Limburg Soil Erosion Model USER MANUAL

Limburg Soil Erosion Model USER MANUAL
Limburg Soil Erosion Model
Windows version 2.x
Victor Jetten
LISEM MANUAL version 2.x - January 2, 2002
This is a first draft of the manual for the new LISEM for windows, currently at version 2.1. It
is a compilation of the website:, on which you can find the latest
LISEM started out as an erosion model that simulates overland flow, splash and rill erosion,
with some extra features. More processes were added and the number of variables and
choices in the interface became prohibitive because initially all combinations were allowed.
Therefore it was decided to have different versions of LISEM that essentially consist of the
Basic version plus an extra set of processes:
LISEM Basic: the original LISEM including parallel infiltration through different types of
surfaces inside a gridcell, as well as grass strips and channels.
LISEM Wheeltracks: the basic version plus the ability to treat semi-permanent
wheeltracks as a network of small channels.
LISEM Multiclass: the basic version with the erosion/deposition changed to a
multiclass system with 6 sediment classes causing spatial changes in texture composition.
LISEM Nutrients: the basic version with the nutrient losses (N and P) in solution and in
suspension, absorbed to the clay particles.
LISEM Gullies (EUROWISE): the basic version with incision in a mult-ilayered soil, in
certain areas that are prone to gullying, according to an empirical landscape analysis.
Not all of the theory and descriptions are included in this manual yet, only of the basic
version and the gully version.
Victor Jetten, January 2, 2002
Jetten, V. (2002). LISEM user manual, version 2.x. Draft version January 2002. Utrecht
Centre for Environment and Landscape Dynamics, Utrecht University, The Netherlands. pp
No warrenties, expressed or implied, are made that the computer programs described in
this publication are free from errors or are consistent with any particular standard of
programming language, or that they will meet a user's requirement for any particular
LISEM MANUAL version 2.x - January 2, 2002
Introduction ............................................................................... 5
LISEM BASIC - Theoretical framework ..................................... 6
Rainfall.................................................................................................................. 7
Interception ........................................................................................................... 7
Infiltration .............................................................................................................. 7
Swatre ........................................................................................................... 8
Holtan method ............................................................................................... 9
Green and Ampt .......................................................................................... 11
Morel-Seytoux and Verdin............................................................................ 12
Subtraction of Ksat....................................................................................... 12
Surface definition on a sub-gridcell level ............................................................. 12
Storage in micro depressions.............................................................................. 12
Surface Storage in LISEM Version 1.63 and earlier ..................................... 15
Overland flow and channel flow ................................................................... 17
Erosion and eposition ......................................................................................... 19
Splash detachment ...................................................................................... 19
Flow detachment and deposition.................................................................. 19
Channel detachment and deposition ............................................................ 21
LISEM-wheeltracks ................................................................. 22
LISEM-multiclass .................................................................... 22
LISEM-nutrients ...................................................................... 22
LISEM-GULLIES (EUROWISE) .............................................. 23
Introduction ......................................................................................................... 23
Ephemeral gullies ........................................................................................ 23
Summer and winter gullies........................................................................... 26
Basic principles ................................................................................................... 26
Critical area analysis - theory....................................................................... 27
Critical area analysis - Implementation in LISEM ......................................... 28
Ephemeral gully formation ........................................................................... 30
Gully incision................................................................................................ 32
THE INPUT DATABASE ......................................................... 33
PCRASTER ........................................................................................................ 33
The input maps ................................................................................................... 34
Catchment maps.......................................................................................... 35
Vegetation maps.......................................................................................... 37
Soil surface maps ........................................................................................ 37
Infiltration related maps................................................................................ 39
Erosion/deposition related maps .................................................................. 43
Channels ..................................................................................................... 43
Additional maps needed for EUROWISE ............................................................ 45
Rainfall data........................................................................................................ 47
LISEM MANUAL version 2.x - January 2, 2002
The interface ........................................................................... 48
Speed buttons & Tabs......................................................................................... 49
Start tab .............................................................................................................. 49
Basic maps ......................................................................................................... 51
Output timeseries................................................................................................ 53
LISEM output .......................................................................... 54
Main output screen and totals ............................................................................. 54
Hydrographs and sedigraphs .............................................................................. 54
Erosion and deposition maps .............................................................................. 55
Timeseries .......................................................................................................... 56
References.............................................................................. 57
Partners .................................................................................. 59
Appendix ......................................................................................... 60
LISEM MANUAL version 2.x - January 2, 2002
LISEM, the LImburg Soil Erosion Model, simulates the hydrology and sediment transport
during and immediately after a single rainfall event in a small catchment. The model has
been used so far in catchments between 10 and approximately 300 ha. LISEM is built to
simulate both the effects of the current land use and the effects of soil conservation
The model was originally made for the Provice of Limburg, the Netherlands, to test the
effects of grass strips and other small scale soil conservation measures on the soil loss. In
the "Limburg" project, three catchments were fully equiped and monitored for 5 years by the
local governement (Waterboard Roer en Overmaas), the Free University of Amsterdam
(Physical Geography), Alterra and the Utrecht Universitfy (Physical Geography). Although it
can be used for planning purposes it is essentially a research tool because of its complexity.
The philosophy behind LISEM is that the model assumes nothing! An example: if a land use
change is modelled there is no way of telling LISEM that it should change all the related
variables because the crop is e.g. winter wheat and not sugar beet. The user must change
all appropriate variables him/herself: infiltration variables, surface roughness, Manning's n
etc. This gives the user more freedom and it is much clearer what happens in the
simulation, however this also means that the user knows what he/she is doing.
Basic processes incorporated in the model are rainfall, interception, surface storage in
micro-depressions, infiltration, vertical movement of water in the soil, overland flow, channel
flow (in man-made ditches), detachment by rainfall and throughfall, transport capacity and
detachment by overland flow. Also, the influence of compaction (e.g. by tractor wheelings),
small paved roads (smaller than the pixel size) and surface sealing on the hydrological and
soil erosion processes is taken into account. The processes are described in detail in the
theory section. The latest developments include the modelling of:
ƒ wheel tracks as small channels,
ƒ multiple sediment classes for erosion and deposition,
ƒ the loss of P, NO3 and NH4 in solution and in suspension,
ƒ the incision and formation of gullies
Although not directly visible to the user, the model is one of the first examples of a
physically based model that is completely integrated in a raster Geographical Information
System, PCRaster (van Deursen & Wesseling, 1992). The source code is a mix of C++
code, GIS operations, mathematical operations and hydrological functions (kinematic wave,
Richards equation). No conversions are necessary between PCRaster and the model. All
input and output maps are raster maps that can be easily displayed and treated with the
PCRaster software.
LISEM MANUAL version 2.x - January 2, 2002
For each gridcell, rainfall and interception by plants are calculated, after which infiltration
and surface storage are subtracted to give net runoff. Subsequently, splash en flow erosion
and deposition are calculated using the stream power principle and the water and sediment
are routed to the outlet with a kinematic wave procedure. Special cases can be defined for
roads and compacted areas, and (man-made) channels can be taken into consideration.
Figure 2.1. Flow chart of LISEM BASIC.
LISEM MANUAL version 2.x - January 2, 2002
2.1 Rainfall
LISEM uses rainfall intensities per time interval (so called break-point data). These are
stored in an ASCII file. Data from multiple raingauges can be entered in an ASCII input data
file with columns for each rainfall gauge. A map with areas having the raingauge ID number
determines the spatial distributed rainfall input. This map can be based for example on
Thiessen polygons or on a geomorphological analysis (assigning valleys to raingauges).
The rainfall is added to the current water height in each cell. However, the slope angle is
taken into consideration: the rain falls on a horizontally projected surface while the actual
terrain has a slope. Thus the rain is spread out over a larger area and the resulting water
height is lower:
h = h + P *cos(a)
Where P is the rainfall depth in that timestep (mm) and a is the slope angle. The
assumption here is that the slope is in one direction only, and the cell has a surface of
2.2 Interception
Interception by crops and vegetation is simulated by regarding the canopy as a simple
storage as a simple storage. The cumulative interception during an event is calculated as
(Aston, 1979):
where S is the cumulative interception (mm), Pcum is the cumulative rainfall (mm), k is a
correction factor for vegetation density (equals 0.046*LAI) en determines the rate with which
Smax is reached, cp is the fraction of vegetation cover and Smax is the canopy storage
capacity (mm) estimated from the LAI (m2/m2), by Von Hoyningen-Huene (1981):
Note that the LAI represents the average leaf area of the fraction of the gridcell that is under
vegetation. For example a gridcell with a single tree has a LAI of e.g. 6 m2/m2, while the
cover fraction cp of the gridcell takes care of the fact that most of the cell is bare.
2.3 Infiltration
Infiltration can be calculated with various sub-models, according to the data available.
Available are the Holtan (Beasley and Huggins, 1982), and Green and Ampt (Li et al., 1976)
equations for one or two layers, or the SWATRE model, a finite difference solution of the
Richard equation (Belmans et al., 1983; Feddes et al., 1978). The latter option includes
vertical soil water transport and the change of matric potential in the soil during a rainfall
event which can be examined by the user. The input is in the form of maps with soil
hydrological properties such as initial moisture content, porosity and saturated conductivity.
For each layer maps are defined. In case of the Richards equation, a map with soil profile
types is linked to a series of tables with values for humidity, θ; matrix potential, ψ; and
hydraulic conductivity, K.
LISEM MANUAL version 2.x - January 2, 2002
The choice of infiltration model is mostly governed by the type of data available and the
experience of the user. It should be noted that the models do not give the same results.
Since initial water content and hydraulic conductivity are among the most sensitive input
parameters of LISEM (De Roo et al., 1996b; Jetten et al., 1998), re-calibration is advisable
when a different infiltration model is used.
Currently the following models are available:
Swatre (finite difference approximation of the Richards equation)
Holtan (as in the Answers model)
Green and Ampt for 1 and 2 layers
Morel and Seytoux
Subtraction of Ksat
Infiltration and soil water transport in soils are simulated by a solution of the well known
Richards equation, which combines the Darcy equation and the continuity equation:
K = the hydraulic conductivity (m/s);
h = the pressure (matric) potential (m);
0 = the volumetric water content (m3/m 3);
z = the gravitational potential or height above a reference level (m);
t = time (s)
Using the soil water capacity C(h)=d0 /dh (the slope of the soil water retention curve ?(h),
the unsaturated flow equation is derived:
The Mualem/Van Genuchten equations (Mualem, 1976; Van Genuchten, 1980) can be used
to predict the soil-water retention curves and the unsaturated hydraulic conductivity, which
are needed to solve the equation above. However LISEM does not use these relations
directly but requires the user to define the water retention curves and the K(h) curves as
tables. The reasoning is that the measured curves hardly ever follow the Mualem-Van
Genuchten relations exactly but usually derivate near saturation of the soil. Thus, for the
catchments soil profile types are defined, and for each characteristic soil horizon, and for
each horizon tables with the measured K-theta-h relations are required. The equations are
solved by explicit linearisation using the so-called Thomas (tridiagonal) algorithm (see e.g.
Remson et al., 1971). The submodel operates with a variable time increment depending on
pressure head changes.
NOTE: SWATRE assumes that the nodes are not in the centre of the layers (as is common)
but between the layers. Thus the user specifies the depth to the bottom of each layer (see
section input database).
NOTE: The average hydraulic conductivity between the nodes can be calculated either as a
arithmetric average (0.5*sum of Ki and Ki+1), or as the geometric average (sqrt(Ki*Ki+1)).
The choice is made in the user interface. This has a large influence as the geometric mean
LISEM MANUAL version 2.x - January 2, 2002
favours the node with the lowest K value and the K(h) relationship is nhighly non-linear:
compare 0.5*(1 + 0.01)=0.5005 with sqrt(1 * 0.01)=0.1.
Figure 2.2. Data structure of the Swatre infiltration module.
Holtan method
For areas without detailed knowledge of the soil physics, different model versions with
empirical infiltration equations can be used, using Green & Ampt and Holtan (e.g. Beasley &
Huggins, 1982; de Roo, 1993). The Holtan model empirical (sub-)model is based on a
storage concept. The main advantage of the model is the capability of recovery in soil
infiltration capacity during periods of light or zero rainfall. The infiltration rate is expressed in
terms of cumulative infiltration, initial soil water content and other soil variables:
f = infiltration rate (mm/h);
FC = infiltration rate at steady state (mm/h);
A = the maximum possible increase in infiltration rate over the steady state rate, FC (mm/h);
St = storage potential of the soil (mm) above the impeding strata: total porosity (TP) minus
the antecedent soil water (ASM):
S = (1-ASM)*TP*DF;
DF = the effective depth to the impeding strata (control zone depth) (mm);
TP = total porosity (% volume);
ASM = antecedent soil moisture (% saturation);
DR = cumulative drainage (mm);
F = cumulative infiltration (mm);
P = dimensionless coefficient relating the rate of decrease in infiltration rate with increasing
soil moisture content.
