, thus
showing that the approach developed
is suitable for the climate change
simulations and low flow predictions.
Further research is underway on
generalising the low flow model for
catchments with larger differences
between high and low flows. Experience
gained so far points towards the
transformation of the quick flow
component of the low flow model
12
R.J. Romanowicz
rather than to building a model based
on the error between the observed flow
and simulated base-flow. The simple
structure of a surface store module may
be extended by including a representation
of off-line effective rainfall, as in the first
model, or can be made more general by
introducing a complex expression for
surface runoff. The important difference
from all previous approaches lies in the
way in which we condition our model
using the available observations. Namely,
we use the logarithmic transformation
of flow to enhance information on low
flows in order to better identify the base
flow and we divide the dataset into
periods characterised by high flows in
order to identify the runoff component
of the model.
REFERENCES
ROBSON A.J., JONES T.K., REED D.W.,
BAYLISS A.C. 1988: A study of national
trend and variation in UK floods,
“International Journal of Climatology”,
165–182, 18.
ROBSON A.J. 2002: Evidence for trends in
UK flooding, “Philosphical Transactions
of the Royal Society London A”, 1327–
–1343, 360.
ROMANOWICZ R.J., 2006: Data Based
Mechanistic model for low flows:
implications for the effects of climate
change, submitted to Journal of
Hydrology.
SMAKHTIN V.U.: 2001 Low flow hydrology:
A review, “J. Hydrol.” 147–186, 240.
YEVJEVICH V.M. 1972 Stochastic
Processes in Hydrology, Water Resources
Publications, Fort Collins, Colorado.
YOUNG P.C. 1989: Recursive estimation,
forecasting and adaptive control, In:
Leondes C. T. (Ed.), Control and Dynamic
Systems: Advances in Algorithms and
Computation Techniques in Dynamic
Systems Control, 30, Academic Press,
San Diego pp. 119–166.
YOUNG P.C. 2000: Stochastic, dynamic
modelling and signal processing: Time
variable and state dependent parameter
estimation. In: W. J. Fitzgerald, A.
Walden, R. Smith, & P.C. Young (eds.),
Nonstationary and Nonlinear Signal
Processing, Cambridge University Press:
Cambridge, 74 –114.
YOUNG P.C. 1998: Data-based mechanistic
modelling of environmental, ecological,
economic and engineering systems,
“Environmental
Modelling
and
Software”, 105–122, 13.
YOUNG P.C. 1999: Data-based mechanistic
modelling, generalised sensitivity and
dominant mode analysis, “Computer
Phys. Communications”, 113–129, 117.
YOUNG P.C. 2003: Top-down and data-based mechanistic modelling of rainfallflow dynamics at the catchment scale,
“Hydrological Processes”, 2195–2217, 17.
YOUNG P.C., BEVEN K.J. 1994: Data-based mechanistic modelling of rainfallflow nonlinearity, “Environmetrics”,
335–363, 5.
YOUNG, P.C., PEDREGAL D., TYCH W.
1999: Dynamic harmonic regression, “J.
Forecasting”, 369–394, 18.
Streszczenie: Modelowanie niskich przepływów
do analizy wpływu zmian klimatu na zasoby wodne. Racjonalna gospodarka zasobami wodnymi,
jak również symulacja scenariuszy zmian klimatu
nie są możliwe bez lepszego zrozumienia bilansu
wody w zlewni, jak również lepszych metod do
modelowania niżówek. Niniejszy artykuł przedstawia zastosowanie Mechanistycznego Bazującego na Danych (DBM) podejścia do modelowania procesu opad-odpływ. Przedstawione są dwa
modele, pierwszy stosuje nieliniową transformację wejść (opadu), drugi, stosuje logarytmiczną
transformację wyjść (przepływu). Obydwa modele wykorzystują liniową Stochastyczną Funkcję Przejścia (STF) do opisu dynamiki procesu.
Wyprowadzone modele są zdekomponowane na
dwa moduły: jeden reprezentujący szybką dyna-
mikę przepływu (wody powierzchniowe) i drugi,
reprezentujący wolną dynamikę zwiazaną z wodami gruntowymi. Zastosowanie stochastycznych
metod identyfikacji i estymacji parametrów modeli pozwala na ocenę niepewności predykcji.
W związku z logarytmiczną transformacją wyjścia, uzyskana wariancja predykcji modelu z
nieliniowością na wyjściu jest zmienna w czasie.
W artykule przedstawiamy zastosowanie obydwu
modeli do modelowania zlewni o podłożu karstycznym w Wielkiej Brytanii (rzeka Thet).
MS. received November 2006
Author’s address
Renata J. Romanowicz
Environmental Centre
Lancaster University
Lancaster, LA1 4YQ
United Kingdom
13
Design hydrological event and routing scheme for flood mapping
in urban area
GIOVANNI RAVAZZANI1, MARCO MANCINI1 and CLAUDIO MERONI2
1
Politecnico
2
di Milano,
MMI s.r.l., Milan (Italy)
Abstract: Design hydrological event and
routing scheme for flood mapping in urban area.
Definition of flood risk maps is a task to which
modern surface hydrology addresses a substantial
research effort. Their impact on the government of
the flood prone areas have increased the need for
better investigation of the inundation dynamics
[Fema 2002]. This identifies open research
problems such as: the definition of the design
hydrograph, the identification of the surface
boundary conditions for the flood routing over
the inundation plan, the choice of the hydraulic
model that is the most close to the physical
behaviour of the flood routing in the specific
environment, such as urban areas or river valley.
Most of academic and commercial mathematical
models resolving the De Saint Venant equations
in mono or bidimensional approach, fail on
complex topography. Steep slopes, geometric
discontinuities, mixed flow regimes, initially dry
areas are just the main problems an hydraulic
model should solve. In this study, we address two
points: the definition of the critical event for an
inundation area and a flood routing modelling
technique for a highly urbanized flat area. For this
latter we show that, in urban areas, a modelling
scheme of a network of connected channels
and storages, gives a better representation of
surface boundary conditions such as aggregation
of buildings and road network and sufficient
accuracy for flood risk mapping purpose respect
to a real 2-D hydraulic routing model.
Key words: design hydrograph, distributed model,
flood hazard maps, quasi-2D model, urban flood.
INTRODUCTION
Flooding is one of the most common
environmental hazard, due to the
widespread geographical distribution of
river valleys and the attraction of human
settlements to these areas.
Floods can be generally considered
in two categories [Castelli. 1994]: flash
floods, the product of heavy localized
precipitation in a short time period over
a given location, and general floods,
caused by precipitation over a longer
time period and over a given river basin.
Although flash flooding occurs
often along mountain streams, it is also
common in urban areas where much of
the ground is covered by impervious
surfaces. Fixed drainage channels in
urban areas may be unable to contain
the runoff that is generated by relatively
small, but intense, rainfall events.
Many modelling approaches exist to
simulate floods [Leopardi et al. 2002].
The choice of the simulation approach
depends on the questions to be answered
[Horrit and Bates 2001, Ferrante et
al. 2000]. In this work we focus on
determination of flood hazard map for
16
G. Ravazzani, M. Mancini, C. Meroni
an urban area located in the Liguria
province in Italy. Traditional approach
is based on determination of flood extent
for a given frequency event, using steady
flow 1-dimensional model. A new rule
has been recently introduced for highly
urbanized area; it identifies the hazard
according to both flow depth and flow
velocity, claiming the necessity to use
more accurate unsteady flow numerical
models.
The aim of the work is to assess the
accuracy of a quasi-2D model versus
pure 2D models for flood prediction in
urban area.
THE CASE STUDY
The study area is located in the western
Italy, in the province of Liguria (Fig.
1). The area approaches the sea and
has an extension of about 1.8 km2 in
which 6 rivers are encountered: Gorleri,
Varcavello, S. Pietro, Pineta, Rodine and
Madonna. The drainage basin ranges
from 0.32 km2 of the river Rodine to
18.05 km2 of the river S. Pietro. The
maximum elevation is reached in the
Evigno Peak at 988.5 m. a.s.l.
The morphology is characterized by
steep slopes in the upper part of the basins,
decreasing while approaching the sea.
The restricted coastal plain has attracted
tourism activity with the effect of an
exponential growth of urban density. As
a consequence the rivers often have been
channelized into artificial drainage. In
this situation, we can frequently observe,
during storm event, water overtopping
the levees and flowing into main roads
and districts, dragging people and cars.
The target area is crossed by a railway
that divides it in two parts: the upper
with a slope of about 1.5% and the lower
with an average slope of about 0.8%.
THE DIRECTIVES FOR BASIN
PLANNING
For the purpose of land use planning
and management, flood risk maps are
required. The Italian legislation leave
the task of the assessment of flood risk
maps to the River Basin Authority with
FIGURE 1. Aerial photo of the study area. The six rivers and the railway are visible
Design hydrological event and routing scheme for flood mapping...
the Basin Plan study. The traditional
approach simulates the inundation
processes with a steady 1-dimensional
model. The analysis is iterated for 50 and
200 year return period peak discharge.
The 50 year flood extent is classified as
class-A hazard and 200 year flood extent
is classified as class-B hazard (Fig.
2a). Class-A areas are subject to more
restrictive rules than class-B areas.
A new scheme recently proposes
[FEMA, 2002; Rosso 2003] to determine
the hazard map on the basis of both
hydraulic depth and flow velocity.
According to this scheme, if an area is
frequently flooded but with low depth
and low velocity, the level of hazard
is reduced (Fig. 2b). Moreover a new
hazard class is introduced: according to
the new scheme, an area can be classified
as class-A, class-B and class-B0 hazard.
The class-B0 area is the less restrictive
one.
The necessity to employ flow
velocity data, requires the use of more
sophisticated hydraulic model, with
a
17
the ability to simulate unsteady flow in
complex urban terrain.
THE DESIGN STORM
HYDROGRAPHS FOR
INUDATION MAPS
There are many types of design
hydrographs that have been developed
over the years [Maidment 1993; Chow et
al. 1988], and the debate is open on the
frequencies of the flood map due to the
difference between the frequencies of
the peak discharge and the hydrograph
volume [De Michele et. al. 2005;
Salvadori and De Michele 2004].
Moreover, if a rainfall runoff model is
used for hydrograph computation, the
hypothesis of the equivalence among the
return period of the peak discharge and
the rainfall is false too.
Inundation maps are generally
computed using peak discharge for
given return period as input variable and
simple steady 1-dimensional analysis
b
FIGURE 2. Evaluation of hazard maps according to the actual rule (a) based on the flooding extent for
given return period events and according to the new rule (b) which considers both hydraulic depth and
flow velocity
18
for the river channel, that means that the
hydrograph volume and the routing on
the inundated areas are not considered.
As well known, especially for plain
urban area, the storaging volume can
significantly affects the routing and the
extension of the inundated areas.
According to this we think that it
is more correct, in the assessing of
inundation map, to take into account the
inundation volume and its statistics as the
key variable for map characterization.
For ungaged basin, when the
hydrograph is determined from rainfallrunoff transformation, we propose
a methodology for the inundation volume
definition based on research of the
critical rainfall event for an inundation
area. This is defined as the one that
gives the maximum value of inundation
volume for a given return period of
the rainfall in the hypothesis that the
intensity duration frequency (IDF) curve
represents the rainfall behaviour. The
maximum inundation volume is defined
as the maximum value of the integral
of the difference between the incoming
hydrograph and the bankfull discharge.
For this purpose, rainfall runoff
distributed model, FEST [Mancini 1990;
Montaldo et al. 2002; Rulli and Rosso
2002; Montaldo et al. 2003; Montaldo
et al. 2004; Salandin et al. 2004],
was employed. FEST is a distributed
hydrologic model especially developed
at the Politecnico di Milano focusing
on flash flood event simulation. As
a distributed model, FEST can manage
heterogeneity in hillslope and drainage
network morphology (slope, roughness,
etc..) and land use [Rosso 1994].
From a family of IDF curves it is
possible to obtain an hydrograph for
any duration at a given frequency and so
a series of hydrographs for every return
period (Fig. 3). Given the bankfull
discharge for the examined river
branch, the hydrograph that presents the
maximum inundation volume identifies
the critical rainfall event (Fig. 4) for the
FIGURE 3. Procedure for the search of the critical event defined as that event which is characterized by
the maximum potential inundation volume. The results refer to river Varcavello for the 50 years return
period
19
FIGURE 4. Series of potential inundation volume for different return period. The design hydrograph for
a given return period is the one characterized by the maximum value of the potential flooding volume
inundation map. In the end, according
to retention pool analysis [Artina et
al.1997], the critical event for a flood
mapping is different from the critical
event for the maximum peak discharge
(Fig. 3) and the correlated hydrograph
peak presents different return period for
a given rainfall frequency (Tab. 1) .
DESCRIPTION OF THE
HYDRAULIC MODEL
Urban areas are usually characterized by
streets and aggregation of buildings (in
the following termed blocks), that can
be schematized, from a flood routing
point of view, as network of channels
where the flow velocity is greater than
zero, and storages where the velocity is
about zero. This latter hypothesis derives
from the consideration on the friction
induced by the macro roughness of the
manmade obstacles present in a block.
So that the implemented network model
presents three main unit: the main rivers,
the channels along the main streets
and the storages for the aggregation of
buildings.
De Saint Venant equations are then
integrated along the river branches and
street channels using the Preissman
implicit numerical scheme [Wallingford
Software 2005]. Energy and continuity
equation is verified at nodes given by the
channel intersection. Water discharge
can flow in every direction in the channel
network according to the hydraulic
gradient. Reservoir equation is used to
model diffusion in the blocks.
The connection between storages and
channels are discretized in the specific
nodes and the weir equation controls
the water flux to and from the reservoir
according to the difference of water
level.
The river branches feed the channel
network where the cross sections are
insufficient respect to the flood discharge.
When the riverbanks are overtopped,
water enters the street channels.
The channel network model is
a very good representation of urban
flood routing when the hydraulic depth
20
TABLE 1. Comparison between maximum peak discharge and peak discharge of the hydrograph with
the maximum inundation volume computed for a given rainfall frequency. In the last column the return
period of the peak discharge of the hydrograph with the maximum inundation volume is reported
Rainfall return
period (years)
50
Return period of the peak
Peak discharge of the
Maximum peak
discharge of the hydrograph
hydrograph with the maximum
discharge (m3/s)
with the maximum inundation
inundation volume (m3/s)
volume (years)
94.04
79.82
25.5
100
111.75
86.37
38
200
132.82
95.61
54
500
161.24
99.7
64.5
and energy is lower than the height of
the surrounding buildings which act as
impervious boundary elements.
The described channel network
model was implemented (Fig. 5) in the
Infoworks-CS software [Wallingford
Software 2005]. Rivers and main roads
are represented by conduits. Manholes
allow the exchange of water between
river and roads and are used to represent
crossroads. Conduits are linked with
storages through weirs.
The main model parameters are the
roughness coefficient which controls
flow velocity in the channels and the
weir discharge coefficient which controls
the amount of water exchanged between
channels and storages. The adopted set
of values are reported in Table 2.
FIGURE 5. Network quasi-2D hydraulic model representing urban area: detail of the river Varcavello
21
TABLE 2. Parameters describing components of the network quasi-2D model
Parameter
Value
Natural river channel Strickler roughness
30 m1/3s–1
Concrete river channel Strickler roughness
50 m1/3s–1
Not asphalted street Strickler roughness
30 m1/3s–1
Asphalted street Strickler roughness
50 m1/3s–1
Weir discharge coefficient for districts upstream the railway (a)
0.1
Weir discharge coefficient for districts downstream the railway (a)
0.28
(a)
50 m
Weir length districts downstream the railway (a)
20 m
Weir length for districts upstream the railway
(a)
Weir formula adopted for discharge computation in the model: Q = Cd BDu g ( Du − Dd )
,
where Q is the discharge [m3/s], Cd is the discharge coefficient, B is the width of the weir [m], Du is
the upstream depth with respect to the crest [m], Dd is the downstream depth with respect to the crest
[m] and g is the acceleration due to gravity [m/s2].
THESIS VERIFICATION USING
2-DIMENSIONAL MODEL
The basic assumption of the network
model is that velocity of flood flow
over the urban area is greater than zero
in the main streets while, in the blocks,
velocity can be neglected. To verify
this assumption, we analysed flood
dynamic in the blocks, by means of
a high resolution 2-D model. A subset
of urban land in proximity to river
Varcavello, has been extracted from the
main domain. It is composed by two
main blocks: one upstream the railway
and the other downstream. The upper
one is characterized by an average slope
of about 1.5%, the lower by a slope of
about 0.8%. These subsets are considered
to be representative of the whole study
area. A full 2-D model was implemented
using the SMS software [www.bossintl.
com] (Fig. 6). A steady analysis has been
performed. Hydraulic depths deriving
from channel network simulation have
been taken as boundary condition of the
blocks.
Buildings have been modelled in
two different ways: as impervious area
(buildings are excluded from the model
domain) and as a high roughness surface
(Strickler coefficient equal to 0.01 m1/3s–1).
The latter way means that water is free
to move in the whole domain, but the
flow through buildings is made difficult
because of an high value of roughness
parameter. Two cases for the Strickler
coefficient for gardens surrounding
buildings were considered: in the first
case its value was fixed to 5 m1/3s–1and
in the second case to 10 m1/3s–1.
Velocity
field
deriving
from
simulations has been mapped to a regular
grid and the cumulative frequency of the
velocity values has been evaluated (Fig.
7). In the upstream block (Fig. 7a), we
can note that most of the cells (67%)
has a value of velocity less than 0.6 m/
s even for the simulation with Strickler
coefficient for gardens equal to 10 m1/3s–1
and impervious buildings.
In the other simulations, the
percentage of cells with velocity under
0.6 m/s increases nearly to 100%. Both
22
a
b
FIGURE 6. Subset of the study area (a) near river Varcavello upstream the railway and (b) representation in the 2-D model
a
b
FIGURE 7. Cumulative frequency of velocity magnitude simulated by means of full 2D model in (a)
upstream district and (b) downstream district; ks_g and ks_b denote, respectively, the Strickler roughness coefficient for gardens and buildings
considering buildings impervious or as
high roughness surface, the peak flow
velocity reaches, anyway, a maximum
value of 1.3 m/s in just a few cells.
The observations are valid for the
downstream block (Fig. 7b) too. As a
consequence of a milder slope, flow
velocity is lower indeed.
We can conclude that motion of
water in the aggregation of buildings is
characterized by low values of velocity.
The assumption to simulate blocks as
storages in the channel network model
seems reasonable.
THE FLOOD HAZARD MAPS
Flow velocity and hydraulic depth values,
computed respectively in the channels
and in the nodes of the network model,
have been interpolated over the entire
study area to obtain a continuous map.
For this purpose the borders of districts
have been considered as barriers. The
overlay of the velocity with the depth
map, produced the hazard map (Fig. 8),
according to the new directive of the
River Basin Authority (§ 3).
A comparison is made with the
previous study performed by Basin
Authority [Regione Liguria] which
made use of steady flow computation
model for delimitation of flood extent,
not regarding flow velocity.
Total amount of flooded area has
increased in this new study (Fig. 9) from
0.97 to 1.27 km2. The class-A areas and
class-B areas have decreased, respectively
from 0.56 to 0.24 km2 and from 0.42 to
0.16 km2. The complementary area is
included in the class-B0 area which was
not defined in the previous analysis.
The differences in the framework of
the present study respect to the previous
23
one, have counterpart in the resulting
flooding map (Fig. 8). In the network
quasi-2D model, overtopping water is
routed along main roads as far as distant
blocks. This is good explanation for those
areas that, in the previous study, are not
even water logged by 200 year flood
event: the pure 1D steady model can’t
predict flood processes in urban area.
The new framework can also predict
water logging due to manmade obstacles
orthogonal to the flow direction. The
railway divides the city on north-south
direction and behaves as an impervious
levee to water flow. This does not seem
to be completely represented by the 1D
model.
CONCLUSIONS
A procedure is presented for estimation
of design hydrographs. It is based on
the search for the critical event which
maximizes the potential flooding
volume. The iterative process makes
use of the FEST model for the rainfallrunoff transformation. The distributed
model permits to well represent
basins characterized by heterogeneous
FIGURE 8. (a) Flood hazard map resulting in this study, compared to (b) actual hazard maps published
by Basin Authority
24
FIGURE 9. Extension of class-hazard area as resulting from this study, compared to previous study
performed by Basin Authority
morphology and land use, as the ones in
this work. The critical event guarantees
that the hydrograph with the maximum
effect on territory is assumed, even
if peak discharge is lower than the
maximum peak discharge for the given
rainfall frequency.
Classical one dimensional models
are poor tools for flood analysis in urban
area. They can be used as long as the main
stream is not overtopped, as they fail to
simulate flow component other than along
river direction. Two dimensional models,
on the other hand, are time consuming
and, for unsteady flow simulations on
complex topography, they fail on steep
slopes, geometric discontinuities, mixed
flow and initially dry areas.
This work proposes an hybrid
approach. The urban and drainage
system is modelled by means of a
network in which both rivers and roads
are modelled as channels linked by
nodes. The basic assumption is that high
density urban blocks can be modelled as
storages in which flow velocity is null.
The assumption is verified by a 2D model
results which confirm that flow velocity
in the blocks is negligible.
The channel network model seems
to well represents flood routing in urban
area, it is not computationally expensive
as 2D model and, above all, is much more
stable on complex topography. On the
other hand it requires a deep knowledge
of the territory and a good skills of the
modeller to set up the hydraulic sketch.
REFERENCES
ARTINA S., CALENDA G., CALOMINO
F., LA LOGGIA, G., MODICA C.,
PAOLETTI A., PAPIRI S., RASULO G.,
VELTRI P. 1997: Sistemi di Fognatura.
Manuale di Progettazione. Hoepli, chap.
X,XVIII, Milan. (In Italian).
Boss International. http://www.bossintl.com
CASTELLI F. 1994: Spatial scales of frontal
precipitation. In Advances in Distributed
Hydrology, R. Rosso, A. Peano, I. Becchi
and G. A. Bemporad Editors, Water
Resources Publications, 87–114.
CHOW V.T., MAIDMENT D.R., MAYS
L.V. 1988: Applied Hydrology. McGrawHill.
De MICHELE C., SALVATORI
G.,
CANOSSI M., SETACCIA A., ROSSO
R. 2005: Bivariate statistical approach to
check adequacy of dam spillway. Journal
of Hydrologic Engineering, 10 (1): 50–57.
FEMA, Federal Emergency Management
Agency,
2002:
Guidelines
and
specifications for flood hazard mapping
partners. FEMA Publications.
FERRANTE M., NAPOLETANO F.,
UMBERTINI L. 2000: Optimization of
transportation networks during urban
flooding. Journal of the American Water
Resources Association, 36 (5): 1115–
–1120.
HORRIT M.S., BATES P.D. 2001:
Predicting floodplain inundation: raster-based modelling versus the finite element
approach. Hydrological Processes, 15,
825–842.
LEOPARDI A., OLIVERI E., GRECO M.
2002: Two-dimensional modelling of
floods to map risk-prone areas. Journal
of Water Resources Planning and
Management, 128(3),168–178.
MAIDMENT D.R. 1993: Handbook of
hydrology. McGraw-Hill.
MANCINI M. 1990: La modellazione
distribuita della risposta idrologica:
effetti della variabilitŕ spaziale e della
scala di rappresentazione del fenomeno
dell’assorbimento. PhD thesis, Politecnico
di Milano. (In Italian).
MONTALDO N., MANCINI M., ROSSO
R. 2004: Flood hydrograph attenuation
induced by a reservoir system: analysis
with a distributed rainfall-runoff model.
Hydrological Processes, 18(3), 545–563.
MONTALDO N., RAVAZZANI G.,
MANCINI M. 2003: The role of the
antecedent soil moisture condition on the
distributed hydrologic modelling of the
Toce alpine basin floods. International
Conference on Alpine Meteorology and
MAP-Meeting 2003, http://www.map2.
ethz.ch/icam2003/ICAM-MAP2003.htm
MONTALDO N., TONINELLI V., MANCINI
M., ROSSO R. 2002: Coupling Limited
Area Models with Distributed Hydrologic
25
Models for Flood Forecasting: the Toce
Basin Study Case. IAHS, 274, 229–236.
Regione Liguria, Autoritŕ di Bacino di
regionale. Piano di bacino stralcio per
la difesa idraulica ed idrogeologica,
(Ambito di Bacino no 7 – DIANESE).
(In Italian).
ROSSO R. 1994: An introduction to spatially
distributed modelling of basin response.
In Advances in Distributed Hydrology,
R. Rosso, A. Peano, I. Becchi and G.A.
Bemporad Editors, Water Resources
Publications, 3–30.
ROSSO R. 2003: Consulenza tecnico
scientifica per la definizione degli ambiti
normativi relativi alle fasce di inondabilitŕ
in funzione di tiranti idrici e velocitŕ di
scorrimento. Abstract on http://www.
regione.liguria.it (In Italian).
RULLI M.C., ROSSO R. 2002: An integrated
simulation method for flash-flood risk
assessment: 1. Frequency predictions
in the Bisagno River by combining
stochastic and deterministic methods.
Hydrology and Earth system Sciences,
6(2), 267–283.
SALANDIN A., RABUFFETTI D.,
BARBERO S., CORDOLA M.,
VOLONTČ G., MANCINI M. 2004:
Monitoraggio e simulazione numerica
del fenomeno finalizzata alla previsione
e gestione dell’emergenza. Neve e
Valanghe, 51. (In Italian).
SALVADORI G., De MICHELE C. 2004:
Analytical calculation of storm volume
statistics involving Pareto-like intensityduration
marginals.
Geophysical
Research Letters, 31 (4).
Wallingford
Software.
http://www.
wallingfordsoftware.com/products/
infoworks_cs/. Site visited on 2005-07-08.
Streszczenie: Konstruowanie hydrogramów
i schematów obliczeniowych do odwzorowania
obszarów zagrożonych powodziami na terenach
zurbanizowanych. Wyznaczanie map ryzyka powodziowego jest jednym z ważniejszych zagadnień we współczesnej hydrologii. Ich wpływ na
zarządzanie terenami zalewowymi zyskał duże
26
znaczenie, co spowodowało potrzebę badań nad
dynamiką powodzi. Było to również przyczyną
intensyfikacji badań nad następującymi zagadnieniami: zdefiniowanie hydrogramu projektowego,
identyfikacja powierzchniowych warunków brzegowych do obliczeń zasięgu powodzi, wybór modeli jak najlepiej opisujących zasięg powodzi w
specyficznym środowisku, jakim jest obszar zurbanizowany lub doliny rzeczne. Większość modeli matematycznych zarówno szkoleniowych,
jak i komercyjnych, rozwiązuje równania De
Sait Venant’a w jednym lub dwu wymiarach, nie
sprawdza się jednak dla skomplikowanych topograficznie powierzchni. Duże spadki, nieciągłości
geometryczne, różne zmieniające się reżimy przepływu, początkowo suche obszary są głównymi
problemami, które powinny być rozwiązywane
przez modele hydrauliczne. W tych badaniach skupiono się na dwóch zagadnieniach: zdefiniowane
zdarzenia krytycznego dla terenów zalewowych
i technice modelowania terenów zalewowych
dla silnie zurbanizowanej płaskiej powierzchni.
W tym przypadku schemat połączonych kanałów
i zbiorników retencyjnych dał lepsze odwzorowanie powierzchniowych warunków brzegowych,
takich jak kompleksy budynków, sieć ulic i wystarczająca dokładność w określaniu obszarów
zagrożonych powodziami w porównaniu do dwuwymiarowego modelu hydraulicznego.
MS received November 2006
Authors’ addresses:
Giovanni Ravazzani
Marco Mancini
Politecnico di Milano
Plazza L. da Vinci 32, 20133 Milano
Italy
Claudio Meroni
MMI s.r.l,
Via Aselli 24, 20133 Milano
Italy
Corresponding author:
Fax +39 02 2399 6207
Estimation of T-year flood discharge for a small lowland river using
statistical method
KAZIMIERZ BANASIK, ANDRZEJ BYCZKOWSKI
Warsaw Agricultural University – SGGW,
Department of Water Engineering and Environmental Restoration
Abstract: Estimation of T-year flood discharge
for a small lowland river using statistical method.
The 34-year series of daily discharges from
a small agricultural river basin of the Zagożdżonka
river at the gauging station of Płachty Stare (A =
= 82.4 km2), located in the center of Poland,
were used in the investigation. The Pearson
type 3 (i.e. 3-parameter gamma) distribution
and the log-normal distribution were considered
to find the best fit with the empirical data. The
IMGW (Institute of Meteorology and Water
Management) computer program, which applies
the method of maximum likelihood, was used to
estimate the parameters of the above distributions.
Also commonly applied in engineering practice,
the Pearson type 3 distribution with the method of
quantile for estimating the distribution parameters
was additionally used. Log-normal distribution
gives better fit with the measured annual flood
flows, applying Akaike criterion, and the best fit
applying subjective visual criterion.
Key words: flood flow, small river basin, peak
flow, frequency curve.
INTRODUCTION
Estimation of flood discharges are
needed in designing of hydraulic or road
structures, as well as in flood protection
activity. It is often assumed that in case
of existence of long term hydrometric
records, the procedure for estimating
T-year flood discharge could be carried
out without larger difficulties, which
however does not always take place, as
selection of various probability function
may produce significant differences in
the results (Maidment 1993, Pilgrim and
Doran 1993). In case of the studied river
basin, which is research river basin of
Department of Water Engineering and
Environmental Restoration of Warsaw
Agricultural University, the results are
also important as reference values for
application and verification of indirect
methods for the T-year flood discharge
estimation, as well as for rainfall-runoff
model verification. In the paper there has
been analysis carried out for examining
the following statistical distributions:
a) the Pearson type 3 – P3 (i.e. 3-parameter gamma) distribution and, b)
the log-normal distribution – LN. For
the statistical calculation the IMGW
(2005; Institute of Meteorology and
Water Management) computer program,
which applies the method of maximum
likelihood – MML, was used to estimate
the parameters of the above distributions.
