MATHEMATICAL MODEL OF

RAINFALL-RUNOFF
TRANSFORMATION - WISTOO

Ozga-Zielińska Maria
Institute of Environmental Engineering Systems
Warsaw University of Technology, Warsaw, Poland

Gądek Wiesław, Książyński Krzysztof,

Nachlik Elżbieta, Robert Szczepanek
Institute of Water Engineering and Water Management,
Cracow University of Technology, Cracow, Poland

ABSTRACT: Mathematical model WISTOO (Wizualizacja

Integralnego Systemu Transformacji Opadu w Odpływ –
Visualization of Integral Rainfall-Runoff Transformation System)
is the complex solution of rainfall-runoff transformation problem
with use of geographical information system. It is integral,
distributed model. Its structure includes main hydrological
processes, like interception, evapotranspiration, infiltration, surface
and subsurface runoff, groundwater runoff and recipients runoff.
Model was readjusted to mountainous and hilly watersheds.
Connection of hydrological processes with geographical
information systems in one model makes easier simulation of
different watershed land-use. Model was created at Cracow
University of Technology in cooperation with Warsaw University
of Technology.

WISTOO model is adopted for simulation with one hour step for
mountainous and hilly watersheds. Long-term simulations can be
conducted – up to 185 days (summer half-year). Model is
implemented for personal computers with Windows operating
system, and uses digital thematic layers (Fig. 1):
• digital elevation model,
• stream network,
• soil structure,
 • land-use structure.
In the algorithm, mainly raster square model of spatial phenomena
description is used. Raster size depends on scale, but computer
program can handle maximum 4 000 000 raster elements (cells).
Computer implementation of algorithm enables tracing of:
• hydrographs in ten arbitrary river cross-sections,
• spatial distribution of: net precipitation, velocity and
depth of surface and subsurface flow, soil moisture,
• influence of water reservoir.

Figure 1. Digital thematic layers used in WISTOO model.

From top – land-use, stream network, soils, topography

Computer program contains many helpful tools for data acquisition

and pre-processing e.g.:
• preparation of digital elevation model from analogue
maps,
• changes in thematic layers content,
 • temporal distribution into hour data of measurement data
collected 2-3 times a day.
General structure of the model is based on standard water cycle in
the watershed (Fig. 2). Input data for the model consist of:
• watershed parameters: terrain slope, coverage retention,
rainfall separation coefficient, surface permeability, soil
porosity coefficient, soil depth, etc.,
• spatial distribution of measuring stations and type of
measured data,
• spatial distribution of gauging stations on streams
(maximum 10),
• hydrometeorological data.
Those data sets are pre-processed by dedicated software into
WISTOO data format.
As WISTOO is physically-based mathematical model, good
watershed identification is an essential issue. None of parameters
used in the model is optimized. Parameters are mainly calculated
from digital thematic layers. Soil parameter determination in small
watersheds with uniform soil pattern is rather easy. When
watershed area increases, complicated soil pattern and soil
diversity makes determination of parameters more difficult. Model
needs also dense net of measuring stations, mainly precipitation
stations.

1. INTRODUCTION

1.1. Place of WISTOO model between other rainfall-

runoff models
In WISTOO model watershed runoff was solved based on
distributed parameters and two-dimensional in plane calculation
space. So it is not typical model simulating runoff in watershed
closing cross-section. Was assumed, that WISTOO model should
enable:
 • correct runoff simulation in watershed closing cross-
section,
• correct runoff simulation in any upstream cross-section,
• representation of spatio-temporal structure of surface and
subsurface runoff.
Those requirements caused, that model is based on hydrodynamic
equations, describing partial runoff processes.
Genesis of model development is related to need of
numerical assistance in flood protection planning for southern part
of Poland. This region is characterized be complex land-use
structure and sparse network of meteorological and hydrological
stations. Flood protection planning in this region requires:
• quantitative evaluation of flood risk, including places
exposed to potential intense water runoff,
• development of suitable behavior strategy and choice of
economically justified protective activities, including
mainly treatments minimizing intense runoff.
Such formulation of problem resulted in development of
simulation model, solving rainfall-runoff transformation in proper
spatial scale. Model structure matching runoff regime and support
of runoff description on its physical characteristics was a key to
success. So far application experiences proved, that such
construction of methodology fulfils established demands.
WISTOO model in details describes runoff process,
concerning its components and spatial scale. It can be seen as
advances rainfall-runoff model. This type of models, besides its
practical application, helps in identification of physical
characteristics of watershed runoff. In practical applications is used
mainly in determination of land-use change on spatio-temporal
runoff structure.

1.2. Limitations in model application

Assumed, high level of detail solution in WISTOO model, limits
its application possibilities. In general, model structure focuses on
mountainous watersheds, but can be also applied for hilly
watersheds. Application criteria can be expressed by following
conditions:
• possibility of surface and subsurface runoff representation
(including stream flow) by kinematic wave,
• possibility of groundwater runoff representation by one
reservoir, which area is equal watershed area and length is
equal to stream network length.
Those conditions are fulfilled in mountainous and hilly watersheds.
Number of meteorological stations determines precipitation supply
precision for analyzed area. It is assumed, that parameters
describing topography of the watershed, its surface and subsurface
structure, land-use – are identified locally on required level of
accuracy.

1.3. Required input data

WISTOO model needs input data in GIS format and adequate
meteorological data. Required thematic layers:
• digital elevation model - in raster format,
• hydrography network with elevations - verified for flow
capacity,
• soil structure and type – according polish or general
classification (chapter 4),
• land-use structure - according described classification
(chapter 4),
• location of measuring stations.
Hydrometeorological data recorded in one hour or daily system are
inputted directly by special procedure. This procedure sorts end
exports data in internal data format. First step is to prepare digital
thematic layers in order to generate watershed parameters. All
thematic layers must be in the same coordinate system with the
same resolution.
2. DESCRIPTION OF PARTIAL PROCESSES

2.1. Structure of the model

Structure of WISTOO model is based on well-known water cycle
in the watershed (Fig.2), where the following processes can be
distinguished: interception, evapotranspiration, infiltration, surface
runoff, subsurface runoff, groundwater runoff, and stream net
runoff.

EVAPOTRANSPIRATION

PRECIPITATION

INTERCEPTION

NET
PRECIPITATION

EXCESS SURFACE
INFILTRATION
RAINFALL RUNOFF

AERATION SUBSURFACE SUBSURFACE

S
O ZONE RECHARGE RUNOFF
I SATURATION GROUNDWATER GROUNDWATER
L ZONE RUNOFF
RECHARGE

TOTAL RUNOFF

Figure 2. Natural water cycle scheme

Developed calculation procedures describe dynamics of particular

process on accepted level of accuracy.
2.2. Spatial distribution of meteorological
parameters
Spatial distribution of meteorological parameters is main source of
information related to meteorological situation at certain moment
in the watershed. Distribution of meteorological values is
calculated based on ground-based measuring systems. Usually,
distribution methods are local and depend on measuring net
density. These methods are not universal, and in most of the cases
transfer to another region is impossible.