LISEM MANUAL version 2.x - January 2, 2002
Huggins and Monke (1968) introduced an expansion of the model, assuming a relationship
between percolation (or 'drainage') and soil water content. Using both equations, the
recovery of infiltration rate as the result of a temporary interruption in rainfall can be
predicted. They assume drainage when soil moisture content in the control zone exceeds
field capacity:
dr = percolation or 'drainage' rate (mm/h);
FP = field capacity (% saturation).
The major advantage of the Holtan model over other infiltration equations is the use of
cumulative infiltration instead of time as the independent variable. This offers several
advantages in catchment-hydrology simulation. Skaggs et al. (1969) evaluated several
infiltration equations and concluded so. Using the Holtan equation, difficulties are not
encountered in computing the infiltration rate at any time during a storm event, even when
the water supply does not exceed the infiltration rate. Furthermore, because the Holtan
model assumes a relationship between drainage and soil water content, the recovery of
infiltration rate as the result of a temporary interruption in rainfall can be predicted.
The major difficulty with the Holtan equation is the control zone depth (DF). Holtan
suggested that the depth to the first impeding layer should be used. If no impeding layer
exists, a value equal to 0.25 to 0.75 of the A-horizon is advised (Beasley & Huggins, 1982).
Beasley and Huggins remarked that of all of the variables used in the ANSWERS model,
the DF is the least well defined and most arbitrary. Huggins and Monke (1966) found that
the effective depth was highly dependent on both surface condition (crusting) and the
cultural practices used in preparing the seed bed. De Roo and Riezebos (1992) also
demonstrated the influence of crusting on the DF variable. Huggins and Monke (1968)
suggested that the DF for deep soils can be determined by the depth necessary for the
hydraulic gradient to approach unity. Skaggs et al. (1969) computed the control zone depth
for fallow soils from the initial soil water content (ASM), the soil porosity (TP), and the
volume of water that had infiltrated (F) at the time the infiltration rate reached a constant
Fsat = Cumulative infiltration (mm) at saturation.
A major disadvantage of this method however, is that the DF variable is made a fitted
variable instead of being physically measurable as the depth to the first impeding layer.
Nevertheless, with this method, Holtan's equation gave an excellent fit (r2 = 0.988, obtained
using non-linear regression) to their experimental data. The Holtan infiltration model
incorporates a minimum infiltration capacity (FC), which equals the saturated conductivity
(Ks) of the topsoil. Because the Ks is always larger than zero, the model always allows for
infiltration, even when the soil profile is saturated. Thus, using this sub-model, only
Hortonian overland flow is simulated. Saturation overland flow is not simulated. Using the
Swatre-based sub-models and the Green/Ampt submodels, one can simulate saturated
overland flow.
LISEM MANUAL version 2.x - January 2, 2002
Green and Ampt
Green and Ampt (1911) first applied the Darcy equation to the wetted zone in the soil,
assuming that a distinct wetting front exist. They produced a one dimensional infiltration
equation used and adapted by many researchers. Generally the equation has the following
 Ω
i = k t + 
I − Ωln t +  = kt
 Ω
in which I = the infiltration rate (m/s), I = the accumulated infiltration over time (m), k = the
hydraulic conductivity in the wetted zone (m/s), t = time since the start of the infiltration and
Ω = potential head parameter (m). Note that k is not necessarily the saturated hydraulic
conductivity but less according to Li et al. (1976). According to Fok and Hansen (1966) this
parameter is defined as:
Ω = (h c + h o )(θ w − θ i )
in which hc = the average capillary suction head at the wetting front, ho = the ponded head
water level at the surface, θw = the water content of the wetted zone (may be smaller than
the porosity) and θi = the antecedent moisture content. To get the infiltration rate at any time
equation 2 has to be solved first for I and inserted into equation 1. This is usually done by
iteration. Li et al. (1976) show that it is possible to use an explicit approximation by
developing a power series expansion of the logarithmic term in equation 2. First the
parameters are rewritten in their non-dimensional form:
I* =
; t* =
and i * =
By using these parameters the equations are now rewritten as:
i* = t +
I * − ln(1 + I* ) = t *
Replacing the logarithmic term with a power series and dropping the higher order terms
yields a quadratic function that can be solved simply by retaining the positive root:
I * = 0.5 t * + t * (t * + 8)
which can be rewritten in the dimensional form as:
dI =
− (2I − kdt) + (2I − kdt) 2 + 8kdt(Ω + I)
This is the relation used in LISEM whereby dI > 0 (if the solution is negative it has no
The Green and Ampt model is very sensitive to the choice of Ksat and initial moisture
content. The initial assumption that the wetting front moves down as a wet body parallel to
the surface with a speed dictated by the Ksat is not correct. Many researchers therefore
suggested "field" variables: a "field porosity or a field "Ksat", or a suction at the wetting front
that is not the matrix potential at a given time but a more soil property.
NOTE: This means that in practice the Green and Ampt solution needs calibration, either by
decreasing the Ksat values or the storage capacity of the soil.
LISEM MANUAL version 2.x - January 2, 2002
Morel-Seytoux and Verdin
not implemented yet
Subtraction of Ksat
The saturated hydraulic conductivity is subtracted from the net rainfall, simulating instant
saturation in a simple way. This option can be used for testing and ‘quick and dirty’
2.4 Surface definition on a sub-gridcell level
LISEM enables the simulation of several types of soil surfaces in a grid cell. These
represent different soil structures. For small sized gridcells (< 100m2) this seems somewhat
superfluous, but the gridcell size in LISEM is user defined and so the option is available.
The following surface types can be defined:
ƒ normal soil surface (tilled)
ƒ crusted
ƒ compacted
ƒ road (impermeable)
ƒ grasstrips
Two maps with information are needed to do this:
1. For each surface type a map is needed that defines the fraction of a cell with that
surface type. For instance where the crust fraction map has values > 0, LISEM assumes
that it has to calculate a separate infiltration for the crusted surface fraction of the
2. In order to do this LISEM needs a map for the saturated hydraulic conductivity of the
crust (note that these names are optional). Note that the SWATRE infiltration module
has a different data structure so the infiltration of different surface types is not done by a
Ksat map but by specific soil physics tables. All other parameters are assumed identical
to the normal surface to avoid too many maps: the infiltration of e.g. a crust is not at all
modelled in a physically based way and the separate ksat should be seen as a
calibration variable.
These surface types can be switched on and off in the interface without the need to change
the maps. In this way scenarios can be modelled that simulate e.g. the effect of severe
crusting or compaction.
The resulting water height at the surface is calculated as a weighted average of the normal
infiltration and the infiltration of the area with a different soil structure, based of the fraction
in the gridcell. Therefore the infiltration is averaged for each timestep.
2.5 Storage in micro depressions
The surface storage is not treated as a sharp threshold: runoff takes place before the water
level reaches some threshold value of average surface storage. This is done because runoff
at a micro-scale is a spatial process of ponds that fill up and overflow into each other. The
LISEM MANUAL version 2.x - January 2, 2002
gridcell in LISEM represents an area of for instance 10x10 m2 and before the average
storage of this surface is filled the water at the edges of this area is moving downstream.
The average depression storage can be measured at a given scale. The start value is more
difficult and in LISEM a pragmatic approach is taken, based on the fraction of the surface
that is ponded.
Surface storage is calculated using the Maximum Depression Storage (MDS). This is the
threshold value of a given area above which surface micro depressions overflow: all
depressions are connected to each other and to an point outside the reference area so that
in principle each additional "drop" of water will flow outside the reference area. The MDS is
determined by Kamphorst et al. (2000) from 221 digital elevation models of various types of
micro relief, in a wide variety of agricultural circumstances and soil types. The analysis is
based on DEMs of roughly 1m2. Figure 1 shows the relation between the MDS and Random
Roughness (RR). Originally the depression storage in LISEM was based on the work of
Onstad (1984). Using the same form of equation Kamphorst et al. (2000) found (R2=0.88,
MDS = 0.243RR + 0.010RR2 + 0.012RRS
in which RR is the standard deviation of surface heights (cm) and S is the terrain slope (%).
They tested 6 different roughness indices but found that the standard deviation of the
heights gave the best relation with MDS (cm).
RR (mm)
2 mm
10 mm
25 mm
30 mm
MDS (mm)
Figure 2.3. Non-linear relation between RR and MDS (Kamphorst et al., 2000), based on
221 DEMs of 1m2 samples of a wide variety of surfaces. The difference in sampling
distance is caused by the application of different tools in the various countries.
Apart from the water available for runoff, the roughness also determines width of the
overland flow in LISEM. Rather then taking the cell width dx, the flow width (and hydraulic
radius) is assumed linearly related to the fraction of ponded surface fpa in the cell. The latter
variable is related to the water depth at the surface h (mm) (Jetten and De Roo, 2001):
fpa = 1-exp(-ah)
where a is an empirical factor between 0.04 and 1.8 for the roughness data set mentioned
above. Figure 2 shows that the factor a appears to be strongly related to RR in mm
(R2=0.99, n=362):
LISEM MANUAL version 2.x - January 2, 2002
a = 1.406*(RR)-0.942
In which RR is in mm. Figure 3 shows the increase of ponded area with water height.
Although not all the surfaces had a random roughness but showed also oriented roughness
elements, the RR appeared to be the best roughness index to explain the variance in the
dataset. Tortuosity based indeices performed less as did the Linden -Van Dooren indices
(Kamphorst et al., 2000).
Figure 2.4. Ponded area shape factor a related to RR (Jetten and de Roo, 2001)
Figure 2.5. Ponded area fraction increase with water height for
different values of RR (increase with 5 mm intervals).
LISEM MANUAL version 2.x - January 2, 2002
Based on the MDS value and the fraction of ponded area the runoff before the water level
reaches the MDS height is calculated as follows. It is assumed that the runoff starts if 10%
of the surface is ponded so that the equation above for fpa reads:
0.1 = 1-exp(-a h)
which means that at h = ln(0.9)/-a, runoff starts. If this threshold for h, called here Start
Depressional Storage or SDS is larger than MDS, 0.9 of MDS is taken. Between this
starting value "SDS" and the MDS value the runoff gradually increases in a non-linear way
to reflect the observation that the surface usually consists of degraded agrregates which
"release" little runoff untill they are fully submerged. After the water level has passed MDS
the runoff height increases linearly with water height:
h < SDS : runoff = 0
h > SDS : runoff = (h-SDS) * (1-exp(-h*(h-SDS)/(MDS-SDS)))
In fact this resembles a Gaussian curve (1-exp(-x2)) as is seen in figure 4.
Figure 2.6. Increase of runoff height with with water height for
different values of MDS and SDS (increase with 5 mm intervals).
Surface Storage in LISEM Version 1.63 and earlier
Version 1.63 and earlier
Storage in micro-depressions is simulated by a set of equations developed by Onstad (1984) and
Linden et al. (1988). The variable Random Roughness is used as a measure of microrelief. Surface
storage in depressions is simulated by (Onstad, 1984):
RETMAX = 0.112RR + 0.031*RR2 - 0.012RRS
LISEM MANUAL version 2.x - January 2, 2002
RETMAX= Maximum Depressional Storage (cm);
RR = Random Roughness (cm);
S = Slope gradient (%).
NOTE: in the Onstad equations the RR is calculated using the method of Almaras (1967), who
eliminated the extreme 10% of the surface height readings and takes the ln of the values to
normalise the dataset.
The rainfall excess (rainfall + overland flow - interception - infiltration) required to fill all depressions
(RETRAIN in cm) is calculated using the equation (Onstad, 1984):
RETRAIN = 0.329RR + 0.073RR - 0.018RRS
Moore & Larson (1979) identified three possible stages during a rainfall event:
a) Micro-relief storage occuring, no runoff;
b) Additional micro-relief storage accompanied by runoff;
c) Runoff only with the micro-relief storage.
To determine the transition from stage a to stage b, the data of Onstad (1984) were re-analyzed.