Also commonly applied in engineering
practice in Poland, the Pearson type 3
distribution with the method of quantile
– MQ, for estimating the distribution
parameters was additionally used. Data
of 34-year daily discharges (1969–2002)
of small river basin of Zagożdżonka
were used for the analyze.
28
K. Banasik, A. Byczkowski
DATA USED
Data of discharge of the Zagożdżonka
river in the gauge of Plachty Stare,
where area of the basin is 82.4 km2,
has been colleted by the Department of
Water Engineering and Environmental
Restoration (former Department of
Hydraulic Structures) of the Warsaw
Agricultural University – SGGW since
1962. Until 1980 water stages had been
measured by observer three times a day,
when additionally water stage recorder
was installed. Location of the river basin
is shown in the Figure 1. The mean annual
precipitation and runoff are estimated at
610 mm and 109 mm respectively. The
Zagożdżonka watershed is of lowland
type. The absolute relief is 37 m. The
mean slops of the main streams are from
2.5 to 3.5‰. The land use is dominated
by arable land (small grain and potatoes).
Sandy soils are the dominant soil types in
the watershed (Byczkowski et al. 2001).
Hydrometric discharge measurements, carried out at the Płachty Stare
gauge up to ten times a year, have
been basis for rating curve estimation
and continue its verification. Annual
maximum flood flows from the period
1969–2002 were used for the statistical
analysis. This period has been accepted,
as the whole 40-year, i.e. 1963–2002
(Banasik et al. 2003) period of records did
not fulfill the postulate of homogeneity
according to Kraskal-Wallis (IMGW,
2005).
RESULTS OF WQp%
(T-YEAR FLOOD DISCHARGE)
ESTIMATION AND
CONCLUDING REMARKS
Using the IMGW computer program
(Ozga-Zielińska et al. 1999; IMGW,
2005), which apply the method of
maximum likelihood (MML) for
parameter estimation, two distribution
function were selected from four for
further analyze, based on the Akaike
criterion. The selected distributions have
been as follow:
FIGURE 1. Map of the upper part of the Zagożdżonka River basin
Estimation of T-year flood discharge for a small lowland river...
•
the Pearson type 3 (P3; i.e. 3parameter gamma) distribution, and
• the log-normal (LN) distribution.
Very close values of the Akaike
criterion for the both distribution (i.e.
167,04 and 166,03 for the Pearson
type 3 and the log-normal distributions
respectively) influence the decision
to include both of the functions in the
analyze.
The statistical series of the 34-year
period have been also approximated
using the most frequently used in Poland
probability distribution according to
the Pearson type 3, with the method of
quantiles (MQ; Kaczmarek, Trykozko
1964). The calculations of the T-year
discharges (or discharges with p%
probability of exceedance) are based on
quantiles of the following range: 10%,
50%, 90% and 100%.
The maximal discharges are estimated
from the equation (Byczkowski 1999):
WQp = WQ50%[1 + cvФ(p; s)]
(1)
where:
WQp – maximal annual discharge with
the probability of exceedance p%,
29
WQ50% – maximal discharge with the
probability of exceedance p = 50%,
cv – coefficient of variation of the series
of maximal discharges,
Ф(p; s) – probability function for a given
type of distribution.
The results of the calculations for the
three distributions are shown on the fig.
2 and given in the Table 1.
Results presented on the Figure
2 and in the Table 1 show relative
significant differences in the flood
discharges of small probability of
exceedance, estimated according the two
different statistical distributions, i.e. the
lognormal distribution and Pearson type
3 distribution. As the difference of the
Akaike criterion for the above statistical
distribution was relatively small (i.e.
166.03 for the lognormal and 167.04
for the Pearson type 3 distribution),
one may consider to select also other
criterion at choosing the propel statistical
distribution. Visual assessment of
agreement of the theoretical distributions
with the empirical data, shown on
the Figure 2, indicates also that the
TABLE 1. Flood flows according various distributions and methods of parameter estimation for the
Zagożdżonka River at Płachty Stare gauge station
Return period
T (year)
Probability p
(%)
(1-CDF)
Flood flows (m3/s) according to distribution and method of
parameter estimation
Lognormal -LN
Pearson type 3 - P3 Pearson type 3 – P3
method of quantiles
method of maximum likelihood – MML
- MQ
71.4
33.3
34.3
1000
0.1
100
1.0
31.1
21.8
22.1
20
5.0
14.9
13.9
13.8
10
10
10.2
10.5
10.3
2
50
2.79
3.20
2.80
1.11
90
0.98
0.94
0.75
1.01
99
0.59
0.63
0.59
30
K. Banasik, A. Byczkowski
1
2
5
10
20
50
100
200
500
T
60
1000
55
50
LN-MML
45
P3-MML
40
Q [m3/s]
35
30
25
P3-Mq
20
15
10
5
100
99
90
80
70
60
50
40
30
20
10
5
2
1
0.5
0.2
p [%]
0.1
FIGURE 2. Annual peak discharges frequency curve for the Zagożdżonka River at Płachty Stare gauge
for the period of 1969–2002
lognormal distribution fit better with the
measurement results.
The similar results of flood discharges
of various probabilities, estimation
according to Pearson type 3 distribution,
with the use of the two various method
of parameter estimation (i.e. method
of maximum likelihood and method of
quantiles) indicate that the method has
had small influence on the results of
computation.
As the five largest flood flows, shown
on the Figure 2, indicated heterogeneity
with the remaining data, an analyze in
which a division of the floods caused
by rainfall events and snowmelt events,
as various genesis for the river basin
responses, seems to be needed.
Acknowledgment. The study described
in this paper has been carried out within
research project Nr 1010/P01/2006/30,
founded by Ministry of Science and
Higher Education. The financial support
provided by this organization is gratefully
acknowledged. The assistance of Mr J.
Gładecki in the process of preparing the
material for the analyses is also gratefully
acknowledged.
REFERENCES
BANASIK K., BYCZKOWSKI A.,
GŁADECKI J. 2003: Prediction of
T-year flood discharge for a small river
basin using direct and indirect methods.
Ann. Warsaw Agricult. University –
SGGW, Land Reclam. No 23, p. 3–8.
Estimation of T-year flood discharge for a small lowland river...
BYCZKOWSKI A. 1999: Hydrologia t. 2.
Wyd. SGGW, Warszawa.
BYCZKOWSKI A., BANASIK K., HEJDUK L., MANDES B. 2001: Wieloletnie tendencje zmian procesu opadu
i odpływu w małych zlewniach nizinnych - na przykładzie rzeki Zagożdżonki
(Long-term tendency in changes of precipitation and runoff in a small lowland river basin of Zagożdżonka). Instytut Meteorologii i Gospodarki Wodnej. Atlasy
i monografie. Warszawa, p. 43–52.
IMGW, 2005: Zasady obliczania największych przepływów rocznych o określonym prawdopodobieństwie przewyższenia – Długie ciągi pomiarowe przepływów (Guidelines for computation of
annual flood discharges with small probability of exceedance – long data sets).
Instytut Meteorologii i Gospodarki Wodnej, Warszawa.
KACZMAREK Z., TRYKOZKO E. 1964:
Application of the method of quantiles
to estimation of the Pearson distribution.
Acta Geoph. Polon. T. 12, z. 1.
MAIDMENT D.R. (ed.) 1993: Handbook
of hydrology. McGraw-Hill, Inc. New
York.
OZGA-ZIELIŃSKA M., BRZEZIŃSKI J.,
OZGA-ZIELIŃSKI B. 1999: Zasady obliczania największych przepływów rocznych o określonym prawdopodobieństwie przewyższenia – przy projektowaniu obiektów budownictwa hydrotechnicznego. (Guidelines for computation
of annual flood discharges with small
probability of exceedance – for designe
of hydrotechnical structures). Materiały
Badawcze, Seria: Hydrologia i Oceanologia. IMGW, Warszawa.
PILGRIM D.H., DORAN D.G. 1993:
Practical criteria for the choice of
method for estimating extreme design
floods. Extreme Hydrological Events:
Precipitation, Floods and Droughts
(Proceedings
of
the
Yokohama
Symposium, July 1993). IAHS Publ. no
213, p. 227–235.
31
Streszczenie. Wyznaczenie przepływów maksymalnych prawdopodobnych w małej zlewni
nizinnej przy zastosowaniu metody statystycznej. Codzienne przepływy rzeki Zagożdżonki w
Płachtach Starych (A = 82,4 km2), położonej na
Równinie Radomskiej z okresu 34 lat były podstawą wyboru wartości maksymalnych rocznych,
wykorzystanych do wyznaczenia przepływów
maksymalnych o małym prawdopodobieństwie
przekroczenia. Rozkłady prawdopodobieństwa
– gamma i logarytmiczno-normalny, wybrano
do opisu własności losowych ciągu maksymalnych przepływów rocznych. Przeprowadzone
obliczenia za pomocą programu komputerowego
IMGW, wykorzystujacego metodę największej
wiarygodności do wyznaczania parametrów, wykazały lepszą zgodność rozkładu lograytmiczno-normalnego z danymi empirycznymi niż rozkładu gamma (Pearsona typu 3), zarówno przy zastosowaniu kryterium Akaike, jak również według
oceny wizualnej (subiektywnej). Przy rozkładzie
tym (logarytmiczno-normalnym) uzyskano znacznie wyższe wartości przepływów maksymalnych
o niskim prawdopodobieństwie przekroczenia
(1% i 0,1%). Wyznaczając przepływy maksymalne prawdopodobne metodą dotychczas powszechnie stosowaną w praktyce inżynierskiej tj.
przyjmując rozkład Pearsona typu 3, z parametrami wyznaczonymi metodą kwantyli, wykazano
znikomy wpływ metody szacowania parametrów
na postać rozkładu prawdopodobieństwa.
Authors’ address:
Kazimierz Banasik, Andrzej Byczkowski
Wydział Inżynierii i Kształtowania Środowiska
SGGW
02-787 Warszawa, ul. Nowoursynowska 166
Poland
[email protected]
Curve Number update used for runoff calculation
DONALD E. WOODWARD*, CLAUDIA C. SCHEER*
RICHARD H. HAWKINS**
*Natural Resources Conservation Service, USA
**University of Arizona, USA
Abstract: Curve Number update used for runoff
calculation. The Natural Resources Conservation
Service (NRCS), formerly the Soil Conservation
Service (SCS), developed the runoff curve
number procedure for estimating direct runoff
from rainfall on ungaged agricultural watersheds
in the late 1950s. This procedure is being used
world wide to estimate direct runoff from
rainfall events and, more recently, applied to
continuous simulation models. In 1990 a work
group of hydraulic engineers from NRCS and
the Agricultural Research Service (ARS) began
an effort to update the procedure for estimating
direct runoff. The results of this joint effort will
be explained in this paper.
Key words: NRCS runoff curve, event analysis,
model fitting.
INTRODUCTION
The NRCS runoff curve number
procedure was developed in the late
1950s to estimate the runoff volumes.
It was developed as a simple procedure
for estimating direct runoff for use in the
design of conservation practices. Curve
numbers represent the runoff potential
of various soil-cover combinations.
Originally, curve numbers were
developed by utilizing the maximum
annual daily runoff and associated rainfall
data from ARS watersheds located across
the United States. The original curve
numbers encompassed a wide variety of
hydrologic soil groups, agricultural land
uses and hydrologic conditions. Curve
number values are found in the NRCS
(NEH-630 1985).
Over the years, additional curve
numbers have been developed for new
land uses, tillage practices and cover
conditions. Since adequate rainfall-runoff
data was not always available for these
new situations, alternative methods of
developing curve numbers were utilized.
Many of the newer curve numbers were
developed by comparison of values
for similar land uses, or by weighting
current land use curve numbers, a method
particularly favored in the development
of urban curve numbers.
In the 1980s, NRCS recognized that
an in-depth review of the runoff curve
number procedure was needed in order
to adequately address concerns with
curve number usage. These concerns
included the development of curve
numbers for additional land use and
cover conditions from limited data, the
need of curve numbers for new land uses
and cultivation practices, regional and
seasonal variation of curve numbers.
At the 1989 ARS/NRCS Hydraulic
Engineers Workshop, the principle
subjects discussed were the use and abuse
34
D.E. Woodward et al.
of curve numbers, the application of
curve numbers in continuous hydrologic
models, and the future of the runoff curve
number procedure. Meeting participants
recommended that an ARS/NRCS Work
Group be established to further the study
of curve numbers and investigate the
potential need to develop curve numbers
for major regional land areas.
At the 1992 meeting, the Curve
Number Work Group decided that the
procedures used to determine the first
curve numbers needed to be revised and
updated. A limited test indicated that
plotting the ordered values rather than the
paired or natural values provided realistic
results. In the initial and subsequent
analyses, annual maximum daily rainfall
and runoff values were used. It was noted
that while paired values matched nature,
ordered values better matched design
and evaluation watershed studies.
The Work Group concluded that they
did not have the resources to analyze all
of the data available for small watersheds
in an orderly and timely manner. Hawkins
(1998) had the most complete, organized
set of ARS data from small agricultural
watersheds and spent considerable time
and effort to remove errors from this
data and was interested in dedicating the
time needed to complete the required
analysis.
About this same time period, the
American Society of Civil Engineers
(ASCE) established a task group whose
purpose it was to review the NRCS
rainfall-runoff estimation procedure
known as the curve number system.
The Curve Number Work Group
revised eight chapters of National
Engineering Handbook (NEH-630)
Revised version l of NEH-630 may be
found at: http://www.wcc.nrcs.usda.
gov/water/quality/common/neh630/
4content.html .
During the revision of the NEH630 chapters incorrect or misleading
statements were corrected and, in some
cases, certain points were reemphasized
for clarity. The most important changes
were:
• The reference to Antecedent Moisture
Conditions (AMC) was removed
and the terminology was changed to
Antecedent Runoff Condition (ARC).
• It was recognized that ARC II
represents the average watershed
condition of the watershed when
flooding occurs. This means the
average watershed conditions could
be different for different geographical
areas when flooding occurs. In other
words, a particular runoff curve
number for ARC II in Arizona
represents a different condition than
the same curve number for ARC II in
Oregon.
• Explicit expression of the Runoff Curve
Number Equation as a transformation
of the rainfall frequency distribution
to runoff frequency distribution was
reemphasized.
• ARC I and ARC III were expressed
as measures of dispersion about the
central tendency (ARC II). This is
a corollary to treating curve number
as a random variable.
• Mathematical proof and demonstration
that S does not include Ia was
documented.
• The desirability of locally determined
curve numbers was reiterated.
Although this was part of the original
documentation in NEH-630, local
calibration has been seldom pursued.
Curve Number update user for runoff calculation
ORDERED PAIRS AND
ASYMPTOTES
One of the goals of the Work Group was to
standardize the procedure for calculating
curve numbers from rainfall-runoff data.
The accepted handbook method was to
plot the annual series of rainfall-runoff
on a scatter diagram and select the curve
number that best fit the data. However,
this method ignored the many storms
that were not the largest annual events.
The runoff curve number equation
is often used to transform a rainfall
frequency distribution into a runoff
frequency distribution. For example, the
100-year rainfall is used to determine
the 100-year runoff, etc. This practice,
called frequency matching, led to the
idea of ordered pairs (Hjelmfelt 1980).
Hawkins (1993) followed this idea by
sorting the rainfall and runoff depths
separately and re-aligning them on a rank
order basis, creating new sets of rainfallrunoff pairs. These rainfall-runoff pairs
can be thought of as having equal return
periods with the individual runoffs not
necessarily associated with the original
causative rainfall.
Using all the storms in the data set,
Hawkins calculated a curve number for
each of the ordered pairs and plotted them
against the rainfall. The curve number
was found to vary with storm depth, but
for most cases approached a constant
value at higher rainfalls. The resulting
curves were fitted with asymptotic
equations to approach this constant
curve number value.
The limiting
curve number, approached as rainfall
approaches infinity, is taken as the best
fit curve number for the watershed.
35
This method of determining a
watershed curve number has the
advantage of being mathematical and
therefore programmable. Results are
mostly influenced by the largest event,
which is in keeping with the usual intended
applications of the curve number. . This
method appears to give results consistent
with the present procedures and curve
number tables in NEH-630. The curve
number Work Group has adopted this
procedure for future determination of
curve numbers from local data.
MODES OF APPLICATION OF
THE CURVE NUMBER METHOD
The Work Group recognized three
distinctly different modes of application
for curve numbers:
1. Determination of runoff volume for a
given return period, given total event
rainfall for that return period.
2. Determination of direct runoff for
individual events. This acknowledges
the variation between events and is
the basis for the initial development.
3. Process models, an inferred
application as an infiltration model,
or a soil moisture-curve number
relationship, or as a source area
distribution.
The first application is the most
widely used in engineering and uses the
curve number to transform the rainfall
frequency distribution into a runoff
frequency distribution. The runoff volume
that is computed is often overlooked
and the peak discharge, which is more
frequently the desired value, is calculated
using a unit hydrograph model.
36
The second application, runoff from
individual rainfall events, is the basis
for the original development, as the
rainfall versus runoff plots led to the
curve number concept. There is a wide
variation of runoff from rainfalls of
the same magnitude which forces us to
acknowledge that curve number varies
among storms for a wide range of reasons.
The original handbook, developed in
large part for conditions in the humid
east, south, and mid-west, designated
Antecedent Moisture Condition, or
AMC, as the most significant variable
in explaining this. Average moisture
condition was called AMC II and
applied to the curve number when
flooding occurs. Dry conditions (AMC
I) applied to the low curve number, and
wet conditions (AMC III) applied to the
high curve number. This condition is
most often determined by prior rainfall
since soil moisture conditions are not
frequently monitored.
The Work Group studied the effect
of soil moisture on curve numbers by
looking at infiltrometer studies (Van
Mullem 1992). Four studies with 162 data
pairs on 86 different soils from across
the United States were used. Although
average curve number increased from
9% to 40% between the studies from the
initial test (dry to average condition) to
the wet condition test (average to wet
condition), no significant relationship
was found between soil moisture and
curve number. The study indicated that
the difference in curve number that
might be related to soil moisture is much
less than the variation between ARC I
and ARC III. Similarly, Hawkins and
Cate (1998) showed that 5-day prior
rainfall was the only consistent factor in
explaining deviations from the central
trend of runoff in 11 of 25 agricultural
watersheds which were studied, and at
levels far below handbook expectations.
Because prior rainfall explains only
part of the variation of curve number,
the terminology has been changed to
Antecedent Runoff Condition, or ARC.
More importantly, the ARC I and ARC
III conditions have been shown to be
the bounds on the distribution of curve
number. Figure 1 shows the ARC I, II, and
III curve numbers plotted on a rainfallrunoff scatter diagram. Upon review of
this by Hjelmfelt (1980) indicated that
there appears to be a correlation between
exceedance limits and ARC. Figure 2
shows that ARC I represents the 10%
exceedance limit and ARC III represents
the 90% exceedance limit for a number
of agricultural watersheds (Hjelmfelt et
al. 1982).
It might be inferred that the curve
number model is an infiltration model
because in its application it is used to
determine runoff incrementally over
the duration of the storm for input into
a unit hydrograph model. With this use
it becomes a surrogate for an infiltration
model. This has created much confusion.
The model doesn’t behave like most
infiltration models/equations because
the curve number losses don’t always
decline with time or with prior rainfall
and may actually increase when rainfall
intensity increases (Hawkins 1980). The
curve number model behavior in this
regard is the same as from a partial area
saturation model. Additionally, with the
curve number model, the ultimate steady
state rate is zero. It does not approach
a steady-state non-zero infiltration rate
with time, as do the Horton or Green-
37
Hastings, Nebraska WS44028 (1941–1954)
5
CN(III) = 94
3
CN(II) = 85
Q(in)
4
2
CN(I) = 70
1
0
0
1
2
3
4
5
P(in)
FIGURE 1. Rainfall-Runoff scatter diagram showing ARC I, II, and III curve numbers
FIGURE 2. The 10% and 90% exceedance values compared to the AMC I and AMC III curve numbers
Ampt Equations. There doesn’t seem to
be any consensus as to whether this is the
way watershed losses occur.
It should be emphasized that the curve
number model is not a point infiltration
model and the difference between
rainfall and runoff is better defined as
watershed “losses.” Watershed indices,
such as the curve number, are lumped
expressions of net basin performance.
In this regard, the runoff curve number
model performs well as an integrator of
38
all the losses from all the processes over
the watershed, which was the original
intention.
It is also sometimes inferred that the
parameter S, defined as the potential
maximum retention, is a physical
property of the site like a soil moisture
storage parameter, and the water in it can
be accounted for. This is has not been
shown with any certainty. The parameter
S (or curve number) is a model variable
and is only constant for a particular storm.
Although it is related to soil and cover
characteristics, it is not an identifiable
physical property.
VARIATION OF CURVE NUMBER
WITH SEASON AND LAND USE
Curve numbers were derived from
rainfall-runoff data for 15 distinct land
uses on 177 small watersheds in the
United States (Rietz and Hawkins 2000).
Curve numbers for each land use on
each watershed were calculated using
the asymptotic method and evaluated
at the local, regional and national scale.
Significant differences at the 5% level
were found between the curve numbers
of almost all of the different land uses
tested. Significant differences in curve
number were also found on grazed and
ungrazed paired watersheds, and on
watersheds that had undergone land use
conversions.
The general magnitudes and rank
order of the average land use curve
numbers were in general agreement with
expected handbook values. Meadows
almost always produced the lowest
curve number at both the local and
regional level. Forestland produced the
lowest overall average curve number
at the national level, but also displayed
the largest variability. No significant
differences could be determined between
curve numbers for pasture and rangeland
at the regional scale or between row
crops and small grain at any scale. Where
comparable, pastures usually had higher
curve numbers than meadows. None of
these comparisons considered hydrologic
soil group or any other soil parameter.
Seasonal variation of curve number
has also been noted. This is seen more
readily in the more humid settings, and
is rare in arid and semi-arid watersheds.
Where evident it follows a pattern with
higher curve numbers in the dormant
season when the ground has less cover
and is likely to be wetter; and lower
curve numbers during the summer when
the ground is dryer and vegetation is
in a high growth stage. Also, seasonal
variation in forest curve numbers may
be associated with leafing stages (Price,
1998).
THE INITIAL ABSTRACTION
RATIO (IA/S RATIO)
The Initial Abstraction Ratio (Ia/S or λ)
was assigned a value of 0.2 in the original
development of runoff curve numbers.
Data from NEH-630 (1985) has been
used to support the value of 0.2. As part
of the effort to develop the required
documentation for the curve number
study, it became apparent that additional
review of λ would be appropriate. Data
analysis from NEH-630 (1985) indicated
the statistical average (mean) of the
plotted values was about 0.05.
Two techniques, Event Analysis and
Model Fitting, were used for determining
Ia/S from field data sets.
Event Analysis. Event analysis
requires concurrent synchronized breakpoint records of both rainfall and runoff
depth. The event rainfall depth recorded
when the direct runoff hydrograph
begins is taken as Ia. Knowing the total
event rainfall P and the direct runoff Q,
equation 1a is solved for S, and the ratio
simply taken Ia/S = λ. Each event gives
a separate value for λ, and the median for
a large number of events is taken as the
representative watershed value.
General Model Fitting: Here the
value of λ is simply determined by
iterative least squares procedure fitting
for both λ and S of the general equation.
Q = (P – λS)2/(P + (1 – λ)S)
39
for P ≤ λS
Q=0
The objective of the fitting is to find
the values of λ and S such that
Σ{Q – [(P – λS)2/(P + (1 – λ)S)]}2
for P > λS
is a minimum. Here each P:Q data set
gives only one value of λ. An illustration
of such fitting is given in Figure 3.
In each of the above two methods, only
larger storms were used. This was done to
avoid the biasing effects of small storms
towards high curve numbers. With Event
Analysis, only events with Pe = P – λS
≥ 1 inch were used. With Model Fitting,
only events with P ≥ 1 inch were used. As
shall be seen, resulting values of Ia were
often quite small, so that this difference
between the two techniques was slight.
For statistical analysis, only watersheds
for P ≥ λS
ARS W S26030 Coshocton, Ohio
6
Direct Runoff Q (inch)
5
4
Q= P
Ordered
3
Natural
2
1
0
0
1
2
3
4
5
6
Rainfall P (inch)
FIGURE 3. Model Fitting by least squares for WS26030 located at Coshocton, OH with a drainage area
of 303 acres. For the natural data (squares): S = 4.0974 inches, CN = 70.8, λ = 0.0179, R2 = 50.50%,
and SE = 0.32 inch. For the ordered data (triangles): S = 2.0943 inches, CN = 82.6, λ = 0.1364, R2 =
= 99.17%, and SE = 0.0372 inches
40
with more than 20 events with P ≥ 1 inch
or Pe ≥ 1 inch were used.
In addition, for the model fitting
determinations, both “natural” and
“ordered” data sets were used. Natural
data pairs the P and Q as they naturally
occurred in time, and thus displays
considerable variety in runoff with
rainfall. Ordered data matches (usually)
rank-ordered P and Q values, so that
each has approximately the same
return period. This is in keeping with
a major application of the method,
which in design work at least - matches
the frequency of the rainfall with the
frequency of the runoff. For example, the
100-year rainfall is assumed to produce
the 100-year runoff.
Rainfall-runoff data from 307
watersheds or plots were used,
originating from USDA-Agricultural
Research Service, US Forest Service,
US Geological Survey, and New Mexico
State University. The data sets covered
23 states, mainly in the east, Midwest,
and south of the United States. There
was no data from the northwestern 1/3 of
the country, from roughly California to
Minnesota. A total of 28,301 events were
available that met the rainfall depth (P
and Pe) criteria. For event analysis, only
ARS data was applicable, insofar as it
alone contained the needed detailed in-
storm break-point information. All others
were only rainfall and runoff depths P
and Q. This is summarized in Table 1.
These 307 watersheds all had 20 or
more events which met the storm size
criteria. The ARS data is available from
Web site ftp://hydrolab.arsusda.gov/
pub/arswater/. The “USLE” plot data
had been used in the development of the
Universal Soil Loss Equation, and was
downloaded from the web site: http://
Forest
topsoil.nserl.purdue.edu/usle/.
Service data was in large part supplied in
reduced form to Hawkins (1980, 1993)
by Hewlett (1977, 1984), who used it
in earlier papers (Hewlett, et al., 1977;
Hewlett and Fortson, 1984). The Jornada
plot data, from a site north of Las Cruces
NM, was supplied by Ward and described
in Hawkins and Ward (1998). The USGS
data was supplied from local sources
for a number of urban and urbanizing
watersheds in the Tucson area.
In general, the results showed that λ
is not a constant from storm to storm,
or watershed to watershed, and that the
assumption of λ = 0.20 is unusually
high.
In 202 out of the 307 watersheds
studied, Ia/S = 0.05 had a lower standard
error in predicted Q that Ia/S = 0.2. ( Jiang
2001)
TABLE 1. Data sets and sources
Data source
ARS
USLE (ARS)
USFS
Jornada (NMSU)
USGS
# Watersheds (w) or plots (p)
134 (w)
137 (p)
26 (w)
6 (p)
4 (w)
Method used
Event Analysis, Model Fitting
Model Fitting
Model Fitting
Model Fitting
Model Fitting
CONCLUSIONS
1. Additional work is needed to develop
a revised set of curve numbers.. An
up-to-date soils maps needs to be
prepared for each of the experimental
watersheds used in the study. NRCS
is reviewing the impacts of changing
the Ia/S ratio.
2. The use of curve numbers as a
moisture counting technique needs
additional study. While it appears
to work, additional documentation
needs to be developed.
REFERENCES
Agricultural Research Service Water Data
Center, 1995: ARS Water Data: ARS/
Access CD. USDA-ARS Hydrology
Lab, Beltsville, Maryland. Agricultural
Research Service Water Data Center,
http://www.hydrolab.arsusda.gov/
arswater.html.
HAWKINS R.H. 1980: Infiltration and
Curve Numbers: Some Pragmatic and
Theoretic Relationships”, Proceedings of
Symposium on Watershed Management
1980, ASCE, 925–37.
HAWKINS R.H. 1993: “Asymptotic
determination of runoff curve numbers
from data”. Journal of Irrigation and
Drainage Engineering. Amer Soc Civ
Eng. 119(2): 334–345.
HAWKINS R.H., WARD T.J. 1998: “Site
and cover effects on event runoff,
Jornada Experimental Range, New
Mexico”. Proceedings from American
Water Resource Association Conference
on Rangeland Management and Water
Resources. Reno, NV, 361–370.
HAWKINS R.H., CATE A. 1998: “Secondary
Influences in Curve Number Rainfall
Runoff”, Proceeding of the International
Water Resources Engineering Conference
ASCE.
41
HEWLETT J.D., CUNNINGHAM. G.B.,
TROENDLE C.A. 1977: “Predicting
Storm Flow and Peak Flow from small
basins in humid areas by the R-index
method”. Water Resources Bulletin.
13(2): 231–253.
HEWLETT J.D., FORTSON J.C. 1984:
“Additional tests on the effect of rainfall
intensity on storm flow and peak flow
from wild-land basins”. Water Resources
Research. 20(7): 985–989.
HJELMFELT A.T. 1980: “Empirical
investigation of curve number technique”.
Journal of the Hydraulics Division.
American Society Civil Eng 106 (HY9):
1471–1476.
HJELMFELT A.T., KRAMER L.A.,
BURWELL R.E. 1982: “Curve Numbers
as Random Variables, Rainfall-Runoff
Relationship´ Resources Publications,
Littleton, CO., 365–370.
HJELMFELT A.T., WOODWARD D.A.,
CONAWAY G., QUAN Q.D., Van
MULLEM J., HAWKINS R.H. 2001:
“Curve Numbers, Recent Developments”.
XXIX IAHR Congress Proceedings,
Beijing, China.
JIANG R. 2001: Investigation of Runoff
Curve Number Initial Abstraction Ratio.