2.2.1. Basic methodical assumptions

Basic method in the model for spatial distribution of
meteorological parameters is inverse distance method IDM. It
enables estimation of meteorological parameters as continuous
function. In calculation, only the closest or all measuring stations
in the region are used, unnecessarily located in the watershed.
Basic assumption of this method is based on observation,
that influence from of measuring station on any point is inversely
proportional to distance from that point to the station. So influence
of any measuring station decreases with distance. Calculated
values for meteorological parameters at any point in the watershed
are determined using formula:

j
1
å(d n * Pk )
k =1
Pi = j
k
(1)
1
å n
k =1 d k

where: Pi - value of calculated parameter at point i, i=1,2,...m,

Pk - measured value at point k, k=1,2,...l,
j - number of measuring stations,
n - index exponent,
dk - distance form point i to measuring station k.
Index exponent n depends on watershed topography. Its value
changes from 1 for lowlands, to 3 for mountainous regions.

P2
P1

P4
x1 x2
x4

P5 P3
x 5 Px x3

Figure 3. Inverse distance method

Long-term observations in mountainous watersheds demonstrated

strong influence on topography on meteorological parameters.

2.2.2. Description of spatial precipitation structure used in

WISTOO model
In general, precipitation depth increases as terrain raises.
Relationship between precipitation depth and elevation is
described by gradient precipitation curve. Another factor
influencing gradient value is precipitation intensity. For intense
rainfall events, inclination of that curve is smaller then for average
events.
Precipitation depth is calculated from formula:

R 'x = R x + A (H 'x - H x ) R x (2)

where: R’x - rectified precipitation depth, [mm],

Rx - precipitation depth calculated by inverse distance
method, [mm],
A - gradient curve inclination coefficient, [1/m],
 H’x - transfer elevation, [m],
Hx - terrain elevation, [m].

Transfer elevation H’x is related to elevation of measuring points

above sea level. Its value is constant for certain measuring stations
net and can also be evaluated from inverse distance method (1)
using elevation of measuring station as parameter.
Gradient curve inclination coefficient is calculated as
follows:

R i - R i+1
A= (3)
(H i - H i+1 ) R i

where: Ri, Ri+1 - recorded precipitation depth for the lowest and the
highest station in the watershed, [mm],
Hi, Hi+1 - elevation above sea level for the highest and the
lowest station in the watershed, [m].

In high mountains rainfall inversion phenomenon can be

observed. Above certain elevation, precipitation depth decreases
with elevation. In such a case, in order to implement precipitation
depth – elevation relation, equation (2) must take into account
inversion process. It can be realized by transformation of elevation
for all the areas above inversion point into corresponding values
calculated from gradient curve. From modified digital elevation
model, spatial distribution of precipitation can be calculated using
only ascending relation (linear or non-linear).

2.2.3. Recommendations for practical application of the model

for meteorological parameters determination
Precipitation. Figure 4 demonstrates spatial distribution of
precipitation calculated using inverse distance method. Figure 5
shows precipitation distribution calculated by gradient IDM
concerning inverse point for the same data from Wielka Puszcza
watershed. Differences in spatial distribution from those two
methods are major – with constant maximum precipitation depth.

Figure 4. Spatial precipitation distribution calculated by inverse distance method

for Wielka Puszcza watershed

Figure 5. Spatial precipitation distribution calculated by gradient inverse

distance method concerning inverse point for Wielka Puszcza watershed
 In presented model, inverse distance method is the main used
for precipitation distribution calculation. This method needs
gradient curve to be elaborated. For many watersheds it is difficult
to archive, due to lack of ground-based measuring stations. But if
such information is available, model enables use of linear or non-
linear function.
Gradient method was successfully implemented in Wielka
Puszcza (Cracow University of Technology experimental
watershed located in Beskid Mały) and Czarny Dunajec (Tatra
Mountains) watersheds. In first watershed inversion point was
located at 550-570 m.a.s.l., in second at 1400 m.a.s.l.
Verification on historical data proved, that gradient curve
method determines spatial distribution of precipitation much better
than inverse distance method. In general, relative precipitation
distribution error is up to 50% smaller for gradient method than for
inverse distance method. It should be pointed out, that proper
location of measuring stations in those watersheds minimizes
errors in inverse distance method.
Other meteorological parameters are calculated using similar
methodology. Wind velocity and air temperature are strongly
related to measurement station location, so gradient method is
recommended rather than inverse distance method. Gradient curve
inclination coefficient A (3) is determined from current
measurements at measurement station. In pre-processing
procedures user can declare stations to calculate this coefficient.
Atmospheric pressure and sun radiation are treated as homogenous
in space.
In general, it is recommended that measurement stations
location should cover whole elevation range. Spatial distribution of
meteorological parameters uses two algorithms depending on
available data:
• only for precipitation,
• for precipitation, air temperature, air relative humidity,
wind velocity, sun radiation.
Selection of algorithm is automatic based on available data.
Second algorithm can be used when at least two measurement
stations collect all necessary meteorological data.

2.3. Interception

2.3.1. Description of the process

Interception process is based on temporal storage of rainfall by
plants. Most often for mathematical description of this process
Rutter model is used.

dc c
= (1 - p ) R - E p dla c < s
dt s
(4)
dc
= (1 - p ) R - E p - D o e b ( c - s) dla c ≥ s
dt

where: R - rainfall intensity, [mm/min],

p - rainfall separation coefficient, [-],
s - plants retention, [mm],
c - current water depth on plants, [mm],
Ep - potential evaporation, [mm/min],
Do - minimal dripping from plants when c = s, [mm/min],
b - dripping vs. water depth on plants curve coefficient,
[1/mm],
e – napierian base.
 precipitation

INTERCEPTION net precipitation

potential
evaporation

Figure 6. Interception calculation scheme

2.3.2. Interception process solution used in the model

In experimental way, simplified model simulating interception
process was elaborated. It was found that dripping process has
minor influence on rainfall-runoff transformation. So dripping
process can be disregarded, simplifying solution of this problem.
Net precipitation is calculated from relation:

Rnett = p R dla c ≤ s
(5)
Rnett = R dla c > s

where: Rnett - net precipitation, [mm/min].

Separation coefficient p defines which part of precipitation

reaches directly ground without wetting plants. Retention s defines
maximum precipitation depth, which wets plants. Value of this
parameter describes precipitation loss. In algorithm, for every
calculation time step current water depth on plants is determined.
In first phase of rainfall event for c <= s, net precipitation Rnett is
defined on the basis of rainfall separation coefficient p. When the
retention s is filled up, net precipitation Rnett has the value of total
precipitation R.
Evaporation process is considered when retention is filling up
and after rainfall event. Current water depth on plants is calculated
from relation:

co
c t = co + [ (1 - p) R - E p ] ∆t (6)
s
where: ct - current water depth on plants at time t, [mm],
co - water depth on plants at time t-1, [mm].