From this analysis the following equation was developed, simulating the starting point of runoff
(RETSTART in cm):
RETSTART = 0.0527 RETMAX RR - 0.0049S
Thus, during stage 1, all excess rainfall becomes depression storage. Then, from point RETSTART to
point RETMAX both overland flow and further depressional storage occur, based on a linear filling of
the depressions until RETMAX. After RETMAX, all excess rainfall becomes runoff. Thus, using these
relationships the actual storage in depressions (RET in cm) can be calculated.
Also, using the same input data, the maximum fraction of the surface covered with water (FWAMAX)
can be calculated (Onstad, 1984):
FWAMAX = 0.152RR - 0.008RR - 0.008RRS
The actual fraction of the surface covered with water (FWA) is calculated using a relationship based
on the work of Moore & Larson (1979) and Onstad (1984):
Ponded area is calculated according to the following equation:
FWA = 1 - exp (-a WH/RETRAIN * FWAmax)
where a is a calibration factor, currently at -5, and WH is the water height of the water layer at the
version 1.53 and earlier
In earlier versions the FWA was defined as a linear increase based on the stages 1, 2 and 3
described above. However the increase was defined in such a way that abrupt changes inj fraction of
ponded area were calculated with only a small increase in water height. Therefore the linear
equations were replaced with the single exponential equation above.
LISEM MANUAL version 2.x - January 2, 2002
Also in earlier versions, a part of the depressions was not included in the FWA because they were
thought to be isolated. The effects of this was hardly noticeable and the fraction of depressions
isolated was based on observations rather than a spatial analysis. Thus this variable became just an
extra parameter to calibrate and it was removed from the source code in order to simplify depression
storaage. The fraction of isolated depressions (FWAISO) was calculated as follows:
Based on the findings of Linden et al. (1988), some depressions are (temporarily) isolated and do not
contribute to overland flow. From their data it was determined that if the storage (RET) was less than
75% of the RETMAX, 20% of the depressions are isolated. If RET is between 75% and 100% of
RETMAX then the following equation was derived:
FWAISO = 0.20 FWA (1- 4(RET/RETMAX-0.75))
Overland flow and channel flow
A gridcell can have more than one type of surface: e.g. a road (smaller than DX) is
impermeable, a wheeltrack is compacted and a field surface may be ponded or not. The
infiltration characteristics vary according to the surface and the infiltration is calculated for
each type. An average water height is than calculated for each gridcell, resulting in an
average hydraulic radius with which the velocity is calculated, The field surface has a
certain roughness and therefore only a part of the water will move.
Figure 2.7. Calculation of average water height and flowwidth
The velocity V (m/s) is calculated with Manning's formula:
V = R2/3 * sqrt(S) / n
In which: R = hydraulic radius (m), calculated with the flowwidth and average water
height(see figure 5); S = sine of the slope (fraction); n = Manning's n (dimensionless)
The discharge Q (m3 /s) per cell is then calculated with (Chow et al., 88):
LISEM MANUAL version 2.x - January 2, 2002
In which:
b = 0.6
A = wet cross section (m2)
There is evidence that alpha is constant for small rills and independent of slope and
resistance (Govers, 1997), because the flow will find a new equilibrium and uses its energy
to form new rill dimensions, while the velocity does not change. Research is currently being
done to see if this should be incorporated in LISEM.
For the distributed overland and channel flow routing, a four-point finite-difference solution
of the kinematic wave is used together with Manning's equation. Procedures of the
numerical solution can be found in Chow et al. (1988) and Moore & Foster (1990). The
kinematic wave is done over the Local Drain Directions map that forms a network which
connects cells in 8 directions.
Some cells may have a channel (ditch, gully) for which a separate kinematic wave is done
(see fig 6). The cells that have a channel receive a part of the overland flow, depending on
the velocity. Thus the velocity is considered the average velocity existing in the cell. The
channel is considered to be in the center of the cell so that the distance from the edge to the
channel is 0.5*(DX-channel width). The part of the water that flows into the channel is
f = dt .V/(0.5*(dx-cw))
The channel width (cw in m) can increase during the run if the channel is not rectangular
and the water level in it rises. In that case it cannot become larger than 0.9 of the cell width
Figure 2.8. Overland and channel flow with LDD structure
LISEM MANUAL version 2.x - January 2, 2002
2.6 Erosion and eposition
Detachment modelling is based on a generalised erosion-deposition as described by
Morgan et al., (1992) and Smith et al (1994), and is identical to the EUROSEM model. It is
assumed that the transport capacity concentration of the runoff reflects a balance between
the continuous counteracting processes of erosion and deposition (Dp). Erosion is sum of
splash detachment by rain drops (Ds), and flow detachment by runoff (Df). The amount of
sediment in suspension (e) is then calculated as:
e = Ds + Df - Dp
Splash detachment
Splash detachment Ds (g.s-1) is simulated as a function of soil aggregate stability, rainfall
kinetic energy and the depth of the surface water layer. The kinetic energy can arise from
both direct throughfall and drainage from leaves. Using splash tests the following general
equation has been derived (unpublished data):
Ds = (2.82/As Ke exp(-1.48 h) + 2.96).P.A
in which Ds is splash detachment (g.s-1), As is the aggregate stability (median number of
drops to decrease the aggregate by 50%), Ke is rainfall or throughfall kinetic energy (J.m-2),
h is the depth of the surface water layer (mm), P is the amount of rainfall or throughfall
under the plant canopy in the timestep (mm), A is the surface over which the splash takes
place (m2). In LISEM splash detachment is calculated as the sum of all splash under and
beside plants (throughfall and rainfall) and on ponded and dry areas:
Ds_rponded : A = (fpa)(1-cover)A, and Ke is Ker
Ds_tponded : A = (fpa)(cover)A, and Ke is Ket
Ds_rdry : A = (1-fpa)(1-cover)A, and Ke is Ker
Ds_tdry : A =(1-fpa)(cover)A, and Ke is Ket
The kinetic energy of free rainfall and throughfall from the plant canopy as respectively:
Ker = 8.95+8.44*log(I)
Ket = 15.8*sqrt(h)-5.87
where I is the rainfall intensity (mm/h) and h is the height of the plants (m). In equations 8a
to 8d f is the splash delivery fraction, a user defined fraction that determines the amount of
splashed soil that is transported through the air from the dry part of the gridcell to the wet
part of the gridcell, so that is can be transported). This means that although most splash
occurs on the dry part of the gridcell because of the exponential decrease of splash with
water height, only a part of it will be transported.
Flow detachment and deposition
The ability of flowing water to erode its bed is assumed independent of the amount of
material it carries and is only a function of the energy expended by the flow. Deposition
takes place at a rate equal to wCVs , where w is the width of flow (m), C is the sediment
concentration in the flow (kg.m-3) and Vsis the settling velocity of the particles (m.s-1). The
concentration at transport capacity (CT) represents the sediment concentration at which the
LISEM MANUAL version 2.x - January 2, 2002
rate of erosion by the flow and accompanying rate of deposition are in balance. In this
condition the net rate of erosion is zero and Df equals the deposition rate (wCTVs). The
equation for soil detachment by flow and deposition during flow, expressed in terms of
settling velocity and transport capacity, then becomes:
D = Y(Tc-C)Vs w dx
in which D is Df or Dp (in kg.s-1), Tc is the transport capacity of the flow (kg.m-3) and Y is a
dimensionless efficiency factor. The latter is included to account for the fact that the
detachment will be limited by the cohesion of the soil material. The pick-up rate for cohesive
soil therefore needs to be reduced by a coefficient whenever C is less than Tc. By definition,
Yis 1 deposition takes place, i.e. when C is larger than Tc , and when erosion takes place it
is calculated as (Rauws and Govers, 1988):
Y = umin/uc = 1/(0.89+0.56Coh)
in which uc and umin are the critical shear velocity and the minimum critical shear velocity
(cm.s-1), and Coh is the cohesion of the wet soil determined with a Torvane (kPa). Cohesion
by plant roots can be accounted for in LISEM by a second cohesion map that is added to
the soil cohesion. Note that dx is added because the unit length of a spatial element in a
raster environment becomes the grid cell size.
The transport capacity of overland flow is modeled as a function of unit stream power
(Govers, 1990):
Tc = ds c(w-wc) d
in which Tc is the volumetric transport capacity (kg.m-3), δs is the material density (2650
kg.m-3), ω is the stream power (calculated as flow velocity*energy slope) and ω c is the
critical stream power defined by Govers (1990) for a fairly wide range of materials to be
approximately 0.4 cm.s-1, c and d are experimental coefficients depending on the median
texture (d50) of the material. This equation describes the transport capacity in rills. In
LISEM equation 9 becomes (with V in m/s and the sine of the slope S):
Tc = 2650 c(VS100 - 0.4) d
Note that currently LISEM does not distinguish in its algorithms between interrill and rill
erosion. In fact all erosion is assumed to be rill erosion, although there are no rills defined
as such. Water simply flows from cell to cell following a drainage network (see below).
Currently the definition of a separate rill network is under development and then rill and
interrill transport capacity can be differentiated. LISEM uses the median of the texture (d50)
as input to represent spatial variability of grain sizes. The model calculates the
corresponding coefficients a and b based on work of Govers (1990):
c = [(D50 + 5)/0.32] -0.6
d = [(D50 + 5)/300] 0.25
NOTE: originally these equations are valid for materials with a grain szie diameter larger
than 32 µ m.
There are two checks carried out in LISEM to avoid the calculation of an incorrect Df or Dp.
The amount of erosion or deposition in a timestep depends on the settling velocity Vs With
a too large timestep it may happen that all sediment has already settled before the end of
LISEM MANUAL version 2.x - January 2, 2002
the timestep. Therefore the deposition is never more than the amount of sediment in
Dp = min(sediment, Dp)
In the case of detachment, the amount of detached soil cannot be more than the remaining
carrying capacity of the flow (Q ihn m3/s):
Df = (Tc - C) Q dt
The net sediment in suspension is transported between gridcell with the kinematic wave.
Channel detachment and deposition
Erosion and deposition in channels is treated in the same way as these processes on land.
There is no separate detachment equation LISEM, although the stream power concept was
initially developed for rills. The same formulas are used, but with the hydraulic radius,
velocity, discharge and transport capacity are based on channel maps for corss sectional
shape, bed-gradient and manning's n. The detachment efficiency is calculated with the
channel cohesion.
LISEM MANUAL version 2.x - January 2, 2002
to be done
to be done
to be done
LISEM MANUAL version 2.x - January 2, 2002
6.1 Introduction
In the MWISED project (EU ENV4-CT97-0687), Modeling Within Storm Erosion Dynamics,
LISEM was extended to cope with the incision and formation of ephemeral gullies. The
model was presented under the name EUROWISE, which is used here. EUROWISE is
synonym to LISEM-gullies.
EUROWISE was designed not to need much more data for gully incision than the original
LISEM. Incision can be modelled in a physically deterministic way by using shear stress of
the flow along the sides and bottom of the gully and compare it to the shear strength of the
profile. This would require a complete knowledge of the soil cohesion and variation of
cohesion with depth, and the capability of simulating the flow conditions in detail. Although
this can be done technically, such a model is very difficult to verify and the data would have
to be extrapolated from a few measurements to a large extent. A different approach was
chosen here: first flow incision takes place in certain areas only, derived from field
observations, and secondly when incision takes place the width of the gully is based on an
empirical relationship with the discharge. The amount of erosion is based on the more
deterministic erosion and deposition equations in LISEM, such as described in the LISEM
BASIC section.
Ephemeral gullies
The following text is copied from the thesis: "A spatial and temporal analysis of the
characteristics, importance and prediction of ephemeral gully erosion" by Jeroen
Nachtergaele (2001), who cooperated in the MWISED project.
Problems related to the definition of ephemeral gullies are two-fold. (1) Traditionally
ephemeral gullies have been defined in the way they differ from rills and (classical) gullies.
Such 'negative' definition is rather vague and open to discussion since it states what an
ephemeral gully is not instead of expressing the essential nature of this erosion feature. (2)
As it is often the case with newly defined concepts, many synonyms or nearly-synonyms
have been used to describe this erosion feature. In what follows, it is attempted to define
ephemeral gully erosion and its related erosion features in such a way that it is at least clear
what within this study is understood by the term ephemeral gully erosion.
Rill erosion is defined as erosion in numerous small channels that are uniformly distributed
across the slope and that can be obliterated by tillage (Hutchinson and Pritchard, 1976).