MS thesis, Watershed Management,
University of Arizona, 120 pp.
PRICE M. 1998: “Seasonal Variation in
Runoff Curve Numbers”. MS Thesis,
Watershed Management, University of
Arizona. 189 pp.
RIETZ P.D., HAWKINS R.H. 2000: “Effects
of land use on runoff curve numbers”.
Watershed Management 2000, American.
Society. Civil Engineers. Proceedings
Watershed Management Symposium,
Fort Collins CO. (CD ROM)
Soil Conservation Service 1985: National
Engineering Handbook, Section 4,
Hydrology (NEH-630).
USDA/USLE data web site: http://topsoil.
nserl.purdue.edu/usle/.
USDA, Agricultural Research Service: ftp://
hydrolab.arsusda.gov/pub/arswater/.
42
Van MULLEM J.A. (1992). “ Soil Moisture
and Runoff–Another Look”. ASCE Water
Forum ‘92, Proceedings of the Irrigation
and Drainage Session, Baltimore, MD.
Streszczenie: Aktualizacja parametru CN metody obliczania opadu efektywnego. W drugiej
połowie lat 50. Służba Ochrony Gleb (ang. Soil
Conservenation Sernice ) w USA (obecnie Służba Ochrony Zasobów Naturalnymi ang. Natural Resources Conservation Service – NRCS)
opracowała procedurę nazywaną metodą Curve
Number do wyznaczania opadu efektywnego dla
zlewni nieobserwowanych. Procedura ta jest sto-
sowana na całym świecie do wyznaczania opadu
efektywnego ze zdarzeń opadowych oraz częściej
w modelach do symulacji ciągłych. W 1990 roku
zespół z NRCS i Służby Badań Rolniczych (ang.
Agricultural Research Service) rozpoczął działania nad aktualizacją procedury określania opadu
efektywnego. W artykule przedstawiono wyniki
przeprowadzonych prac.
MS received July 2006
Authors address:
Donald E. Woodward
Natural Resources Conservation Service, USA
Assessment of hydrologic regime changes induced by the Jeziorsko
dam performance and morphodynamic processes in the Warta river
TOMASZ DYSARZ, JOANNA WICHER-DYSARZ
Department of Hydraulic Engineering, Agricultural University of Poznań, Poland
Abstract: Assessment of hydrologic regime
changes induced by the Jeziorsko dam
performance and morphodynamic processes
in the Warta river. In the paper the overview of
morphological processes occurring in the selected
Warta river reach is given. The processes were
induced by the performance of the Jeziorsko dam,
which was built in 1986. The main purpose of
the presented research is the assessment of these
processes impact on the hydrologic conditions in
the river system. First time the Jeziorsko reservoir
it was filled up to the maximum water level in
1991. In 1995 the hydropower plant was put into
operation. The sediment particles transported
with flowing water were accumulated in the inlet
part of the reservoir. In this area very convenient
conditions for wild life, especially for water plants
and birds, appeared as the result of water stages
increase. However, the risk related to seasonal
floods and dike break possibility have threaten
the local societies until now. The historical data
and hydrodynamic simulation model are used
for the assessment of the hydrologic changes.
The data consists of field measurements of the
river morphology and hydrologic data from the
selected river gages. The hydrodynamic model
was used to reconstruct the flow conditions in
particular states of the river system evolution. The
statistical analysis of the obtained results enabled
the assessment of the hydrologic characteristics
transformation in the selected Warta river reach.
Key words: hydrodynamic, simulation model,
river morphology, river hydrology.
INTRODUCTION
In the paper the overview of
morphological processes occurring
in the selected Warta river reach is
given. The processes were directly or
indirectly induced by the performance
of the Jeziorsko dam, which was built
in 1986. The main purpose of the
presented research is the assessment of
these processes impact on the hydrologic
conditions in the river system. Since the
dam was built the velocity characteristics
in the analyzed Warta river reach were
changed what induced the deposition
of the transported sediment in the inlet
part of the reservoir. Next the sediment
accumulation in the river channel induced
one more process affecting hydrologic
regime. Due to the increase of floodplain
inundation frequency, the luxuriant
vegetation start to grow there. These
two processes, sediment accumulation
and vegetation growth, cause additional
increase of floodplain inundation. The
processes induced by the dam built and
its performance seems to be self-driven
cycle causing irreversible alterations of
hydrologic regime.
The problem under consideration
draws attention of many researchers
and engineers today. The relationships
between hydrologic regime conditions
and the riparian vegetation are crucially
important for the robust river management.
The examples of such studies were given
by Keeland et al. (1997), Surtevant
(1998), Stromberg (2001), Nilsson and
44
T. Dysarz, J. Wicher-Dysarz
Svedmark (2002) and many others.
Amongst many anthropogenic actions in
the water environment the damming of
rivers plays special role due to complexity
of induced ecological responses. One
of the most interesting aspects of these
responses is the dam impact on the
upstream hydraulic conditions and
related changes of hydrologic regime.
The example of studies on this aspect
was given by Mumba and Thompson
(2005). They analyzed the changes in
the floodplains located along the Kafue
river (Zambia) reach between two dams,
Itezhi-tezhi (upstream) and Kafue Gorge
(downstream). The Authors proved that
dams totally changed the hydrological
and ecological conditions in the region.
Almost permanent inundation of the area
located between the dams was the direct
reason. Such conditions were imposed
by parallel influences of releases from
upper dam and high water levels in the
lower reservoir.
The methodology used for assessment
of hydrologic regime changes ranges
from simple approach based on equal
discharge studies (Pinter and Heine
2005) to complex analysis including
studies of discharges and water table
frequency (e.g. King et al. 1998, Maingi
and Marsh 2002, Magilligan and Nislow
2005, Wellmeyer et al. 2005). This
is also quite common to simulate the
behavior of the water system by means
of mathematical model (e.g. Kite 2001,
Maingi and Marsh 2002). In each case
the methodology applied should match
the specific features of the system,
processes occurring there as well as data
availability.
The main purpose of the presented
paper is to present the hydrological
regime changes in the Warta river
reach located upstream of the Jeziorsko
reservoir. According to the performed
observations the hydrologic conditions
are caused by three factors. The primary
reason is water management in the
reservoir. Two next elements are (1)
sediment accumulation in the river
and (2) vegetation growth. The field
measurement data and hydrodynamic
model are used for our studies. Steady
and unsteady flow simulation were run
to reconstruct the flow conditions in the
system before and after the analyzed
changes. The final results are presented in
two way. First method consists of water
stages in the selected river cross-section
related to some flows determined for the
inflow gauge station. The chosen flows
are related to the specific probabilities of
excedance, e.g. 1%, 10%, or maximum
observed discharge in the system is used.
The second method is the statistical
analysis of unsteady flow simulations
presented as relative and cumulative
frequency functions related to the
particular stages of river evolution.
The paper includes into six main
parts, acknowledgments, references and
two appendixes. In the first part brief
introduction to the investigated problem
is given. In the second part the study
area and processes occurring there are
described. The data collected are briefly
presented in third part. Then the basic
elements of the applied methodology
are presented in fourth sections. In fifth
part the obtained results are presented
and discusses. Finally the conclusions
are formulated in the sixth part. The
attached appendixes includes the details
of applied methodology.
Assessment of hydrologic regime changes induced by the Jeziorsko...
CASE STUDY SYSTEM
Jeziorsko reservoir was built in 1986
in central part of Poland. It is located
in the Warta river, between the Sieradz
(upstream) and Uniejów (downstream)
gauge stations. The reservoir inlet with
Warta river reach are shown in the attached
photo map (Fig. 1). The Jeziorsko
reservoir is relatively new object. First
time it was filled up to the admissible
maximum water level (121.50 m a.s.l.)
in 1991. The hydropower plant was put
into operation in 1995.
The Warta catchment area in the
inlet of the reservoir is 8450 km2. The
45
average discharge in the nearest Sieradz
gauge station is about 45 m3/s, but the
discharges variability ranges from 10 m3/s
to 440 m3/s (observed in 1997). The wet
season during the year is January – May.
The lowest discharges are observed
during the summer and autumn time,
from September to December.
The
Jeziorsko
reservoir
is
multipurpose reservoir performing in
the annual working scheme. During the
months from September to December the
water stages are kept on the minimum
level of 116 m a.s.l. In spring months
the reservoir is filled with water. The
maximum water level 121.5 m a.s.l. is
km 499
km 500
km 501
km 502
km 503
km 504
km 505
km 506
FIGURE 1. Photo-map of Jeziorsko reservoir inlet part and Warta river
46
hold during April – June period. In the
period from July to August the water
stored at the reservoir is used and the
reservoir water stages gradually come
back to the minimum level. The main
purpose of the reservoir performance is
the flood protection during wet season.
The reservoir is also the source of
water for the big cities in Great-Poland
province: Poznań, Koło, Śrem and
Konin. The water stored is also used for
irrigation
purposes and to preserve biological
life below the dam. The Pątnów-Konin
and Adamów power plants used the
water from the reservoir for cooling.
Other goal of the Jeziorsko performance
is production of hydro-power, though,
this is not significant due to the small
difference between upper and lower
water levels in the dam (about 10 m).
The reservoir is also used for recreation
and inland fishery.
Since the very beginning of the
reservoir performance the changes of
the Warta river bed and floodplains are
observed in the inlet part of the Jeziorsko.
Due to decrease of flow velocities there
the capacity of sediment transport is
lower and material is deposited. The
most important processes occur in the
river reach below the bridge in Warta
town, between km 503 and km 500 (see
Fig. 1).
The annual scheme of reservoir
performance foster the vegetation
growth in this area. It enabled to
establish a nature reserve for birds in
the upper part of the reservoir in 1998.
Since this time the important reservoir
performance purpose is conservation of
water conditions sufficient for biological
life in this area. Due to the existence of
the nature reserve and Environmental
Protection Ministry act valid since 1998,
the river engineering actions are not
allowed in the reservoir. The purpose
of this regulation was the protection of
birds habitats and biological life in the
upper part of the considered object.
The regulation causes also some
problems indicated in previous papers,
e.g. Wicher (2004), Wicher-Dysarz and
Przedwojski (2005). The water levels
in the backwater part are gradually
increasing. Long-term results of these
processes may cause degradation of the
river in the inlet part of the reservoir as
well as the reservoir capacity decreasing.
Second risk is related to destruction of
flood protecting and backwater dikes
by overtopping. The losses occurring in
such case are difficult to overestimate
but the probability of this danger seems
to increase.
However, the Ministry act enabled the
presented research in some sense. In other
conditions the authorities responsible
for the reservoir effectiveness and safe
use should take steps preventing from
these processes. Now, it is possible to
observe the hydrologic regime evolution
in the pure form. The only influences
are reservoir performance, sediment
accumulation and vegetation growth.
Hence, the chosen study area properly
fits the research purposes.
COLLECTED DATA
The set of available data consists of (a)
system geometry, (b) discharges and
water stages in the system, (c) state of
the vegetation cover.
The channel and floodplain elevations
are observed for long period of time in the
area under consideration. The most basic
set of data consists of the river regulation
design prepared in 1975 (Matan 1975).
The regulated river seemed to be stable
and was not subject to any changes until
the reservoir started to perform in 1991.
Hence, this set of data may be considered
as the original terrain and river bottom
without any significant influence of the
analyzed processes. The another sets of
applied data consists of measurements
done in 1997 and 2004. After a few years
of reservoirs performance the Warta river
changes were checked and assessed by
the team from Department of Hydraulic
Engineering, Agricultural University of
Poznań. The measurements results were
described by Wicher (2004) and Wicher-Dysarz and Przedwojski (2005). Hence,
the system geometry data consists of
three sets of river bottom and floodplain
measurements in 37 cross-sections
covering the river reach from Sieradz
gauge station (km 520+850) to Jeziorsko
dam (km 486+500). These data were
used to describe three stages of the river
bottom and floodplain evolution.
Other data used in our analyses are
discharges and water stages observed
in Sieradz gauge station as well as
water levels measured at the Jeziorsko
dam headwater. The set of data from
Sieradz gauge station consists of daily
measurements done in periods 1963–
–1970, 1973–1983, 1993–1995 and
1997–2001. These are 27 years of
hydrological data. Some of the data were
measured and published by Institute of
Meteorology and Water Management
(1963–1970, 1973–1983). Others were
collected by Regional Board for Water
47
Management in Poznań. The later data are
made available under the circumstances
of cooperation between Department of
Hydraulic Engineering and Regional
Board for Water Management. The data
collected from Sieradz gauge station
are used to describe the hydrologic
condition in the inlet of the analyzed
river reach which are not affected
by reservoir performance. The water
stages at Jeziorsko headwater were
observed during the period of reservoir
performance, from 1992 to 2001. This set
of data was collected by Regional Board
for Water Management. The data were
used to design the average scenario of
reservoir performance during the year.
The impact of vegetation on the
roughness coefficients was assessed on
the basis of measurements including
plants localization and vegetation
cover density. The stalk thickness was
estimated in the same way. The necessary
measurements were done by researchers
from
Department
of
Hydraulic
Engineering during 2004–2005.
On the basis of collected data hydromorphological stages of the system
evolution is classified as four stages of
river bottom. The stages of river bottom
enable the design of five stages of
hydraulic conditions in the water system.
First analyzed state of the system is the
primary bottom described in Matan
(1975). This geometry was considered
as pre-dam conditions. The second
geometry was prepared on the basis of
measurements done in 1997 (Wicher
2004). This stage of the system correspond
to the system state a few years after the
beginning of reservoir performance.
The roughness in the particular crosssections is the same as in the 1975. It
48
does not describe the slight changes of
vegetation cover in this period. Hence,
the impact of pure sediment deposition
may be assessed. The third and fourth
geometry of the system was prepared on
the basis of measurements done in 2004
and corresponds to current conditions.
However, the third scheme does not
reflex to the roughness changes due to
the vegetation growth. This element is
included in the fourth system geometry.
Hence, the current state of the system
may be assessed in two ways: with and
without vegetation growth. The prepared
stages of river bottom and hydraulic
conditions are as follows:
1) initial bottom reflecting pre-dam
conditions (1975) with free outflow
in the outlet;
2) initial bottom reflecting pre-dam
conditions (1975) with outflow
corresponding to the average annual
scheme of reservoir performance;
3) bottom measured in 1997 with the
initial roughness and outflow as
above;
4) bottom measured in 2004 with the
initial roughness and outflow as
above;
5) bottom measured in 2004 with
roughness reflecting current state of
vegetation and its seasonal variation,
outflow corresponding to the
average annual scheme of reservoir
performance.
In the fifth stage the concept of
roughness changes due to the vegetation
growth is applied. The riparian vegetation
impact on the flow conditions may be
modeled as the changes of Manning
roughness
coefficients.
Following
Klopstra et al. (1997) concept of velocity
changes due to the presence of non-
submerged vegetation the roughness
coefficient may be expressed as
2
n = h3
Cw
md
2g
(1)
where:
g – acceleration of gravity,
h – the depth,
Cw, m and d – the parameters describe the
characteristics of riparian vegetation,
Cw – the coefficient reflecting seasonal
changes of leaves cover.
Its values were taken as Cw = 1.05
for winter period and Cw = 1.40 for
summer (Armanini et al. 2005). m is
average number of plants in one square
meter and d is average diameter of
bushes stalks. The presented approach
was used to describe the changes of
roughness in the floodplain. Parameters
m and d were determined on the basis of
measurements done in 2004 and 2005.
The listed stages of river system were
used for the simulation of steady and
unsteady flow conditions along the reach
under consideration.
APPLIED METHODOLOGY
The main basis for the assessment
presented here is the set of collected
data described in previous part of the
paper. To estimate the hydrologic regime
characteristics related to five stages of
river system several deterministic and
stochastic techniques are used. The
methodology applied consists of three
main elements. These are (1) maximum
flows probability of excedence, (2)
hydrodynamic model simulations for
steady and unsteady conditions, (3)
analysis of relative and cumulative
frequencies of water stages in selected
river cross-sections. The above elements
are described shortly in the following
parts of the paper.
In the first part of the analysis the
maximum flows probability of excedence
is elaborated. The curve is determined
for the Sieradz gauge station located
in the inlet of the system. The curve is
shown in Figure 2. Crosses represent
the annual maximum flows observed
and probabilities of excedance assigned
to them according to Weibull formula.
The continuous line is the Pearson
type III curve estimated by means of
quintile method (Ozga-Zielińska and
Brzeziński 1994). The estimated curve
is used to determine the characteristic
flows occurring in the system with some
specified probability of excedance. The
chosen flows are commonly used for the
assessment and design purposes. They
are presented in Table 1.
100
observations
Pearson III curve
probability [%]
80
60
40
20
0
0
100
200
300
400
maximum discharge [m3/s]
500
FIGURE 2. Maximum flows probability of excedance for Sieradz gauge station
49
TABLE 1. Selected flows with specific
probabilities of excedance determined for the
Sieradz gauge station
Probability
of excedance
10 %
5%
1%
0.5 %
Observed
maximum
Denotation
Value [m3/s]
Q10 %
Q5 %
Q1 %
Q0.5 %
245.40
263.28
294.34
304.89
Max Q
440.00
The determined probable flows
together with minimum and maximum
water stages in the reservoir are used
to simulate the steady flow conditions
in the analyzed Warta river reach. The
computations are done for each of five
river bottom stages listed in previous part
of the paper. The hydrodynamic model
is used in this purpose. The routines
included in steady flow module of the
classic HEC-RAS package are applied.
The main basis of the used methodology
is energy balance equation written for
the compound channel. The complete
description of HEC-RAS package may
be found in Brunner (2002).
The results of the simulations are
changes of water levels in selected river
cross-sections related to the maximum
flows and stages of river bottom. They
are presented and discussed in the next
part of the paper.
The second part of the performed
analyses consists of inflow scenarios,
unsteady flow simulations and frequency
analysis. The used inflow scenarios are
the observed discharges in the Sieradz
gauge station. They are combined with
the average annual scheme of water
management in the Jeziorsko reservoir.
The unsteady flow module included in
50
HEC-RAS package is used to simulate
the behavior of the system related
to five stages of the river. Then the
obtained results are analyzed by means
of water stage relative and cumulative
frequencies determined in several river
cross-sections. The basic definitions
of frequency curves are used. More
information may be found in Chow et al.
(1988).
The selected cross-sections are
located between km 503 and km 500,
where the most intensive morphological
and vegetation changes are observed.
The results are presented and discussed
in the next part of the paper.
elevation
[m a.s.l.]
124
Q1% in 2004
123
Max Q in 1975
Q1% in 1975
122
121
120
cross-section
distance
[m]
119
0
The described analyses are done for the all
cross-sections in the river reach between
the bridge in Warta town (Fig. 1) and the
Jeziorsko reservoir. These are sections
located at km 503+560, 503+380,
503+140, 502+800, 502+300, 501+270,
500+970, 500+510 and 499+890. The
examples of obtained results are shown
in Figures 3–10. Figures 3–6 presents the
results related to the first part of analyses
done with the help of steady flow module.
The rest of mentioned figures consists of
relative and cumulative frequency curves
elaborated on the basis of unsteady flow
simulations.
The Figures 3 and 4 present the water
stages related to analyzed flows Q1% and
Max Q. The water stages are estimated
for the pre-dam conditions (denoted as
‘in 1975’) and current conditions (‘in
2004’). These stages are drawn on the
original cross-sections contours. The
100
200
300
400
500
600
FIGURE 3. Cross-section shape and characteristic water stages in km 503+560
elevation
[m a.s.l.]
124
ANALYSIS OF SIMULATION
RESULTS
Max Q in 2004
Max Q in 2004
123
Q1% in 2004
122
Q1% in 1975
Max Q in 1975
121
120
cross-section
119
distance
[m]
118
0
100
200
300
400
500
600
700
FIGURE 4. Cross-section shape and characteristic water stages in km 503+140
given examples present the results for
the cross-sections located in km 503+560
and km 503+140. It is clearly visible
that the morphological and vegetation
changes in the river system caused crucial
increase of water stages in the system.
The maximum observed flows should are
close to the maximum dikes elevations
in the system. It is important to indicate
that further changes in this area may
cause the dike break by overtopping.
The increase of water stages related to
specific maximum flows is also presented
in Figures 5 and 6. In this graphs the
estimated changes of water stages related
to the flows Q10%, Q1% and Max Q are
plotted. The squares, circles and triangles
represent values estimated for particular
hydraulic conditions. Horizontal axis
describes the time. The dotted lines
and empty figures reflects the changes
of water stages without increase of
vegetation cover. It is clearly visible that
water stages are raising. The vegetation
impact causes dramatic intensification of
these processes.
123.5
water stage
[m a.s.l.]
51
water stage
[m a.s.l.]
123.0
122.7
122.4
Max Q
Q1%
122.1
Q10%
year
121.8
1975
1980
1985
1990
1995
2000
2005
FIGURE 6. Changes of the water stages related to
characteristic flows in cross-section km 503+140
124
water surface
elevation
[m a.s.l.]
123.2
123
122.9
effect of impacts,
bottom 2004
Max Q
122.6
122
Q1%
122.3
Q10%
floodplain
minimum
121
year
122.0
1975
1980
1985
1990
1995
2000
free outflow,
bottom 1975
2005
FIGURE 5. Changes of the water stages related to
characteristic flows in cross-section km 503+560
In the Figures presenting relative
frequency functions (Figs. 7 and 9) the
estimated initial stage of hydrologic
regime is compared with the estimated
current conditions. The initial stage
is represented by Pearson type III
probability density function fitted to the
results of simulation. In these figures
the minimum floodplain elevations are
denoted as dashed line. The Figures 8
and 10 shows the gradual changes
of cumulative frequency functions
120
0.00
0.02
0.04
0.06
relative
frequency
0.08
0.10
FIGURE 7. The relative frequency function corresponding to the initial and current state of the
hydrologi-cal regime in cross section located at
km 503+560
related to analyzed stages of the hydromorphological conditions in the river.
Presented results shows two common
trends in cross-sections located between
km 503 and 501. The most frequent
stages become higher and their relative
frequency is greater. This means greater
frequency of floodplain inundation, what
favors the riparian vegetation growth.
52
124
123
water surface
elevation
[m a.s.l.]
123
water surface
elevation
[m a.s.l.]
bottom 2004
+vegetation
bottom 2004
+vegetation
122
bottom 2004
bottom 2004
bottom 1997
122
bottom 1997
121
free outflow, bottom 1975
121
free outflow, bottom 1975
cumulative
frequency
bottom 1975
120
0.20
0.40
0.60
0.80
1.00
FIGURE 8. The cumulative frequency function
corresponding to the analyzed states of the hydrological regime in cross section located at km
503+560
123
120
0.00
0.00
water surface
elevation
[m a.s.l.]
effect of impacts,
bottom 2004
122
121
floodplain
minimum
free outflow,
bottom 1975
relative
frequency
120
0.00
0.04
0.08
0.12
FIGURE 9. The relative frequency function corresponding to the initial and current state of the
hydrologi-cal regime in cross section located at
km 503+14
CONCLUSIONS
The obtained results shows that potential
hydrologic regime in the analyzed area
is significantly changed. The impact
of each single process as well as the
cumulated impact of all processes have
cumulative
frequency
bottom 1975
0.20
0.40
0.60
0.80
1.00
FIGURE 10. The cumulative frequency function
corresponding to the analyzed states of the hydrological regime in cross section located at km
503+140
one common trend. The higher water
stages become more frequent and their
frequency is greater. This is well visible
in Figures 7–10. The changes are crucial
even in the cross-section located in
the longest distance from reservoir
(km 503+560). The areas located along
river reach were not flooded frequently
there (Fig. 7). Now, the inundation is
almost permanent.
The risk related to the dike break by
overtopping is raising, too. This aspect
of the hydrologic regime changes is
clearly visible in Figures 3–6. The
channel capacity was huge enough to
pass flows as Q1% and Max Q in the past.
Now, the changes of channel geometry
and roughness cause huge resistance and
increase of water stages. It is quite easy
to predict, that further changes will result
in natural catastrophe and damages.
The presented results shows that the
probability of such an event dramatically
increases.
ACKNOWLEDGEMENT
This work was supported by Polish
Committee for Scientific Research under
grant “Analysis of morphodynamic and
vegetation impacts on the hydraulic
conditions in backwater part of lowland
reservoirs”, project no. 2 P06S 034 30.
REFERENCES
ABBOTT M.B. 1979: Computational
Hydraulics, Elements of the Theory of
Free Surface Flows. London: Pitman.
ARMANINI A., RIGHETI M. & GRISENTI
P. 2005: Direct measurement of vegetation
resistance in prototype scale. Journal of
Hydraulic Engineering, 43 (5): 481–487.
BARKAU R.L. 1982. Simulation of the
July 1981 flood along Salt River, Report
for CE695BV, Special Problems in
Hydraulics. Fort Collins, CO.: Department
of Civil Engineering, Colorado State
University.
BRUNNER G.W. 2002: HEC-RAS,
River Analysis System hydraulic
reference manual, computer program
documentation. Davis, CA.: US Army
Corps of Engineers, Hydrologic
Engineering Center.
CHOW V.T., Maidment D.R. & Mays L.W.
1988: Applied Hydrology. McGraw-Hill
International Editions, Civil Engineering
Series, McGraw-Hill Book Company.
CUNGE J.A., HOLLY F.M.Jr & VERVEJ A.
1980: Practical aspects of computational
river hydraulics, Pitman Advanced
Publishing Program.
KEELAND B.D., CONNER W.H. &
SHARITZ R.R. 1997: A comparison
of wetland tree growth response to
hydrologic regime in Louisiana and
South Carolina. Forest Ecology and
Management, 90: 237–250.
KING S.L., ALLEN J.A. & McCOY J.W.
1998: Long-term effects of a lock and
dam and greentree reservoir management
53
on a bottomland hardwood forest. Forest
Ecology and Management, 112: 213–226.
KITE G. 2001: Modelling the Mekong:
hydrological simulation for environmental
impact studies. Journal of Hydrology,
253: 1–13.
KLOPSTRA D., BARNEVELD H.J.,
NOORTWIJK J.M. von & VELZEN E.H.
von 1997: Analytical model for hydraulic
roughness of submerged vegetation.
The 27th Congress of the International
Association for Hydraulic Research,
Proceedings of Theme A, Managing
Water: Coping with Scarcity and
Abundance, San Francisco, California.
LIGGETT M.B. & CUNGE J.A. 1975:
Numerical methods of solution of the
unsteady flow equations, in Mahmood
and Yevjevich (1975).
MAGILLIGAN F.J. & NISLOW K.H. 2005:
Changes in hydrologic regime by dams.
Geomorphology, 71: 61–78.
MAHMOOD K. & YEVJEVICH V. (eds.)
1975: Unsteady flow in open channels.
Fort Collins, Colorado: Water Resources
Publications.
MAINGI J.K. & MARSH S.E. 2002:
Quantifying hydrologic impacts following
dam construction along the Tana River.
Kenya. Journal of Arid Environments,
50: 53–79.
MATAN J. 1975: Jeziorsko reservoir in the
Warta river, Hydro-engineering and land
improvements structures in the reservoir
backwater. Poznań: Hydroprojekt (in
Polish).
MUMBA M. & THOMPSON J.R. 2005:
Hydrological and ecological impacts
of dams on the Kafue Flats floodplain
system, southern Zambia. Physics and
Chemistry of the Earth, 30: 442–447.
NILSSON C. & SVEDMARK M. 2002: Basic
principles and ecological consequences
of changing water regimes: riparian
plant
communities.
Environmental
Management, 30 (4): 468–480.
OZGA-ZIELIŃSKA M., BRZEZIŃSKI
J. 1994: Applied hydrology. Scientific
Publishing PWN, Warsaw (in Polish).
54
PINTER N. & HEINE R.A. 2005:
Hydrodynamic and morphodynamic
response
to
river
engineering
documented by fixed-discharge analysis,
Lower Missouri River, USA. Journal of
Hydrology, 302: 70–91.
STROMBERG J.C. 2001: Restoration of
riparian vegetation in the south-western
United States: importance of flow
regimes and fluvial dynamism. Journal of
Arid Environments, 49: 17–34.
SURTEVANT B.R. 1998: A model of wetland
vegetation dynamics in simulated beaver
impoundments, Ecological Modeling,
112: 195–225.
WELLMEYER J.L., SLATTERY M.C.
& PHILLIPS J.D. 2005: Quantifying
downstream impacts of impoundment on
flow regime and channel planform, lower
Trinity River, Texas. Geomorphology,
69: 1–13.
WICHER J. 2004: Sediment accumulation
in lowland reservoirs. Ph.D. thesis in the
Agricultural University of Poznan (in
Polish).
WICHER-DYSARZ J., PRZEDWOJSKI B.
2005: Modeling of sediment accumulation in the inlet part of Jeziorsko reservoir, Annual Reviews of Agricultural
University of Poznan, 26: 483–493.
Streszczenie: Ocena zmian warunków hydrologicznych wywołanych działaniem zapory Jeziorsko oraz procesami morfodynamicznymi
w rzece Warcie. W artykule zostały opisane procesy morfodynamiczne zachodzące na wybranym
odcinku rzeki Warty. Zjawiska te zostały wywołane działaniem zapory Jeziorsko wybudowanej
w 1986 roku. Głównym celem badań jest oszacowanie wpływu tych procesów na aktualne warunki hydrologiczne panujące w rzece. Pierwsze
napełnienie zbiornika do poziomu maksymalnego
piętrzenia odbyło się w 1991 roku. W 1995 roku
została uruchomiona elektrownia wodna „Jeziorsko”. W wyniku przerwania ciągłości transportu
rumowiska, w górnej części zbiornika Jeziorsko
nastąpiło akumulacja transportowanych przez
rzekę cząstek rumowiska. W efekcie osadzania
się rumowiska w górnej części zbiornika, zaczęła
intensywnie rozwijać się roślinność oraz ptactwo
wodne, które znalazło idealne warunki do rozwoju. Czynniki te spowodowały spiętrzenie wody
w tej części zbiornika. Dane historyczne oraz model hydrodynamiczny zostały wykorzystane, aby
oszacować zmiany warunków hydrologicznych.