Potential evaporation depth is calculated evapotransipration

model. Water depth on plants c at the moment of filling retention s
is stabilized on s level. For early spring and late fall (April,
October) retention and separation coefficient values independently
on plants are taken arbitrary as s=0.1[mm], p=0.05.

2.4. Evapotranspiration

2.4.1. Description of the process used in the model

Evapotranspiration is the process, in which water changes phases
from liquid to gaseous (evaporation) or through plant’s metabolism
(transpiration). Mechanism of this process is very complex. As a
consequence, its descriptions are dependent on significance of
factors in specific cases. In presented model only two elements are
calculated:
• evaporation from wetted surfaces – potential
evapotranspiration,
• plant transpiration – current evapotranspiration (Fig.7).
Other types of evaporation like evaporation from water and soil
surfaces are not taken into account in the model.
 high plant's low plant's
transpiration transpiration

soil
moisture
change
subsoil

Figure 7. Current evapotranspiration process scheme

2.4.2. Description of evapotranspiration process used in the
model
Penman-Monteith equation is the basic describing
evapotranspiration process in the model:

∆ (Rn - G ) + ρ cp (ε - e e ) / ra
Ea = (7)
λ [∆ + γ (1 + rc /ra )]
where: Ea - current evapotranspiration, [mm],
∆ - inclination of water vapor pressure curve vs.
temperature, [hPa/K],
Rn - net radiation, [W/m2],
G - soil heat flux, [W/m2],
ρ cp - volumetric dry air heat capacity, [J/m3K],
ε, ee - maximal and current water vapor pressure, [hPa],
ra - aerodynamic resistance function, [s/m],
rc - superficial resistance function, [s/m],
λ - latent heat of evaporation, [J/kg],
γ - psychrometric constant, [hPa/K].

Potential evaporation is determined from equation (7) with

assumed superficial resistance coefficient rc equal zero. Value of
superficial resistance coefficient rc is calculated from:

ba
rc = a + (8)
R n + ca

where: a, ba, ca –empirical parameters determined from

experiments for two plant types – high plants (forest)
and low- cultivable plants: a [s/m], ba [sW/m3], ca
[W/m2].
Aerodynamic resistance coefficient ra was taken after
[Stigter C. J.,1980]:

2
é æ z v - d h öù
êlnçç ÷÷ú
è z øû
ra = 8.06 ë
o
(9)
1 + 0.864 v w

where: zv - altitude of wind velocity measurement, [m],

dh - base altitude dh =0.63 hr, [m],
zo - height of plants roughness zo = 0.13 hr, [m],
hr - average height of plants, [m],
vw - wind velocity, [m/s].

Calculation procedure starts when input data, like air

temperature and relative humidity, wind velocity and sun radiation
are prepared. Sun radiation can be substituted by cloud cover
values. Additionally, in pre-processing procedure, measuring
stations for determination of gradient curve coefficient A have to
be defined.
Often, location of meteorological stations eliminates
possibility of using this procedure due to lack of measured
meteorological parameters or inappropriate location. In order to
determine well evaporation values, meteorological parameters
must become from mountain tops and valleys. For most of
watersheds such observations are not conducted. In such a case
values of potential and current evaporation are set obligatory for
every month. Those values come from regional measurement
calculations or transferred from other similar watersheds.
In the model authors assume, that low plant’s transpiration
directly influences soil layer. For forest, evenly water intake from
soil surface and subsurface is assumed. This process starts in
periods without precipitation. During precipitation and just after
event evaporation values are calculated from plants wetting rate:
 c
Ea = Ea - Ea (10)
s

2.5. Infiltration

2.5.1. Process description and model assumptions

Infiltration is the process of percolation of part of net precipitation
through soil surface and displacement of water in aeration zone. It
is the most important process in every rainfall-runoff
transformation model. It decides on quantitative separation of net
precipitation into: surface runoff, subsurface runoff and
groundwater flow. From description of this process depends
accuracy of runoff hydrograph representation in watershed closing
cross-section.
Necessity of model implementation for every cell increased
model complexity. Every model has to take into account soil
retention and precisely describe transformation of precipitation in
soil profile.

net precipitation SOIL

excess rainfall
INFILTRATION

subsoil supply

soil supply

SUBSOIL
subsurface
INFILTRATION supply

gruntwater
supply

Figure 8. Scheme of procedure simulating infiltration process

 In hydrology, many infiltration models are implemented,
from empirical ones to more complicated using conductivity and
diffusion theory. Selection of appropriate model is always
compromise between expected result and computation
possibilities. Many computation cells highly limits use of diffusion
models due to number of calculation parameters and soil
characteristics. Not always there are available data.
From conductive type models, piston Green-Ampt model
was selected. Model assumes, that every precipitation event forms
wetting front described by two variables: location z and volumetric
humidity Θ. Humidity for every front is constant. Saturated front
movement velocity is described by Darcy equation.

2.5.2. Application of the model for infiltration

Front movement velocity.
Described by Darcy equation, saturated front movement velocity
has a form:

æ h + hk ö
v = K Sh = K ç 1 + w ÷ (11)
è z ø

where: v - infiltration rate, [m/s],

hk - capillary height, [m],
hw - water depth on soil surface, [m],
K - soil hydraulic conductivity, [m/s],
z - wetting front depth, [m],
Sh - hydraulic gradient, [-].

Hydraulic conductivity.
Hydraulic conductivity is estimated from formula:
 α
æ Θ - Θs ö
K(Θ) = K o çç ÷÷ (12)
è Θo - Θs ø

where: Ko - maximal soil conductivity (filtration coefficient),

[m/s],
Θs - volumetric humidity of permanent plant wilting, [-],
Θo - maximal volumetric humidity – soil porosity index, [-],
α - exponent index dependent on soil type: α ∈ á3; 4ñ, in
model α is equal 3.5.

Aeration zone structure and its description.

In natural watersheds streams are supplied from three layers:
surface, aeration zone (subsurface) and saturation zone
(groundwater). In the model, aeration zone is divided into two sub-
layers: soil and subsoil. This is not typical approach in
conductance models. Algorithm in presented model enables
calculations for those two sub-layers with water exchange between
them.
Solution of this problem was derived from infiltration
models using diffusion theory. In diffusion models water exchange
between nodes or layers is modeled by moisture flux q depending
on diffusion potential, usually described by main component of
soil suction pressure:
æ ∂ψ ö
q f (z, t) = - K(ψ ) ç - 1÷ (13)
è ∂z ø

where: qf - moisture flux, [m3/m2 min],

ψ - soil suction pressure, [m].

Relation between volumetric humidity Θ and soil suction

pressure ψ is usually presented in form of pF = log(-ψ) curve,
where ψ[m] is soil suction pressure. This relation is one of basic
soil characteristics.
 In the model, simplified, linear description of this
characteristic is used (Fig.9), with 0 assigned for humidity equal to
porosity coefficient and maximal value for dry soil. Water
displacement velocity is defined as:

æ ψ g -ψ d h ö
v = K çç + 1 + w ÷÷ (14)
è z z ø

where: Ψg - suction pressure in upper node, [cm],

Ψd - suction pressure in bottom node, [cm].