According to the Soil Science Society of America (1997) rill erosion is characterized by
numerous and randomly occurring small channels of only several centimetres in depth. Rills
form on sloping fields, mainly on cultivated soils. Rills can follow tillage marks, or they may
develop much like a drainage network of rivers in a large basin (Foster, 1986).
(Classical) gully erosion is defined as erosion in channels that are too deep to cross with
farm equipment (Hutchinson and Pritchard, 1976). According to the USDA Soil
Conservation Service (1966), classical gullies are channels formed by concentrated flow of
water, removing upland soil and parent material and of a size to large to be obliterated by
normal tillage operations.
Bank gullies form where a concentrated flow zone, a rill or an ephemeral gully crosses and
erodes an earth bank, e.g. a terrace, a river bank (Poesen, 1993; Poesen and Hooke,
1997). Bank gullies may develop upslope by head-cut migration.
LISEM MANUAL version 2.x - January 2, 2002
Ephemeral gully erosion was first comprehensively discussed by Foster (1986). The
introduction of ephemeral gullies as a separate erosion class resulted from the fact that in
the 1980's soil conservationists in the US became progressively aware that if only rills and
classical gullies are considered in soil erosion assessments, an important erosional area
and source of sediment within fields is being overlooked (Foster, 1986). The topography of
most fields causes runoff to collect and concentrate in a few major natural waterways before
leaving the fields. Erosion occurring in these channels is what is known as ephemeral gully
erosion (Foster, 1986). The 'ephemeral' nature of this erosion feature results from the fact
that ephemeral gullies are ploughed in and tilled across annually (or more frequent),
therefore restoring the original hollow, but leaving a potential zone for subsequent
ephemeral gully erosion. Poesen (1993) added to the view of Foster (1986) that ephemeral
gullies may also form where overland flow concentrates along (or in) linear landscape
elements (e.g. drill lines, plough furrows, parcel borders, access roads).
Definition-wise, incisions in linear landscape elements can be classified as rills since their
formation is associated with the micro-relief generated by tillage or landforming operations
(Haan et. al, 1994). However, incisions in linear landscape elements are a form of
concentrated flow erosion that will never develop in a standard USLE runoff plot used to
quantify interrill and rill erosion. According to the original idea behind the introduction of
ephemeral gully erosion as a separate erosion class (Foster, 1986), it seems sound to
consider incisions in linear landscape elements as part of ephemeral gully erosion.
Synonyms used to describe (erosion due to) ephemeral gullies are: concentrated flow
erosion (Foster, 1986; Auzet et al., 1995), talweg erosion (Papy and Souchère, 1993),
thalweg gullies (De Ploey, 1990), megarills (Foster, 1986), rills in valley floors (Evans and
Cook, 1987), valley-bottom rills (Boardman, 1992).
With respect to the differences between rill, ephemeral gully and (classical) gully erosion, it
is clear that ephemeral gullies and classical gullies can be discerned based on their
persistence. In contrast to the permanency of classical gullies, ephemeral gullies are shortlived. Definitions of ephemeral and classical gullies often refer to the means by which these
gullies are (resp. are not) obliterated (i.c. normal tillage operations). Defining the exact
operation or equipment, however, is less relevant since this is variable in both space and
time. Relevant is the fact that, once formed, a classical gully evolves by erosion of the gully
floor, head-cut migration and erosion of the gully walls, which implies a combination of
processes (e.g. water erosion and mass movement). As ephemeral gullies form, are
disguised and potentially form again on the same spot, their evolution requires mainly a
repetition of the incision process, while the relative importance of head-cut migration and
erosion of the gully walls is less significant.
Differences between rills and ephemeral gullies are not always very clear. Foster (1986)
gives two firm arguments to distinguish rills from gullies. (1) Flow in rills is usually classified
as a part of overland flow that is assumed to occur uniformly across a slope even though it
is concentrated in rills. In contrast, flow in ephemeral gullies is clearly channelized. Also,
erosion due to rills will affect the entire slope, while erosion due to ephemeral gullies is
much more confined. However, net erosion due to rills/ephemeral gullies should be
assessed by taking into account the combined effect of water erosion (formation) and tillage
erosion (removal). (2) Subsequent ephemeral gullies will occur in the same spot, while the
position of rills is variable from time to time since it is strongly influenced by microtopography (e.g. tillage marks). In addition to the arguments of Foster (1986), it can be
remarked that rills and gullies also differ in the way they contribute to the drainage pattern of
a watershed. While rills are usually limited by field borders, ephemeral gullies often extend
over multiple fields. In terms of erosion this implies that rills normally only redistribute soil
within one field, whereas ephemeral gullies transport soil material to completely different
parts of the watershed. The difference between rills and ephemeral gullies with respect to
LISEM MANUAL version 2.x - January 2, 2002
the way they affect a watershed's hydrology, is also clearly illustrated by Steegen et al.
(2000). They showed that there exists a positive relation between ephemeral gully
development in a given watershed and measured suspended sediment concentrations at
the outlet of that watershed for comparable rainfall events. In other words, ephemeral
gullies do not only act as sediment sources, but once established they also function as
efficient sediment transport ways. Despite the aforementioned arguments, there still exists a
small overlap between rills and ephemeral gullies. For example, when a slope shows
several clear rills, that concentrate gradually downslope and finally form a clear ephemeral
gully (Figure 1), where should the critical point between rill and ephemeral gully be placed?
It is clear that there exists a transition zone between rills and ephemeral gullies. While an
overlap between the two cognate concepts does not impede the existence of each
individual concept, problems may arise when assessing rills/ephemeral gullies in the field.
In order to have an objective measure to distinguish rills from gullies in such dubious cases,
Poesen (1993) proposed a critical cross-section of 929 cm2 or one square foot, a criterion
first used by Hauge (1977). Within this study this criterion was used to ensure that the data
sets assessed in all study areas and by all persons involved are fully comparable.
Figure 6.1. Photograph illustrating a transition from rill to ephemeral
gully erosion (January 1998, Guadalentin, south-east Spain, foto: J.Nachtergaele)
Finally, questions about the true need for a distinction between rills and ephemeral gullies
can often be heard from the side of soil erosion modellers. First of all the need for such a
distinction with respect to erosion modelling originates from the fact that some erosion
models explicitly state that they (only) cover interrill and rill erosion (e.g. USLE). What falls
beyond the scope of these models needs to be named and defined since it requires another
model or modelling approach. Concerning the erosion processes involved, modellers often
raise that rills and ephemeral gullies can be treated equally since the initiation of both linear
erosion features require soil material to be detached and transported by flowing water.
LISEM MANUAL version 2.x - January 2, 2002
Summer and winter gullies
The cross section criterion can be applied to classify both so called "summer gullies" in
central Belgium (shallow and very wide), and winter gullies (deep and narrow), as
ephemeral gullies (figure 2).
Figure 6.2. winter (left) and summer gully (right) in belgium (photos: I.Takken, KULeuven)
EUROWISE has to be able to simulate both types of gullies: shallow gullies that widen
when more erosion takes place and deep gullies that deepen and stay narrow. The
difference is caused by the soil cohesion: in summer the situation is often a seedbed above
a more compacted layer and sideways erosion encounters less resistance and costs less
6.2 Basic principles
EUROWISE is an extension of LISEM: a feedback loop is added whereby the simulated
flow detachment is recalculated to a loss of soil volume and causes a decrease of the
surface height in the digital elevation model. This in its turn causes a change in local slope
and flow path, which influences the erosion in the next time step.
LISEM MANUAL version 2.x - January 2, 2002
The modelling of gully incision and formation takes a three step approach:
1. An analysis of the landscape to determination likely locations of a gully as a result of
one or more rainfall events, called here "critical areas". These areas are selected on
the basis of topographic features such as slope and upstream drainage area.
2. The simulation of splash and flow detachment on a physically deterministic basis
everywhere in the catchment, using the principles of the LISEM BASIC.
3. Incision and change in gully dimensions as a result of flow detachment inside the
critical areas, using an empirical discharge-width relationship. The rest of the
catchment is assumed to have rill erosion only, which is not causing gullying. This is
an extension of the code to simulate the incision in a 2 layer soil based on bulk
density and cohesion, and an empirical width-discharge relationship.
Critical area analysis - theory
The position of the gullies is determined as the areas where topological critical thresholds
are exceeded. The algorithms in this section are based on the work of Nachtergaele et al.,
2001. He found that ephemeral gullies in the Belgian loessbelt and the Spanish
Guadelantine area often occur on the same locations. These zones of likely incision, termed
“critical areas” here, can be derived from an analysis of the area contributing runoff (A) to a
given location, and the local slope of that location (S). Different types of critical thresholds
for the Belgian loess belt are found in the literature. For the EUROWISE model, four
empirical critical threshold values are evaluated:
Critical threshold value according to Vandaele (1996)
Critical threshold value according to Moore et al. (1988)
Critical threshold value according to Vandaele et al. (1996)
Critical threshold value according to Desmet and Govers, (1997)
The critical threshold values are area specific. The equation of Moore et al., (1988) is a
more uniform relationship whereby A·S>18 is related to the erosive power concentrated
Hortonian overland flow, while the relation “ln(A/S)>6.8” is assumed to be a measure of soil
saturation and therefore is more related to the process of saturated overland flow.
According to Poesen et al. (1997), gully development in the Belgian loess belt is essentially
caused by concentrated Hortonian overland flow. Therefore Moore’s equation was
considered at first less appropriate for the area.
LISEM MANUAL version 2.x - January 2, 2002
After analysis it appeared that a second threshold is needed (Nachtergaele, 2001): when a
critical area is defined according to the algorithms above, the contributing area will soon
dominate the analysis and all area towards the outlet will be seen as critical. Thus a second
threshold is needed to give a lower boundary that will make the gully formation stop. For the
Belgian loessbelt it was shown that no gullying occurred on the lower slopes and thalwegs
when the slope angle decreased below 4%. Below this angle deposition processes become
Critical area analysis - Implementation in LISEM
Since the EUROWISE model uses raster maps created with the Geographical Information
System PCRaster (Van Deursen and Wesseling, 1998). This GIS combines the regular
spatial expression of any GIS with a modelling language that enables the analysis of flow of
material over a landscape and is therefore very suitable for this work package. The
language follows simple rules that can be combined in a "script", a kind of macro that is
executed with a compiler (PCRCALC.EXE). In PCRaster a network can be generated from
the digital elevation model (DEM) that defines the direction of flow over the landscape. The
network is called LDD, the Local Drain Direction. The LDD can be made directly from the
DEM by using the steepest angle of flow (see figure 1). To delimit critical areas, the
contributing area draining towards a given point has to be generated. In PCRaster the
command to do this is: =
accuflux(, cellarea());
This command accumulates the cellarea over the network and gives therefore in each cell
of the output map the cumulative area of all cells draining towards that cell, so e.g. the value
of the outlet of the catchment has the total area of the whole catchment above the outlet.
This is the PCRaster equivalent of the variable A in the algorithms above.
In an agricultural environment the direction of flow is often the tillage direction (at least for a
part of the year). Takken (2001) in her work on the influence of tillage on runoff, created a
PCRaster script to generate automatically a valid LDD based on the tillage direction per
field, while including features such as headlands, dead furrows and topography where
necessary. (see Fig. 3). It can be seen that the effect is very different: the concentration
zones follow the field pattern much more in the right hand image than in the left hand image
(Fig. 4). The main concentration lines are formed by the four small side valleys. The slope is
calculated by taking the difference in elevation between a grid cell and the cell towards it
drains, divided by the distance.
Figure 6.3. PCRaster LDD maps for a 40ha catchment (Limburg, NL): a) drainage direction (lines) and slope (grayscales)
according to the steepest slope, and b) according to tillage direction. Note that the field boundaries are preserved.
LISEM MANUAL version 2.x - January 2, 2002
Figure 6.4. PCRaster contributing area from accuflux (identical to variable A in the critical zone algorithms, see text): a)
patterns using the steepest slope, and b) patterns using the according the tillage direction.
However, the implementation of the above algorithms in PCRaster needs a certain
interpretation of the thresholds. Firstly, when the algorithms are implemented directly the
critical area is calculated on a cell to cell basis. That means that if a cell has a local slope
angle that does not meet the criteria, that cell may not be seen as a critical cell. This causes
fractioning of the critical zone. It must therefore be assumed that once a gully starts forming
it will not stop at a single location and continue downslope of that location simply because
the local slope is not sufficient. Therefore a more continuous zone has to be created.