Zestaw wykorzystanych danych zawiera: dane
z pomiarów terenowych geometrii i morfologii
koryta oraz pomiary stanów i przepływów z wybranych wodowskazów. Model hydrodynamiczny
został wykorzystany do odtworzenia warunków
przepływu w wybranych fazach zmian systemu
wodnego. Analiza statystyczna wyników symulacji umożliwiła ocenę zmian charakterystyk hydrologicznych na wybranym odcinku rzeki Warty.
Authors’ address:
Tomasz Dysarz, Joanna Wicher-Dysarz
Akademia Rolnicza w Poznaniu
ul. Wojska Polskiego 73a
60-625 Poznań
Characteristics of the particulate matter PM10 concentration field
and an attempt to determine the sources of air pollution in the living
district of Ursynów
GRZEGORZ MAJEWSKI, WIESŁAWA PRZEWOŹNICZUK
Warsaw Agricultural University – SGGW
Department of Water Engineering and Environmental Restoration
Abstract: Characteristics of the particulate
matter PM10 concentration field and an attempt
to determine the sources of air pollution in the
living district of Ursynów. The paper presents
characteristics of the imission field of PM10
particulate matter with the use of basic statistical
data such as mean values, variability ranges,
numbers of exceedences of the allowable
concentration, frequency distributions of
concentration occurrence in individual periods,
and time courses. Mean values of basic
meteorological elements were calculated, circular
graphs for percentiles of dust concentration were
prepared, and percentiles of pollution plume rates
were made in order to analyse meteorological
conditions of dispersal and spreading out of air
pollution. A synthetic WZ indicator of pollution
has also been calculated that encompasses the
joint impact of various meteorological elements
on pollution concentration levels.
Key words: particulate matter PM10, pollution
plume rate, meteorological conditions.
INTRODUCTION
First, identification of areas where
allowable standards of pollution
concentration are exceeded, and second,
identification of pollution sources
being the threat to the environment
are basic tasks for the atmospheric
air quality monitoring system. The
knowledge on the concentration field
and emission sources is a basis for
setting correction programmes necessary
for the improvement of environmental
conditions.
The order by Minister of Environment
as of 5 July 2002 (DzU, No 115, pos.
1003) settles in detail requirements
which the air protection programs should
answer to. It is said inside that act,
among others, that apart of assessing the
imission stated on a given area, the inflow
of pollution on that area should also be
estimated. Discrimination between the
emission influence and the impact of
meteorological conditions on the rate
of pollution concentration registered
(Walczewski et al. 2000) is extremely
important for proper execution of all
tasks of environment protection services,
because the increase or the decrease of
pollution concentration values depend on
both emission and dispersion conditions.
The meteorological characteristics of the
measurement period makes possible to
evaluate properly the imission situation.
Therefore, basic meteorological data
should be taken into account at the
interpretation of imission measurement
results, and a detailed analysis of the
wind rose from the measurement period
would allow conclude on whether the
measured concentrations came either
from local emission sources or are rather
56
G. Majewski, W. Przewoźniczuk
caused by an inflow of pollution from
outside the measuring network.
The living quarter of Ursynow is
one of the three districts of Warsaw,
the Polish capital, for which the air
protection correctional programs were
made. That program was prepared
because the allowable level of PM10
particulate matter had been exceeded
(Order No 62).
The particulate matter is a commonly
occurring
pollution
matter,
the
concentrations of which maintain in
urban agglomerations at a rather high
level exceeding frequently allowable
values. Miniparticles dispersed in the
air are considered to be one of more
essential potential threats to human
health as related to air pollution (i.a.
Juda & Chróściel 1974, Stern 1994,
Dockery 1996, Quarg 1996, Niecko
et al. 1998, Warych 1999, Jabłońska
2003). It is assessed that about one third
of the population in Poland exposed to
inhalation of dust pollution is bound to
chronic diseases of the breathing system,
that leads to circulatory system and
cancer diseases (Warych 1999).
Ursynów is a district of Warsaw,
capital of Poland, situated almost entirely
on the Warsaw Lowland, elevated 20–30
m up of the level of water in the Vistula
River, and it encompasses the southern
regions of the Warsaw city. It covers the
area of 44.6 km2 constituting 8.6% of
the Warsaw city total area, being at the
third place as for the size within the city.
Three nature reserves occur within the
quarter mentioned; they are: the Natolin
Wood, the Ursynów Slope – being the
main topographic element shaping the
local climate of the region mentioned,
and the Kabaty Wood called the Stefan
Starzyński Wood – a forest tract over
900 ha in size lying within the Warsaw
region; the latter wood is the most
essential natural element of the southern
aeration system for the capital.
The industrial plants and structures
equipped in their own boiler-rooms
influencing the air quality on the area
of the Ursynów quarter are as follows:
Oncology Centre, Warsaw Plant of
Aggregate Exploitation, Midas Polska
Ltd, Volvo Auto Polska, Underground
Standstill Station, PBM Południe Co,
Radio Ceramics Plant, Stolbud Ltd,
Geant Hypermarket, Unikon-Beton,
Tesco Hypermarket, and others. The
streets with the greatest intensity of car
traffic in this quarter are: Pulawska,
KEN Avenue, Rodowicza Anody,
Rosoła, Płaskowickiej, Ciszewskiego,
Roentgena, and Poleczki street. The areas
with the occurrence of low communalliving emission are as follows: ‘Green
Ursynów’ along the Puławska Avenue,
borders of the Kabacki Wood from the
Zolna street up to the city boundary,
the western side of the Puławska
Avenue, the area encompassing the
streets of Krasnopolska, Baletowa,
Karczunkowska, Sarabandy, Farbiarska
and Gawota, the area along the Ursynów
Slope, and finally the area along the
Prawdziwka street.
MATERIAL AND METHODS
An attempt has been undertaken
in this paper to assess the imission
situation in the Ursynow quarter, in the
MzWarszUrsynow station area, as well as
an attempt to point out the reasons for the
existing state. The data for calculations
Characteristics of the particulate matter PM10 concentration...
were taken from the automatic station
of atmospheric air monitoring called
the MzWarszUrsynow station (φN =
= 52o09’39’’, λE=21o02’03’’, 102 m a.s.l.)
working within the regional network of
the Masovia Province (WIOS database).
Measurements of SO2, NO, NO2, NOx,
CO, and PM10 concentrations were made
at that station, as well as measurements
of basic meteorological elements.
The
calculations
made
from
measurement data as averages in onehour intervals concerned the period
Oct. 2003 – Sept. 2005 that was divided
into winter half-years (the period from
October to March, covering heating
57
seasons) and summer half-years (the
period from April to September) .
The MzWarszUrsynów measuring
station under discussion is representative
for the general urban ambience and it
characterises well imission of pollution
dust in the area of living quarters exposed
to the impact of traffic and communal
and industrial emissions. Figure 1
illustrates the location of the station.
The concentration of PM10 particulate
matter is measured there in continuously
with MLU TEOM1400 analyser using
the oscillation microweight method. The
principle of the TEOM function consists
in measuring the frequency of vibration
FIGURE 1. The Air Pollution Monitoring Stations in Ursynów
58
of a conical element on which the filter
is fixed. The increase of the filter mass
due to deposition of dust particles on it
causes the change in the frequency of
measured vibration. This method ensures
the detectability threshold below 0.06
[μg·m–3] for mean one-hour values at
the throughflow 3 [l⋅min–1]. The device
has got the certificate of conformity with
the PM10 measurement relay method
according to the European standard
(EN12341).
The following calculations were the
basis for characterising the concentration
field of PM10 particulate matter:
1) mean concentration of dust in
calculation periods,
2) number of exceedences of the
allowable concentration,
3) variability ranges of concentration
values in individual calculation
periods,
4) concentration histograms based
on the Air Quality Scale rates
recommended by WIOS,
5) percentiles 25, 50, 70, and 98 of
concentration values,
6) mean annual 24-hour courses of dust
concentration values.
The
data
registered
at
MzWarszUrsynów and MzWarszSGGW
stations were used for evaluation of
meteorological conditions of pollution
dispersal. Both those stations laid 800
m apart from each other (correlation
coefficient
between
temperatures
measured on both stations is 0,99)
are equipped in the automatic outfit
registering basic meteorological elements
in 10-minute periods. Measurements
of air temperature, wind speed, and air
humidity at several levels are made at
the MzWarszSGGW station. The air
temperature values from heights 5 cm and
22 m were used to calculate gradients.
The analysis of meteorological
conditions of pollution dispersal and
spreading out covered the following
points:
1. Setting mean monthly values of basic
meteorological elements (Table 1),
2. Comparing the course of air
temperature in calculation periods
with the mean long term course,
circular
graphs
for
3. Making
percentiles of particulate matter
concentration as related to air inflow
directions,
circular
graphs
for
4. Making
percentiles of pollution plume rate in
order to get additional information
on location of emission sources,
synthetic
WZ
5. Calculating
indicator of pollution for assessing
meteorological
conditions
of
pollution dispersal.
The vicinity of the station was divided
into equal sectors with the central α
angle constituting 1/16 part of the full
angle in order to prepare circular graphs
for percentiles of particulate matter
concentration. The percentiles of the
order p = 25, 50, 70, 98 were calculated
basing on the sets of dust concentration
values occurring at the air inflow from
the direction related to the given sector
α. Circular graphs were made for
percentiles calculated for all sectors
of wind directions; they are graphical
presentations of concentration value
distributions at various wind directions.
Pollution flows were analysed for to
get more information on directions of
dust inflows over the area under study
(Klis, Matejczyk 2002). The pollution
flow is a vector. The scalar size of this
Calendar year
2005
2004
2003
Mean monthly cocentration of
PM10 [μg·m-3]
34,7
41,9
37,1
42,5
33,3
40,5
42,0
27,5
28,7
28,5
34,9
38,7
38,4
29,1
41,2
29,7
43,5
36,3
46,7
28,1
28,5
29,0
34,6
47,3
Month
X
XI
XII
I
II
III
IV
V
VI
VII
VIII
IX
X
XI
XII
I
II
III
IV
V
VI
VII
VIII
IX
Mean air temperature [0C]
5,5
5,1
1,0
-4,9
0,0
3,7
8,9
12,2
16,0
18,1
19,5
14,0
10,2
3,4
1,7
0,8
-3,1
0,2
9,5
14,0
16,4
20,8
18,2
16,2
Mean wind speed [m·s-1]
1,5
1,4
1,9
1,7
1,8
1,6
1,2
1,1
1,0
1,1
1,2
1,3
1,5
1,7
1,6
1,9
1,7
1,6
1,3
1,2
1,1
1,0
1,0
0,8
Mean air humidity [%]
84,3
90,5
88,1
88,5
83,4
78,4
63,7
69,8
66,2
70,3
67,6
71,0
81,5
89,1
90,5
84,5
84,8
76,4
59,5
66,6
64,6
58,8
62,9
62,9
Precipitation [mm]
67,2
20,7
41,7
30,9
34,6
40,9
52,2
54,9
42,3
61,3
53,7
20,1
39,7
62,1
11
33,1
27,2
40
18,8
56
45,3
4,7
38,8
33,6
181
180
182
181
6
7
7
10
7
9
10
1
7
7
2
7
4
7
6
12
4
5
14
9,0
7,7
33
35
9,9
5,2
Min. S24 [μg·m-3]
18
46
Mean concentration PM10 [μg·m-3]
Date of max.
Max. S24 [μg·m-3]
98,0 03.04 35,7
119,2 11.12 36,3
75,2 16.04 33,3
156,4 29.01 38,3
Standard deviation [μg·m-3]
17,2
20,7
12,3
22,7
The 75th percentile [μg·m-3]
Percentile of 1-hour
23,3 30,6 44,5 79,9 19,0 29,5 46,4 102,5
21,8 31,5 46,5 100,5 18,9 29,8 46,7 116,2
24,0 31,2 41,7 62,0 19,5 29,1 42,4 93,9
21,2 35,6 50,5 97,7 18,2 32,5 52,0 114,8
Percentile of 24-hour
The number of exceedance of D24
TABLE 1. The selected statistical characteristics of PM10 concentration and meteorological elements from October 2003 to September 2005, station
MzWarszUrsynów
Winter, summer half-year period
winter half-year
period
winter half-year
period
summer half-year
period
The number of concentrationS24
summer half-year
period
60
vector is equal to the amount of pollution
flowing in a time unit through the area
unit perpendicular to the flow direction.
This is the flow intensity [μg ·m–2·s–1],
being the measure of the inflow or outflow
of a substance over an area unit situated
within a territory (Stull 1995). Circular
graphs were made for percentiles of the
order p = 25, 50, 70, 98 after calculating
the pollution flow intensities.
A synthetic WZ indicator of pollution
was calculated for discriminating
between the impact of the emission and
the influence of atmospheric conditions
on the size of the imission registered.
This indicator encompasses all impacts
of various meteorological elements
influencing the increase or decrease of
air pollution during winter (Walczewski
1997). It is expressed in the formula:
m above the ground level at the time 12
UTC (10-minute mean), C(IN) = 0 when
there is a lack of lower inversion layer at
12 UTC, and 2 when the layer is present.
The data necessary for verifying the
existence of the inversion layer were got
from measurements of air temperature
at two levels on the mast installed in the
MzWarszSGGW station.
D(d) = 0 for d ≥ 4 [mm], 1 for 0.8 < d ≤4
[mm], 2 for 0 < d ≤ 0.8[mm],
E(p) = 1.5 when there is the anticyclone
on a given day, and 1 in the rest of
cases.
The minimum WZ value may amount
to 0, and the maximum is 16.5. The WZ
index was used with success for the first
time by Walczewski (1997) in Kraków
for identification of the changes in the
level of air pollution with SO2.
WZ = [A(T) + B(v) + C(IN) + D(d)] · E(p)
where:
A(T) – component describing the
temperature impact,
B(v) – component describing the wind
speed impact,
C(IN) – component describing the
inversion layer impact,
D(d) – component describing the
precipitation impact,
E(p) – coefficient taking into account the
atmospheric air pressure.
The individual articles for the WZ
indicator as adopted in the paper were
defined in the following way:
A(T) = 0 for T > 0 [°C], 1 for –5 < T ≤
[0°C], 2 for –10 < T ≤ –5 [°C], 3 for –15
< T ≤ –10 [°C], 4 for T < –15 [°C], where
T is the mean 24-hour temperature of air,
B(v) = 0 for v > 2[m·s–1], 2 for 1< v ≤2
[m·s–1], 3 for 0 < v ≤1 0[m·s–1], where
v is the wind speed at the height of 22
RESULTS
The series of 1-hour concentration
measurements of PM10 particulate matter
gained in the period under study, i.e.
1.09.2003–30.10.2005 contained 94%
of correct results with variability range
from 0.8 to 241.8 [μg·m–3]. One can
notice the occurrence of a clear cycle of
daily dust concentration variability, with
two local peaks in morning and evening
time – around 9h00 and after21h00, and
the hours of their occurrence vary during
the year, on the figure 2 presenting the
mean hourly concentrations of PM10
particulate within 24 hours. In the winter
season the concentration maximums
occur later in the morning and earlier
in the evening as related to the summer
season. A clear drop of PM10 particulate
matter concentration is observed during
w arm season
61
cold season
60
50
[μg•m-³]
40
30
20
10
23:00
22:00
21:00
20:00
19:00
18:00
17:00
16:00
15:00
14:00
13:00
12:00
11:00
10:00
09:00
08:00
07:00
06:00
05:00
04:00
03:00
02:00
01:00
00:00
0
FIGURE 2. Mean 1-hour concentrations of PM10 in the sequence of 24-hours in the station MzWarszUrsynów, 2004
the day around 13 h 00 and in the night
around 5 h 00. The occurrence of these
relationships results from not only the
daily course of pollution emission to the
atmosphere (e.g. traffic peaks) but also
24 hour cycle of changes in the height
of mixing layer and further on from
the development and disappearing of
convection processes, and general and
local circulation.
Daily concentrations of PM10
particulate matter in the period under
study (Tab. 1) acquired the values from
the interval 5.2 – 156.4 [μg·m–3], that
is from 10.4 to 312.8% respectively
of mean daily allowable value (D24
= 50 [μg·m–3]) (Order by Minister of
Environment of 6 June 2002). This value
has been exceeded in 132 cases within
the all measuring period, including 60%
exceedences in heating seasons.
The mean daily concentrations of
dust acquired most frequently the values
from 25 to 50 [μg·m–3] as it is shown
in the frequency histogram of daily
concentration occurrence (Fig. 3), based
on intervals of the Air Quality Scale
(WIOS) in both winter and summer
seasons. The highest concentration
values, exceeding D24, occurred most
frequently in winter months, this being
probably linked with the dropping air
temperature and with more intensive work
of heating facilities. In summer seasons,
the values from the intervals 50–75 and
75–100 [μg·m–3] appeared as well, but the
frequency of their occurrence was lesser
than in winter seasons, but the values
over 100 [μg·m–3] did not appear at all.
In the warm season of year, exceedences
of mean daily allowable value occurred
most frequently in April and September
(Tab. 1), but in the months V–VIII they
appeared only sporadically. No one case
of exceedence of the allowable standard
occurred in the periods: V, VI, VIII of
2004, and VII–VIII of 2005.
The phenomenon of the occurrence
of a considerable number of exceedences
in April and September requires a deeper
analysis. In the years under study, the
reasons of such a state can be seen in low
mean speed of wind, low mean humidity
of air, low sums of precipitation, and
low diversity in the inflows of air masses
(source: Daily Synoptic Bulletin of the
IMGW 2003–2005).
62
50
results [%]
warm half-year
56,1%
60
cold half-year
43,1%
35,1
40
29,3%
30
20
13,0%
16,6%
10
1,7%
3,3%
1,9%
0
0–25
25–50
50–75
75–100
>100
–3
[μg•m ]
FIGURE 3. Frequency distributions in percents of 24-hours particulate matter concentration at MzWarszUrsynów, X 2003–IX 2005
The mean monthly values of PM10
concentration were from 27.5 [μg·m–3]
(in May 2004) to 47.4 [μg·m–3] (in
September 2005), while the values
averaged for winter and summer year
halves differed only very slightly from
each other, and they were 38.3 [μg·m–3]
and 36.3 [μg·m–3] in winter year halves
and a bit less – 33.3 [μg·m–3] and
35.7 [μg·m–3] in summer year halves
respectively.
Circular graphs of concentration
percentiles related to individual sectors
of wind rose were made and analysed
in order to identify the directions of
inflows of particulate matter (Fig. 4a).
This figure shows the difference in the
distribution of percentiles in summer
and winter seasons over individual
sectors. They were more even in summer
seasons, and this fact could evidence the
impact of pollution sources located at
various directions around the measuring
point. The emission from traffic sources
prevailed in those periods because a
thick network of streets with a high
intensity of car traffic occurred around
the measuring point.
In the summer 2004, when no
overstandard values were observed, no
dominant direction of pollution inflow
was seen. It results from the analysis
of pollution plumes that the sources of
emission intensities being the greatest in
the period mentioned existed from the
measuring station toward the South in
the sectors 67–270oN (Fig. 4b).
A bit otherwise was in the summer
2005. It is seen on Fig. 4a that the
highest concentrations occurred at the air
inflow from NE direction (the value of
percentile of 98% level in the NE sector
was by 20% higher than in the remaining
sectors), and this can suppose an impact
of a large point source of emission. The
analysis of pollution plumes has shown
that the intensity of the emission was
the greatest in the sectors 110–180oN,
and moreover in the sectors NE, N,
WNW, where the following industrial
plants were located: EC Siekierki
Heating
plant, Urban Construction
Percentiles of PM10 pollution plume rate - 2003/2004
Perecentiles concentrations of PM10 - 2003/2004
N
NNW
a)
200
NNE
[μg•m ]
150
NW
N
NNW
b)
-3
63
NE
250
[μg•m -2 •s-1]
200
NW
NNE
NE
150
100
WNW
ENE
WNW
100
50
W
ENE
50
E
0
W
E
0
50%
50%
70%
70%
WSW
ESE
WSW
ESE
98%
98%
SW
SW
SE
SSW
SE
SSW
SSE
SSE
S
S
Perecentiles concentrations of PM10 - 2004/2005
N
NNW
200
-3
[μg•m ]
150
NW
Percentiles of PM10 pollution plume rate - 2004/2005
N
NNW
NNE
[μg•m-2 •s -1 ]
200
NW
NE
250
NNE
NE
150
100
WNW
ENE
WNW
100
50
W
ENE
50
E
0
W
E
0
50%
70%
50%
70%
WSW
ESE
WSW
ESE
98%
98%
SW
SW
SE
SSW
SE
SSW
SSE
SSE
S
S
Perecentiles concentrations of PM10 - 2004
Percentiles of PM10 pollution plume rate - 2004
N
N
NNW
200
[μg•m -3 ]
NNW
NNE
150
NW
NE
250
NNE
[μg•m-2 •s -1 ]
200
NW
NE
150
100
WNW
ENE
WNW
100
50
W
E
0
W
50%
70%
ESE
98%
SW
WSW
SE
SSW
ESE
98%
SW
S
SSE
S
Percentiles of PM10 pollution plume rate - 2005
N
N
NNW
200
[μg•m -3 ]
NNE
NE
WNW
NE
ENE
50
50
W
E
0
50%
E
0
-1
100
ENE
W
NNE
-2
[μg•m •s ]
150
100
WNW
250
200
NW
150
NW
SE
SSW
SSE
Perecentiles concentrations of PM10 - 2005
NNW
E
0
50%
70%
WSW
ENE
50
70%
50%
WSW
70%
WSW
ESE
ESE
98%
SW
SE
98%
SE
SW
SSW
SSW
SSE
SSE
S
S
FIGURE 4. Circular graphs of percentiles concentrations of PM10 (a) and percentiles of PM10 pollution plume rate (b) in the cold and warm half-years from October 2003 to September 2005 at MzWarszUrsynów station
64
Enterprise, Radio Ceramics Plant, and
the Construction Woodworking Plant. A
part of the pollution could also flow from
the city centre.
In the winter season 2003–2004
the distribution of concentrations in
individual sectors of the wind rose was
almost evenly, with a slight overweight
(under 10%) of concentration values
occurring in the sectors 180–203oN and
338–360oN. This shows an impact of
some sources on the imission field, the
sources that were dispersed in various
directions around the measuring station.
It results from the analysis of pollution
plumes that the greatest intensity of
emissions came from the sources located
in the sector 67–160oN and in the sector
315–360oN where point emission
sources existed as it was mentioned in
the preceding paragraph.
In the season 2004–2005 the highest
concentrations occurred at the air inflow
from northern direction (Fig. 4a). The
analysis of the plumes, similarly as
for the preceding season, revealed the
most intensive emission from the sector
67–180oN with local heating facilities
and areas with low communal and
living emissions (Fig. 1). Moreover, in
that season one could note an inflow of
greater amounts of dust from the sector
248–290oN. The diagram of plumes
showed the existence of sources with
great concentrations of emission from
that direction where pollution sources
worked intensively.
The calculation of the WZ pollution
index described in the preceding section
was the next stage of the analysis of
meteorological conditions important
for pollution dispersal. The calculations
were made for both heating seasons
mentioned. However, only the results
from January 2004 and 2005 were put
down in the paper due to its limited size.
It is seen on Figure 5 presenting
the monthly courses of mean daily
PM10 concentrations compared to the
analogous course of the WZ index that
the maximums of concentration values
cover the maximums of the index values.
Sometimes a slight retardation of the
concentration maximum as compared to
the highest index value can be noted, this
fact being well grounded because the
concentration resulted from unfavourable
meteorological conditions.
In January 2004 meteorological
conditions were less favourable for
pollution spreading out than in January
2005, because the mean WZI.2004
index amounts to 5.3 (max = 12), and
WZI.2005 is 3.5 (max 8.2), and the mean
concentrations were: 42.5 [μg·m–3]
and 29,7 [μg·m–3] respectively. A more
detailed analysis of those winter months
has shown that high air pressure areas
shaped the weather in January 2004.
Either frosty and dry polar continental air
flew from NE and E or arctic air from the
north. At that time an increased emission
of pollution from heating sources could
occur at very low air temperatures. Not
only the pressure pattern but also a small
amount of precipitation (7 days with
precipitation over 0.8 mm) favoured
increased concentrations of pollution.
However in January 2005 the mean
monthly air temperature was higher than
in 2004 and that pattern could cause a
bit lesser emission from heating sources.
The following impacts: occurrence of
low air pressure patterns and the fronts
accompanying them, the inflow of polar
marine air, the increase of wind speed,
180
65
14
160
12
140
10
8
100
80
6
WZ indicator
PM10 [μg·m-3]
120
60
4
D24
40
2
20
0
0
2
3
4
5
6
7
8
January 2004
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
PM10
D24 - 24-hour permissible concentration
WZ
70
14
60
12
D24
PM10 [μg·m-3]
50
10
40
8
30
6
20
4
10
2
0
WZ indicator
1
0
1
2
3
January 2005
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
PM10
WZ
FIGURE 5. Comparison of monthly variabilities of air pollution WZ indicator and PM10 concentration
(24-hours) at MzWarszUrsynów station
and the frequency of precipitation
had contributed to better pollution
spreading out and they could cause a
drop in registered values of pollution
concentration.
STATEMENTS AND
CONCLUSIONS
The high values of PM10 particulate
concentrations
occurring
in
the
Ursynow living quarter were caused, at
a considerable rate, by a low emission
coming from main communication
66
avenues and from dispersed sources
(local heating facilities, individual
domestic fire places) occurring in the
sectors 67–180oN and 245–360oN. Low
sources were of local importance but
their impact decided on the scale of the
threat from pollution. This threat results
just from the fact that the low sources are
located near human living areas; therefore
dust aerosol breathed in is a ‘fresh’ one
with a high share of miniparticles, and
at the same time it is very dangerous
for human health. This threat has been
confirmed by introductory results of the
quality analysis of PM10 found at the
MzWarszSGGW station that showed
the presence of heavy metals in the dust,
such as Ni, Pb, Cr, Zn, Fe, Cu (archival
data of the Meteorology and Climatology
Department of the Warsaw Agricultural
University).
Since the year 2003, i.e. from
starting the measurements of PM10
particulate matter concentration at the
MzWarszUrsynow station, a decreasing
tendency of this pollution is observed,
although the number of allowable
exceedences of mean daily values
maintains still at a high level (mean daily
allowable value of the particulate matter
concentration was exceeded in 132 cases
in the whole measuring period).
The
differences
in
seasonal
concentrations
were
caused
by
emissions of particulate matter from
burning out fuels in low sources being
greater in winter seasons at relatively
worse conditions of aeration and air
self-cleaning that occur in the cold
season. Meteorological conditions were
the factor deciding on the imission
registered in all winter months under
study. It was impossible to found for the
study period that a pollution emission
decrease caused by implementation of
the air quality improvement program
had been the reason for the betterment of
the imission situation in the area under
analysis. According to the opinion of
the authors, the decrease of particulate
matter amount registered in the period
2003–2005 was caused by the betterment
of meteorological conditions.
REFERENCES
DOCKERY D.W., CUNNINGHAM J.,
DAMOKOSH A.I., NEAS L.M.,
SPENGLER J.D., KOUTRAKIS P.,
WARE J.H., RAIZENNE M, SPEIZER
F.E. 1996: Health effects of acid aerosols
on North American children: respiratory
symptoms. Environ. Health Perspect.
104, 500–505.
EN12341 – Determination of the PM10
fraction of suspended particulate matter –
Reference method and field test procedure
to demonstrate reference equivalence of
measurement methods. November 1998.
JABŁOŃSKA M. 2003: Skład fazowy pyłów
atmosferycznych w wybranych miejscowościach Górnośląskiego Okręgu Przemysłowego. Wydawnictwo Uniwersytetu
Śląskiego. Katowice. Juda J., Chróściel
S. 1974: Ochrona powietrza atmosferycznego.
NIEĆKO J., NIEĆKO M., NIEĆKO U. 1998:
Charakterystyka pyłu zawieszonego
i jego wpływ na organizm ludzki. Air
protection in theory and application.
Section I, s. 29-49.
KLIŚ CZ., MATEJCZYK M. 2002: Ocena wpływu źródeł na jakość powietrza
w świetle ustawy „Prawo ochrony
środowiska”. Ochrona Powietrza i Problemy Odpadów, vol. 36, nr 3: 95–98.
Rozporządzenie Nr 62 Wojewody Mazowieckiego z dnia 8 grudnia 2003.
Rozporządzenie Ministra Środowiska z dnia
6 czerwca 2002 r. w sprawie dopuszczalnych poziomów niektórych substancji
w powietrzu oraz marginesów Tolerancji
dla dopuszczalnych poziomów niektórych substancji (DzU z 2002r. nr 87 poz.
798).
STERN A.C. et al. 1994: Fundamentals of air
pollution. Academic Press, San Diego.
STULL R.B. 1995: Meteorology Today For
Scientists and Engineers, West Publishing
Comp. New York.
Quarg, 1996: Airborne Particulate Matter in
the United Kingdom. Third Report of the
Quality.
WALCZEWSKI J. et al. 2000: Wykorzystanie danych meteorologicznych w monitoringu jakości powietrza. Biblioteka
Monitoringu Środowiska, Warszawa.
Walczewski J. 1997: Wskaźnik meteorologiczny określający prawdopodobieństwo
wzrostu zanieczyszczenia powietrza
w okresie zimowym. Wiad. IMiGW, t.
XX (XLI).
WARYCH J. 1999: Zanieczyszczenie powietrza cząstkami aerozolowymi i wynikające stąd problemy. Ochrona powietrza
i Problemy Odpadów. Vol. 33, nr 3, majczerwiec.