Expression Ψg - Ψd stands for capillary height hk. This placement

of potential difference between soil and subsoil in stead of
capillary height enables automatic control of infiltration process.
Water movement in new formula depends on bottom layer
humidity. When upper layer humidity decreases comparing to
bottom layer, hydraulic gradient decreases and as the result
infiltration intensity decreases. When humidity difference between
layers is high, water movement in soil layer is stopped until
humidity gradient decreases. In such situation infiltration rate has
negative values, it means, humidity transfer goes from subsoil to
soil layer. Lack in connection between soil and subsoil due to
higher soil hydraulic conductivity causes over-drying of soil layer.
In longer dry periods, humidity can reach values close to plant
wilting point.
 7
6
5
4
pF

3
2
1
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
humidity

Figure 9. Linear relation between suction pressure expressed in pF = log(-Ψ)

function and soil humidity, where Ψ, [cm]

Runoff occurrence conditions

Both surface and subsurface runoff appear when water inflow
exceeds infiltration capacity. Runoff starts when v < R (where: v –
infiltration rate, R – precipitation intensity). Excess height (R – v)
is input value for surface and subsurface runoff transformation
models. Subsurface runoff starts when wetting front passes through
surface part of the soil. Water velocity in this layer depends also
from water depth hw. This depth is determined from subsurface
runoff transformation model considering soil filtration coefficient
µ.
In every time step, moisture volume balance in soil and
subsoil is calculated. Supply volume of every layer is equal:

V = R ∆t (15)

and total volume of absorbed moisture:

3 3
V = å vi ∆t ∆Θi = å z i ∆Θi (16)
i =1 i =1
where: ∆Θi – width of humidity front, [-].

Real velocity of humidity front movement is equal:

K (Θ i ) - K (Θ i-1 )
vi = (17)
Θ i - Θ i-1

Based on field experiments, three fronts were used for

description of humidity movement: one basic and two variables
for movement during rainfall event description. Creation and
removal of fronts is based on assumptions:
• new fronts are created always when existing fronts are not
able to get supply water and maximal humidity is smaller
then soil porosity coefficient,
• if supply water volume exceeds absorption possibilities of
existing fronts, two fronts (with maximal moisture) are
connected into one, with volume equal sum of source
fronts and humidity equal humidity of removed front,
• in dry periods, moisture is transferred between fronts,
• if front with higher moisture content runs ahead of
another, those two fronts are connected,
• variable front is removed when reaches soil or subsoil
boundary (groundwater table) and removed front is
connected with basic front,
• front width (∆Θ) can not be smaller than 0.001[cm3/cm3],
in such a case, this front is added to adjacent front and
outcome front volume is a sum of two source fronts.

Outputs from the model are: exceed rainfall, input values for
subsurface runoff and groundwater runoff.
Initially conditions.
Key element is determination of initially conditions of the process,
i.e. initial humidity. This problem is solved with assumption that
volume of water supplying groundwater reservoir is equal
watershed runoff:
 nz nz
Q k = å Fi Rg i = å Fi Rp i (18)
i =1 i =1

where: Qk - groundwater runoff in watershed closing cross-section,

Fi - cell area,
Rgi - groundwater table supply,
Rpi - subsoil layer supply,
nz - number of elements supplying groundwater table.

Determination of initial humidity distribution starts from subsoil

layer in iterative mode. Initial condition is that humidity in every
element is equal porosity coefficient (maximal volumetric
humidity). Iteration process is over when volume of water
supplying groundwater reservoir is close to watershed runoff.
Calculations are conducted with 24 hours time step. Similar
procedure is used for soil layer humidity distribution. Initial
humidity is equal subsurface humidity. Balance of inflow and
outflow from the sub-soil layer is evaluated for each computational
element.

2.6. Surface and subsurface runoff

Surface runoff is the water flow process over terrain surface. It is
formed as result of excess rainfall over certain area, i.e. when
infiltration soil capacity is smaller than net rainfall intensity.
 excess rainfall

outflow from previous SURFACE

element runoff
RUNOFF

parcolation

supply from
subsurface runoff

Figure 10. Surface runoff calculation scheme

outflow from soil

surface runoff
supply

outflow from previous SUBSURFACE

element RUNOFF runoff

infiltration into
subsoil

Figure 11. Subsurface runoff calculation scheme

Subsurface runoff process covers horizontal water movement in
aeration zone. It is caused by water excess in this zone. This
phenomenon is a consequence of hydraulic gradient between soil
and subsoil layers. Soil hydraulic conductivity is two orders of
magnitude higher subsoil conductivity. Due to difference on
boundary between those layers, subsurface runoff from water
excess is formed.
2.6.1. Detailed description of runoff process
Surface and subsurface runoff processes are described by
kinematic wave equation:

∂ qx ∂ q y ∂ h
+ + = RE
∂x ∂ y ∂t (19)
q = αk h N

where: q - unit flow, [m3/m s],

h - water depth, [m],
RE - excess rainfall or subsurface supply, [m/s],
αk, N - kinematic wave coefficient.

In used model parameters are set like for turbulent flow, according
Manning theory:

Sz
αk =
nh
(20)
5
N=
3

where: Sz - terrain slope, [-],

nh - hydraulic resistance coefficient, called roughness
coefficient, [1/m1/3].

Assumption was made that water exchange between cells takes

place on the boundary between them. In presented procedure,
implicit four-point Preismann scheme was used. Partial
derivatives are written in the form:

∂ qx q i+1 - q i2+1 qi - qi
≈β 4 + (1 - β ) 4 2 (21)
∂ x ∆x ∆x
 ∂ qy q i+1 - q i3+1 qi - qi
≈β 1 + (1 - β ) 1 3 (22)
∂ y ∆y ∆y

∂ h h iśr+1 - h iśr
≈ (23)
∂t ∆t

RE = β R iE+1 + (1 - β ) R iE (24)

where: β - weight coefficient [0,5 ÷ 1,0],

RE - excess rainfall or subsurface supply, [mm/s].

q3

y k+1

q4 q2
hśr
X

yk
y
q1

xi x i+1
x

Figure 12. Cell supply scheme used in surface runoff procedure

Developing equation (19) for hśri+1 and assuming, that ∆x is equal

∆y we get nonlinear equation:
5/3
a h śr + b h śr + c = 0 (25)

αβ
where: a = -
∆x
 1
b =
∆t
β q id+1 (1 - β ) (q id - q io ) h iśr
c= + - - β R iE+1 - (1 - β ) R iE
∆x ∆x ∆t
qd - inflow sum for cell,
qo - outflow sum for cell.

For solution of equation (25) Newton-Raphson method was used,

where hśri+1 value is calculated from relation:

h iśr+1 = h io + ∆h (26)

f(h śr )
where: ∆h = - ,
f(h o )
ho – water depth prom previous time step.