Secondly, the gully formation will not continue endlessly and the algorithms show a
dominance of the drainage area A above the local slope S. In reality gully formation may
stop because the gradient at the bottom of the slope is low and deposition occurs. This has
been observed for the Belgium loess belt although not for the Mediterranean area.
The implementation in PCRaster is done as follows:
1) A critical zone is made according to the algorithms
2) the map edge is excluded from the zone because the flow path created in PCRaster
will consider the map edge as a boundary (so often a gully will form there);
3) This zone is accumulated along the network to give a continuous area;
4) A critical threshold slope angle is applied when necessary (depending on the area);
5) Isolated "critical" pixels are removed;
6) Isolated "non-critical" pixels are classified as "critical" if the pixels upstream and
downstream are also "critical" (repair holes in the critical zone).
===========begin of file=============
# Critical area analysis #
Fcrit1 =;
Fcrit2 =;
Fcrit3 =;
Fcrit4 =;
Fcrit5 =;
# create ldd
LISEM MANUAL version 2.x - January 2, 2002
report Ldd=lddcreate(Dem,1E25,1E25,1E25,1E25);
# create upstream area and slope map
# critical threshold according to Vandaele, 1996
report Fcrit1=if((Slope*((Upstream/B)**0.4)) gt 0.5 and Slope gt
# critical threshold according to Moore et al., 1988
report Fcrit2=if(Slope*(Upstream/B) gt 18 and ln((Upstream/B)/(Slope)) gt
# critical threshold according to Vandaele et al., 1996
report Fcrit3=if(Slope gt (0.025*((Upstream/10000)**-0.4)),0,nominal(10));
# critical threshold according to Desmet and Govers
report Fcrit4=if((Slope*((Upstream/B)**0.4)) gt 0.72,0,nominal(10));
# critical threshold according to Vandaele et al., 1997, Portugal
report Fcrit5=if(Slope*(Upstream/B) gt 40 and ln((Upstream/B)/(Slope)) gt
===========end of file=============
An example: critical areas for the Kinderveld test site (BE)
Figure 5 shows 6 images of the Kinderveld test area near Leuven (BE) where gullies where
monitored in the MWISED project (see introduction). The test site has a size of 13.7 ha and
a grid cell size of 5x5 m. The measured gullies are shown as a black line while the white
areas are the 5 critical areas susceptible to gully erosion. It can be seen that all critical
areas include the gullies except for algorithm 5 which begins too far downslope. Also the
figure shows that the algorithms 1 and 3 produce too large areas while 2 and 4 are closer to
reality. Since the main overland flow process is Hortonian overland flow, criteria 4 seems
the to give the best result.
Ephemeral gully formation
The section above results in a water flux and sediment flux on the surface, that is distributed
using the kinematic wave principle. The sediment is a result of erosion which may or may
not cause gully formation, depending on the location in the field. As is explained above the
gullying is assumed to take place ONLY in the critical areas. The following approach is
taken to simulate the incision processes.
Gully width
The gully formation, i.e. the deepening of the critical area, is not necessarily over the whole
grid cell width: the gully width depends on the peak discharge of the runoff in a non-linear
way: w = a*Qb. The parameters a and b are predefined for Belgium from laboratory and field
For areas which exceeds the critical threshold value, gully width is calculated according to
the relation: W = a·Qpeakb, with a and b as empirical constants, W is the gully width in meters
and Qpeak is the peak discharge of the runoff in m3/s. For the Belgium loess belt these
relationships are found by KULEUVEN for different kind of gullies in field and in
laboratories: the best relation is:
Note that the user can define the parameters a and b in the interface.
LISEM MANUAL version 2.x - January 2, 2002
Figure 6.5. Critical areas according to the 5 algorithms (white). The lines are the actual mapped gullies.
The gully width has to be initialized in order for the erosion to start. A sensitivity analysis
shows that this value is not very sensitive and an arbitrary value of 0.1 m is chosen. The
width of the gullies increases with increasing runoff discharge till the point of the maximum
peak discharge (Fig. 6). At that point gully width is assumed to be constant.
Q (l/s)
width (m)
LISEM MANUAL version 2.x - January 2, 2002
time (min)
discharge (l/s)
w=2.37Q^0.37 (m)
gully width (m)
Figure 6.6. Simulated gully width when the peak discharge is passing a pixel which exceeds the critical threshold.
Gully incision
For each gridcell and timestep the processes above result in the water discharge Q (kg.m-3),
the corresponding gully width w (m) and the sediment discharge Qs (kg). The eroded mass
of the soil is recalculated to a depth depending on the situation:
1. Homogeneous soil: only the depth d of the gully increases over the gully width w using
the amount of flow detachment Df and the bulk density D of the soil: d = Df/(w*dx*D).
Note that these parameters are represented by maps that can be spatially variable, only the
depth is assumed homogeneous.
2. Two layers: in case the soil has a compacted layer with a higher cohesion (e.g. a plough
layer), both the gully width and depth have to be adjusted. In this case the gully often
becomes shallow and wide because the sideways erosion is easier than the downward
erosion. This effect is not accounted for in the Q-w relationship. Thus the following
system is devised: if the depth of the gully reaches the second layer AND the water level
in the gully still "touches" the first layer the flow detachment is divided over the walls and
bottom of the gully as follows:
where: h is the water depth (m), Pw is the wet perimeter (m), and Y1 and Y2 are the
efficiencies based on the cohesions of the 1st and 2nd layer. The gully width w and depth
d are increased according to:
The gully width cannot become larger than the grid cell width. During the calculations the
digital elevation model changes. For each time step a new DEM is calculated in the
dynamic loop of PC Raster. This new DEM is derived from old DEM with subtraction of the
erosion or addition of the sedimentation. Changes in the DEM causes changes in local drain
direction (LDD), which allows fluctuations in the stream pattern of the gullies.
LISEM MANUAL version 2.x - January 2, 2002
The input database consists of a series of raster maps created with or converted to the GIS
PCRaster. Furthermore LISEM needs event based (so called breakpoint) data of rainfall
intensity for one or more rainstorms. Note that LISEM does not yet simulate
evapotranspiration so that this is not taken into consideration when dry periods between
rainfall events are calculated.
In the next sections you will find an overview of the input data and the commands needed to
create them with PCRaster. Note that a PCRaster script file is available from the website
which generates the complete input dataset from a few basic maps.
All input and output maps in LISEM are in the format of the PCRaster Geographical
Information System. PCRaster is a free GIS is produced by PCRaster environmental
software, and can be obtained through their website:
There you will find a complete list of functions, research and coding examples and literature.
PCRaster is not only a normal GIS but a generic programming language for dynamic
landscape modelling. If you are interested in using PCRaster for modelling purposes, a
distance learning course is available through the internet:
PCRaster is used in two contexts here:
1. LISEM is largely built in a mixture of PCRaster commands and C++
2. The modelling capabilities of PCRaster are used to create the input maps with one
or more macros (scripts).
LISEM is developed in an environment that is a mix of C++ and PCRaster libraries. A large
part of the source code is written in the macro language of PCRaster The advantage of this
close integration is primarily for the model developer: the PCRaster macro-language
includes a powerful 'calculator' that is able to parse virtually any mathematical or logical
equation. Equations that are commonly found in erosion models can be typed almost
literally from a text to include them into the model. An example: the Manning's equation can
be written in PCRaster as a string that is given to a 'calculator' (the statements are
translated to C++ at compilation):
V=if(n gt 0, R^(2/3)*sqrt(max(slope(DEM), 001))/n,0);
where V, R, DEM and n are raster maps, and slope is the GIS operation to calculate the
gradient from the DEM but it cannot be smaller then 0.001. The instruction is carried out for
all cells in the raster maps. As a precaution, the logical "if" statement sets the Velocity to 0
where n is 0. The advantage of this approach is that the code is very compact and that
several developers can work with the code at the same time, without the need of knowing a
higher programming language such as C++ or Pascal.
LISEM MANUAL version 2.x - January 2, 2002
Note that for processes where iteration is required (such as the kinematic wave and
Richard's type infiltration) special functions are developed. The advantage for the user lies
in the fact that the output maps of runoff and erosion/deposition can be viewed directly with
PCRaster without the need of conversions.
PCRaster can also be used to vcreate input maps using separate statements for each map,
examples of which are given in the "input maps" section. These commands have the
general syntax:
pcrcalc [options] =
where a large number of opertaions are available for the Note that PCRaster is very
strict on the variable types which have to be specified explicitely if the type type of resulting
map is ambiguous. The types are:
boolean = 0/1
directional = 0-360
ldd = network direction (1-9)
nominal = class integer
ordinal = class integer
scalar = real value
Alternatively all input maps can be created in one operation with the script that can be found
on the LISEM website. The script is run with the command:
pcrcalc -f lisemmaps.mod
which produces the LISEM input maps based on 4 basic catchment maps.
NOTE: PCRaster has no capabilities of digitizing maps. It can however import maps from
IDRISI and ARC-INFO (basic Arc-grid format). Also PCRaster has extensive capabilities of
geostatistical interpolation.
7.2 The input maps
All input and output maps in LISEM are in the format of the PCRaster GIS. This raster GIS
is produced by PCRaster environmental software, and can be obtained through their
website, where you will find a complete list of functions and examples.
LISEM needs a minimum of 24 maps depending on the input options selected in the
interface (e.g. enabling the simulation of ditches requires additional maps, and some
infiltration option require more maps than others). If needed all maps can be derived from 4
base maps:
digital elevation model,
land use,
soil type and
impermeable areas such as tarred roads (this is needed because the road can be
narrower than the gridcell size and LISEM needs to know what type of soil or land use
exists adjacent to the road).
All input data can be derived form these four base maps, but off course other sources can
be used to make the input maps (geostatistical interpolation, remote sensing etc.).
LISEM MANUAL version 2.x - January 2, 2002
The spatial input data for LISEM can be structured as follows:
Catchment maps
Vegetation maps
Soil surface maps
Infiltration related maps
Erosion/deposition related maps
In this part all input maps are listed with the PCRaster commands to make them. Since
PCRaster is also capable of running macros and even spatial models, all commands given
here are combined in a script, so that the complete database can be constructed in one go.
The script can be downloaded from the website.
NOTE: the map names given here are NOT compulsory but can be specified individually by the user
in the interface. However, to enable compatibility with the older versions of LISEM the map names
used here are the default names. Also the extension ".map" is NOT compulsory, just convenient.
Catchment maps
The actual digital elevation model is not used in LISEM, only maps derived from it. This has
the advantage that a number of important maps such as the surface drainage network
(called Local Drain Direction or LDD in PCRaster) and Gradient are not automatically
derived with GIS operations, but can be supplied by the user.
Map name
Unit Range
Local drain direction
1-9 (see below)
Catchment boundaries
Area covered by raingauges
1- n (= # of raingauges)
Slope gradient (sine of slope angle)
Location of outlet and suboutlets
must be > 0 and <= 1
LDD: Local Drain Direction, this map gives for each cell the direction of the surface runoff
(the number corrspond to a keypad on your keyboard: 1=drainage to lower left, 6 is
drainage to the right etc.) the number 5 is reserved for the catchment outlet. Note that this
map cannot contain local depressions, the entire catchment has to drain towards the outlet.
PCRaster provides commands that can automatically generate such a map, with removal of
the local depressions ('pits') based on constraints given by the user. In this case the LDD is
based on the steepest slope. The command is:
The 4 large values are thresholds to remove the pits and create a continuous LDD.
It is also ppossible to generate a valid LDD map based on the tillage direction. A PCRaster
macro has been constructed by Takken et al. (2000) that combines major topographic
concenration lines with tillage direction per field, including 'headlands' and 'dead furrows'.
NOTE: errors with LISEM related to the kinematic wave where MVs are reported (missing
values) are often because the LDD does not have the same dimensions as the other maps.