67
Streszczenie: Charakterystyka pola imisji oraz
próba określenia źródeł zanieczyszczenia powietrza pyłem zawieszonym PM10 na obszarze
dzielnicy mieszkaniowej Ursynów na podstawie danych pomiarowych z automatycznej stacji
monitoringu atmosfery. W pracy przedstawiono
charakterystykę pola imisji pyłu zawieszonego
PM10 za pomocą podstawowych statystyk, takich
jak wartości średnie, zakresy zmienności, liczby
przekroczeń dopuszczalnego stężenia, rozkłady
częstości występowania stężeń w poszczególnych
przedziałach, przebiegi czasowe. W celu przeanalizowania warunków meteorologicznych dyspersji i rozprzestrzeniania się zanieczyszczeń, obliczono średnie wartości podstawowych elementów
meteorologicznych, sporządzono wykresy kołowe
percentyli stężenia pyłu, oraz percentyli natężenia
strumieni zanieczyszczenia. Obliczono również
syntetyczny wskaźnik zanieczyszczenia WZ, ujmujący łączne oddziaływanie różnych elementów
meteorologicznych na poziom stężeń.
MS. received July 2006
Authors’ address:
Wydział Inżynierii i Kształtowania Środowiska,
SGGW,
02-787 Warszawa, ul. Nowoursynowska 159
Runoff volume and slope gradient relationship – laboratory
investigations
ANNA BARYŁA
Department of Environmental Improvement, Warsaw Agricultural University – SGGW
Abstract: Runoff volume and slope gradient
relationship – laboratory investigations. The
objective of the research was to analyze influence
of the slope gradient on the volume of surface
runoff. Runoff is the primary driving variable in
the water – induced erosion process. The overland
flow process is strongly affected by the slopes
because of the effective rainfall rate at the surface.
The laboratory physical model diameters were
120 cm long, 15 cm wide and 120 cm deep. The
experiment the measurements of water erosion
was done for different slopes: 5%, 10%, and
15%. The rainfall was artificial simulated and his
intensity for experimental separate sets. Based on
performed research it can be concluded that the
volume of the runoff depends on the gradient of
the slope. As the gradient increase the volume of
the runoff also increase.
Key words: rainfall simulation, physical model,
experiment.
INTRODUCTION
The runoff of rainfall water on the slope
depends on several factors such as:
decline, form, length and plant cover
of the slope, amount ant intensity of the
rainfall as well as the water properties of
the soil (Carson and Kirbky 1972, Słupik
1981, Morgan 1986, Józefaciuk 1996,
Kinnell 1997).
Among physical soil properties it is
the soil infiltration ability that influences
the runoff the most. The impact of
infiltration on the possibility of surface
runoff occurrence is revealed particularly
when the surface lay is hardly permeable.
The structure and the porosity of the soil
have a decisive impact on the amount of
infiltration (Józefaciuk A. and Józefaciuk
C. 1996). Basic causes of variation
of infiltration speed are changes of
retentive capacity in the ground, the type
of the gradient and the primary capacity
value, treated as moisture in a time given
(Słupik 1981, Schmidt 1997).
Theoretical base of infiltration was
formed by Horton (1993); according to
him the surface runoff forms on a slope
when the intensity of rainfall exceeds the
infiltration capacity of the soil. In view
of gradual filling of the soil pores with
the water, the infiltration capacity of the
soil diminishes as the rainfall lasts. When
the intensity of the rainfall exceeds the
infiltration process, a so called superinfiltration runoff occurs. However,
when the surface ground lay is totally
saturated with the water, a so called
saturated zone runoff is formed (Słupik
1981, Kinnell 1997, Józefaciuk A. and
Józefaciuk C. 1996, Rejman and Usowicz
1999). A particular type of it is a runoff
in the initial phase of intense rainfall
formed thanks to a dried-off and hardly
permeable shell or a blockage laying
on the surface against the infiltrating
water which is created by the air bound
in the soil. This type of runoff is called
a pre-infiltration runoff (Schmidt 1997,
70
A. Baryła
Kinnell 1997). In this case the surface
runoff lasts until the soil is softened
and the soil moisture is complemented.
This type of runoff occurs the most often
on heavy soils, thus short and of little
intensity rainfalls very often can entirely
constitute the runoff (Kinnell 1997). The
process of forming the surface runoff
depends in a big measure on, among
others, the degree of covering of the soil
surface. The practice has shown that it is
more intense on open soils.
FIGURE 1. Schema of laboratory experiment
In natural conditions because of
their big differentiation, it is extremely
difficult to estimate precisely the impact
of each of the factors mentioned above on
the course of the runoff and its amount.
The aim of the research was to determine
the impact of the initial soil moisture, the
decrease and the intensity of the rainfall
on the value of surface runoff. The results
of laboratory experiments are presented
in the paper.
Runoff volume and slope gradient relationship...
MATERIALS AND METHODS
The research was carried out on a ground
model with the following diameters: 136
cm of length, 120 cm of altitude, 15
cm of width (Fig. 1). The inside oh the
model was filled with clay sand firmly
poured layer by layer and alternately
compacted.
In order to enable the observation
of water flow in the soil the front wall
of the model was made of a transparent
polymethacrylate board. Inside the model
16 probes TDR were installed in order
to measure the soil moisture. 0.8 mm
of diameter each of them, the distance
between them was 5 mm. They were
connected to the TDR (Malicki 1996)
set ant to the computer. All the time
during the experiment (in ten-minutes
intervals) changes of soil moisture were
registered. The rainfall was simulated by
microsprinklers; the intensity of outflow
depended on the water pressure, which
enabled modeling of rainfalls of different
intensity.
The surface runoff water was collected
at the end of the model by a trough
having an outlet directly connected to
the measuring vessel. The measurements
of outflow and moisture were made in
the same intervals. 12 experiments were
carried out on the model with three slopes
of the terrain surface – 5, 10 and 15%,
TABLE 1. Physical properties of the soil used in
experiment
Depth [m]
Density
[g·cm–3]
Porous
[%]
Filtration
coefficient
[cm·s–1]
0.00–0.15
1.68
34.1
15.2*10–4
0.15–0.45
1.62
35.6
7.84*10–4
0.45–0.80
1.58
36.8
4.2*10–4
71
and different rainfall intensity: 41 mm
h–1, 46 mm h–1, 52 mm h–1 and 54 mm
h–1. The duration of every experiment
was 1 hour.
Surface runoff in the time from the
physical model was collected at the end by
trough, from of which one accompanied
to dishes measuring. Physical properties
of the soil are set together in the Table 1.
Lesser and lesser values of filtration
coefficient in subsequent layers may
evidence about more compacted lower
parts of the ground in the model.
RESULTS AND DISCUSSION
Knowledge and proper determination
of interdependencies between surface
runoffs and factors determining their
value, intensity and frequency of
occurrence is indispensable in the
modeling of the value of surface runoff
on a slope. On the Figure 2a are shown
the results of surface runoff for the
slope of 5%. The analyze of the results
obtained can lead to the conclusion that
at the rainfall intensity of 41 mm h–1 and
a slope of 5% the summary runoff was
0.8 mm h–1, instead at a slope of 15% the
summary runoff was 5.6 mm h–1. When
the rainfall intensity was increased by 5
mm h–1 (46 mm h–1) at a slope 5% the
summary surface runoff amounted to 1.2
mm h–1 at 15 mm h–1 – 11.8 mm h–1.
The time of starting-up the surface
runoff at the intensity of 41 mm h–1 was
22 minutes and was the longest time of
starting-up the surface runoff at a slope
10%. The shortest time was observed at
the intensity of rainfall of 54 mm h–1 and
then the runoff started 10 minutes after
the experiment had started.
A. Baryła
Surface runoff [mm]
12
a)
5%
41 mm/h
10
46 mm/h
8
52 mm/h
54 mm/h
6
4
2
0
0
Surface runoff [mm]
12
10
b)
20
30
40
50
60
30
40
50
60
30
40
50
60
10%
10
8
6
4
2
0
0
10
12
Surface runoff [mm]
c)
20
15%
10
8
6
4
2
0
0
10
20
Time [min]
FIGURE 2. Dynamics of runoff surface with rain
intensity and different slope (a – 5%, b – 10%,
c – 15%).
As at the 5 and 10% slopes, the highest
value of surface runoff during 1 hour was
obtained at the rainfall intensity of 54
mm and equaled to 10.86 mm; this value
was twice higher than surface runoff
obtained at the intensity of 41 mm (3.68
mm). At the intensity of rainfall equal 41
mm h–1 the surface runoff increased after
10 minutes, and at the intensity of 54 mm
h–1 – 2 minutes. The research also proved
that in the initial rainfall phase there is an
intensified runoff volume until a balance
is obtained between the amount and the
energy of runoff. The surface runoff
speed stabilisation depends mainly on
the rainfall intensity, which makes a very
thin layer on the soil surface. For that
reason the runoff value remains at the
similar level, although the surface layer
is almost fully saturated.
The performed research confirmed
that there is an important impact of the
land slopes on the course and value of
surface runoffs (Fig. 3). At a slope of
5% surface runoffs’ value was more
than five times less than at a slope of
15% at the same rainfall intensity. The
research performed by Szafrański and
others (1998) revealed that on grounds
at a slope 3–6% surface runoff value is
low and therefore the erosion danger
is very slight. A severe water erosion
danger may occur only on the lands at
a slope of more than 10%. None the less
during violent summer high-intensity
thunderstorms the phenomenon of linear
erosion can occur even at slopes of less
than 6% (Rejman 1999).
30
Coefficients of runoff C [%]
72
slope 5%
slope 10%
slope 15%
20
= 99
;R
6
6*P
R=9
646
*P;
3
,
0
434
0
=
3
C
,2
C=0
*P; R=91
03
90
66
C=0,0
10
0
0
10
20
30
40
Rainfall P [mm]
50
60
FIGURE 3. Relationship between rainfall intensity and coefficients of runoff
Runoff volume and slope gradient relationship...
The Figure 3 presents a relationship
between rainfall intensity and coefficients
of runoff for the results of the research.
The determination coefficient amounted to respectively: for a slope 5%–99,
for a slope 10%–96 and for 15% –91. The
research performed revealed that when
the rainfall intensity volume increases,
surface runoff increases too.
CONCLUSIONS
1. The research performed on a ground
model confirmed an important role
of land slopes played in a course
and value of surface runoffs. At a
slope equal 5% the runoffs’ value
was almost five times less than at
slopes equal 15%. It was ascertained
that there is a rectilinear dependency
between surface runoffs volume and
land slopes.
2. Surface runoffs volume described
by the runoff coefficient was 0,02 to
0,05 at a land slope 5% and 0,13 to
0,21 at a land slope equal 15%.
REFERENCES
CARSON M.A., KIRBKY M.J. 1972:
Hillslope form and process. University
Press, Cambridge.
JÓZEFACIUK A., JÓZEFACIUK C. 1996:
Mechanizm i wskazówki metodyczne
badania procesów erozji. Biblioteka
Monitoringu Środowiska w Warszawie.
KINNELL P.I.A. 1997: Runoff ratio as
a factor in the empirical modeling of
soil erosion by individual rainstorms.
Australian Journal of Soil Research No
25, p 1–6.
MALICKI M.A., PLAGGE R., ROTH
C.H. 1996: Improving the calibration of
dielectric TDR soil moisture determination
73
taking into account the solid soil. European
J. Soil Sci. 47, 57–366.
MORGAN R.P.C. 1986: Soil erosion and
conservation. Longman Scientific &
Technical, Essex, UK.
REJMAN J., USOWICZ B. 1999: Quantitative
description of water ad soil transport in
process of water erosion. Acta Agrophysica,
no 23, p. 143–148 (in Polish).
SCHMID B. 1997: Critical rainfall for overland
flow from an infiltrating plane surface.
Journal of Hydrology 193, 45–60.
SŁUPIK J. 1981: Rola stoku w kształtowaniu
odpływu w Karpatach Fliszowych. Prace
Geograficzne, 142, 1–94.
SZAFRAŃSKI C., FIEDLER M., STASIK
R. 1998: Rola zabiegów melioracyjnych
w ochronie przeciwerozyjnej gleb
terenów bogato urzeźbionych Bibliotheca
Fragmenta Agronomica tom 4B/98.
WISCHMEIER W.H., SMITH D.D. 1978:
Predicting rainfall erosion losses. USDA
Agric. Handb. 537. U.S. Gov. Print.
Office, Washington DC, p. 1–58.
VAN DIJK A.I.J.M., BRIJNZEEL L.A.
Modeling runoff and soil loss from bench
terraced hillslopes in the volcanic uplands
of West Java, Indonesia.
YU B., SOMBATPANIT S., ROSE C.W.,
CIESIOLKA, C.A.A. and COUGHLAN
K.J. 2000: Characteristics and modeling
of Runoff Hydrographs for Different
Tillage Treatments. Soil Science Society of
American Journal No 64, p. 1763–1770.
Streszczenie: Zależność pomiędzy spływem powierzchniowym a spadkiem terenu – doświadczenie laboratoryjne. Spływ powierzchniowy jest
procesem złożonym, a czynniki od których zależy
są zmienne zarówno w czasie, jak i w przestrzeni.
Poznanie i właściwe określenie zależności pomiędzy spływami powierzchniowymi a czynnikami
wpływającymi na ich wielkość, natężenie i częstotliwość występowania jest niezbędne m.in. dla
oceny istniejących modeli spływu powierzchniowego. Badania laboratoryjne wykonano na modelu gruntowym o wymiarach: długość 136 cm, wysokość 120 cm, szerokość 15 cm. Wnętrze modelu
wypełnione zostało piaskiem gliniastym mocnym
74
A. Baryła
zagęszczanym warstwowo. Badania dowiodły, że
w początkowej fazie opadu następuje zwiększenie
wielkości spływu do momentu ustalenia się stanu
równowagi pomiędzy intensywnością i energią
opadu a wielkością i energią spływu. Prędkość
ustabilizowania się spływu powierzchniowego
zależy głównie od intensywności opadu, który
powoduje tworzenie się na powierzchni gleby
cienkiej warstewki. Wartość odpływu utrzymuje się przez to na zbliżonym poziomie, mimo że
wierzchnia warstwa osiągnęła prawie pełny poziom nasycenia.
Author’s address:
Katedra Kształtowania Środowiska – SGGW
ul. Nowoursynowska 166
02-787 Warszawa
Influence of sprinkling irrigation and nitrogen fertilization on health
status of potato grown on a sandy soil
DARIUSZ PAŃKA1, ROMAN ROLBIECKI2, CZESŁAW RZEKANOWSKI2
1
Department
2
of Phytopathology, University of Technology and Agriculture, Bydgoszcz, Poland
Department of Land Reclamation and Agrometeorology, University of Technology and Agriculture,
Bydgoszcz, Poland
Abstract: Influence of sprinkling irrigation and
nitrogen fertilization on health status of potato
grown on a sandy soil. The field experiment was
carried out in the years 2001–2003 at Kruszyn
Krajeński near Bydgoszcz on a sandy soil
(Typic Hapludolls). The field experiment was
done using the split-plot method, in a dependent
system with two variable factors (sprinkler
irrigation and nitrogen fertilization) and three
replications. ‘Drop’ early potato cultivar was
taken into consideration. It is the most frequent
cultivated cultivar in Kujawy-Pomerania region.
The following levels of experimental factors were
used: water: O – without sprinkler irrigation and
W – sprinkler irrigation according to tensiometer,
nitrogen fertilization: N1 – 75 and N2 – 125 kg
N ha–1, respectively. Potato tubers were analyzed
at once after harvest from the point of view of
occurrence of common scab, black scurf, potatotuber dry-rot and potato-tuber soft-rot symptoms.
Estimation of degree of infestation was carried
out on 50 tubers taken at random from the each
plot. Nine-degree scale was used (0–8°), where:
0° means lack of infestation symptoms (sound
tubers) and 8° means over 50% of a tuber surface
with symptoms of common scab or black scurf.
In case of other diseases a percentage of tubers
with symptoms of infestation was estimated. On
the basis of the investigation results it was found
that sprinkler irrigation had a significant influence
on the occurrence of common scab symptoms.
Stronger infestation was noted on irrigated
treatments. A strong influence of the rainfall
course in the vegetation period of the potato on
the infestation of tubers by Streptomyces sp. was
observed. An influence of nitrogen fertilization on
the occurrence of common scab and black scurf
symptoms was not found. The cultivar tested was
characterized by a higher resistance to infestation
by Rhizoctonia solani than Streptomyces sp.
Key words: potato, sprinkling irrigation, fertilization,
nitrogen, Streptomyces sp., Rhizoctonia solani,
common scab, black scurf, sandy soil.
INTRODUCTION
High quality yield is at present the main
aim of the table potato production.
The quality of potato is affected most
strongly by such the factors like cultivar,
course of weather conditions during the
vegetation period, mineral fertilization,
soil cultivation and plant protection
against agrophages [3], [7], [9], [16].
Good quality yield can be produced when
optimal conditions for the potato growth
and development are created. In such the
conditions the pressure of agrophages is
limited to minimum.
Pathogens developing on tubers are
included to the most dangerous pathogens.
They often decreased the suitability of
potato for food processing. In addition
these pathogens decreased the quality
of seed-potatoes and they can cause
considerable storage losses [1], [6], [15].
Common scab, rhizoctoniose, tuber dryrot and tuber soft rot are included to the
most dangerous diseases in cultivation
of potato. They occur commonly. Their
76
D. Pańka, R. Rolbiecki, Cz. Rzekanowski
development are influenced by a number
of such the factors like for example
moisture, temperature or mineral
fertilization [8], [10], [11], [14]. In
Poland cultivation of potato is conducted
mostly on light soils characterized by a
low fertility and limited water-holding
capacity. Because of this, higher doses
of mineral fertilization and irrigation
of potato plantations should be used,
especially in the regions characterized
by periodical rainfall deficits. All of
this with connection of other factors
can influence the development of potato
pathogens.
The aim of the study was to determine
the influence of nitrogen fertilization and
sprinkler irrigation on the potato tuber
health of ‘Drop’ cultivar, grown on the
light soil.
MATERIAL AND METHODS
The field experiment was carried out
in the years 2001–2003 at Kruszyn
Krajeński near Bydgoszcz on a sandy soil
(Typic Hapludolls). The water reserve to
1 m depth of soil at field capacity was
87 mm and the available water 67 mm.
The field experiment was done using the
split-plot method, in a dependent system
with two variable factors (sprinkler
irrigation and nitrogen fertilization) and
three replications. ‘Drop’ early potato
cultivar was taken into consideration. It is
the most frequent cultivated in Kujawy-Pomerania region.
The following levels of experimental
factors were used:
– water: O – without sprinkler
irrigation and W – sprinkler irrigation
according to tensiometer,
nitrogen fertilization: N1 – 75 and
N2 – 125 kg N ha–1, respectively.
The plot area for harvest was 15 m2.
The mineral fertilization was applied
at following doses: 80 kg P2O5 ha–1
(superphosphate) and 140 kg K2O ha–1
(potash salt). Fertilization was used presowing in early spring.
Chemical protection against potato
late blight, colorado potato beetle was
conducted in particular years of the
study. Potato tubers were analyzed at
once after harvest from the point of view
of occurrence of common scab, black
scurf, potato late blight, potato-tuber dryrot and potato-tuber soft-rot symptoms.
Estimation of degree of infestation was
carried out on 50 tubers taken at random
from the each plot. 9-degree scale was
used (0–8°), where: 0° means lack of
infestation symptoms (sound tubers) and
8° means over 50% of a tuber surface
with symptoms of common scab or
black scurf. In case of other diseases
a percentage of tubers with symptoms
of infestation was estimated. Infection
degrees were transformed into infection
indexes (II) according to Townsend and
Heuberger formula [17]. Obtained data
were statistically analyzed using analysis
of variance. Mean values were verified
with Tukey’s test.
–
In all of the study years, symptoms
of infestation by Streptomyces sp. and
Rhizoctonia solani, occurred on potato
tubers most often. Other pathogens were
noted sporadically and they amounted
less than 1% tubers with disease
symptoms.
Influence of sprinkling irrigation and nitrogen...
Significant influence of sprinkler
irrigation on the occurrence of common
scab symptoms on tubers was observed
(Tab. 1). Tubers from irrigated plots
were significantly higher infected
by Streptomyces sp. every year. The
strongest symptoms of infestation
occurred in 2001, on plots fertilized with
the higher nitrogen dose. The lowest
infestation by this pathogen occurred in
exceptional dry year 2003 (Tab. 2). In this
year the infection index on irrigated plots
was lower than 24,5%. Similar results
obtained Sadowski et al. [10], [11].
They observed a decrease of common
scab symptoms occurrence on the very
light soils, in years characterized by
small amounts of rainfall. In such the
77
conditions, irrigation was favourable
to the development of Streptomyces sp.
An infestation increase of tubers by this
pathogen on irrigated plots, in years with
lower rainfall, noted also Gładysiak and
Czajka [5] as well as Borówczak and
Gładysiak [2].
In years with higher rainfall amounts
they observed a decrease of the common
scab occurrence. The first two years
of the study (2001 and 2002) were
characterized by higher rainfall amounts,
especially year 2001 (Tab. 2). In spite of
this, the infection increase of tubers by
Streptomyces sp. on irrigated treatments
was obtained. Reason of this phenomenon
can be a low rainfall in June in these
years. Amounts of rainfall in this month
TABLE 1. Effect of nitrogen fertilization and irrigation on cv. ‘Drop’ tubers infection with Streptomyces
sp. (Infection index in %). Kruszyn Krajeński 2001–2003
Years
Water variant
Non irrigated
2001
Sprinkler irrigation
Mean
Non irrigated
2002
Mean
Non irrigated
2003
Mean
Non irrigated
2001–2003
Mean
1
Nitrogen fertilization
125 kg·ha–1
75 kg·ha–1
18.7 a1
18.5 a
a
a
52.8 a
53.3 a
b
b
35.8 a
35.9 a
15.7 a
8.7 a
a
a
43.4 a
50.1 a
b
b
29.6 a
29.4 a
1.4 a
2.5 a
a
a
21.3 a
24.5 a
b
b
11.3 a
13.5 a
11.9 a
9.9 a
a
a
39.2 a
42.6 a
b
b
25.6 a
26.3 a
Mean
18.6 a
53.1 b
12.2 a
46.8 b
1.9 a
22.9 b
10.9 a
40.9 b
Mean values followed by the same letter in columns and rows are not significantly different at α = 0.05,
according to Tukey’s test.
78
TABLE 2. Rainfall conditions recorded over vegetation period
Period
Rainfall (mm)
April
May
June
July
August
Total
1951–2000
26
40
56
70
48
240
2001
45
30
49
106
27
257
2002
13
50
44
108
41
256
2003
18
18
30
106
18
190
2001–2003
25
33
41
107
29
234
were, like in 2003, lower than the manyyear average. According to Sawicka [13]
potato is the most susceptible to infection
in the period from the beginning of tuber
formation to the stage when diameters of
tubers achieve 1,5–2 cm. Small amounts
of rainfalls in this period are favourable to
the development of common scab. ‘Drop’
is the very early cultivar, characterized
by a short vegetation period, and
because of this its largest susceptibility
to infection occurred in the years,
exactly in the period with low rainfall.
Sprinkler irrigation was then favourable
to the pathogen development. According
to Szutkowska [14], the occurrence of
common scab is mostly affected by the
soil moisture and the soil temperature in
the period of the early growth of tubers.
The highest infection was observed when
the soil temperature was about 21°C,
and its moisture around 6%. Favourable
arrangement of these factors in the years
2001 and 2002 in the remaining months
of vegetation was probably the reason
of the stronger infection of tubers on
irrigated treatments as compared to the
year 2003.
Significant influence of the nitrogen
fertilization dose on the common scab
occurrence on tubers was not observed.
Czajka et al. [4] observed an increase
of common scab infection between
treatments with doses 0 and 80 kg N ha–1.
Doses which were increased above 80 kg
N ha–1 had no larger effect. In the own
experiments the occurrence of common
scab was also on a similar level in case
of both the doses tested.
Significant influence of sprinkling
irrigation on infection of tubers by
Rhizoctonia solani was observed
only in 2003 (Tab. 3). This year was
characterized by the lowest amount of
rainfall during the vegetation period
of potato. Sprinkler irrigation in
such the condition was favourable to
the development of the pathogen as
compared to control treatments (without
irrigation). According to Weber [16],
the high moisture of soil is favourable
to sclerotia formation on tubers. This
dependence was observed also in our
earlier investigations [8]. Similar results
were obtained also by Sadowski et al.
[12]. In the remaining years of the study,
with higher rainfall during the period of
the potato growth, no differences in black
scurf occurrence between irrigation and
control treatments were observed.
A significant effect of the nitrogen
fertilization dose on black scurf of
79
TABLE 3. Effect of nitrogen fertilization and irrigation on cv. ‘Drop’ tubers infection with Rhizoctonia
solani (Infection index in %). Kruszyn Krajeński 2001–2003
Years
Water variant
2001
2002
2001–2003
Mean
75 kg·ha–1
125 kg·ha–1
Non irrigated
21.81
24.2
23.0
20.1
20.8
20.5
Mean
21.0
22.5
Non irrigated
13.0
13.7
13.4
17.1
15.8
18.4
Mean
Mean
14.4
12.0 a
a
20.6 a
b
16.3 a
16.1
10.9 a
a
25.2 a
b
18.1 a
Non irrigated
2003
Nitrogen fertilization
11.5 a
22.9 b
Non irrigated
15.6
16.3
16.0
18.9
21.5
20.2
Mean
17.2
18.9
1
Mean values without letters or followed by the same letter in columns and rows are not significantly
different at α = 0.05, according to Tukey’s test.
potatoes was not observed, too. In
investigation of Czajka et al. [4] the
significantly highest level of black scurf
was noted on treatment with 80 kg
N ha–1, and the significantly lowest level
of black scurf of potatoes – on treatments
with 0 and 160 kg N ha–1, in the years
with average amount of rainfall. Higher
nitrogen fertilization decreased the
pathogen occurrence in warm and dry
years.
A strong influence of the rainfall
course in the vegetation period of the
potato on the infestation of tubers by
Streptomyces sp. was observed. An
influence of nitrogen fertilization on
the occurrence of common scab and
black scurf symptoms was not found.
The cultivar tested was characterized
by a higher resistance to infestation by
Rhizoctonia solani than Streptomyces
sp.
CONCLUSIONS
REFERENCES
On the basis of the investigation results
it was found that sprinkler irrigation had
a significant influence on the occurrence
of common scab symptoms. Stronger
infestation was noted on irrigated
treatments.
1. BOLIGŁOWA E., ŁABZA T., GLEŃ K.,
PUŁA J. 1999: Infection of potato tubers
by storage diseases depending on organic
fertilization. Progres in Plant Protection,
Vol. 39 (2): 898–901.
2. BORÓWCZAK F., GŁADYSIAK S.
1999: Disease infestation of potato tubers
80
depending on irrigation and cultivation
system. Progress in Plant Protection/
Postępy w Ochronie Roślin, Vol. 39 (2):
786–788.
3. CZAJKA W. 1988: Badania nad
występowaniem
ważniejszych
bakteryjnych i grzybowych chorób
ziemniaka na tle wybranych czynników
agrotechnicznych
oraz
zabiegów
chemicznych. Acta Acad. Agricult.,
Tech. Olst. Agricultura Supplement, C
44: 1–58.
4. CZAJKA W., CWALINA B., CZAJKA
M., FABISIEWICZ M. 1999: Porażenie
bulw ziemniaka patogenami w zależności
od nawożenia mineralnego. Progress in
Plant Protection/ Postępy w Ochronie
Roślin, 39 (2): 852–855.
5. GŁADYSIAK S., CZAJKA M. 1996:
Sprinkling influence on potato cultivars
infestation by diseases. Nowe Kierunki
w Fitopatologii, Materiały z Sympozjum,
Kraków: 217–221.
6. JABŁOŃSKI K., CZERKO Z. 1995: Zbiór
i przechowywanie ziemniaków. Fundacja
Rozwój SGGW, Warszawa: 62–69.
7. MATKOWSKI K., PROŚBA-BIAŁCZYK
U., PLĄSKOWSKA E. 2002: Influence
of cultivar mixture of potato and chemical
control against Phytophthora infestans
(Mont.) De Bary on the health status of
tubers. Advances of Agricultural Sciences
Problem Issues, 489: 261–268.
8. PAŃKA D., PIŃSKA M. 2004: Influence
of potassium fertilization on health status
of Barycz and Triada potato cultivars.
Prace Komisji Nauk Rolniczych i Biologicznych, BTN, 39, Seria B, 52: 261–269.
9. RUDKIEWICZ F., SIKORSKI L.,
ŚLĄZAK I. 1983: Wpływ rodzaju gleby,
nawożenia i zwalczania Phytophthora
infestans na rozwój niektórych chorób na
roślinach i bulwach ziemniaka. Biul. Inst.
Ziemn., 30: 157–170.
10. SADOWSKI CZ., GRABARCZYK S.,
RZEKANOWSKI CZ. 1988: Wpływ
nawadnianianawystępowanieStreptomyces
scabies (Taxter) i Rhizoctonia solani Kühn
na bulwach ziemniaków uprawianych
na glebie bardzo lekkiej. Acta Acad.
Agricult. Techn. Olst., Agricultura 47:
45–54.
11. SADOWSKI CZ., PESZEK J.,
RZEKANOWSKI CZ., SOBKOWIAK
S. 1996: Effect of irrigation and
different nitrogen fertilization rates on
the occurrence of Streptomyces scabies
(Taxter) on potato cultivated on very
light soil. Plant Breed. and Seed Sci. 40,
No. 1–2: 45–49.
12. SADOWSKI CZ., KORPAL W., LENC
L., KAWALEC A. 2003: Health status of
tubers of potato cultivated under organic
and integrated conditions. In: Obieg
Pierwiastków w Przyrodzie, Gworek, B.
and Misiak, J. (eds). Monografia tom II:
682–686.
13. SAWICKA B. 1995: The influence of
selected elements of meteorological
conditions on potato tubers with common
scab. Zesz. Probl. Post. Nauk Roln. 419:
89–93.