Calculations are conducted until ∆h is smaller than arbitrary

accuracy ε.

2.6.2. Practical solution of the problem

Two numerical procedures are used for problem solution: for
surface runoff and subsurface runoff. Both procedures are based on
pre-processed data set describing calculation order. In topological
structure of the watershed cells forming it have:
• supplying cells,
• draining cells.
Important feature of those procedures is connection between each
other and relation to infiltration process. In subsurface runoff
transformation process, water depth can exceed layer depth. Part of
the water in this case moves from aeration zone to surface forming
secondary surface runoff.
In the model this problem was solved by determination of
maximal supply, which will not exceed runoff considering layer
depth in one time step. Excess runoff is transferred to surface
runoff procedure. Transformed water depth is calculated from soil
depth and filtering coefficient µ, according equation (27).

R gr =
( )
(q grj - q dj ) β + q oj-1 - q dj-1 (1 - β )
+
Hg µ - h śrj-1 (1 - β ) R Ej-1
- (27)
∆x β ∆t β β

where: qgr – maximal outflow from cell calculated from soil layer
depth Hg and filtering coefficient µ.

Water flowing on surface and in aeration zone encounters

different soil types and obstacles. For example, when water flows
from impermeable into natural surface (soil) infiltration process
starts. In this case direct influence between runoff and infiltration
takes a place. In general, checking soil infiltration capacity at every
cell solved this problem. When water supply is smaller than
infiltration capacities, and there is water on the surface, part of it is
taken. Part of the water taken is calculated from balance, with
assumption that the outflow from cell in time step tj is equal zero.
Then:

Rw =
( )
- q dj β + q oj-1 - q dj-1 (1 - β ) h śrj-1 (1 - β ) R Ej-1
- - (28)
∆x β ∆t β β

Calculated from equation (28) soil and subsoil infiltration depth is

source part of infiltration process. When sum depth (supply R and
infiltration Rw) exceeds infiltration capacity of the layer, only part
of the water needed for balance compensation resulting from
potential infiltration is taken. Infiltration intensity is negative part
of equation describing surface and subsurface transformation
process.

2.7. Groundwater runoff

Groundwater runoff consist of streams supply from saturation
zone. In general, for groundwater table determination in saturation
zone one or two-dimensional Boussinesqe’ models are used.
Geological structure of mountainous watersheds is usually so
complex, that it is difficult to define water table. In WISTOO
model this process is simplified. Groundwater is described by one
reservoir with area equal to watershed area and width equal to the
total of all lengths of river channels in the watershed.
Total outflow from this reservoir is equal watershed runoff at
time t0, i.e. from beginning of calculations. Value of side supply
for stream cells was divided proportionally to stream length.
Groundwater supply depth is sum of partial supply from all cells.
Accepted solution of stream supply from saturation zone simplifies
complicated calculation procedure to formula:

q g = K f h g Sgr (29)

where: Kf - filtration coefficient, [m/s],

hg - groundwater reservoir depth, [m],
Sgr - average watershed slope, [-].

One stream cell is supplied according following formula:

qg
qi = (30)
Nz

where: qi – side stream supply for one cell,

Nz – total number of stream-cells.
Filling up of groundwater reservoir was calculated from relation:

L r K f Sgr ∆t j R g ∆t
h gj+1 = 1 - hg + (31)
µ Fg µ
where: Lr - total streams length, [m],
Fg - groundwater reservoir supply area, [m2],
Rg - average groundwater supply from aeration zone, [m/s].

Base on groundwater reservoir outflow, initial condition for

infiltration process in subsoil zone is calculated. Total supply in
time t0 is equal total watershed runoff.

2.8. Recipients system runoff (streams)

Runoff from recipients system (streams) is the process of water
flow in streams supplied by: groundwater runoff, surface and
subsurface runoff (Fig. 12). This unsteady flow in whole range of
water levels, from low to high. Total runoff along streams was
described by one-dimensional kinematic wave equation:

∂Q ∂F
+ = qb
∂x ∂t (32)
Q = α k R hN-1F

where: Q - runoff, [m3/ s],

Rh - hydraulic radius, [m],
qb - side supply, [m3/ m s],
αk, N - kinematic wave parameters,
F - flow cross-section area, [m2].
 surface runoff subsurface runoff

supply from previous TRANSFORMATION

element IN STREAMS runoff

groundwater
runoff
Figure13. Transformation process in streams calculation scheme

For the model purposes those parameters were accepted for

turbulent flow according Manning theory. Solution and calculation
schemes are analogue to surface and subsurface runoff
transformation. Stream channel cross-section was approximated as
rectangular, with width equal cell size. Side supply is sum of
surface, subsurface and groundwater runoff. Initially conditions are
determined from groundwater zone side supply.

3. STRUCTURE OF COMPUTER PROGRAM

FOR WISTOO MODEL

Computer application of WISTOO model consists of: procedure

responsible for data preparation (INPUT) and main calculation
procedure (Fig.14).
Pre-processing can be divided into following procedures:
• digital elevation model preparation support,
• watershed parameters from thematic layers generation,
• meteorological data preparation.
First procedure has auxiliary character. Is used only if digital
elevation model is not available. Two following procedures,
described bellow, are important for proper model operation.
 INPUT
INPUT DATA
- meteorological paramerers,
- measuring points location,
- watershed topography,
- stream network,
- land-use
- soil structure

DIGITAL ELEVATION
MODEL PROCEDURES
TOOL WATERSHED PARAMETERS
PROCEDURES PROCEDURES

METEOROLOGICAL DATA
PROCEDURES

CALCULATION PROCEDURES
RAINFALL-RUNOFF
TRANSFORMATION WITH
HYDROGRAPH VISUALIZATION

RAINFALL-RUNOFF
TEMPORAL TRANSFORMATION WITH DATA
VISUALIZATION
SPACIAL VISUALIZATION OF STORAGE
OF PROCESSES SELECTED PROCESSES

RUNOFF TRANSFORMATION
THROUGH
WATER RESERVOIR

Figure 14. Internal structure of computer program

3.1. Watershed parameters procedure

WISTOO model is the distributed mathematical model, using
dense calculation grid. Every cell must have defined parameter set
needed for calculations. Those parameters are determined from
thematic layers like: digital elevation model, stream network
topography, soil type and structure, watershed land-use. Special
tool procedure generates important watershed parameters from that
information. Basic data flow is based on water flow direction, thus
one of principal issues is determination of supply order for every
cell in the watershed. Sequence of cell supply is created (Fig.15).