LISEM MANUAL version 2.x - January 2, 2002
AREA: all maps are checked against this map for the number and location of non-missing
value cells. If maps do not have the same size and number of valid cells, an error occurs.
pcrcalc, pit(
GRAD: this is the gradient in the direction of the LDD. PLEASE NOTE: this must be the
SINE of the slope (not the tangent). Because the slope can occur in the denominator of
various algorithms, it cannot be 0 (zero). Use the slope command to create the slope from
the DEM:
or make a gradient map with the tangent of the slope in the direction of the flow (it has to be
covered with the slope derived from the DEM to account for edge cells):
convert this map to a sine and make sure it is larger than 0::
ID: this map can be made in various ways, using Tiessen Polygons or zones based on
relief. The numbers in this map must correspond to the number of rain gauges in the rainfall
file. An example is to create a map according to Thiessen polygons from a map with pxiels
that have the numbers of the rain gauges:
Alternatively the relief can be taken more into account by taking the slope as a weighing
or by using the mean elevation line between two gauges as a separation line (gauges have
id numbers 1, 2 and 3 here, dem is a map):
pcrcalc a1 = maptotal(if(p1 eq 1,dem,0))
pcrcalc a2 = maptotal(if(p2 eq 2,dem,0))
pcrcalc a3 = maptotal(if(p3 eq 3,dem,0))
pcrcalc a4 = (a1+a2)/2
pcrcalc a5 = (a2+a3)/2
pcrcalc = nominal(if(dem lt a4,1,if(dem lt a5 and dem
ge a4, 2, 3)))
NOTE: that the gauges must be situated inside the area otherwise it will not be
included in the spread operation.
OUTLET: a map with values 0 (background) and 1 which is the main outlet. On this cells a
hydrograph and sedigraph are stored as ASCII files. Two additional cells can be given
(numbers 2 and 3 compulsory) to store hydrographs and sedigraphs for additional locations
(e.g. sub catchments). The main outlet is found by:
LISEM MANUAL version 2.x - January 2, 2002
Vegetation maps
Vegetation and/or crops are mainly used by the interception and splash erosion processes
in LISEM. The maps are usually made by reclassifying the field boundary map. vegetation
influences also the soil cohesion for which a separate map must be defined (see below).
Note that LISEM has never been tested for forest areas and should be used with care in
these circumstances.
Map name
Leaf area index
scalar -
Fraction of soil covered by vegetation
scalar -
Vegetation height
scalar m
LAI: leaf area index, used to calculate the storage capacity of the canopy. The value
represents the LAI of the vegetated part of the gridcell. For instance if a grid cell has a
single tree with a LAI of 6 in it, the LAI of that cell is 6.
PER: fraction of soil cover by canopy, but also litter or other types of soil cover (except for
CH: crop height, used to calculate throughfall kinetic energy in the splash erosion part.
These maps can be made in similar operations if a table is available that has as the first
column the id numbers of the landuse map-units, and in the other columns values for the
LAI, cover fraction and vegetation height (the option "--matrixtable" is needed because of
the table layout conventions in PCRaster):
pcrcalc --matrixtable = lookupscalar(, 1,
pcrcalc --matrixtable = lookupscalar(, 2,
pcrcalc --matrixtable = lookupscalar(, 3,
The table may look like this (four landuse types with different IDs not necessarily
======begin of file======
1.23 0.35 0.4
16 6.1
0.95 1.8
======end of file======
NOTE: LISEM assumes no logic in this table! It will not check if a LAI of 6 is combined with
a cover fraction of 0.
Soil surface maps
LISEM MANUAL version 2.x - January 2, 2002
This is a mixed catagory with maps that determine several processes: infiltration, surface
storage, overland flow (veloccity). In an agricultural catchment these maps are based on the
landuse mostly, or on a combination of landuse and soils. Note that in a crop rotation cycle
these values will change according to for instance crusting stage, soil cover and tillage
It also includes different types of surface for which the processes are calculated separately.
Map name
Manning's n
scalar -
0.001 - 10
Random Roughness
scalar cm
0.05 - 20
Fraction covered with stones
scalar -
Fraction cover with a crust
scalar -
width of impermeable roads
scalar m
N: Surface resistance to flow, expressed as Manning's n.
RR: 'random roughness' used here as the standard deviation of the micro relief heights (do
not use transformations to obtain a more 'normal' distribution).
STONEFRC: fraction of stones for which no splash erosion is calculated (used only in this
CRUSTFRC: fraction of the soil crusted. On gridcells with non-zero values the infiltration is
calculated according to soil physical properties supplied by the user, depending on the
infiltration option chosen.
ROADWIDT: width of imermeable roads for which no infiltration is calculated (is used in all
infiltration modules).
pcrcalc --matrixtable = lookupscalar(, 1,
pcrcalc --matrixtable = lookupscalar(, 2,
pcrcalc --matrixtable = lookupscalar(,
3, surface.tbl)
pcrcalc --matrixtable = lookupscalar(,
4, surface.tbl)
pcrcalc = scalar(if ( neq 0, *width*,
where "surface.tbl" is a table such as used with the vegetation related maps, and *width* is
a value for the road width in m.
LISEM MANUAL version 2.x - January 2, 2002
Infiltration related maps
There are currently 6 infiltration options in LISEM, based on 3 different models. Depending
on the choice of the user in the interface, a different set of input maps is needed.
no infiltration: no maps needed
SWATRE (including crusts, wheel tracks and grass strips)
Green & Ampt: 1 and 2 layers and additional surface types
Morell-Seytoux & Verdin
No infiltration
The entire catchment is assumed impermeable. Can be used for testing.
To operate the SWATRE infiltration module , the user has to set up a series of maps and
tables with soil physical properties that describe the profile (max 20 layers) of each land
unit. These can be e.g. soil type units or fields or parts of fields.
Map name
ASCII lookup file describing soil profiles of map units
ASCII tables with soil physical data
map with profile ID numbers
nominal -
>= 1
map with ID number of crust profiles
nominal -
>= 1
ID number of wheel track profile
nominal -
>= 1
ID number of grass strip profile
nominal -
>= 1
locations of matric potential output
nominal -
0, 1-4
positive initial matric potential of each
soil layer
PROFILE.INP: when the SWATRE infiltration model is used LISEM needs a 3D structure of
the soil. This structure is described in the text file PROFILE.INP. In this file the number of
nodes are given for the solution of the Richards equation (using a finite difference
approximation) and their depths. These are independent of the actual soil horizons. NOTE
that the node depths are used in LISEM as boundaries between simulation layers (not
centres). Below the nodes the soil profiles are given (see syntax below) with 'pointers' to the
names of the tables describing the soil physical relations (pF curve and hydraulic
conductivity curve) of a particular soil type.
NOTE that in this file ALL information is stored, including that of the units for crusts, wheel
tracks and grass strips. So if these features are simulated the file PROFILE.INP has to be
adjusted to reflect their presence.
LISEM MANUAL version 2.x - January 2, 2002
The file PROFILE.INP has the following layout:
Begin of file ==============
number of soil layers
1:depth of boundary between first and second layer (cm)
2:depth of boundary between second and third layer
3: etc.
14: depth of final soil layer (cm)
description of map unit 1. Used in Swatre if
the name of the soil physics table to be used
the depth down to which the above table is used (cm)
use this table for depth 15-30
description of map unit 2. Used in Swatre if
End of file ==================
name.TBL: in the soil above three horizons are defined, each of them with seperate soil
phyisical characteristics, referring to the specific tables. The horizons used here are: 0-15
cm, 15-30 cm and 30-100 cm. The tables used for these layers contain 3 columns: soil
moisture content (theta, fraction), pressure head (h in cm), and unsaturated conductivity (k
in cm/day). An example:
LISEM MANUAL version 2.x - January 2, 2002
0.01 -4.86E+24
0.02 -1.04E+07
0.03 -9.02E+05
0.04 -2.15E+05
0.31 -3.19E+01
0.32 -2.33E+01
0.33 -1.54E+01
PROFILE.MAP: a map based on soils, fields or a combination, with ID numbers that are
used to look up the corresponding soil physics tables in PROFILE.INP.
PROFCRST.MAP: a map with ID numbers on those places where there is a crust, to look
up the soil profiles for crusted soils in the file PROFILE.INP. The areas that are crusted are
defined in CRUSTFRC.MAP. There can be more crust profiles if there are for instance more
crop types with a different developing surface. Simply include more crust profile descriptions
PROFGRAS.MAP: a map with ID numbers corresponding to the soil profile for grass strips.
The area where strips are present are given in GRASSWID.MAP.
PROFWHL.MAP: a map with ID numbers corresponding to the soil profile for wheel tracks.
The area where tracks are present are given in WHEELWID.MAP. Note that there are more
maps to define wheeltracks.
HEADOUT.MAP: map with up to four cells numbered 1 to 4 (de rest having value 0). When
the option matric output is checked in the infiltration options of the input menu, an ASCII
table with matric potential for each layer at each timestep is created.
INITHEAD.001 - INITHEAD.020: initial matric potential for each layer (if for instance you
have defined 6 layers in PROFILE.INP, there are 6 maps 001 to 006). The values are
positive matric potentials in cm.
The Holtan infiltration model is a one layer empirical model. The maps need no further
description. Note that the extra infiltration options (wheel tracks, crusts and grass strips) are
not implemented for the Holtan model.
Map name
infiltration rate saturation
difference initial and saturation rate
soil porosity
depth infiltration control zone
coefficient infiltration equation
initial soil moisture content
soil moisture at field capacity
LISEM MANUAL version 2.x - January 2, 2002
Green & Ampt, 1 and 2 layers and additional surface types
G&A demands the hydraulic properties of the first layer. NOTE that the soil depth is only of
consequence if the option "impermeable sub soil" is checked, meaning that the soil storage
is limited to the product of the moisture deficit and the depth: (thetas1-thetai1)*soildep1. If
this option is not checked the soil is assumed of unlimited depth and at moisture content
thetai1 for its entire depth.
Map name
Saturated hydraulic conductivity
Saturated volumetric soil moisture content
Initial volumetric soil moisture content
Soil water tension at the wetting front
Soil depth
G&A1: Additional surfaces
In addition separate ksat maps have to be defined for crusted soils, wheel tracks or grass
strips. These maps are required when any (or all) of the extra options in the infiltration
options menu are checked.
Map name
wheeltrack Ksat, used with infiltration option
"include wheel tracks"
grass strip Ksat, used with infiltration option
"include grass strips"
crust Ksat, used with infiltration option
"include crusts"
NOTE: in addition the user has to define the maps WHEELWID.MAP, GRASSWID.MAP
These options are only used for the first layer of the Green and Ampt module. If a 2 layer
Green and Ampt is used, the second layer will be identical for all option.
G&A2: Two layer Green & Ampt
Requires all maps of Green and Ampt 1 layer plus an identical set for 2nd layer. The depth
of the second layer (soildep2) is only of consequence when the "impermeable sub-soil"
option is checked (see remark above).
Map name
see KSAT1
see KSAT1
see PSI1
see PSI1
LISEM MANUAL version 2.x - January 2, 2002
Morel-Seytoux & Verdin
Not implemented yet.
Erosion/deposition related maps
Erosion and deposition in LISEM is based on the deficit transport capacity (capacity surplus
equals deposition and transport deficit equals erosion). The particles in flow have a settling
velocity that is estimated by the settling velocity of the median fraction (D50 in mu). To
calculate more sediment classes the LISEM-MULTICLASS SEDIMENT version has to be
Map name
Aggregate stability
0.00001-200; -1
Cohesion of bare soil
COH+COHADD >= 0.196
Additional cohesion by roots
COH+COHADD >= 0.196
D50 value of the soil
(advise:) 25-300
AGGRSTAB: aggregate stability is used for the splash erosion. It is the median number of
the drops needed to decrease the size of the aggregates by half on a sieve (LOWE test).
COH: Cohesion of the soil (kPa) measured with a torvane. The cohesion value is used in an
empirical formula to decrease the transport deficit with a factor between 0 (no erosion) and
1 (cohesionless soil).
COHADD: Additional cohesion (kPa) to be added by the user to simulate the effect of plant
roots on the soil strength. Note that this is a calibration value as the Torvane test is not
suited to measure the effect of roots.
D50: median of the texture of the soil (µm) used to simulate the settling velocity. This value
determines strongly the depostion.
WARNING: if water with sediment with e.g. D50=25 flows over a gridcell with a D50=150, all
sediment will deposit, even if the sediment in flow has a D50 that is less. To avoid these
kind of effects use a homogeneous D50 map or better use the version LISEM
Channels in LISEM are meant to be man-made ditches with a width much smaller than the
grid cell (typically a ditch of less than 1.5 m wide in a grid cell of 10x10 m). The channel
network MUST be a continuous network connected to the outlet. Water that enters the
channel in LISEM does not come out again (the channels do not overflow) and if the
channels are not connected to the outlet they are effectively sinks. LISEM does a separate
kinematic wave for the channel.