14. SZUTKOWSKA M. 1999: Czy można
ograniczyć porażenie bulw parchem
zwykłym. Ziemn. Pol. 3: 5–9.
15. ŚNIEG L. 1992: Wpływ nawadniania
i nawożenia azotem na występowanie
niektórych chorób bulw ziemniaka i ich
odporność na uszkodzenia mechaniczne.
Fragmenta Agronomica, 3/92: 49–57.
16. WEBER Z. 1976: Wpływ przedplonu
i innych czynników na występowanie
rizoktoniozy ziemniaka (Rhizoctonia
solani Kuhn). Rocz. Nauk Roln., Ser. E,
2: 45–65.
17. WENZEL H. 1948: Zur Erfassung des
Schadenausmasses in flanzenschutzversuchen. Pflanzenschutzberichte 15:
81–84.
Streszczenie: Wpływ deszczowania i nawożenia azotowego na zdrowotność ziemniaka uprawianego na glebie bardzo lekkiej. Ziemniak jest
bardzo często atakowany przez liczne patogeny,
mogące powodować znaczne straty plonu. Do
najgroźniejszych chorób należą: Phytophthora
infestans, Rhizoctonia solani, Streptomyces sp.,
Erwinia carotovora subsp. carotovora i Fusarium
spp. Ich rozwój, a tym samym szkodliwość, zależy od wielu czynników. W największym stopniu na przebieg procesu chorobowego wpływają
opady, temperatura, dostępność składników pokarmowych i stosowane zabiegi pielęgnacyjne.
Celem przeprowadzonych badań było określenie
wpływu nawadniania i zróżnicowanego nawożenia azotowego na porażenie bulw ziemniaka
przez najgroźniejsze patogeny. Doświadczenie
z uprawą ziemniaka odmiany ‘Drop’ wykonano
w latach 2001–2003 jako dwuczynnikowe, na
glebie lekkiej, należącej do kompleksu żytniego,
w trzech powtórzeniach. Pierwszym czynnikiem
było nawadnianie deszczowniane stosowane
w zależności od potrzeb roślin oraz występujących
niedostatków wody, a drugim nawożenie azotowe
na poziomie 75 i 125 kg N ha–1. Bezpośrednio po
zbiorach analizowano bulwy pod kątem występowania objawów parcha zwykłego, rizoktoniozy,
zarazy oraz suchej i mokrej zgnilizny.
W doświadczeniu zaobserwowano w kolejnych latach trwania badań duże nasilenie występowania na bulwach parcha zwykłego i rizoktoniozy. Deszczowanie powodowało wzrost porażenia bulw przez S. scabies i spadek zanieczyszczenia bulw sklerotami R. solani. Stopień porażenia
bulw różnicowało także nawożenie azotowe.
Pozostałe choroby notowano sporadycznie, a ich
nasilenie zależało od przebiegu warunków atmosferycznych w poszczególnych latach trwania
eksperymentu.
Dariusz Pańka
Department of Phytopathology,
University of Technology and Agriculture,
Bydgoszcz, Poland
Kordeckiego 20, 85-225 Bydgoszcz, Poland
Roman Rolbiecki
Department of Land Reclamation
and Agrometeorology,
Bydgoszcz, Poland
Bernardyńska 6, 85-029 Bydgoszcz, Poland
Czesław Rzekanowski
Department of Land Reclamation
and Agrometeorology,
Bydgoszcz, Poland
Bernardyńska 6, 85-029 Bydgoszcz, Poland
81
Heat balance and climatic water balance in vegetation period
of spring wheat
ELŻBIETA MUSIAŁ1, JOANNA BUBNOWSKA1, EDWARD GĄSIOREK1,
LESZEK ŁABĘDZKI2,
1
Agricultural
2
University of Wrocław, Departament of Mathematics
IMUZ – Bydgoszcz
Abstract: Heat balance and climatic water
balance in vegetation period of spring wheat. This
paper characterizes the changes of heat and climatic
water balance structure during the growing season
of spring wheat in four observatories: WrocławSwojec 1964–2000, Bydgoszcz 1946–2004,
Gorzów Wielkopolski 1970–1995 and Łódź 1954–
–1995. Study of changes and trends contains the
following elements of heat balance: net radiation,
latent heat flux,sensible heat flux, soil heat flux
and their contribution in net radiation. Climatic
water balance is defined as a difference between
precipitation and potential evapotranspiration
reckoned by the Penmann method.
Key words: heat balance, spring wheat, sensible
heat flux, latent heat flux, climatic water balance.
INTRODUCTION
Energy and water relations in different
ecosystems are best described by water
and heat balance structure. Heat and
water balance are connected through
a streamvapour, that transports a huge
quantity of energy to the atmosphere.
Because of it any change in water balance
influences heat balance and vice versa.
Climate
changes
with
their
consequences are connected with heat
exchange between active surface and
atmosphere. Therefore, searching for
climate variability concentrates on
studying heat and water balance of
different ecosystems. Different surfaces
transform this energy in a particular way
[Bubnowska, Gąsiorek, Łabędzki, Musiał
2005], [Bubnowska, Gąsiorek, Łabędzki,
Musiał, Rojek 2005], [Kędziora, Olejnik,
Kapuściński 1989], [Kapuściński 2000],
[Olejnik 1996]. The knowledge of heat
balance for various kinds of surface
allows characterizing changes in flux
distribution.
METHODS
The calculation of components of active
surface heat balance was carried aut by
using the BMC model, worked out by
Olejnik and Kędziora [1991, 1999]
The heat balance of ecosystems is
described by the following equation
[Kapuściński 2000; Kędziora 1989]:
Rn + LE + H + G = 0
where:
Rn – net radiation [Wm–2]
G – soil heat flux [Wm–2]
H – sensible heat flux [Wm–2]
LE – latent heat flux [Wm–2]
The exact description of applied
calculation method could be found in
a previous paper by the same authors
[Musiał 2001; Bubnowska, Gąsiorek
84
E. Musiał et al.
Łabędzki, Musiał, 2005] and in [Bowen
1926], [Karliński, Kędziora 1968],
[Shuttleworth, Wallace 1985].
Water balance in this paper is defined
as a difference between precipitation and
potential evapotranspiration reckoned by
Penmann method. Penman showed that
the latent heat flux could be expressed as:
Δ
( Rn + G ) + Ea
γ
LE =
⎛ Δ⎞
⎜1 + γ ⎟
⎝
⎠
where:
Ea – ability of air evapotranspiration
[Wm–s2],
Ea = 7,44(1 + 0,54v)d
v – wind speed at 2 m height [ms–1],
d – vapour pressure deficit [hPa],
γ – psychrometric constant
γ = 0,655 [hPaK–1],
∆ – mean rate of change of saturated vapour
pressure with temperature [hPaK–1],
LE, Rn, G – like above.
A simple relationship exists between
evapotranspiration ETP expressed in
[mm] and latent heat flux LE expressed
in [Wm–2]:
ETP = n
LE
28,34
Where n – number of days in decade, in
month.
RESULTS
The study was carried out on data
from the following meteorological
observatories: Wrocław-Swojec (1964–2000), Bydgoszcz (1946-2003), Łódź
(1954-1995) and Gorzów Wielkopolski
(1975–1995).
The variability of heat and water
balance components was analyzed in
given perennials.
HEAT BALANCE IN VEGETATION
PERIOD OF SPRING WHEAT
Mean values of net radiation calculated
for vegetation period of a spring wheat
(IV–VIII) in ssubsequent years in
Gorzów Wielkopolski fluctuated between
94 and 112 Wm–2. The lowest value
was obtained in 1962 when the sum of
precipitation in vegetation period was
also the lowest. Latent heat flux absorbed
from 61 to 68% of net radiation, and
sensible heat flux – from 24 to 61%. The
soil heat flux accounted for around 8% of
net radiation.
Values of heat balance components in
other observatories had similar variability.
Net radiation in Łódź changed from
90W/m2 in 1960, when the precipitation
was the highest (476.7 mm), to 115 W/m2
in 1983, when the precipitation was the
lowest (171.7 mm). In Wrocław and
Bydgoszcz net radiation fluctuated from
89 to 112 [Wm–2] and from 89 to 109
[Wm–2] respectively. Latent heat flux
absorbed from 58% in Bydgoszcz to
69% of net radiation in Łódź.
The highest values of net radiation
during the growing season were observed
in Łódź and Gorzów Wielkopolski,
whereas in Wrocław and Bydgoszcz the
values of Rn were lower.
Latent heat flux values were the
highest in Łódź, where net radiation was
the highest. This may be due to the fact
that in Łódź more energy was supplied
and therefore, more of it could be used
for evapotranspiration.
Heat balance and climatic water balance...
85
TABLE.1. Components of heat balance during the growing season of spring wheat (IV–VIII) in
Gorzów Wielkopolski (1970–1995)
Rok
Rn
-LE
-G
-H
-LE/Rn
-H/Rn
-G/Rn
P
1970
101
66
8
27
0,65
0,27
0,08
215,3
1971
105
64
8
33
0,61
0,31
0,08
256,4
1972
97
62
7
28
0,64
0,29
0,07
306,2
1973
105
66
8
31
0,63
0,30
0,07
273,7
1974
94
59
8
27
0,63
0,29
0,08
330,6
1975
110
71
8
31
0,65
0,28
0,07
233,0
1976
110
72
8
30
0,65
0,28
0,07
172,7
1977
98
65
8
25
0,66
0,26
0,08
460,0
1978
101
64
8
29
0,63
0,29
0,08
279,0
1979
102
65
8
29
0,63
0,29
0,08
225,5
1980
95
59
8
28
0,62
0,30
0,08
320,1
1981
104
63
9
32
0,61
0,31
0,08
306,9
1982
112
70
9
33
0,63
0,30
0,07
153,6
1983
109
71
8
30
0,65
0,28
0,07
214,4
1984
95
60
7
28
0,63
0,29
0,08
353,1
1985
103
65
8
30
0,63
0,29
0,08
229,6
1986
106
68
8
30
0,64
0,28
0,08
255,2
1987
94
62
7
25
0,66
0,27
0,07
365,3
1988
95
59
8
28
0,62
0,30
0,08
259,7
1989
107
70
8
29
0,65
0,27
0,08
190,9
1990
108
66
9
33
0,61
0,31
0,08
285,4
1991
102
69
8
25
0,68
0,24
0,08
221,0
1992
108
72
8
28
0,67
0,26
0,07
170,0
1993
100
64
8
28
0,64
0,28
0,08
268,0
1994
107
71
8
28
0,66
0,26
0,08
259,0
1995
108
69
8
31
0,64
0,29
0,07
320,0
–2
–2
–2
Rn – net radiation [Wm ] , LE – latent heat flux [Wm ], G – soil heat flux [Wm ], H – sensible heat
flux [Wm–2], P – sum of precipitation in period IV–VIII [mm]
Regarding the Figure 3, the following
conclusion could be drawn: in Łódź and
Gorzów Wielkopolski more energy was
transferred to the atmosphere from the
active surface of spring wheat than in
Wrocław and Bydgoszcz. In 1982–1995
perennial, common for all observatories,
there was a distinct increasing tendency
for sensible heat flux. Thus, the amount
of energy used for heating atmosphere is
growing.
Due to the fact that the values of heat
balance components depend on the net
radiation value, it is worthy looking into
contribution of each component to net
radiation (Rn).
86
E. Musiał et al.
FIGURE 1. Variation of mean ten-days values of net radiation (Rn) during the growing season
of spring wheat in Wrocław – Swojec, Bydgoszcz, Gorzów Wielkopolski and Łódź
FIGURE 2. Variation of mean ten-days values of latent heat flux (LE) during the growing season
FIGURE 3. Variation of mean ten-days values of sensible heat flux (H) during the growing season
87
FIGURE 4. Variation of mean ratios of latent heat flux and net radiation (LE/Rn) during the growing
season of spring wheat in Wrocław – Swojec, Bydgoszcz, Gorzów Wielkopolski and Łódź
FIGURE 5. Variation of mean ratio values of sensible heat flux and net radiation (H/Rn) during the
growing season of spring wheat in Wrocław – Swojec, Bydgoszcz, Gorzów Wielkopolski and Łódź
Looking at the course of mean values
of latent heat flux contribution in net
radiation, distinct decreasing tendency
for this contribution in the last 20 years
is seen. Therefore, less and less energy
is used for evapotranspiration, especially
in Wrocław and Bydgoszcz.
Increasing share of sensible heat flux
in net radiation in all observatories stress
that more and more energy in all regions
is used for heating atmosphere.
The above mentioned results are
concordant with those by Musiał [Musiał,
Gąsiorek, Rojek 2004], Trepińska[1997],
[Ryszkowski Kędziora, 1995] and
[Kożuchowski 2004].
The temperature increase is a
consequence of the growing contribution of
latent heat flux to net radiation. Regression
equations in Table 2 confirm this.
CLIMATIC WATER BALANCE
The net climatic water balance determines
conditions of plant vegetation [Bac,
Rojek, 1979, 1982]. This index may be
the source of information on climate
change effects.
88
E. Musiał et al.
TABLE 2. Mean parennial air temperature values for periods IV–VIII and IV–IX with regression
equations
Obserwatory
Tweg
[°C]
Tpc
[°C]
ssr
[°C]
Linear regression equation
Tendency
[°C/10 years]
Bydgoszcz
15.22
15.0
0.92
y = 0.0176x + 14.7
0.18*
Gorzów
Wielkopolski
14.57
14.4
0.85
y = 0.0477x + 13.9
0.48*
Łódź
14.32
14.1
0.88
y = 0.0129x + 14.0
0.13**
Wrocław
14.93
14.7
0.83
y = 0.0347x + 14.3
0.35*
Tweg – mean seasonal yearly air temperature (IV–VIII),
Tpc – mean seasonal, yearly air temperature (IV–IX),
ssr – standard devation of temperature
*) – statistically significant for α = 0.05
**) – statistically significant for α = 0.3
FIGURE 6. Variation of climatic water balance during the growing season of spring wheat (IV–VIII)
in Bydgoszcz (1946–2003)
FIGURE 7. Variation of precipitation (P) and climatic water balance (CWB) during the growing season
of spring wheat (IV–VIII) in Wrocław (1964–2000)
In Bydgoszcz climatic water balance
was positive in two years during the
period 1946–2004. Those two years
were chracterized by high sums of
precipitation: 599,1 mm in 1980 and
506,6 mm in 1985.
In years 1980 and 1985 the spring
wheat vegetation period was characterized
by rather low (390,56 and 419,35 mm,
respectively) evapotranspiration values,
whereas mean evapotranspiration value
for years 1946–2004 was 456,5 mm.
89
In Wrocław climatic water balance
was positive in 1980 (11,3 mm) and in
1986 (9,2 mm).
During the spring wheat vegetation
period in 1980, the precipitation
value was 438 mm and the potential
evapotranspiration reached the level of
426,7 mm. The year 1997 is noteworthy
due to the flood in Wrocław. During
that spring wheat vegetation period
the precipitation sum (494 mm) was
52% higher than the mean value. In
spite of such high precipitation value,
of spring wheat (IV–VIII) in Łódź (1954–1995)
of spring wheat (IV–VIII) in Gorzów Wielkopolski (1970–1995)
90
E. Musiał et al.
TABLE 3. Evapotranspiration tendencies in 4 examined regions
Tendency
[°C/10 years]
P
[mm]
ETP
[mm]
Linear regression equation
Bydgoszcz
281
457
y = 0.29x + 443.4
2,9
Gorzów Wielkopolski
266
477
y = 4.91x + 466.7
49,1*
Łódź
312
505
y = 2.0x + 470.6
20**
Wrocław
326
500
y = 2.0x + 468.2
20*
Obserwatory
P – mean seasonal, yearly precipitation (IV–VIII)
ETP – mean seasonal, yearly evapotranspiration (IV–VIII),
**) - statistically significant for α = 0.01
the net climatic water balance in the
1997 growing season remained negative
(–31 mm).
In Łódź, during the 1954–1995
perennial, there were 3 years with
positive net climatic water balance: 1960
(precipitation 477 mm, climatic water
balance 20 mm), 1980 (458 mm and
38,9 mm, respectively) and 1985 (464
mm and 17 mm, respectively). Mean
precipitation sum in the growing season
for spring wheat in Łódź was 312 mm.
In Gorzów Wielkopolski, the climatic
water balance was positive only once and
reached the value of 48,2 mm in 1977,
when the precipitation sum in vegetation
period was the highest (460 mm).
All
analyzed
regions
were
characterized by continuous water
shortage in vegetation period. During
the period 1970–1995 the worst water
conditions were observed in Gorzów
Wielkopolski (the highest water shortage
sum – 5474 mm), whereas the best terms
were seen in Wrocław (water shortage
sum of 4640 mm was 15% lower than
that in Gorzów). During the wheat spring
growing season the values of climatic
water balance were positive only in years
with the highest precipitation sums.
CONCLUSIONS
The enlargement of negative net climatic
water balance, along with increasing
potential evapotranspiration, confirms
diminishing atmospheric precipitation.
The increase of temperature during
the spring wheat growing season in all
discussed regions is a consequence of
enlarging sensible heat flux, used for
heating atmosphere.
The growing temperature values
enlarge the saturation deficiency, thus
allowing more water to evapotranspirate
(PET increase). Consequently, growing
precipitation enlarge the negative net
climatic water balance.
The tendencies observed among heat
balance components are concordant with
changes seen among the components of
climatic water balance.
The research supported by KBN grant
in years 2004–2007.
REFERENCES
BAC S., ROJEK M. 1979: Klimatyczny
bilans wodny a odpływy w Polsce, Przegl.
Geofiz. 24(3) 293–297.
BAC S., ROJEK M. 1977: Metodyka oceny
stosunków wodnych obszarów rolniczych
na podstawie danych klimatycznych,
Zesz. nauk ART Olszt. Nr 21, 13–24.
BAC S., ROJEK M. 1982: Klimatyczne
bilanse wodne w Polsce [w:] Bac S.
(red.) Agrometeorologiczne podstawy
melioracji wodnych w Polsce. PWRiL
Warszawa.
BOWEN I.S. 1926: The ratio of heat losses
by conduction and by evaporation from
any water surface. Phys. Rev., 27, p.
779–787.
BUBNOWSKA J.,
GĄSIOREK E.,
ŁABĘDZKI L., MUSIAŁ E., ROJEK
M. 2005: Bilans cieplny lasu iglastego
w latach o ekstremalnych opadach i jego
wieloletnie zmiany w rejonie Bydgoszczy
i Wrocławia. Woda-Środowisko-Obszary
Wiejskie t. 5, z. spec. (14) s. 69–82.
BUBNOWSKA J., GĄSIOREK E., ŁABĘDZKI L., MUSIAŁ E. 2005: Struktura bilansu cieplnego łanów w ekstremalnych warunkach opadowych na tle
wielolecia. Woda-Środowisko-Obszary
Wiejskie, t. 5 z. 2(15) s. 31–52.
ŁABĘDZKI L., BĄK B. 2004: Standaryzowany klimatyczny bilans wodny jako
wskaźnik suszy. Acta Agrophys. Vol.
3(1) s. 117–124.
KAPUŚCIŃSKI J. 2000: Struktura bilansu
cieplnego powierzchni czynnej na tle
warunków klimatycznych środkowozachodniej Polski, Rocz. Nauk. AR Pozn.
Rozpr. Nauk. AR Poznań, 303, s. 250.
KARLIŃSKI M., KĘDZIORA A. 1968:
Rozważania metodyczne przy układaniu
kalendarza przyrody dla województwa
szczecińskiego, Pr. Kom. Nauk Roln.
Kom. Nauk Leśn. PTPN 24, s. 159–175.
KĘDZIORA
A.
1999:
Podstawy
agrometeorologii, Poznań, PWRiL.
KĘDZIORA A., OLEJNIK J., KAPUŚCIŃSKI J. 1989: Impact of landscape structure on heat and water balance. Ecol.
Intern. Bull. 17 ss. 1–17.
KOŻUCHOWSKI K. 2004: Skala, uwarunkowania i perspektywy współczesnych
91
zmian klimatycznych w Polsce, Łódź,
Biblioteka s. 170.
MUSIAŁ E. 2001: Modelowanie procesu
ewapotranspiracji rzeczywistej i prognozowanie jego tendencji. Zesz. Nauk. AR
Wroc., nr 412, Rozpr. 182, ss. 116.
MUSIAŁ. E. , GĄSIOREK E., ROJEK M.S.
2004: Zmienność temperatury powietrza
w obserwatorium Wrocław–Swojec w
latach 1964-2001. Acta Agroph. nr 105
Vol. 3(2) s. 333–342.
OLEJNIK J., KĘDZIORA A. 1991: A model
for heat and water balance estimation and
its application to land use and climate
variation., Earth Surface Processes
Landforms Vol.16, ss. 601–617.
OLEJNIK J. 1996: Modelowe badania
struktury bilansu cieplnego i wodnego
zlewni w obecnych i przyszłych
warunkach klimatycznych, Rocz. Nauk.
AR Poznań z. 268, ss. 125
RYSZKOWSKI L., KĘDZIORA A. 1995:
Modification of the effects of global
climate change by plant cover structure
in an agricultural landscape. Geogr. Pol.
Vol. 65, s. 5–34.
RYSZKOWSKI L., KĘDZIORA A. 1993:
Rolnictwo a efekt szklarniowy. Kosmos
42 s. 123–149.
SHUTTLEWORTH W.J., WALLACE J.S.
1985: Evaporation from sparse crops-an
energy combination theory, Quart. J. R.
Met. Soc. 111, pp. 839–895.
TREPIŃSKA J. 1997: Wahania klimatu
w Krakowie (1792–1995). Kraków Inst.
Geogr. UJ, s. 1–204.
Streszczenie: Bilans cieplny i klimatyczny bilans
wodny dla okresu wegetacji pszenicy jarej. Praca
charakteryzuje zmiany struktury bilansu cieplnego
i klimatycznego bilansu wodnego w czasie okresu
wegetacji pszenicy jarej dla czterech stacji:
Wrocław-Swojec 1964–2000, Bydgoszcz 1946–
2004, Gorzów Wielkopolski 1970–1995 i Łódź
1954–1995. W pracy przeanalizowano zmiany
i trendy takich składowych bilansu cieplnego,
jak: saldo promieniowania i strumienie ciepła
jawnego, ciepła utajonego przeznaczonego na
parowanie, ciepła wymienianego z podłożem oraz
ich udział w saldzie promieniowania.
92
E. Musiał et al.
E. Musiał, J. Bubnowska, E. Gąsiorek
Agricultural University of Wrocław
Departament of Mathematics
ul. Grunwaldzka 53
50 357 Wrocław
Poland (071)3205659
Leszek Łabędzki
ul. Glinki 60
85 174 Bydgoszcz
Poland
[email protected]
Climatic and agricultural water balance for grasslands in Poland
using the Penman-Monteith method
WIESŁAWA KASPERSKA-WOŁOWICZ, LESZEK ŁABĘDZKI
Institute for Land Reclamation and Grassland Farming,
Regional Research Centre in Bydgoszcz
Abstract: Climatic and agricultural water
balance for grasslands in Poland using the
Penman-Monteith method. The aim of the paper
was to estimate the spatial variability of climatic
and agricultural water balance for two-cut
meadows (for two hay yield: 5 and 7 Mg ha-1)
in different agro-climatic regions in Poland. It let
to estimate meadow water needs under different
The climatic water balance is a difference
between precipitation and reference evapotranspiration calculated according to the PenmanMonteith formula. The agricultural water balance
is a difference between precipitation and potential
crop evapotranspiration. Potential grassland
evapotranspiration is the result of multiplication
crop coefficient and reference evapotranspiration
according to the Penman-Monteith formula. The
crop coefficient values were calculated on the
basis of multiannual lysimeter experiments of
meadow evapotranspiration.
The analysis was carried out in the growing season
(from April to September) in the years 1970–1995
for 17 meteorological stations located in different
regions of Poland. The balances were done for
ten-day periods, months and the whole growing
season.
The climatic water balance indicates the risk
of agro-climatic drought. The agricultural water
balance indicates water deficit and irrigation water
needs of two-cut meadows in different regions of
Poland. The negative values of this balance mean
the scarcity of water for grasslands. The values
of agricultural water balance are shown on the
maps.
Key words: climatic and agricultural water
balance, two-cut meadow.
INTRODUCTION
The estimation of drought hazard and
irrigation needs for plants can be estimated
using crop coefficients method according
to the Penman-Monteith formula.
This method for calculation reference
evapotranspiration ETo is commonly
recommended and used in the word To
estimate crop water requirements, one
can relate potential evapotranspiration
from the cropped soil with an optimum
water supply under consideration to an
estimated reference evapotranspiration
by means of crop coefficient [1, 3].
Evaluation of crop water demands,
agricultural droughts and irrigation
requirements can be made with the
help of climatic water balance and
agricultural water balance. The climatic
water balance can be interpreted as the
indicator of climate dryness [7]. The
agricultural water balance indicates
the values of irrigation water needs for
plants. These two balances are needed to
the preliminary estimation of irrigation
needs.
94
W. Kasperska-Wołowicz, L. Łabędzki
The aim of the paper was to estimate
the spatial variability of climatic and
agricultural water balance for grasslands
in different agro-climatic regions in
Poland and to show these values on
the maps. These maps let to estimate
water needs for 2-cut meadow with the
different yield in the years with different
The climatic water balance CWB is the
difference between precipitation sum P
and reference evapotranspiration ETo
calculated according to the PenmanMonteith formula [1, 6]:
(1)
TABLE 1. The crop coefficients kc for the PenmanMonteith equation for 2-cut meadow in Poland
Month
IV
V
VI
VII
VIII
IX
Ten-day
period
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
AWB = P – ETP
kc coefficient for hay
yield:
7 Mg⋅ha–1 5 Mg⋅ha–1
0.50
0.45
0.75
0.70
0.95
0.80
1.00
0.90
1.15
1.00
1.20
1.10
1.30
1.20
0.55
0.45
0.65
0.55
0.80
0.70
0.90
0.80
1.10
1.00
1.30
1.15
1.20
1.10
1.35
1.25
1.10
1.10
1.10
1.10
1.10
1.10
(2)
where:
ETP – is evapotranspiration of wellwatered crop calculated from the
formula:
ETP = ETo⋅kc
METHODS AND MATERIAL
CWB = P – ETo
The agricultural water balance AWB is
the difference between precipitation sum
P and potential crop evapotranspiration
ETP:
(3)
where:
kc – crop coefficient,
ETo – reference evapotranspiration
(mm).
The kc coefficients values were
calculated on the basis of multiannual
lysimeter experiments and depended on
the hay yield (Tab. 1) [4, 6, 8].
The analysis was carried out in
the growing season (from April to
September) in the years 1970–1995
for 17 meteorological stations. The
chosen stations are situated in different
agroclimatic regions of Poland (Tab. 2).
According to Bac, Koźmiński, Rojek
[2] taking into account water conditions
one can distinguish regions: A – dry,
B – moderately wet, C – wet. Taking
into account thermal and solar energy
conditions one can distinguish regions: 1
– hot and sunny, 2 – hot and moderately
sunny, 3 – hot and cloudy, 4 – moderately
hot and sunny, 5 – moderately hot and
moderately sunny, 6 – moderately hot
and cloudy, 8 – cold and moderately
sunny.
The agricultural water balance AWB
was calculated for 2-cut meadows with the
hay yield of: 7 Mg⋅ha–1 and 5 Mg⋅ha–1.
The negative values (–) of climate
and agricultural water balance showed
Climatic and agricultural water balance...
TABLE 2. Meteorological stations and agroclimatic regions according to Bac [2]
Number of
the station
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Name of the
station
Koszalin
Bielnik
Biebrza
Chojnice
Szczecin
Frydrychowo
Poznań
Warszawa
Zielona Góra
Łódź
Sosnowica
Wrocław
Opole
Częstochowa
Zamość
Kraków
Rzeszów
Agro-climatic
region
C8
B4
B8
B8
A6
A2
A3
A2
A1
A5
A1
B2
B3
B6
B5
C5
C8
the drought risk and water deficit for
grasslands irrigation, the positive values
(+) – showed the excess of water.
The calculation of the balances have
been done for ten-day periods, months
and the whole growing season (from
April to September). The probability
distributions of climatic water balance
and agricultural water balance were
determined. It was assumed that the
examined period was classified as wet
at the probability of exceedence equal to
25%, mean – 50% and dry – 75%.
RESULTS
Climatic water balance
The climatic water balance CWB is
different in Poland and depends on
95
agrometeorological conditions in the
examined regions. The mean multiannual
values of this balance (at the p = 50%)
for growing season (April to September)
were negative in all examined stations.
In the north (stations 1–3) and south
(stations 14, 16, 17) parts of the country
the values of climatic water balance
were negative but higher than –40 mm.
The values of CWB less than –100 mm
were in the central Poland, in the regions
represented by the stations numbered
5–10. The lowest values of the balance
(less than –150 mm) were in the region
between the Odra and left side of Wisła
rivers, especially represented by the
stations 5–7 and 9–10. The highest risk
of agro-climatic drought in Poland was
observed on the area from the north-west
to the central part of Poland (Fig. 1).
The lowest monthly values of CWB
in central Poland were in May, June
and July. It resulted from the monthly
distribution of precipitation and reference
evapotranspiration. For instance in
Poznań countryside (station 7), the
multiannual mean monthly sum of ETo
reaches the highest values in July (about
100 mm), a little less in June, May and
August (about 80 mm) [5].
The higher values of CWB in the
average year were found in the south
(stations 14, 16, 17) part of Poland,
the region situated near the Karpaty
mountains and in the north part of Poland
– near Bałtyk sea (station 1) and in the
Biebrza river valley (station 3) – in the
north-east part of the country.
Agricultural water balance
In the wet growing season (at 25%
probability) agricultural water deficit
did not occur in the north and south
parts of Poland. The negative values of
96
FIGURE 1. Mean multiannual precipitation (1) and the climatic water balance (2) in the growing season (IV–IX) in different regions of Poland
agricultural water balance for low and
high yield meadows were observed in
the Warta river basin (stations 6, 7, 10),
in the west part of Poland (stations 5,
9) and in the Polesie Lubelskie region
(station 11). It means that even in the wet
year the big area of Poland is threaten by
droughts. The highest water deficit for
meadow was in the Szczecin countryside
(Fig. 2a).