0 0 0

0 1 1 1 0

2 3 2 0

Figure 15. Water exchange between cells scheme

All source cells are of order 0. Those cells supply order 1

cells, they supply order 2 cells etc. At the end of this chain streams
cells are located. Stream cells have similar assigning sequence.
Every cell gets unique identifier. Created file describes calculation
sequence in successive time steps.
From thematic layers (DEM, soil type and structure, land-
use) information is read and processed into following parameters:
• location - x, y coordinates from DEM,
• identifiers of cells supplying every cell,
• slope from DEM,
• velocity modulus α for kinematic wave, where hydraulic
resistance coefficient is determined from land-use,
• precipitation separation coefficient p and plants retention s
used in net precipitation procedure from land-use,
 • soil porosity coefficient and maximal soil and subsoil
conductivity from soil type and structure,
• soil depth and types,
• soil permeability parameter. This parameter is divided into
eight categories:
- 0 - natural,
- 1 – water outlet ex. catch basin gate, water intake,
- 2 - impermeable,
- 3 - 75% impermeable,
- 4 - 50% impermeable,
- 5 - 25% impermeable,
- 6 – surface without soil layer, ex. field way,
- 7 – barrier, making water transfer impossible.

Table 1. Identifiers of soil types used in WISTOO program

Identifier Soil type
101 Rocks
102 Gravel and stony
103 Fine gravel
104 Fine to coarse sand
105 Fine sand
106 Dusty sand
107 Sandy clay
108 Light clay
109 Dust
110 Clay
111 Stiff clay
112 Silt
113 Peat light decomposed
114 Peat decomposed
115 Skeletal soils

Soil permeability parameter was defined from land-use:

• impermeable surfaces for bituminous roads,
• urban areas are divided into four categories: compact
urban (75% impermeability), compact rural (small towns
or small urban constructions - 50% impermeability),
 sparse urban (peripheral housing estates – 25%
impermeability), sparse rural. Last category is treated in
the model as natural.

Table 2. Identifiers used for land-use

Identifier Land-use type
1 Asphalt or concrete roads
2 Pavement roads
3 Macadam or partially permeable roads
4 Field or forest ways
5 House
6 Dense urban area
7 Sparse urban area
8 Dense rural area
9 Sparse rural area
10 Sidewalks or pavements partially permeable
11 Concrete or impermeable areas
12 Areas 50% impermeable
13 Pastures with low grass
14 Pastures with high grass
15 Bushes or weeds
16 Sparse bushes or trees
17 Coniferous forest
18 Deciduous forest
19 Mixed forest
20 Fields – different cultivation
21 Fields – grains
22 Meadows, lawns
23 Sink basins or other surface water intakes
24 Barrier for horizontal water flow in aeration zone
25 Impermeable rocks
26 Fractured rocks– permeability like for rock waste
27 75% impermeable areas
28 50% impermeable areas

Thematic layers prepared in GIS format must fulfil certain

conditions according map attributes. As mentioned in paragraph 1,
those conditions are related to digital elevation model and
hydrographical network. For soil type and structure, land-use
conventionally numbers are assigned to different classes. Classes
are presented in Tables 1 and 2. For soil type and structure two
classifications are presented: according polish regulations (82
categories) and general soil division. Table 1 shows this second
classification.

3.2. Meteorological data procedure

In meteorological data preparation procedure as first the
calculation period for simulation is selected. It is based on data
introduced from measuring stations or climatic stations. Saved data
can be visualized – it helps in period selection.
Data can be introduced into database as hour or temporal
sum values (for precipitation and runoff as daily values). One of
the procedures converts this type of data into hour values.

3.3. Calculation procedures

Calculation procedures of WISTOO model consist of three
modules:
• rainfall-runoff transformation with hydrographs
visualization (maximal ten cross-sections including
observed),
• rainfall-runoff transformation with spatial visualization of
selected processes (net precipitation, surface or subsurface
water depth and velocity, soil humidity),
• hydrograph transformation through water reservoir.

3.3.1. Rainfall-runoff transformation procedures

First two procedures use previously prepared watershed parameters
and meteorological data. Last one uses results of rainfall-runoff
transformation model.
Input for rainfall-runoff transformation procedure are data
files containing:
• watershed parameters,
• precipitation stations location,
• calculation cross-sections location; this can be done
before program start,
• meteorological data.
Results are presented in form of graphs or bitmaps on-line during
calculation on computer screen. In rainfall-runoff transformation
procedure hydrographs for all selected cross-sections and net
rainfall spatial distribution are displayed. Hydrological processes
are displayed as bitmaps.
Procedures responsible for hydrological processes simulation
are managed by main program. Those procedures correspond to
hydrological processes described in chapter 2. In main program
connection between procedures and data exchange is established
and controlled. Only processes determining rainfall-runoff
transformation at certain moment are calculated. It highly reduces
computation time.
When the calculation starts all input data are checked from
homogeneity point of view. If by accident, data files from wrong
layer are selected, program displays message and calculation is
stopped. User must select correct data. Use of first two procedures
is simple and no additional computer knowledge is needed.

3.3.2. Runoff transformation through water reservoir

Hydrotechnical infrastructure (water reservoir) is also implemented
into model. Special subroutine is here used. Use of runoff
transformation through water reservoir requires water management
knowledge. Besides inflow hydrograph, many additional
parameters characterizing reservoir (use capacity, flood reserve,
etc.) and work mode must be introduced. This procedure is
adopted for polish conditions and regulations.
 Four policies for normal and water deficiency modes and
three for flood modes are used. Any mode can be selected and
influence its influence on watershed closing cross-section
hydrograph can be analyzed.
For normal conditions four policies can be used:
• standard – it is the simplest mode, assuming that one
water consumer is taken into consideration. Water
resources are used for current reservoir work without
limitations.
• standard parametric I type – it is modified standard
policy, with two groups of consumers. Modification
enables better water use. Initial reservoir filling, water
inflow, used demands and target filling is taken into
account.
• standard parametric II type – it is also modification of
standard policy, but critical filling term is introduced, i.e.
minimal filling below which restrictions in water
consumption are introduced
• ‘N-days’ type – used in deficiency periods. One or group
of consumers is concerned. Assuming low inflows during
n days, reservoir can not become empty.
Flood period is the period from moment when water inflow to
reservoir exceeds permitted inflow and use capacity is filled up, to
the moment of flood reserve formation. Flood reserve
reconstruction starts when inflows is smaller than restricted inflow
Qdoz. For such conditions there are three policies:
• fixed policy – assumes, that during peak flow storage
outflow is equal restricted outflow. Cut off of peak flow
occurs only when peak flow volume is smaller than flood
reserve. When flood wave volume is bigger than flood
reserve, restricted outflow will be exceeded. Can be used
when flood reserve is sufficient.
• partly fixed policy of type I- when flood reserve is
filling, restricted outflow is increased α (Q - Qdoz) where:
Q is inflow. When flood reserve is filled up, outflow is
determined like for fixed policy. Parameter α is
 optimized, α ∈ [0,1]. For α = 0 partly fixed policy of type
I becomes a fixed one.
• partly fixed policy of type II - For this policy procedure
is according rules:
- when flood reserve is filling, outflows can not exceed
inflow,
- when flood reserve is filling, next outflow cannot be
smaller then previous one,
- when flood reserve is filling, outflow can not be
smaller than restricted outflow.