LISEM MANUAL version 2.x - January 2, 2002
Simulations can also be done considering a concentrated flow line, such as a thalweg, to be
a channel. The PCRaster operation accuflux can be used to determine a channel based
on the ldd:
pcrcalc = scalar(if(accuflux(ldd,cellarea()) lt
10000, 1))
This command creates a mask of a channel (having value 1 as the channel and MV
outside) based on the contributing area: all cells that have 10000 m2 of drainage area
upstream are considered a channel. The mask itself is not used in LISEM but it can be used
to create the channel input maps:
CAUTION: The channel form (hydraulic radius) and resistance (Manning's n) determine to a
large extent the shape of the hydrograph. Channels are considered impermeable.
Map name
Units Range
Local drain direction of channel network
Channel gradient
0.0001-10 6)
Manning's n for the channel
Cohesion of the channel bed
> 0.196
Width of channel
Channel cross section shape
LDDCHAN: the LDD map based on the channel mask alone. This map can opnly have one
pit that has to be the same as the pit (outlet) of the LDD. A way to create this map is:
CHANGRAD: the gradient of the channel bed based on the channel mask alone. A way to
create this map is:
CHANMAN: the Manning's n (resistance to flow of the channel). Usually this is not known
and a literature value (ditches with vegetation) is assumed. If the channel is used to
simulate a thalweg, the manning's n of the land use could be used (, see above):
CHANCOH: the cohesion (kPa) of the channel bed, resistance to flow erosion. Usually this
is not known and a literature value (ditches with vegetation) is assumed, or of a concrete
channel is simulated, a very high cohesion could be assumed. If the channel is used to
simulate a thalweg, the cohesion of the land use could be used ( and,
see above):
CHANWIDT: the channel bed width in m. Because the channel shape is not necessarily
rectangular, the top width at water level, and the hydraulic radius, is calculated from the side
angle and the water height in the channel. This value should be smaller than the grid cell
LISEM MANUAL version 2.x - January 2, 2002
width (not more than 90% of the cell width). The width is ussually a field measured value
pcrcalc ( ne 0, *width*)
CHANSIDE: tangent of the angle between the channel side and the vertical: 0 is a
rectangular channel, 1 is a side angle of 45 degrees, 10 is a very wide channel (84
degrees). Note that the width of the top of the water level depends on this side angle, if the
water rises the stream becomes wider, but it cannot become wider than the gridcell.
pcrcalc ( ne 0, 0)
7.3 Additional maps needed for EUROWISE
All input and output maps in LISEM are in the format of the PCRaster GIS. This raster GIS
is produced by PCRaster environmental software, and can be obtained through their
website, where you will find a complete list of functions and examples.
In order to run EUROWISE (LISEM-GULLIES) a few additional maps are needed, on top of
the regular maps needed with the LISEM BASIC version.
Note that the names are not compulsory but default in the current version. They can be
changed individually in the interface.
EUROWISE needs a Digital Elevation model for the critical area assessment, and a few
additional maps that define the structure of the layers to be eroded.
Map name
Digital elevation model
0- top of Mount Everest
Initial width of gully
must be > 0
Resistance of gully bottom
Bulk density first layer
kg/m3 must be > 1000
Depth to second layer
must be > 0
Cohesion of second layer
must be > 0.2
Bulk density second layer
kg/m3 must be > 1000
DEM: Digital elevation model, used to delimit the critical areas according to the algorithm
chosen in the interface by the user. The analysis is done inside EUROWISE.
GULWIDTH: initial width of the gully for those areas where incision takes place (in practice
this is the critical area). This parameter is maybe not necessary and will disappear in the
future based on a sensitivity analysis of various areas.
LISEM MANUAL version 2.x - January 2, 2002
GULLYMAN: resistance to flow at the bottom of the gully based on Manning's n. This wilol
be purely based on assumptions, but a value is needed (like the "channel" maps in the
lisem basic version)
BULKDENS: this map can be made by interpolation of measured values or by reclassifying
a soil or land use map. The classification can be done with a lookup table as is explained in
the LISEM BASIC input map section (soil surface maps):
pcrcalc, soil.tbl, 1)
where soil.tbl is a text file containing bulk density values of each map unit.
NOTE: that the gauges must be situated inside the area otherwise it will not be
included in the spread operation.
GULDEPTH: the depth to a second layer with a higher bulk density and cohesion. Although
in reality there can be a gradual increase in bulk density, EUROWISE assumes an abrupt
change to simulate the situation of a loose seedbed on a more compact sub-soil. This map
determines to a large extent whether the widening of the gully. The depth can be assumed
pcrcalc ( eq 1,8,12))
gives a map with a depth of 8 cm for unit 1 and 12 cm for all other units. Also a random
change (noise) around a mean value could be assumed:
gives a map with a depth of 8 cm with a normal distributed noise that has a standard
deviation of 1 cm.
GULCOH2: Cohesion of the second layer, see the map COH.MAP.
GULBD2: see the bulk density map above.
LISEM MANUAL version 2.x - January 2, 2002
7.4 Rainfall data
The file containing the rainfall data is in ASCII-format and can have any name. The rainfall
file should have the following structure (the second column is for explanation only and
should be ignored in the real file).
========begin of file========
Fixed text, 1 indicates the number of
name of gauge 1, if more gauges are
present then put each name on a
seperate line.
empty line, required!
0.000 00.00
20.00 60.00
indicate 60 mm/hours rainfall intensity
from minute 0 to 20
250.0 00.00
========end of file==========
If two or more rain gauges are used, e.g. 3 (referring to the numbers used in ID.MAP ), the
file should have the following structure:
========begin of file========
955.00 0.00 00.00 0.00
960.00 2.53 20.00 0.50
960.30 2.51 10.00 0.50
960.90 2.49 05.63 0.50
961.97 2.44 10.00 0.50
965.87 2.27 60.00 6.66
1300.00 0.00 00.00 0.00
========end of file==========
The first column is: time (in minutes). The second column is the rainfall intensity (in mm/h)
of rain gauge number 1, the third column is the intensity of raingauge number 2 etc.
So, from time 955 to 960 (minutes), the rainfall intensity at raingauge 1 is 2.53 mm/h. From
time 965.87 to time 1300 there is no rainfall.
The last time entered should be large enough to allow all runoff to reach the catchment
outlet. This time should be equal to or larger than the time at which the simulation is
stopped, specified in the interface.
Note:the rainfall file is not in standard PCRaster format. This format was an experimental
format and is only used in LISEM. It may change in the future.
LISEM MANUAL version 2.x - January 2, 2002
The five version of LISEM are contained in a single executable:
When it does not exist, a file lisemwin.ini is created in which the last workspace is saved.
LISEM opens with the following screen:
Pressing one of the buttons will generate a LISEM interface that only contains the options
needed to run that particular version of LISEM. The basic interface looks as follows:
LISEM MANUAL version 2.x - January 2, 2002
8.1 Speed buttons & Tabs
The icons on top of the interface are:
load a run file
save a run file
save a run file under a different name
save the directory name of the workspace in the lisemwin.ini file
NOTE that the lisemwin.ini file also contains the LISEM Type (0 =
Basic, 1=Wheeltracks etc.) and the names upto 4 runfiles (full pathnames needed). If the variable LisemType = -1, LISEM starts with the
opening screen (top picture), else it starts straight away with the
input screen, and starts executing the runfiles in the list.
dump the active (visible) screen to a printer (e.g. the simulation screen)
dump the active screen to a GIF file
resize the screen so that it is entirely visible (800x600 pixels)
the icons:
are the run buttons:
o the arrow executes the run-file list
o the pause icon pauses the run and it can be resumed with the arrow
o the cross halts the current run and starts the next
o the last symbols halts all runs
On the top of the screen a number of tabs are visible:
1. Start : general directories and options
2. Basic Maps : input map names
o Swatre infiltration options
3. Simulation : water balance totals, hydrograph and sedigraph
4. Output maps : output timeseries
8.2 Start tab
Input directories
Names of workspace, input directories and rainfall file. Clicking on the icons brings up a
directory tree to change the dir or it can be types. Clicking on the "text page" icon brings up
a window with the rainfall file for inspection (not edit).
Output directories
Specification of the output directories and file names. The result directory is made
automatically if it doesn't exist. Limit the number of runoff map characters: the runoff maps
are stored as: try00010.500, try00010.750 etc. with simulation time minutes before the dot
and fractions of minutes as extension. Selecting the PCRaster timeseries option creates
sequential map names: try0000.001, try0000.002 etc.
The runoff maps are optionally stored as discharge (l/s) or unit discharge (l/s/m).
The erosion (positive values) and deposition (negative values) maps are stored in
LISEM MANUAL version 2.x - January 2, 2002
Simulation time
Begin and end times of the simulation in minutes: these have to be present in the rainfall
file. The timestep is in seconds: be sure to make it not too large (in general equal to the
gridcell size in m). A good rule of thumb is to use simulation intervals in sec that are 0.2-2
times the grid cell size in m. Do not make it too small either, this causes instabilities
(apparently!). A smaller timestep results in a higher peek because of a greater accuracy of
the kinematic wave solution.
Output times
Definition of output times to store the runoff timeseries. There are two options: give an
interval (e.g. every 4 timesteps) or define individual timesteps.
Runfile list
List of runfiles loaded with the 'load file' button. More files are added until the "trashcan"
button is pressed and the list is emptied. LISEM will start executing the runfiles one by one.
DOUBLE CLICKING on a runfile name will activate it and load the options stored in it. The
list of runfiles is executed from the file you double clicked to the end of the list. You can
select several run files in the "file open" window by holding 'shift' or 'ctrl' and pointing the
An example of a run file is given in the appendix. It can be edited by hand but the variable
names should not be edited because LISEM uses them to recognize the them.
Model options
No erosion, runoff only: all erosion calculations are skipped.
Simulate main channels: channel calculations are done, additional maps are
Channel infiltration: ponded infiltration in channel using simple subtraction of a Ksat
map. The infiltration will stop if the storage (given in mm) is filled up. Additional maps
NOTE: this option does not work properly at the moment!
LISEM MANUAL version 2.x - January 2, 2002
Additional options
A few parameters in LISEM do not need to be defined in a spatial way but as a single
constant in the interface:
Splash delivery ratio: the relative amount of sediment that is splashed from the dry
part of a gridcell to the wet part of a grid cell so that it can be transported. The
remainder of the splash sediment settles again as splash deposition.
Manning's n grass strips: a large Manning's n value is given here to simulate the
friction of flow on a grass strip. As the strip is generally smaller than a gridcell, the
surrounding Manning's n is given in the Manning's n map, while the strip itself is
specified here.
Surface types
Activating this option will enable the user to define surface types that "occupy" only a
fraction of the pixel. This influences infiltration and in the case of grass strips also the flow
resistance. At least two additional maps are needed: one that defines for each cell the
fraction of the cell containing this type (e.g. contains values 0-1) and other
maps that define the infiltration characteristics (e.g. tell LISEM how much the
ksat is for the crusted fraction of a cell). When the SWATRE infiltration model is chosen the
structure is different).
Clicking on an infiltration model will activate certain map input options. Additional choices
for infiltration are:
subsoil impermeable: the layer below the last layer (depends on the infiltration
method) infiltration zone is assumed impermeable. Thus the soil will gradually fill up.
This has not been tested for the Holtan equation (otion 3)
callibration Ksat: the Ksat value in the soil physics tables (see SWATRE) or in the
ksat maps (other infiltration methods) is multiplied by this factor: Ks_new = Ks_table
* factor/100.
8.3 Basic maps
LISEM Basic needs a series of input maps organized as described in the "input maps"
Catchment and land use etc.
In these windows the maps can be selected. The run file contains a complete list of maps
abd their paths so it is easier to adapt an existing run-file of a different project than clicking
on each map name in turn. Clicking on a name opens a directory window and the name
field can be linked to a specific map. There are no compulsory names in LISEM but a series
of default names are used that date from LISEM versions that did not have this name
LISEM MANUAL version 2.x - January 2, 2002
selection feature. Thus old databases can still be used. The restore names button will
replace all mapnames with their defaults, and will fill up empty places.
Similar screen exist for the other categories. Note that when the maps are not selected the
mapnames are grayed out and not accessible (e.g. swatre maps are grayed out when
Green and Ampt is selected).
SWATRE options
1. minimum timestep: to adjust for mass balance errors SWATRE calculates internally
with a flexible timestep, smaller than the timestep used in LISEM. The timestep
cannot become smaller than the value given here. A smaller value gives greater
accuracy. Generally 1 sec gives sufficient accuracy.