In the mean growing season (at 50%
probability) the values of agricultural
water balance were negative for 7
Mg⋅ha–1 hay yield meadow in the whole
area of Poland. The high water deficits
for grasslands (more than 100 mm) were
observed on the area from the northwest to the central part of Poland. The
highest scarcity of water (more then 150
mm) was observed in Warta and middledown Odra river basins. In mean year, in
all regions of Poland, water deficit for 5
Mg⋅ha–1 hay yield meadow was 40–47
mm less than for 7 Mg⋅ha–1 hay yield
meadow (Fig. 2b).
In the dry growing season (at 75%
probability) agricultural water deficit was
observed in the whole area of Poland.
The highest values of water deficit were
in the area between Odra and the left side
of Wisła and they exceeded 200 mm for
7 Mg⋅ha–1 yield meadow and 160 mm for
5 Mg⋅ha–1 yield meadow. Water deficit
for high yield meadow was less than
100 mm only in south part of Poland in
Kraków countryside (Fig. 2c).
The values of climatic water balance
(CWB) and agricultural water balance
(AWB) for the 7 Mg⋅ha–1 hay yield
meadow did not differ very much in the
whole mean growing season (at 50%
probability). These values differed in
particular months. They depended on
agrometeorological conditions and the
time of first and second cut of meadow.
The monthly distribution of climatic
a
97
b
c
FIGURE 2. Agricultural water balance for 2-cut
meadow at the yield 7 Mg·ha–1 (upper number,
mm) and 5 Mg·ha–1 (lower number, mm) in different agroclimatic regions of Poland in the growing
season (IV–IX) a) wet, b) mean, c) dry
• 01 – number of meteorological station, 01 – Koszalin
and agricultural water balance is shown
on the example of Wielkopolska region
(dry and hot) – represented by Poznań
agrometeorology station (7) and the
Biebrza river valley (moderately wet and
cold) – represented by station 3. (Fig. 3).
The highest water deficits for
meadow were in May and September
– the months of intensive grass growth.
In Wielkopolska region, in these months,
they exceed 50 mm and in Biebrza river
catchment – 30 mm. In the other months
the values of climatic water balance
were higher than values of agricultural
water balance. In Biebrza region the
values of climatic and agricultural water
balance were positive in September. In
every month they were higher than in
Wielkopolska region.
CONCLUSIONS
1. The climatic water balance CWB is
different in Poland and depends on
agrometeorological conditions. The
highest risk of agro-climatic drought
in Poland was observed in the region
98
a
b
FIGURE 3. Mean monthly multiannual climatic water balance (1) and agricultural water balance for
7 Mg·ha–1 yield 2-cut meadow (2) in the growing season (IV–IX) in a) Poznań, b) Biebrza
2.
3.
4.
5.
between the Odra and left side of
Wisła rivers.
In the wet year the big area of Poland
is threaten by droughts. The negative
values of agricultural water balance
for low and high yield meadows
were observed in the west, central
and east part of Poland.
In the mean year the values of
agricultural water balance were
negative for 7 Mg⋅ha–1 hay yield
meadow in the whole area of Poland.
The highest scarcity of water (more
than 150 mm) was observed in Warta
and middle-down Odra river basins.
Water deficit for 5 Mg⋅ha-1 hay yield
meadow was 40–47 mm less than for
7 Mg⋅ha–1 hay yield meadow.
In dry year meadow water deficit was
observed in the whole area of Poland.
The highest water needs were in the
area between Odra and left side of
Wisła and they exceeded 200 mm
for 7 Mg⋅ha–1 yield meadow and 160
mm for 5 Mg⋅ha–1 yield meadow.
Water deficits should be satisfied
by efficiently performing irrigation
systems.
REFERENCES
1. ALLEN R.G., PEREIRA L.S., RAES D. &
SMITH M. 1998: Crop evapotranspiration
– Guidelines for computing crop water
requirements. FAO Irrigation and
drainage no 56, pp. 300.
2. BAC S., KOŹMIŃSKI C., ROJEK M. 1993:
Agrometeorologia. [Agrometeorology]
Warszawa. PWN, pp. 250 (in Polish).
3. FEDDES R.A., LENSELINK K.J. 1994:
Evapotranspiration. ILRI Publication 16.
Drainage Principles and Applications.
Wageningen, p. 145–173.
99
4. KACA E., ŁABĘDZKI L., CHRZANOWSKI, S,. CZAPLAK I., KASPERSKA-WOŁOWICZ W. 2003: Gospodarowanie zapasami wody użytecznej
gleb torfowo-murszowych w warunkach
regulowanego odpływu w różnych regionach agroklimatycznych Polski. Woda,
Środowisko, Obszary Wiejskie. Rozprawy naukowe i mono-grafie nr 9, ss.
118. Falenty, Wydaw. IMUZ. [Managing
useful water resources in peat-moorsh
soils at a regulated outflow in different
agro-climatic regions of Poland. Water-Environment-Rural Areas, Treatises &
Monographs 9. Falenty: IMUZ, Poland,
pp 108] (in Polish, summary in English).
5. KASPERSKA-WOŁOWICZ W., ŁABĘDZKI L. 2004: Porównanie ewapotranspiracji wskaźnikowej według Penmana i Penmana-Monteitha w różnych
regionach Polski. Woda, Środowisko,
Obszary Wiejskie t. 4 z. 2a (11) s. 123–
–136. [A comparison of reference evapotranspiration according to Penman and
Penman-Monteith in various regions of
Poland. Water-Environment-Rural Areas
4(2a). Falenty: IMUZ, Poland, p. 123–
136] (in Polish, summary in English).
6. ŁABĘDZKI L. 1997: Potrzeby nawadniania użytków zielonych – uwarunkowania przyrodnicze i prognozowanie.
Rozpr. Habil. Falenty: Wydaw. IMUZ ss.
120. [Grasslands water requirements –
nature conditions and forecasting. Falenty: IMUZ, Poland, pp. 120] (in Polish).
7. ŁYKOWSKI B. 1986: Wskaźnik
suchości klimatu P – E w nowym ujęciu.
Problematyka melioracji w nauczaniu
i badaniach naukowych. Warszawa:
Wydaw. SGGW s. 59–68. [Indicator of
climate dryness P – E in a new approach.
Reclamation problems in teaching and
research studies. Warszawa: SGGW, p.
59–68] (in Polish, summary in English).
8. ROGUSKIW., SARNACKAS., DRUPKA
S. 1988: Instrukcja wyznaczania potrzeb
i niedoborów wodnych roślin uprawnych
i użytków zielonych. Mat. instruktażowe
nr 66, Falenty: IMUZ ss. 90. [Guidelines
100
for estimating water needs and water
deficits for cultivated plants and
grasslands. Instruction Materials no 66,
Falenty: IMUZ, pp. 90] (in Polish).
Streszczenie: Klimatyczny i rolniczy bilans wodny dla użytków zielonych w Polsce z wykorzystaniem metody Penmana-Monteitha. Celem pracy
była ocena przestrzennego zróżnicowania wodnych bilansów klimatycznego i rolniczego dla
dwukośnych użytków zielonych (dla plonów 5
i 7 Mg ha–1). Autorzy przeprowadzili analizę bilansów sporządzonych w okresie dekady, miesiąca oraz wegetacji dla 17 stacji meteorologicznych
w różnych regionach Polski. W przeciętnym okresie wegetacji IV–IX wartość rolniczego bilansu
wodnego jest ujemna dla plonów w wysokości
7 Mg ha–1 w całym kraju. Największe niedobory (powyżej 150 mm) obserwuje się w dorzeczu
Warty. Niedobór dla plonów wysokości 5 Mg ha–1
jest o 40–47 mm mniejszy w porównaniu z plonami użytków zielonych w wysokości 7 Mg ha–1.
MS. received December 2006
Authors’ address:
Wiesława Kasperska-Wołowicz
Leszek Łabędzki
IMUZ
Glinki 60, 85-174 Bydgoszcz
Poland
Variation of climatic water balance and heat balance for various
ecosystems in Wrocław in the years 1964–2000
ELŻBIETA MUSIAŁ, JOANNA BUBNOWSKA, EDWARD GĄSIOREK
Agricultural University of Wrocław, Department of Mathematics
Abstract: Variation of climatic water balance and
heat balance for various ecosystems in Wrocław
in the years 1964–2000. The study contains a
description of water relations, characterized by
cumulative climatic water balance, in the region
of Wrocław-Swojec. Heat relations during the
vegetation period were described by heat balance
of various active surfaces. The evaluation of
climatic water balance variation was based on
monthly sums of precipitation and potential
evapotranspiration. The latter was calculated
according to the Penman’s model for the warm
half of the year (IV–IX) and according to the
Tichomirow’s and Iwanow’s model for the cold
half of the year (X–IX). The values of climatic
water balance were assessed for the period from
X to IX in every year, as well as for all years
in common during the period 1964–2000. The
perennial 1964–2000 in Wrocław-Swojec was
characterized by potential evapotranspiration
increase and a decreasing trend in yearly sums of
atmospheric precipitation. The calculated values
of cumulative climatic water balance proved the
deepening water shortage in this area. Due to
aggravating water deficit in Wrocław latent heat
flux decreases and sensible heat flux, used for
heating the atmosphere, increases, thus causing
the warming up in this region.
Key words: potential climatic water balance,
potential evapotranspiration, heat balance, latent
heat flux, sensible heat flux, Bowen’s ratio.
about changes of climatic water balance
in this region. The importance of climatic
water balance in meteorology is due to
the fact that net climatic water balance
defines conditions for plant vegetation,
influences changes of outflow indexes
in river cachments and retention.
The investigation was performed
in years 1964–2000. Every analyzed
year started in October and ended in
September. Creation of cumulative
climatic water balance from October
to September gave the possibility to
evaluate water overflow or deficit after
autumn and winter, at the beginning
of vegetation period and during this
period until October. Due to the fact that
climatic water balance is connected with
heat balance through the stream vapour,
the values of balance components were
evaluated for coniferous forest, spring
wheat and potatoes. The main aim of
the study was to describe variations of
water overflow and deficit after autumn
and winter in subsequent years and
consequences of water deficits in heat
balance component values for examined
ecosystems.
INTRODUCTION
Investigations on variation of heat
balance components for various
ecosystems, performed in the WrocławSwojec region, made the authors think
METHODS
The term of climatic water balance
was first used in Polish literature by
Bac and Rojek in 1970 [Bac, Rojek
102
1977, 1979, 1982]. The authors
defined climatic water balance as a
difference between precipitation and
indicatory evapotranspiration [Rojek
1994], [Rojek, Wiercioch 1994, 1995].
The name climatic water balance is
mainly connected with net balance
assessed for the warm period (IV–IX).
Contradictory to the above mentioned
investigations, the authors of this study
created a term of potential climatic water
balance (PCWB). PCWB is defined as
a difference between precipitation and
potential evapotranspiration calculated by
Penman, assessed in the warm half of the
year (IV–IX). When the cold half of the
year (X–III) is taken into account, PCWB
is a difference between precipitation and
potential evapotranspiration calculated
by Iwanow and Tichomirow [Kędziora
1999], [Olejnik, Kędziora 1991].
Penman [Penman 1948, 1950, 1956,
1963] assumed that the density of latent
heat flux used for evapotranspiration can
be calculated in the following way:
LE =
⎤
γ ⎡Δ
( Rn + G ) + Ea ⎥ =
⎢
Δ+γ⎣γ
⎦
Δ
( Rn + G ) + Ea
γ
=
⎡ Δ⎤
⎢1 + γ ⎥
⎣
⎦
where :
Rn – net radiation [Wm–2],
G – soil heat flux [Wm–2],
H – sensible heat flux [Wm–2],
LE – latent heat flux [Wm–2],
Ea – ability of air evapotranspiration
[Wm–2],
Ea = 7.44(1 + 0.54v)d,
v – wind speed at 2 m height [ms–1],
d – vapour pressure deficit [hPa],
∆ – mean rate of change of saturated
vapour pressure with temperature
[hPaK–1],
γ – psychrometric constant γ = 0.655
[hPaK–1].
A simple relationship exists between
potential
evapotranspiration
ETP
expressed in [mm] and latent heat flux LE
expressed in [Wm–2]: ETP = nLE/28.34,
where n is a number of days in decade,
in month.
Climatic water balance, through
the stream vapour transporting huge
amount of energy to the atmosphere,
is connected with active surface heat
balance. The components of heat balance
are not independent due to the rule of
evapotranspiration priority in nature.
Therefore, both deficit and excess in
water, defined by cumulative potential
climatic water balance at the beginning
and during the vegetation period, decide
on values of various heat balance streams.
Thus, heat balance should be the next
step of investigation. Heat balance of
any active surface may be expressed by
the following equation:
Rn + LE + H + G = 0
where H – sensible heat flux [Wm–2],
Rn, LE, G – like above.
Heat balance components for various
ecosystems were calculated according
to the MBC [Kędziora 1999], [Olejnik,
Kędziora 1991].
Seasonality of precipitation and potential
evapotranspiration in the perennial
1964–2000 in Wrocław-Swojec is
Variation of climatic water balance...
shown in Figure 1 as sums of potential
evapotranspiration and precipitation
in the following months of the year. In
our study we assume that one season
is the period from October, when
vegetation period ends, until September.
Characteristics of precipitation and
potential evapotranspiration seasonality
is as follows: the highest sums of
atmospheric precipitation and potential
evapotranspiration in Wrocław-Swojec
during the period 1964–2000 were seen
in June, July and August. Moreover, apart
from January, February, October and
November, atmospheric precipitation was
103
lower than potential evapotranspiration.
The analysis of mean sums of potential
climatic water balance in this perennial
revealed that net PCWB was negative in
March, April, May, June, July, August and
September. This may be due to the fact
that the vegetation period in the above
mentioned perennial was characterized
by continuous water deficits.
The description of 1964–2000
perennial in Wrocław-Swojec by
yearly precipitation and potential
evapotranspiration sums show that the
last two decades of the XXth century were
characterized by increasing potential
FIGURE 1. Sums of precipitation (P) and potential evapotranspiration (ETP) in Wrocław-Swojec
(1964–2000)
FIGURE 2. Mean monthly sums of potential climatic water balance (PCWB) in Wrocław-Swojec
(1964–2000)
104
evapotranspiration and diminishing
atmospheric precipitation (Fig. 3).
Precipitation decrease and potential
evapotranspiration increase resulted in
enlarging net potential climatic water
balance during the years 1964–2000
(Fig. 4).
Tendencies of yearly changes
in
precipitation
and
potential
evapotranspiration sums as well as
basic statistical characteristics are
shown in Table 1. In the 37-year period
of observation in Wrocław-Swojec,
potential evapotranspiration has shown
a growing tendency both during the
whole year and during the period from
X–III. The last two decades of the XXth
century revealed quicker increase of
potential evapotranspiration than in the
earlier period. However, the potential
evapotranspiration rise in the period
X–III was less pronounced than in the
preceding years. Decreasing trends of
yearly atmospheric precipitation sums
and the sums in the period X–III, suggest,
the aggravating atmospheric precipitation
shortage in the last 20 years.
The next step was the characteristics
of the period 1964–2000 in WrocławSwojec, conducted in the perennial for
every season from X to IX separately, by
assessing cumulative potential climatic
water balance. This characteristics
looks as follows: in years 1964–1986
net PCWB, depending on precipitation,
was alternately positive and negative
FIGURE 3. Yearly sums of potential evapotranspiration and precipitation in Wrocław-Swojec (1964–
–2000)
FIGURE 4. Yearly sums of potential climatic water balance in Wrocław-Swojec (1964–2000)
105
σp
Linear regression
equation for
precipitation
Tendency
mm/
/10 year
ETP
(mm)
Linear regression
equation for
evapotranspiration
Tendency
mm/10 year
σETP
1980–
–2000
P
(mm)
1964–
–2000
season
Years
TABLE 1. Basic statistical charateristics of evapotranspiration and precipitation in Wrocław-Swojec
(1964–2000)
X–IX
I–XII
X–III
X–IX
I–XII
X–III
571
570
200
550
555
196
92
95
55
83
104
52
y = –1.8x+604
y = –1.9x+606
y = –0,5x+210
y = –2.3x+580
y = –3.6x+600
y = –0,7x+205
–18*
–19*
–5
–23*
–36*
–7
746
745
180
777
772
204
y = 4x+670
y = 3,9x+672
y = 2,8x+128
y = 5,7x+670
y = 6,9x+640
y = 1,4x+190
40**
39**
28**
57**
69**
14**
69
73
45
73
79
38
P – mean precipitation, σp – standard deviation for P, ETP-mean evapotranspiration
σETP – standard deviation for ETP
**) – statistically significant for α = 0.01
throughout the period X–IX. It was often
seen that within the analyzed season net
PCWB was positive until April, which
means that precipitation covered the
needs of potential evapotranspiration.
Contrarily, since May net PCWB
became negative, which was the sign
of deepening water deficit. The detailed
analysis starts in the year 1964/1965.
Net potential climatic water balance
(PCWB) was positive for the whole
period from X to IX (precipitation 661
mm). Other seasons with positive net
PCWB were as follows: 1967/1968
(precipitation 706 mm), 1970/1971 with
693.5 mm of precipitation and 1976/1977
(precipitation 722.6 mm). In the above
mentioned years precipitation was higher
than the mean value in the perennial
(571 mm). The following seasons had
negative net potential climatic water
balance 1969/1970 with 525.6 mm of
precipitation, 1972/1973 with 425 mm of
precipitation, 1973/1974 with 431.4 mm
of precipitation, 1977/1978 with 599.2
mm of precipitation and 1983/1984
with 544.3 mm of precipitation. In other
seasons of the perennial 1964-1986
potential climatic water balance was
positive until April, whereas after that
period became negative. Since 1986/1987,
when net potential climatic water balance
was positive, in every coming year water
deficit was growing. Even in 1997, the
year of flood in Wrocław, precipitation in
July did not manage the water shortage
to vanish. Deepening water deficits are
clearly seen in the last two decades of the
XXth century (Fig. 5a, b, c, d, e, f).
Water deficit or excess during
vegetation period have significant
influence on the values of heat balance
components. It follows from the rule
of evapotranspiration priority, that if
there is sufficient amount of water in the
substrate, the excess of energy is used
for evapotranspiration prior to heating
air and soil. Therefore, this rule connects
106
a
1986/1987
150
100
50
0
X
X-XI
X-XII
X-I
X-II
X-III
X-IV
X-V
X-VI
X-VII
X-VIII
X-IX
X-V
X-VI
X-VII
X-VIII
X-IX
X-V
X-VI
X-VII X-VIII
X-IX
X-V
X-VI
X-VII X-VIII
X-IX
-50
b
1987/1988
100
50
0
-50
-100
-150
-200
-250
X
X-XI
X-XII
X-I
c
X-II
X-III
X-IV
1988/1989
0
-100
X
X-XI
X-XII
X-I
X-II
X-III
X-IV
-200
-300
-400
-500
d
1996/1997
0
X
-50
-100
-150
-200
-250
X-XI
X-XII
X-I
X-II
X-III
X-IV
e
107
1997/1998
100
0
X
X-XI
X-XII
X-I
X-II
X-III
X-IV
X-V
X-VI
X-VII X-VIII
X-IX
X-V
X-VI
X-VII X-VIII
X-IX
-100
-200
-300
f
1999/2000
50
0
-50
-100
-150
-200
-250
-300
X
X-XI
X-XII
X-I
X-II
X-III
X-IV
FIGURE 5. Changes of potential climatic water balance (PCWB) in Wrocław-Swojec for the period
X–IX in years 1986–2000
climatic water balance and heat balance.
Thus, in the following investigations
heat balance components and their
contribution in net radiation were
calculated for ecosystems of coniferous
forest, potatoes and spring wheat (Fig.
6, 7, 8). The consequences of deepening
water deficits in Wrocław-Swojec during
the last 20 years are as follows :
1. Decreasing latent heat flux used for
evapotranspiration;
2. Slow increase of sensible heat flux
used for heating atmosphere.
Increasing values of Bowen’s ratio
(H/LE) [Bowen 1926] for coniferous
forest, potatoes and spring wheat prove
sensible heat flux enlargement in the last
20 years (Fig. 9). The indicated above
tendencies suggest the climate warming
up in this area.
CONCLUSIONS
1. The 1964–2000 perennial in Wrocław-Swojec is characterized by increased yearly sums of potential
evapotranspiration and decreasing
trend in yearly precipitation sums.
2. The analysis of a yearly course of
potential cumulative climatic water
balance in the season from October
to September revealed growing
water deficits in this region, more
evident in the last decades of the
XXth century.
3. Growing water deficits in this area
are the cause of increasing sensible
heat flux in the vegetation period
for coniferous forest, spring wheat
and potatoes, as well as decreasing
tendency for latent heat flux.
108
FIGURE 6. Variation of mean ten-days values of sensible heat flux (H) during the growing season four
coniferous forest, potatoes and spring wheat in Wrocław-Swojec (1964–2000)
FIGURE 7. Variation of mean values of sensible heat flux and net radiation ratio (H/Rn) during
the growing season for coniferous forest, potatoes and spring wheat in Wrocław-Swojec (1964–2000)
109
FIGURE 8. Variation of mean values of latent heat flux and net radiation ratio (LE/Rn) during the growing season for coniferous forest, potatoes and spring wheat in Wrocław-Swojec (1964–2000)
FIGURE 9. Variation of mean ten-days values of the Bowen’s ratio (H/LE) during the growing season
for coniferous forest, potatoes and spring wheat in Wrocław-Swojec (1964–2000)
110
4. Due to increasing values of sensible
heat flux used for heating the
atmosphere, the air temperature in
this region rose 0,03oC per year,
which means warming up in this
area.
REFERENCES
BAC S., ROJEK M. 1977: Metodyka oceny
stosunków wodnych obszarów rolniczych
na podstawie danych klimatycznych,
Zesz. nauk ART Olszt. Nr 21: 13–24.
BAC S., ROJEK M. 1979: Klimatyczny
bilans wodny a odpływy w Polsce, Przegl.
Geofiz. 24(3): 293–297.
BAC S., ROJEK M. 1982: Klimatyczne
bilanse wodne w Polsce. [w:] Bac S. (red.)
Agroklimatyczne podstawy melioracji
wodnych w Polsce. PWRiL, Warszawa.
BOWEN I.S. 1926: The ratio of heat losses
by conduction and by evaporation from
any water surface. Phys. Rev., 27, p.
779–787.
KĘDZIORA A. 1999: Podstawy agrometeorologii, PWRiL, Poznań.
OLEJNIK J., KĘDZIORA A. 1991: A model
for heat and water balance estimation and
its application to land use and climate
variation., Earth Surface Processes
Landforms vol. 16 ss. 601–617.
PENMAN H.L. 1948: Natural evaporation
from open water, bare soil and grass.
London: Proc.Royal Soc. Vol. 193, 120–
146.
PENMAN H.L. 1950: Evaporationover the
BritishIsles, Q.J. Roy. Met. Soc., 76
372–83.
PENMAN H.L. 1956: Evaporation: an
introductory survey, Netherlands J. Agric.
Sci. 4, –29.
PENMAN H.L. 1963: Vegetation and
Hydrology, Tech. Comm. Nr 53, Comm.
Bur. of Soils, Harpenden.
ROJEK M.M., WIERCIOCH T. 1995:
Zmienność czasowa i przestrzenna
parowania wskaźnikowego , ewapotranspiracji aktualnej i niedoborów opadowych w Polsce nizinnej w okresie 1951–
–1990, ZNAR Nr 268, Monografie VI.
ROJEK M. 1994: Time variablity of climatic
water balances in selected meteorological
stations in Poland. Zesz. Probl. Post.
Nauk Rol., 405, 147–153.
ROJEK M., WIERCIOCH T. 1994: Indicatory
evapotranspiration os summer half-year
in various long-term periods. Zesz. Probl.
Post. Nauk Rol., 405, 155–161.
Streszczenie: Zmienność klimatycznego bilansu
wodnego i bilansu cieplnego różnych ekosystemów
dla Wrocławia w latach 1964–2000. Praca zawiera
charakterystykę składowych bilansu klimatycznego bilansu wodnego we Wrocławiu w latach 1964–
–2000. W analizowanym okresie obserwuje się
trend wzrostu rocznych sum ewapotranspiracji
potencjalnej oraz spadku rocznych sum opadów
atmosferycznych. W związku ze wzrostem strumienia ciepła jawnego, ogrzewającego atmosferę
obserwuje się wzrost średniej rocznej temperatury powietrza o 0,03°C. Ten wzrost temperatury
związany jest ze wzrostem deficytu wody dla lasu
liściastego, pszenicy jarej i ziemniaków oraz rosnącym strumieniem ciepła jawnego przeznaczonego na ogrzanie atmosfery.
MS received December 2006
Authors address:
Elżbieta Musiał, Joanna Bubnowska,
Edward Gąsiorek
Department of Mathematics
Agricultural University of Wrocław
ul. Grunwaldzka 53, PL-50357 Wrocław
Polska
The research supported by KBN grant in
years 2004–2007.
Investigation for biological nitrogen removal from wastewater using
simultaneous nitrification/denitrification technology
GIEDRÉ VABOLIENÉ
Department of Water Supply and Management, Vilnius Gediminas Technical University, Lithuania
Abstract: Investigation for biological nitrogen
removal from wastewater using simultaneous
nitrification/denitrification technology. Biological
nitrogen removal from the wastewater is based on
the nitrification and denitrification processes at
the biological treatment plant with the activated
sludge. Different technological schemes can
be used for the above mentioned processes.
“BioBalance” technology as the newest way
of nitrogen and phosphorus removal has been
applied at Utena Wastewater Treatment Plant.
“BioBalance Symbio” technology for the nitrogen
removal is based on the active sludge technology
with simultaneous nitrification/denitrification.
The aeration zone and the anoxic zone are in
one tank. The nitrification and denitrification are
carried out during the aeration switching on and
off. Nitrification and denitrification processes
have been estimated during five experiments in
the aeration tanks, when durations of aeration
and low aeration were from 120 to 180 min. The
impact of aeration regime on biological nitrogen
removal has been estimated during this scientific
work.
Key words: biological nitrogen removal,
nitrification, denitrification, nitrate, ammonium
nitrogen, biological active potential (BPA).
INTRODUCTION
Biological nitrogen removal from the
wastewater is based on the nitrification
and denitrification processes at the
biological treatment plant with the
activated sludge (Berţinskienë 1999).
The nitrification is carry out by the
nitrification bacteria, which are divided
into two groups. The first group
bacteria (Nitrosomonas, Nitrosospira,
Nitrosococcus, Nitrosolobus genera)
oxidise the ammonium hydrate to nitrite.
The second group bacteria (Nitrobacteria,
Nitrospira, Nitrococcus genera) oxidise
nitrite into nitrate (Bitton 1994). The
energy is produced during both stages
of the nitrification. Nitrification bacteria
consume the energy for CO2 assimilation.
Thus nitrification bacteria are autotrophic
to the carbon. Nitrification bacteria grow
significantly slower than other bacteria
during the intensive nitrification process
(Droste 1997).
The nitrogen cycle ends by its return
back to the atmosphere in nature i.e. by
denitrification. Denitrification is one of
anaerobic respiration variation (nitrate
respiration), when NO3 ions are being
used as a final electrons acceptor for
the oxidation of organic substances.
The denitrification is oxygen split from
the nitrate, after that – from nitrite by
denitrification bacteria. The denitification
is being carried out by heterotrophic
bacteria: pseudomonas, spirillium,
tiobacillus, alkaligenes, bacillus. They
use organic pollution, existing in the
wastewater as carbon source. There are
two main stages in the denitrification.
None of the stages will take place till
there will be any sufficient amount of the
dissolved oxygen. The first is carrying out
112
G. Vaboliené
nitrate reduction up to nitrite; the second
is reducing nitrite up to nitrogen gas.
During the second stage nitrites change
nitrate as electrons acceptors, and the
nitrogen gas changes nitrite as a product
to finish the denitrification reaction.
Different technological schemes,
following the source of carbon oxidation
during denitrification processes, can be
used for the above mentioned processes
(Henze et al. 1995). The 1st scheme,
i.e. carbon source – carbon of raw
wastewater. The scheme is composed of
the denitrification tank, nitrification tank,
and secondary clarifiers. The 2nd scheme,
i.e. carbon sources – additionally added
easily biodegradable organic substances
(methanol, ethanol, etc.). The scheme
is composed of the nitrification tank,
denitrification tank, additional aeration
tank, and secondary clarifiers. The 3rd
scheme, i.e. carbon source – carbon
of raw wastewater. According to this
scheme the aeration zone and the anoxic
zone are in one tank. The nitrification
and the denitrification is carry out
during the aeration switching on and off
(Matuzevičius et al. 1998).
Usually schemes of the nitrogen
removal are combined with phosphorus
removal schemes. Different new
technologies are used to decrease
phosphorus and nitrogen quantity in
the wastewater. The mostly advanced is
“BioBalance” technology, where nitrogen
is removed following the 3rd nitrogen
removal scheme, and the anaerobic
zone is equipped before the nitrification/
denitrification tank the phosphorus
biological removal. The nitrogen removal
according to “BioBalance Symbio”
technology carried out in a simultaneous
way of nitrification and denitrification;
controlled the oxygen supply by NADH
sensor according to sludge activity.
Commonly,
there
exist
co-ferments that function as suppliers
of hydrogen and electrons and are
nicotinamideadeninedinucleotide NAD+
and its phosphoresced compound
derivative NADP+. Reduced NAD(P)H
form, can be oxidised once more during
the formation of ATP. During the
processes of metabolism co-ferments
NADH and NADPH are produced in
all microorganisms. NAD(P)H quantity
in the microorganisms produced
depend on their activity. The abovementioned activity depends on existing
sludge loads, supplied in the form of
nutritious matters. This means that,
for example, NAD(P)H quantity will
remain stable under the conditions of
constant quantity of nutritious matters
supplied to the active sludge system,
the constant load and sludge activity.