4. RANGE OF MODEL APPLICATION

As mentioned in introduction, WISTOO model is developed for

mountainous and hilly watersheds, where main source of water
supply is surface and subsurface runoff. Model can be applied to
watersheds with areas from few to few thousands square
kilometers. Increasing watershed area, calculation time and
computation demands increases too.
Input parameters for the model are calculated from digital
thematic layers. Basic format used for digital layers, for both raster
and vector maps, is compatible with IDRISI program developed at
Clark University - USA [Estman J.R., 1995]. Use of this
geographical information system in many polish universities in
didactic process and its simplicity were main reasons for this
format selection.
Basic element for calculations is square cell with size
specified by user. Correctness of cell size determines runoff
process simulation accuracy. I general, better accuracy can be
archived by decreasing of cell size. Increasing of cell size causes
deterioration of watershed surface representation and consequently
description of water transport dynamics. Bigger cell produces
deformations of watershed topography by averaging. Also
parameters responsible for rainfall-runoff transformation changes:
slope, surface water pathways. Moreover, uncontrolled
deformation of impermeable and semi-impermeable areas takes a
place. Too wide streams width (corresponding to cell size) changes
flow dynamics.
From field experiments result was stated, that the biggest
recommended for WISTOO model cell size is 25x25 meters. Up to
now, positive and well documented by measuring data model
applications covered:
• determination of flood runoff hydrographs in uncontrolled
watersheds in different cross-sections,
• determination of influence of land-use change on runoff
hydrograph in selected stream cross-sections,
• determination of areas with intense surface runoff,
including secondary surface runoff causing surface
erosion,
• determination of small water reservoir influence on flood
risk reduction,
• determination of potential location for inlet into storm
sewerage system.
Model, thanks to its flexibility according required measuring data
sets, can be applied both for well and not fully gauged watersheds.
For watershed with areas above 100 km2, data with one-day time
resolution can be used. In polish conditions with sparse network of
precipitation monitoring and one-day recording resolution for
precipitation and runoff, WISTOO models meets demands.
For small watersheds with areas below 100 km2 temporal
and spatial resolution of data must be high. For smaller watershed
input data demands are higher. Additionally, in small watersheds
geological parameters must be taken into account: water layer
shape and subsoil layer depth in aeration zone. This additional
information is used for determination of groundwater part in runoff
supply. For most of mountainous regions, underground watersheds
do not coincide with topographical watershed. Value of
groundwater supply has influence on infiltration initial conditions
accuracy.
WISTOO model enables use of daily resolution data records.
Special function distributes daily values into hour values. Hour
hyetograph assumes 24-hour term of precipitation event. That is
why it should be used mainly for long-term events. This possibility
must be used with care.
Theoretical hyetographs not always reflect real flood wave
shape. Lack of information concerning hyetograph total time,
maximal intensities time and values is main source of errors.

5. EXAMPLE OF WISTOO MODEL

APPLICATION FOR ŁOSOSINA RIVER
WATERSHED

Model WISTOO was applied for different watersheds,

among others: Ropa, Czarny Dunajec, Łososina, Wielka Puszcza,
Leśnianka watersheds. Those watersheds are diverse in respect of
runoff regime and available input data. Physical foundation of
WISTOO model and spatial (distributed) description of supply and
runoff parameters enabled use of this model for all different cases.
As an example, application for Łososina river watershed
(413 km2 area) is shown. Only two gauge stations and two
precipitation stations recording observations once a day, are
located in this watershed. Most of hydrological models can not
operate with this number of hydrometeorological data. With
WISTOO model simulation results are satisfactory.

5.1. Watershed characteristics

Łososina river watershed is located in Beskid Wyspowy region and
is typical mountainous watershed for Upper Vistula river basin. It
has big slopes and dense streams network (Fig.16). Maximal
elevations exceeds 1000 m.a.s.l., (Mogielnica 1170 m), and outlet
is situated on 225 m.a.s.l. For this are digital elevation model was
prepared based on square cells 25x25 m. Slopes (Fig.17), surface
flow directions and cell order was determined.
 Figure 16. Streams network in Łososina watershed with marked gauge and
precipitation stations. 1-3 precipitation stations; 4-5 gauge stations

Figure 17. Slopes in Łososina watershed – lighter areas correspond to flat terrain
(minimal slope)
 There are five types of soil in the watershed: clayey, skeletal,
fen, dust, sandy. Skeletal soils are in initial phase of evolution.
Most of the area is covered by clay and silt soils with rock waste,
stones, fine stone bits, sand.
Land-use (Fig.18) was divided into five classes: fields and
croplands (1), forest (2), sparse rural areas (3), dense rural areas
(4), main asphalt roads (width above 15m) (5).

Figure 18. Spatial distribution of land-use

5.2. Simulation results for Łososina watershed

Simulations were performed for monthly periods: July 1970 and
June 1974. In precipitation distribution calculations, two additional
stations from neighboring watersheds were used. The stations are
situated in the Dunajec watershed. Łososina is a left hand side
tributary of Dunajec. Figure 19 shows five hydrographs. Two of
them are from gauge stations. Calculated hydrographs are from
gauge cross-sections (2) and watershed closing cross-section.
Figure 19. Observed and calculated hydrographs (31 May - 30 June 1974)
where:
1. stream network with calculation marked cross-sections,
2. calculation and gauge cross-sections co-ordinates,
3. procedure name bar (rainfall-runoff transformation with hydrograph
visualization),
 4. calculation period,
5. maximal net precipitation for every calculation step and current
calculation time,
6. spatial net precipitation distribution with color legend,
7. runoff [m3/s],
8. hydrographs - observed and calculated,
a/ calculated for watershed closing cross-section,
b/ calculated in cross-section 5,
c/ observed in cross-section 5,
d/ calculated in cross-section 4,
e/ observed in cross-section 4,
9. time [h].

Figure 20 demonstrates results of runoff simulation for year

1970. Simulation was performed for two gauge cross-sections and
in watershed closing cross-section. First three curves are for
simulated values, last two for observed. Cross-sections direction is
upstream.
Besides standard runoff hydrographs simulations,
simulations of surface runoff ware prepared. Four processes can be
simulated and visualized by the program: net precipitation, soil
moisture, surface and subsurface runoff. Two last processes in
standard version are displayed as transformed water depth and flow
velocity. Solving detailed problems, spatial runoff distribution is
presented by unit runoff (runoff for unit width e.g. running meter).
Secondary surface runoff from aeration zone can also be traced.
Examples of spatial distribution of surface and subsurface runoff
velocities in different time steps are shown on Figures 21-24.
Two other procedures displaying spatial distribution of soil
moisture and net precipitation are used sporadically, mainly for
didactic purposes.
Figure 20. Calculated and observed runoff hydrographs (7-31 of July 1970)
Figure 21. Surface runoff velocity distribution (1 of June 1974, 23:00 h)

Figure 22. Surface runoff velocity distribution (2 of June 1974, 6:00 h)

Figure 23. Subsurface runoff velocity distribution (1 of June 1974, 23:00 h)

Figure 24. Subsurface runoff velocity distribution (2 of June 1974, 6:00 h)

6. A FEW WORDS ABOUT THE MODEL

WISTOO model is continuously developed and improved. More

flexible land-use description is of special interest. One of possible
solutions if use of grid with variable size. It enables more detailed
representation of features small in size, but influencing runoff
formation.
Second direction of model development is related to
hydrotechnical constructions on streams and in the watershed. It
will be focused mainly on water reservoirs and stream cascades.
Module aiding in hydrotechnical construction design can be on of
results of model development.
Basement of the model on geographical information systems
and physical processes of runoff formation should facilitate
achievement of those two goals. Sophisticated studies should be
possible to perform using WISTOO model:
• Simulation analysis of historical runoff conditions,
• Model verification and identification of its parameters,
• Influence of land-use change of runoff formation forecast.
Fast development of computer sciences, especially data processing,
enables further development of the model. Today WISTOO is a
very reliable model even with minimal set of input data, comparing
to other models. Those facts and new ideas how to improve the
model, make hope, that WISTOO has big chance to be even more
reliable and competitive.
Symbols used in text
A - gradient curve inclination coefficient, [1/m],
a –empirical parameters determined from experiments for two
plant types – high plants (forest) and low, [s/m],
α - exponent index dependent on soil type,
αk – kinematic wave coefficient.
b - dripping vs. water depth on plants curve coefficient, [1/mm],
ba –empirical parameters determined from experiments for two
plant types – high plants (forest) and low, [sW/m3],
β - weight coefficient,
c - current water depth on plants, [mm],
ca –empirical parameters determined from experiments for two
plant types – high plants (forest) and low, [W/m2],
ct – current water depth on plants at time t, [mm],
co – water depth on plants at time t-1, [mm],
Do – minimal dripping from plants when c = s, [mm/min],
dh – base altitude, [m],
dk – distance form point i to measuring station k,
∆ - inclination of water vapor pressure curve vs. temperature,
[hPa/K],
Ea – current evapotranspiration, [mm],
Ep – potential evaporation, [mm/min],
e – napierian base,
ee – current water vapor pressure, [hPa],
ε – maximal water vapor pressure, [hPa],
F – flow cross-section area, [m2],
Fi – cell area,
Fg – groundwater reservoir supply area, [m2],
G – soil heat flux, [W/m2],
γ - psychrometric constant, [hPa/K],
Hi, Hi+1 – elevation above sea level for the highest and the lowest
station in the watershed, [m],
Hx - terrain elevation, [m].
H’x – transfer elevation, [m],
h - water depth, [m],
hg - groundwater reservoir depth, [m],
hk - capillary height, [m],
ho - water depth prom previous time step,
hr - average height of plants, [m],
hw - water depth on soil surface, [m],
K- soil hydraulic conductivity, [m/s],
Kf - filtration coefficient, [m/s],
Ko - maximal soil conductivity (filtration coefficient), [m/s],
Lr - total streams length, [m],
λ - latent heat of evaporation, [J/kg],
N - kinematic wave coefficient,
Nz - total number of stream-cells,
n – index exponent,
nh - hydraulic resistance coefficient, called roughness coefficient,
[1/m1/3],
nz - number of elements supplying groundwater,
Pi - value of calculated parameter at point i, i=1,2,...m,
Pk - measured value at point k, k=1,2,...l,
p - rainfall separation coefficient, [-],
Ψ - soil suction pressure, [m],
Ψd - suction pressure in bottom node, [cm],
Ψg - suction pressure in upper node, [cm],
R - rainfall intensity, [mm/min],
RE - excess rainfall or subsurface supply, [mm/s],
Rg - average groundwater supply from aeration zone, [m/s],
Rgi - groundwater table supply,
Rh - hydraulic radius, [m],
Ri, Ri+1 - recorded precipitation depth for the lowest and the
highest station in the watershed, [mm],
Rn - net radiation, [W/m2],
Rnett - net precipitation, [mm/min],
Rpi - subsoil layer supply,
Rx - precipitation depth calculated by inverse distance method,
[mm],
R’x - rectified precipitation depth, [mm],
ra - aerodynamic resistance function, [s/m],
rc - superficial resistance function, [s/m],
ρ cp - volumetric dry air heat capacity, [J/m3K],
Q - runoff, [m3/ s],
Qk - groundwater runoff in watershed closing cross-section,
q - unit flow, [m3/m s],
qb - side supply, [m3/ m s],
qd - inflow sum for cell,
qf - moisture flux, [m3/m2 min],
qgr - maximal outflow from cell calculated from soil layer depth
Hg and filtering coefficient µ,
qi - side stream supply for one cell,
qo - outflow sum for cell,
Sgr - average watershed slope, [-],
Sh - hydraulic gradient, [-],
Sz - terrain slope, [-],
s - plants retention, [mm],
Θo - maximal volumetric humidity – soil porosity index, [-],
Θs - volumetric humidity of permanent plant wilting, [-],
v - infiltration rate, [m/s],
vw - wind velocity, [m/s],
z - wetting front depth, [m],
zo - height of plants roughness, [m],
zv - altitude of wind velocity measurement, [m].
 REFERENCES

1. Ahuja I.R. – Applicability of the Green and Ampt approach to

water infiltration through surface crust; Soil Sci.,1974
2. Beven K.J and Moore I.D (editors) – Terrain Analysis and
Distributed Modelling in Hydrology, John Wiley & Sons,
1993
3. Chow V.T. – Open channel hydraulics, McGraw-Hill Book
Company, 1959
4. Chow V.T. (editor) – Handbook of Applied Hydrology,
McGraw-Hill Book Company, 1964
5. Eagleson P.S. – Dynamic Hydrology, McGraw-Hill Book
Company, 1970
6. Eastman J.R. – IDRISI for Windows, Clark Labs for
Cartographic Technology and Geographic Analysis, Clark
University, USA 1995
7. Gądek W., Banach Wł., Bodziony M., Książyński K.,
Łabuda P., Szczepanek R., Szypułka P. – Integral distributed
rainfall-runoff model; International Scientific Conference
„Forest and Water”, Cracow 1998
8. Książyński K., Nachlik E. – Discrete model of infiltration and
groundwater runoff for mountainous watershed (in polish);
Archiwum Hydrotechniki 1-2, 1990
9. Morel-Seytoux H.J., Khanji J. – Derivation of an eqation of
infiltration; WRR 1974/4
10. Ozga-Zielińska M., Brzeziński J. – Applied Hydrology (in
polish); Wydawnictawa Naukowe PWN, 1994
11. Soczyńska U. – Dynamic Hydrology (in polish);
Wydawnictwa Naukowe PWN, 1997
12. Stigter C.J. – Assessment of the quality of generalized wind
functions in Penman’s equations; Journal of Hydrology 45,
1980
13. Szczepanek R., Szypułka P., Łabuda P. – A modification of
digital elevation model for hydrological modelling purposes;
International Scientific Conference „Forest and Water”,
Cracow 1998