The sludge activity and at the same time
NAD(P)H production increase gradually
at the same time with an increase of the
sludge load because of a more intensive
supply of nutritious matters. Differently,
NAD(P)H production decreases during a
decrease of the nutritious matter supply.
During research this was determined
by linear dependences (Norgard et
al. 1996). “BioBalance” measuring
equipment NADH fluorescensor is
based on the process mentioned above.
The fluorecsensor fixes only NADH
that gives information on how much
energy the microorganisms possess. The
information on the energy is defined as the
biological active potential (BPA). NADH
fluorescensor controls denitrification
and nitrification process according to
the sludge activity in a single aeration
Investigation for biological nitrogen removal...
tank. Ammonia is oxidised to nitrates by
nitrifying bacteria during the nitrification
phase. The nitrates are reduced to
molecular nitrogen (N2) by denitrifying
bacteria during the denitrification
phase. These processes are carried out
periodically. Organic matter is oxidised
bacterially during both the nitrification
and denitrification phases, with oxygen
and nitrates, respectively, as the oxidising
agents and phosphorus is absorbable with
special microorganisms.
The aim of the work was to evaluate the
impact of aeration regime on biological
nitrogen removal from wastewater using
simultaneous nitrification/denitrification
technology.
MATERIALS AND METHODS
Researches have been carrying out
during period (July–November, 2005)
at Utena Wastewater Treatment Plant
in Lithuania. Two aeration tanks have
been used for the biological wastewater
treatment in Utena wastewater treatment
plant. The nitrogen removal according to
“BioBalance Symbio” technology carried
out in a simultaneous way of nitrification
and denitrification; controlled the oxygen
supply by NADH sensor according to
sludge activity.
113
Two aeration tanks worked at the same
time. Wastewater-flow after mechanical
treatment was divided into both aeration
tanks simultaneously. However, the
regime of aeration in each aeration tank
was different. Aeration can be carried
out in one aeration tank, in the meantime
low aeration can be carried out in the
other aeration tank. The term of aeration
and low aeration can be different in both
aeration tanks. Because of the abovementioned reasons, this research’s
results were analysed of both aeration
tanks. Low oxygen concentration
0.1–0.5 mg O2 l–1 has been held until the
denitrification carries out and the higher
oxygen concentration 0.5–1 mg O2 l–1
has been maintained for the nitrification
processes carry out periodically.
Conditions for the nitrification process
were held from 120 to 180 min, and
after that 120–180 min were spent for
the denitrification. The duration of the
nitrification and the denitrification had
been held in five different regimes during
experiments (Tab. 1). The researches
from 2 to 9 occasions in each aeration
tank have been carrying out during
different aeration regime.
The total nitrogen concentration
in wastewater after the mechanical
treatment, ammonium nitrogen and
nitrate concentration in the beginning of
aeration and by the end of aeration rate
TABLE 1. The duration of the nitrification and the denitrification during experiments
Number of experiment The duration of aeration rate, min
The duration of low aeration rate, min
1
150
150
2
4
180
180
120
120
150
120
120
180
5
150
180
3
114
G. Vaboliené
in the both aeration tanks, total nitrogen
concentration in wastewater after
biological treatment had been measured
during all experiments. The active sludge
concentration and the volatile suspended
solids of active sludge in the aeration tank
had been measured. The average flow in
the aeration tank had been fixed. The
efficiency of total nitrogen removal and
the active sludge load in the aeration tank
had been estimated. Total nitrogen had
been estimated adding Kjeldahl nitrogen
to nitrites and nitrates nitrogen. All
analysis was carried out using standard
methods (LST ISO 5815:1989).
RESULTS
Total nitrogen, [mgN/l]
The total nitrogen concentration in
wastewater after the mechanical treatment
fluctuated from 37 to 52 mg/l, average 44
mg/l during five experiments (Fig. 1).
The first experiment carries out 21
day during period (July 19 – August 8).
Conditions for the nitrification carried
out were held 150 min, and 150 min – for
the denitrification. The researches were
carrying out 9 times in the both aeration
tanks. The gathered results proved that
the efficiency of total nitrogen removal
fluctuated from 87 to 91%, average 88%.
At start of experiment, the concentration
of ammonium nitrogen in the beginning
of aeration rate fluctuated from 0.84 to 1.4
mg/l, by the end of aeration rate – from
0.47 to 0.91 mg/l (Fig. 2). However, the
concentration of ammonium nitrogen 1
mg/l was estimated in one aeration tank
by the end of aeration rate after 7 day of
experiment start.
After that, aeration tanks worked at
mentioned above aeration regime yet one
week. Then it was estimated that partial
nitrification was carried out in the both
aeration tanks by the end of aeration rate.
The concentration of ammonium nitrogen
by the end of aeration rate only increased
from 1.4 to 2 mg/l till experiment was
finished. Following experimental results
can be seen, that the duration of aeration
was not enough. The efficiency of total
55
50
45
40
35
30
1
2
3
4
5
6
7
8
9
aeration and reduced aeration rate 150 min
aeration rate 180 min, reduced aeration rate 150 min
aeration rate 180, 120 min, reduced aeration rate 120 min
FIGURE 1. Total nitrogen concentration in wastewater after the mechanical treatmen
N-NH4 , N-NO3,[mg/l]
a
115
3
2,5
2
1,5
1
0,5
0
1
2
3
4
5
6
7
8
9
10
b
N-NH4 , N-NO3 [mg/l]
N-NH4 in the beginning of aeration rate
N-NO3 in the beginning of aeration rate
N-NH4 by the end of aeration rate
N-NO3 by the end of aeration rate
2,5
2
1,5
1
0,5
0
1
2
3
4
5
6
7
8
9
FIGURE 2. The concentration of ammonium nitrogen and nitrate in the beginning and by the end of
aeration rate at first (a) and second (b) aeration tank when aeration regime: aeration and reduced aeration rate – 150 min
nitrogen removal decreased from 91% at
experiment start to 87% at the experiment
finish. The concentration of nitrate was
changing in small interval during first
experiment: in the beginning of aeration
rate 0.1÷0.6 mg/l, by the end of aeration
rate 0.45÷1.2 mg/l (Fig. 2). A low
concentration of nitrate in the beginning
of aeration rate provided that 150 min.
reduced aeration rate was enough for
complete denitrification carry out.
Consequently results of first
experiment indicated the condition
of second experiment. The second
experiment carries out 11 day during
period (August 9–19). Conditions
for the nitrification carried out were
held 180 min, and 150 min – for the
denitrification. It was extended duration
of aeration for complete nitrification.
The researches were carrying out 8 times
in the both aeration tanks. The obtained
results proved that the efficiency of
116
G. Vaboliené
total nitrogen removal fluctuated from
83 to 89%, average 85%. During two
days of the second experiment the
concentration of ammonium nitrogen in
the beginning of aeration rate decreased
from 3,1 to 0.8 mg/l (Fig. 3). Process
of nitrification carries out completely.
Duration of aeration was enough till the
end of experiment. The concentration of
nitrate changed more than during first
experiment. The concentration of nitrate
by the end of aeration rate exceeded 1
mg/l several time. The efficiency of
a
total nitrogen removal decreased from
89% at experiment start to 83% at the
experiment finish.
The third experiment carries out
during 7 days period (August 22–28).
Aeration tanks worked in follow aeration
regime: aeration rate – 180 min, reduced
aeration rate – 120 min first 5 day of
experiment. Later aeration regime
was changed: aeration rate – 120 min,
reduced aeration rate – 120 min. The
researches were carrying out 5 times in
first aeration regime and 2 times in second
N-NH4 , N-NO3 , [mg/l]
3,5
3
2,5
2
1,5
1
0,5
0
1
2
3
4
5
6
7
8
b
N-NH4, N-NO3 , [mg/l]
2,5
2
1,5
1
0,5
0
1
2
3
4
5
6
7
8
aeration rate at first (a) and second (b) aeration tank, when aeration regime: aeration rate 180 min, reduced aeration rate – 150 min
aeration regime in the both aeration
tanks. The gathered results proved that
the efficiency of total nitrogen removal
fluctuated from 73 to 80%, average 77%.
Though the concentration of ammonium
nitrogen in the beginning of aeration rate
fluctuated from 0.13 to 2.2 mg/l, but it
already decreased and fluctuated from
0.01 to 0.96 mg/l by the end of aeration
rate (Fig. 4). However nitrate increased
signally by the end of aeration rate. The
concentration of nitrate was estimated
from 0.47 to 11 mg/l in the beginning
of aeration rate and 1.2÷17 mg/l by
a
117
the end of aeration rate. The reduced
aeration rate was preferred for complete
denitrification, so the efficiency of total
nitrogen removal decreased to 73%.
Then the aeration regime was changed
and aeration tanks worked in other
regime (aeration rate – 120 min, reduced
aeration rate – 120 min) last 2 day of
experiment for reduction of nitrate. Then
aeration rate was changed from 180
min to 120 min concentration of nitrate
decreased to 0.75 mg/l in the beginning
and by the end of aeration rate during 2
18
16
14
12
10
8
6
4
2
0
1
2
3
4
5
6
7
b
18
16
14
12
10
8
6
4
2
0
1
2
3
4
5
6
7
aeration rate at first (a) and second (b) aeration tank, when aeration regime: aeration rate 180, 120 min,
reduced aeration rate – 120 min
118
G. Vaboliené
day of experiment. The efficiency of total
nitrogen removal increased to 73%.
Next experiment was carrying
out interchanging aeration regime.
Aeration rate reduced from 180 min.
to 120 min., and reduced aeration rate
prolonged from 120 min. to 180 min.
for reduction of nitrate. The fourth
experiment carries out 7 day during
period (August 29–September 4). The
researches were carrying out 7 times in
the both aeration tanks. The gathered
results proved that the efficiency of total
nitrogen removal fluctuated from 73 to
a
88%, average 83%. The nitrification
carried out completely first five day
of experiment. The concentration of
ammonium nitrogen fluctuated from
0.31 to 1.4 mg/l, 0.02–0.7 mg/l in the
beginning and by the end of aeration rate
respectively (Fig. 5). However later (two
last day of experiment) the concentration
of ammonium nitrogen increased in the
beginning and by the end of aeration
rate. Only partial nitrification carried out.
Consequently concentration of nitrate
decreased progressively and decreased
until 0.5, 0.31 mg/l in the beginning of
N-NH4, N-NO3, [mg/l]
7
6
5
4
3
2
1
0
1
2
3
4
5
6
7
b
10
8
6
4
2
0
1
2
3
4
5
6
7
aeration rate at the each aeration tanks
respectively and 0.89 and 0.33 mg/l by the
end of aeration rate at the each aeration
tanks respectively during third day of
experiment. Enough duration of reduced
aeration for completely denitrification
had positive impact on efficiency of total
nitrogen removal, which increased to
88% later decreased to 82%.
The fifth experiment carries out
74 day during period (September 5
–November 18). Conditions for the
nitrification carried out were held 150
a
119
min, and 180 min – for the denitrification.
The researches were carrying out 7 times
in the both aeration tanks. The gathered
results proved that the efficiency of
total nitrogen removal fluctuated from
80 to 89%, average 85%. Following
concentration of ammonium nitrogen
0.44–4.6 mg/l and 0.06÷0.65 mg/l in
the beginning and by the end of aeration
rate respectively, nitrification process
carry out completely (Fig. 6). However
concentration of nitrates (3.8 mg/l by
the end of aeration rate) showed partial
5
4
3
2
1
0
1
2
3
4
5
6
7
b
4
3,5
3
2,5
2
1,5
1
0,5
0
1
2
3
4
5
6
7
120
G. Vaboliené
denitrification process. It is proved that
the duration of reduced aeration was not
enough for denitrification. The efficiency
of total nitrogen removal decreased from
89% at experiment start to 80% at the
experiment finish.
The efficiency of total nitrogen
removal fluctuated from 73 to 91%
during five experiment when aeration
tank worked in different aeration regime
(Fig. 7). The total nitrogen concentration
in wastewater after the biological
treatment fluctuated from 3.9 to 12 mgN/
/l during five experiments.
CONCLUSIONS
Efficiency of nitrogen removal,%
1. Using biological nitrogen removal
technologies, when nitrification
and denitrification are in process
changing aeration intensiveness, it
is very important properly evaluate
aeration and low aeration duration.
2. When total nitrogen concentrations
of mechanically treated wastewater
ranged from 37 till 52 mgN/l, it
will be proper to use two aeration
regimes: 150 minutes aeration and
150 minutes low aeration rates,
similarly 180 minutes aeration and
150 minutes low aeration rates.
3. During 120 minutes conditional
short aeration rate and prolonged low
aeration rate, presently ammonium
nitrogen increase, following partial
denitrification outcome. Working in
this regime it is possible to obtain
completed nitrification.
4. During experimental research it was
obtained, when aeration and low
aeration rates are similar or slightly
different approximately 30 minutes,
nitrogen removal effectiveness is the
biggest.
ACKNOWLEDGEMENT
This research study is dedicated to EC
financed FP6 project MAPO „Enhancing
Research and Development Projects
to find Solutions to Struggle against
various Marine Pollutions”. The author is
involved into deliverable on innovation
roadmap for technologies based on
95
90
85
80
75
70
1
2
3
4
5
6
7
8
9
aeration and reduced aeration rate 150 min
aeration rate 180, 120 min, reduced aeration rate 120 min
FIGURE 7. The efficiency of total nitrogen removal when aeration tank worked in different aeration
regime
gaps within current state-of-the-art.
I’m gratefully to all Small Medium
Enterprises participated into MAPO
project activities; more Small Medium
Enterprises working into related to
marine pollutions areas are welcome to
join future activities.
REFERENCES
BERŢINSKIENË
J.
1999:
Water
microbiology.
Textbook
Vilnius:
Technika, p. 144.
HENZE M., HARREMOES P., JANSEN C.,
ARVIN E. 1995: Wasterwater Treatment.
Biological and Chemical Processes.
Springer-Verlag, p. 383.
BITTON G. 1994: Wastewater microbiology.
Wiley-Liss, New York, p. 456.
DROSTE R. 1997: Theory and practice of
water and wastewater treatment, p. 954.
MATUZEVIČIUS A., PAULAUSKIENË
Z. 1998: Experimental researches of
phosphorus and nitrogen consumption
and nitrogen removal from the
wastewater in the biological treatment
plant. 3rd International Conference.
Cities Engineering and Environment,
VGTU, “Technika”, p. 137–142.
NORGARD P., HELMO K., SORENSER E.
1996: Purification process for nitrogen
removal controlled by NADH. Vand og
Jord, Danish, vol 3, p. 126–129.
TSAI M.W., WENZEL M.C., EKAMA G.A.
2003: The effect of residual ammonia
concentration under aerobic conditions
on the growth of Microthrix parvicella in
biological nutrient removal plants. Water
Research 37, p.p. 3009–3015.
HATZICONTINOU G.J., ANDREADKIS A.
2002: Differences in nitrification potential
between fully aerobic and nitrogen removal
activated sludge systems. Water Science &
Technology. Vol. 46, No 1/2, p. 297–189.
JENICEK P., SVEHLA P., ZABRANSKA J.,
DOHAYOS M. 2004: Factors affecting
nitrogen removal by nitrification/
121
/denitrification. Water Science &
Technology. Vol. 49, No 5/6, p. 73–79.
Lithuanian Ministry of Environment Protection.
1994: Unified methods of the wastewater
and surface water quality researches.
Chemical analysis methods. Part I. 224 p.
Lithuanian Ministry of Environment
Protection, 2002: LAND 47-1:2002.
(ISO 5815:1989), 17 p.
Protection, 2000: LAND 32-2000.10 p.
Protection, 2005: LAND 66-2005. 7 p.
Protection, 2005: LAND 58:2003. 24 p.
Streszczenie: Badania biologicznego usuwania
azotu ze ścieków za pomocą technologii jednoczesnej nitryfikacji i denitryfikacji. Metoda biologicznego usuwania azotu ze ścieków oczyszczanych w komorach osadu czynnego opiera się na
procesach nitryfikacji i denitryfikacji. Do wyżej
wymienionych procesów można stosować różne
ciągi technologiczne. Technologia „BioBalance”
zaliczana do najnowszych metod usuwania azotu
i fosforu została wdrożona w oczyszczalni ścieków w Utenie (Litwa). Technologia „BioBalance
Symbio” do usuwania azotu wykorzystuje metodę
osadu czynnego z jednoczesnym prowadzeniem
procesów nitryfikacji i denitryfikacji. Strefy tlenowa i beztlenowa znajdują się w jednej komorze.
Procesy nitryfikacji i denitryfikacji są prowadzone poprzez naprzemienne włączanie i wyłączanie
napowietrzania komory. W pracy przedstawiono
wyniki badań przebiegu procesów nitryfikacji
i denitryfikacji podczas pięciu cykli badań w komorze napowietrzania w warunkach intensywnego i wolnego napowietrzania. W pracy oszacowano wpływ warunków napowietrzania na skuteczność usuwania biologicznego azotu.
MS. received July 2006
Author’s address:
Giedré Vaboliené
Vilnius Gediminas Technical University,
Dept. of Water Supply and Management,
Saulëtekio al. 11, AIF, LT-10228, Vilnius-40,
Lithuania.
Development of technologies used in agricultural engineering work
on an example of selected stages of land consolidation process
URSZULA LITWIN*, JAROSŁAW JANUS*, MARIUSZ ZYGMUNT**
*Department of the Geodesical Arrangement of Rural Settlements
**Department of Geodesy, Agricultural University of Cracow
Abstract: Development of technologies used in
agricultural engineering work on an example of
selected stages of land consolidation process.
The land consolidation process is one of the most
time-consuming and complicated agricultural
engineering tasks performed by geodesists. Their
automation and computerisation is not as advanced
as for other types of geodetic works. The reason
for that is a limited number of works that have
been performed in this range since the mid 1980s,
i.e. during the period of intensive IT (Information
Technology) implementation in geodesy. At
present one can observe again the growth in
interest in performance of land consolidations
processes. The purpose of this elaboration is to
introduce the latest solutions IT applied in land
consolidation process on an example of the land
consolidation in Wojków village.
Three characteristic stages of the process are
presented herein: creation of assessment register
before consolidation, preliminary selection of plots
for designed blocks and designing. Proposal for a
method for automation of the above-mentioned
activities using a digital map developed by means
of Bentley MicroStation software, retaining the
requirements imposed by land consolidation
instruction, was discussed.
The above-mentioned stages are illustrated
by examples from the land consolidation carried
out in Wojków village in years 2001–2004. The
work was performed on these premises using the
technology presented herein. It allowed reduction
of the time of land consolidation process by
approx. 30%.
Key words: spatial structure of village, land
consolidation.
INTRODUCTION
The efficient use of rural areas has
become one of the key problems in
the age of Poland’s membership of the
European Union. To use the potential
of rural areas effectively, among other
things a significant improvement in the
structure of the areas is required. One of
the problems to be solved is an excessive
dispersion and fragmentation of farms
[Urban 1981, Hopfer and Urban 1984,
Harasimowicz 1995]. The stagnation in
performance of the land consolidation
process that occurred in the 1980s
has intensified the problem. Only the
increased interest in land consolidation
process and related prospects for
significant growth in the number of
consolidations resulted in appearance
of the first IT tools to automate some of
the process’s stages [Janus and Zygmunt
2005]. This project describes selected
functionalities of the system supporting
modern land consolidation process,
which is the only system in Poland to
treat this process in a comprehensive
manner.
124
U. Litwin, J. Janus, M. Zygmunt
CREATION OF DESIGNED
BLOCKS
The land consolidation process the new
system was tested on was carried in
Wojków village, Mielec district in years
2001–2004. The area of the consolidated
land was 670 ha and comprised of 1503
plots. There was no digital map for the
area, therefore the first thing was to
convert the contents of analogue map
into the digital form. Bentley Systems
software was used for this purpose. The
digital map was checked for correctness
by the software the purpose of which was
to pick up any topological errors. Another
stage of the check was to compare a set
of plots on the map to the plots included
in the land register survey. The checks
ensured that register of lands before
consolidation was created properly. To
facilitate the work, the software had the
functionality of semi-automatic creation
of maps with comparative assessment.
The borderlines of assessment
contours occurred as a result of automatic
FIGURE 1. Division of object into 9 parts
selection of appropriate borders of plots
and facilities, such as: roads, rivers and
railway lines. For a digital map prepared
like that the software automatically
generates register of lands before
consolidation. The register is prepared
so that it can be printed in A3 format
[Litwin and Zygmunt 2005].
To enable multi-station operation
the object was divided into 9 designed
areas. The software has functionalities
allowing this operation to be performed
automatically (Fig. 1).
Each of the newly created designed
areas was then divided into designed
blocks (Fig. 2). Division into blocks may
also be carried out automatically.
Information on how the lines of
assessment contours run was used to define
boundaries of designed blocks. Relevant
function of the software defines a new
boundary upon the assessment contour
number is indicated. When designing
the technology, complete latitude in
definition of the boundaries of designed
blocks was adopted. The only condition
Development of technologies used in agricultural engineering...
125
FIGURE 2. Division of designed blocks within area no 01-Wojków
to be met by the designed block’s
boundary is topological correctness. The
topological correctness is controlled at
the operator’s request and any errors are
immediately communicated with their
precise location.
PRELIMINARY SELECTION OF
PLOTS FOR DESIGNED BLOCKS
Another stage of a land consolidation
process that has been automated in a
comprehensive manner is the preliminary
selection of plots for designed blocks.
This procedure is one of more important
stages of a land consolidation process. It
is performed after the land assessment
register is approved and before the stage
of designing a new land arrangement is
commenced. The purpose of the stage
is to make the initial assignment of
lands in individual farms to dedicated
designed blocks. The number and
sizes of the blocks most often result
from the run of boundaries of designed
invariants, such as boundaries of roads,
watercourses, administrative boundaries
and boundaries of building sites. The
essential problem that occurs during this
operation is the necessity of considering
all participants’ requests regarding the
location of the newly divided lands. The
appropriate performance of preliminary
selection activities is impeded by the
following factors:
– large number of participants
in the process (in case of land
consolidations carried out over large
areas it may reach a few thousands),
most often exceeding the number of
separated designed blocks by several
dozen times;
– necessity of considering diversification in designed blocks of classes
and arable lands when collecting
requests;
– necessity of as even collection of
requests for individual blocks as
possible so that further design of a
new land arrangement in accordance
with this selection would be
feasible.
126
To facilitate this operation the process
participants submit their requests in
more than one version to increase the
likelihood of proposing such a land
arrangement that the expectations of
most participants could be come up.
In the time of domination of analogue
technologies, this stage of the process
was extremely time-consuming and
any errors made during the stage were
decisive for any other operations carried
out during the whole land consolidation
process. In particular, making changes
to the already submitted requests and
control of acquired data took a lot of
time.
The purpose of the solutions designed
is a comprehensive support for the initial
selection stage, in particular:
– supporting the negotiations with
parties to the land consolidation
process by visualisation of the
farm’s lands against the demarcated
boundaries of designed blocks;
– ensuring that designed blocks are
filled by value assignments as
proposed by participants in the land
consolidation process;
– ensuring that the data being entered
are verified for conformity to
requirements of the consolidation
instruction and the land consolidation
and exchange act;
delivery
of
any
– immediate
information on the registration
unit and designed blocks that may
be useful on this stage of the land
consolidation process, including
data from the assessment register
before consolidation;
– preparation of final summary of
equivalents to be designed for every
farm and for every block.
The following items are used as input
data for designed solutions:
– data from assessment register before
consolidation;
– geometrical data to define the
boundaries of separated designed
blocks;
– layer of former condition plots;
– layer of classification contours;
– layer of designed blocks.
From the user’s point of view, the main
element of the system to support the
initial selection for designed blocks is
the specially designed form.
The most important components of
this form are:
– information panel containing, among
other things, data on the unit, former
condition plots included in it, data
on area and value of the unit, and a
lot of other information required by
the consolidation instruction;
– table containing information on
collected requests for the unit;
– figures and graphical information
showing the current filling status of
individual designed blocks;
– graphical window showing the
arrangement of former condition
plots, lands of active farm, boundaries
of designed blocks and classification
contours.
The following method for collecting
requests for a single farm was proposed:
– the first step is to select the active
registration unit, which results in
filling relevant information fields
with data to describe the unit;
– the next step is to select the value
for which a request in its basic and
alternative version is submitted;
– for every version the appropriate
designed block is selected;
Development of technologies used in agricultural engineering...
the last step is confirmation of any
decisions made.
It is possible to cancel each of the
decisions for any farm at any time,
which makes the specific land value
return to the appropriate designed block.
The final effect of this stage of a land
consolidation process is generating a
number of summaries that are required
further in the design work carried out
in the digital map environment and the
purpose of which is to design the target
land arrangement.
–
DESIGN FOR A PRESET VALUE
Designing starts with block selection.
It requires the number of block to
be indicated by the mouse. Next the
designing direction is defined and score
and number of designed plot are entered.
Data are entered using a special form.
Coordinates of the newly designed
plot are calculated then and boundaries
and number are inserted into the block.
The designing direction may be altered
at any time. The designed boundaries
may be deleted and the block that occurs
after such a boundary is deleted may be
subject to another division.
CONCLUSIONS
The approach to the land consolidation
process presented herein allows
achievement of significant economic
advantages resulting from, among
other things, reduction in duration of
individual operations. Choosing Bentley
MicroStation software as a system
platform enabled using a rich set of CAD
127
tools and allowed significant acceleration
of work performance.
The functions allowing quick redesign
of the plot arrangement made it possible
to transfer a lot of design variants to the
land consolidation process participants,
which effected in lack of complaints,
and thus completing the process on
time. Simultaneously with reduction in
duration of individual stages, conditions
were achieved for improvement in
quality of a land consolidation process as
regards its adjustment to the participants’
requests and through automatic detailed
verification of the project for conformity
to requirements of the consolidation
instruction and the land consolidation
and exchange act. The implementation
of the discussed IT technologies for land
consolidation process in Wojków village
confirmed the practical usefulness and
correctness of designed solutions. It also
needs to be mentioned that individual
elements of the proposed method for
carrying out this stage are subject to
constant modifications as the described
technology is implemented in successive
land consolidation processes in Poland.
REFERENCE
HARASIMOWICZ S. 1995: Wpływ cech
działki i gospodarstwa na wartość
dochodową gruntów. (Influence of plot
and farm features on revenue value of
lands). Zesz. Nauk. AR w Krakowie, ser.
Geodezja No 16, 77–86 (in Polish).
HOPFER A., URBAN M. 1984: Geodezyjne
urządzanie terenów rolnych. PWN.
Warszawa (Geodetic development of
agricultural areas) PWN, Warsawa (in
Polish)
JANUS J., ZYGMUNT M. 2005: Technologia kompleksowej automatyzacji prac
128
scaleniowych (Technology of comprehensive automation of land consolidation
process). Materials from the 17th Scientific and Technical Session in the series
“Current issues in geodesy and cartography”, Nowy Sącz (in Polish).
LITWIN U., ZYGMUNT M. 2005: Nowa
technologia generowania rejestru gruntów
przed scaleniem (New technology
for generation of land register before
land consolidation). Zesz. Nauk. AR
w Krakowie, ser. Geodezja No 21 (in
Polish).
URBAN M. 1981: Ekonomika i organizacja
gospodarstw rolnych (Economics and
organisation of agricutural farms), PWN,
Warszawa (in Polish).
W pracy przedstawiono trzy charakterystyczne etapy procesu scalenia: tworzenie rejestru
szacunkowego przed scaleniem, etap wstępnego naboru do kompleksów projektowych oraz
projektowanie. Omówiono propozycję sposobu
automatyzacji wymienionych czynności z wykorzystaniem mapy numerycznej opracowanej przy
pomocy programu Bentley MicroStation, z zachowaniem wymogów narzucanych przez instrukcję
scaleniową.
Wymienione etapy zilustrowano przykładami pochodzącymi ze scalenia przeprowadzonego
w latach 2001–2004 na gruntach wsi Wojków.
Prace na tym obiekcie wykonano z zastosowanej
przedstawionej technologii. Pozwoliło to skrócić
czas scalenia o około 30%.
Streszczenie: Rozwój technologii wykorzystywanych w pracach urządzeniowo-rolnych na przykładzie wybranych etapów scalenia gruntów. Prace
scaleniowe należą do najbardziej czasochłonnych
i skomplikowanych zabiegów urządzenioworolnych wykonywanych przez geodetów. Proces
ich automatyzacji oraz informatyzacji nie jest tak
zaawansowany jak w przypadku innych rodzajów prac geodezyjnych. Przyczyną takiego stanu
jest ograniczenie ilości prac z tego asortymentu
wykonywanych od połowy lat osiemdziesiątych,
czyli w okresie intensywnego wdrażania technik
informatycznych w geodezji. Obecnie można
zaobserwować ponowny wzrost zainteresowania
wykonywaniem scaleń gruntów. Celem autorów
było przybliżenie najnowszych rozwiązań informatycznych stosowanych w procesie scalenia
gruntów, na przykładzie obiektu scaleniowego
Wojków.
Urszula Litwin, Jarosław Janus,
Katedra Geodezyjnego Urządzania Terenów
Wiejskich
Akademia Rolnicza w Krakowie
30-059 Kraków
Al. Mickiewicza 24/28
[email protected]
Mariusz Zygmunt
Katedra Geodezji
Akademia Rolnicza
30-059 Kraków
A. Mickiewicza 24/28
[email protected]

Land Reclam. 37.indb - Annals of Warsaw University of Life

Transkrypt

Podobne dokumenty

Dimethicone impact on aeration of suppository mass in Unguator

the role of downpours in transformation of slopes in the polish

Time series analysis of reference crop evapotranspiration for Bokaro

pdf - Information Systems In Management