NOTE: in old LISEM versions (older than version 1.60) this value was given in days
because SWATRE was initially designed for daily timesteps. This has been
changed to seconds for ease of use. In the run file this value is still in days to have
downward compatibility.
2. save matric head: this option needs a map "" that has numbered
locations. For these locations ascii tables are written with the temporal change of the
hydraulic head during the run.
3. geometric mean: the average conductivity K between nodes is calculated as:
Kavg = sqrt(Ki *Ki-1), else the average K is calculated as: Kavg = (Ki +Ki-1)/2
NOTE: the conductivity between the surface and the first node is always calculated
with the geometric mean (Feddes et al, 1978).
LISEM MANUAL version 2.x - January 2, 2002
8.4 Output timeseries
The results of LISEM can be stored in "timeseries" these are maps that give a flux or status
at regular intervals or at timesteps chosen by the user (type a series of timesteps in minutes
in the "user defined" area). Simply select the type of output you want. Depending on the
LISEM types chosen, additional timeseries can be stored (second list in picture below).
The maps are stored in the "result" directory given by the user in the "Start" screen.
The maps are stored as: name0001.250; name0006.500 etc. for the output at 1 min 15
seconds or respectively 6 minutes and 30 seconds. In order to have compatibility with
PCRaster, which stores the maps by timestep number (name0000.001,
name0000.002,...etc), the "store as PCRaster timeseries" can be chosen. Additionally the
runoff can be stored in unit discharge: l/sec/m flow width.
LISEM MANUAL version 2.x - January 2, 2002
LISEM generates a wide variety of outputs that enables the user to evaluate any erosion
scenario in detail. Not only total values of the water balance variables and of erosion ,
deposition and soil loss from the catchment are given, also so called "timeseries" can be
generated, which are a series of output maps at regular intervals or at times chosen by the
user. These timeseries can be shown in the display (2D) or aguila (3D) program of
PCRaster as a sequence of images, giving the impression of a film of the variable.
9.1 Main output screen and totals
The picture below shows the output screen with total values of all water balance variables,
as well as erosion, depostition and soil loss. The hydrograph, sedigraph and sediment
concentration as shown as graphs.
9.2 Hydrographs and sedigraphs
Hydrographs and sedigraphs of up to three locations (usually the outlet and two
subcatchments) can be stored as comma delimited files, that is one of the standard input
formats of MS- Excel (extension ".csv", see example below). The names of these files are
specified on the LISEM start screen. Below is the basic output. If LISEM multiclass is
chosen all suspended sediment classes are stored as separate columns.
LISEM MANUAL version 2.x - January 2, 2002
==========begin of file===========
==========end of file=============
Example hydrograph data, stored as comma delimited file.
9.3 Erosion and deposition maps
LISEM produces two maps that contain the erosion (positive values) and deposition
(negative values) in PCRaster format. The names of these files are specified on the LISEM
start screen. The values are expressed in tonnes/ha, and represent the total
erosion/deposition in a gridcell (including channels and/or wheel tracks). In case of
multiclass erosion the values are the sums of the erosion and deposition of all the texture
classes. Note that It is not so easy to interpret these maps as material that is eroded
somewhere in the catchment may deposit further downstream. Below an example is shown
of these output maps, left is erosion (mostly in the thalweg), right is deposition (on the
LISEM MANUAL version 2.x - January 2, 2002
9.4 Timeseries
The following timeseries output maps can always be generated:
runoff as discharge in (l/s) or unit discharge (l/s/m)
water height at the surface (mm)
runoff height (mm)
concentration of suspended sediment (g/l)
erosion (ton/ha)
deposition (ton/ha)
LISEM Wheeltracks has the same output map options, except that the discharge and
runoff/water height will include the flow in the wheeltracks.
LISEM Multiclass has as additional output the timeseries of suspended sediment
concentration of all 6 classes.
LISEM Nutrients has as additional output the timeseries of the NO3, NH4 and P losses in
solution, suspension and infiltration, as well as the suspended clay concentration.
LISEM Gullies has as additional output the timeseries of the gully with, depth and volume.
LISEM MANUAL version 2.x - January 2, 2002
note: these references are copied from LISEM literature and not checked yet !
Aston, AR , 1979, Rainfall interception by eight small trees: J. Hydrol., 42: 383-396.
Auzet, AV, Boiffin, J and Ludwig, B, 1995, Concentrated flow erosion in cultivated catchments:
influence of soil surface state: Eath Planet Sci. Lett,. 20: 759-767.
Beasley, DB and Huggins, LF, 1982, ANSWERS Users Manual. U.S. Environmental Protection
Agency, Region V, Chicago, Illinois. Purdue University, West Lafayette, Indiana, 54p.
Belmans, C, Wesseling, JG and Feddes, RA, 1983, Simulation model of the water balance of a
cropped soil: SWATRE: J. Hydrol., 63: 271-286.
Beven, KJ and AM Binley, 1992, The future of distributed models: model calibration and uncertainty
prediction: Hydrol. Proc., 6, 279-298.
Boardman, J and Favis-Mortlock, DT, 1998, Modelling Soil Erosion by Water, Springer-Verlag
NATO-ASI Series I-55, Berlin.
Boardman, J, Ligneau, L, De Roo, APJ., and Vandaele, K, 1994, Flooding of property by runoff from
agricultural land in northwestern Europe: Geomorph., 10: 183-196.
Chow, VT, Maidment, DR, and Mays, LW, 1988, Applied Hydrology. McGraw-Hill, 572p.
De Roo, APJ, Offermans, RJE, and Cremers, NHDT, 1996b, LISEM: a single event physically-based
hydrologic and soil erosion model for drainage basins. II: Sensitivity analysis, validation and
application: Hydrol. Proc., 10: 1118-1127.
De Roo, APJ, Wesseling, CG. and Ritsema, CJ, 1996a, LISEM: a single event physically-based
hydrologic and soil erosion model for drainage basins. I: Theory, input and output: Hydrol. Proc.,
10: 1107-1117.
Favis-Mortlock, DT, 1998, Validation of field-scale soil erosion models using common datasets: in:
Modelling Soil Erosion by Water (J. Boardman and DT Favis-Mortlock, eds.) Springer-Verlag,
NATO-ASI Series I-55: 89-128.
Govers, G., 1990, Empirical relationships on the transporting capacity of overland flow: Int. Assoc.
Hydrol. Sci. Pub. 189: 45-63.
Hoyningen-Huene, J von, 1981, Die Interzeption des Niederschlags in landwirtschaftlichen
Pflanzenbeständen. Arbeitsbericht Deutscher Verband für Wasserwirtschaft und Kulturbau,
DVWK, Braunschweig, 63p.
Jetten, V, De Roo A, and Favis-Mortlock, DT. 1999. Evaluation of field-scale and catchment-scale
soil erosion models: in: Modelling of Soil Erosion by Water on a Catchment Scale (APJ de Roo,
ed.) GCTE Focus 3 workshop,14-18 April 1997, Utrecht University. CATENA Special Issue.
Jetten, VG, De Roo, APJ, and Guerif, J, 1998, Sensitivity of the LISEM model to parameters related
to agriculture. in: Global Change: Modelling Soil Erosion by Water (J. Boardman, ed.) NATO ASI
Jetten, VG, Boiffin, J, and De Roo, APJ, 1996, Defining Monitoring strategies for runoff and erosion
studies in agricultural catchments: a simulation approach: Eur. J. Soil Sci., 47: 579-592.
Kamphorst, E, Jetten, VG, Guerif, J, Pitkanen, J, Iversen, B, Douglas, J, and Paz Gonzales, A, 2000,
Predicting depressional storage from soil surface roughness. Soil Sci. Soc. Am. J. 64: 1749-1758.
Li, R, Stevens, MA, and Simons, DB, 1976, Solutions to the Green and Ampt infilttration equation: J.
Irrig. Drain. Div. 2: 239-248.
LISEM, 2000, Limburg Soil Erosion Model, Faculty Geogr. Sci., Utrecht University, The Netherlands;;
Ludwig, B, Boiffin, J, Chadoeuf, J, and Auzet, AV, 1995, Hydrological structure and erosion damage
caused by concentrated flow in cultivated catchment: Catena, 25: 227-252.
Merriam, RA, 1960, A note on the interception loss equation: J. Geophys. Res., 65: 3850-3851.
Moore, ID and Foster, GR, 1990, Hydraulics and overland flow: in: Process Studies in Hillslope
Hydrology (MG Anderson and TP Burt, eds.) John Wiley, 215-254.
Morgan, RPC, Quinton, JN, Smith, RE, Govers, G, Poesen, JWA, Auerswald, K, Chisci, G, Torri, D,
Styczen, ME, and Folly, AJV, 1998, The European Soil Erosion Model (EUROSEM):
Documentation and User Guide. Silsoe College, Cranfield University.
Nearing, MA, 2000, Evaluating soil erosion models using measured plot data: Accounting for
variability in the data: Earth Surf. Prop. Landf., (accepted).
Onstad, CA, 1984, Depressional storage on tilled soil surfaces. Trans. Am. Soc. Agric. Eng., 27:
Rauws, G and Govers, G, 1988, Hydraulic and soil mechanical aspects of rill generation on
agricultural soils. J. Soil Sci., 39: 111-124.
LISEM MANUAL version 2.x - January 2, 2002
Risse, LM, Nearing, MA, and Laflen., JM, 1991, Assessment of error in the Universal Soil Loss
Equation using natural runoff plot data. Am. Soc. Agric. Eng. Paper 91-2558. Am. Soc. Agric.
Eng. Winter Meeting, Chicago, IL, Dec. 17-20.
Takken, I, Jetten, VG , Govers, G, Nachtergaele, J, and Steegen, A, 2001, The effect of tillageinduced roughness on runoff and erosion patterns. Geomorph., 37, (in press).
Takken, I, 2000, Effects of Roughness on Overland Flow and Erosion. PhD Thesis, Catholic
University Leuven. p. 137-181.
Takken, I, Beuselinck, L, Nachtergaele, J, Govers, G, Poesen, J, and Degraer, G, 1999, Spatial
evaluation of a physically-based distributed erosion model (LISEM): Catena 37: 431-447.
Van Deursen, WPA and Wesseling, CG, 1992, The PC-Raster Package: Department of Physical
Geography, Utrecht University;
Wendt, RC, Alberts, EE, and Hjelmfelt, AT Jr., 1986, Variability of runoff and soil loss from fallow
experimental plots: Soil Sci. Soc. Am. J., 50: 730-736.
LISEM MANUAL version 2.x - January 2, 2002
The LISEM project is an ongoing project with input of many organizations. The model is
constantly improved (hopefully) and expanded in EU projects. The following institutions
contribute or have contributed to its development. The development is in charge of Dr.
Victor Jetten
Victor Jetten, Rudi Hessel and Daniel
van der Vlag, Department of Physical
Geography, Utrecht University (NL)
Cees Wesseling, PCRaster
environmental software
Coen Ritsema, Jannes Stolte and
Simone van Dijck, Alterra (NL)
Ad de Roo, JRC (IT)
Gerard Govers, Jean Poesen, Jeroen
Nachtergaele, Ingrid Takken,
Laboratory for Exprimental
Geomorphology, Catholic University
Leuven (BE)
The Erosion research group, Unite
d'Agronomie de Laon-Peronne, INRA
Roy Morgan, John Quinton,
EUROSEM, Institute of Water and
Environment, Silsoe College, Cranfield
University (UK)
Dino Torri, Lorenzo Borselli, Institute
Limburg Waterboard - Roer en
University of Amsterdam, Physical
LISEM MANUAL version 2.x - January 2, 2002
The Run file
All options and map names of a single run are stored in the "run" file which can be edited by
hand. These include the names and option for all LISEM types, wheeltracks, multiclass
sediment etc. Where no values are found, default values are used. The rainfall file and
maps are stored with their full path names so they are linked individually and can be found
in different directories (e.g. infiltration maps for different seasons).
==========begin of file========================
[LISEM for WINDOWS run file]
[Work Directory]
Map Directory=C:\data\china\10m2\
Table Directory=C:\data\china\10m2\
Rainfall Directory=C:\data\china\
Rainfall file=p990720l.txt
Result Directory=C:\data\china\res10m1\
Main results file=res.csv
Outlet 1 file=outt.csv
Outlet 2 file=out1.csv
Outlet 3 file=out2.csv
[Simulation times]
Begin time=1
End time=99
[General options]
No Erosion simulation=0
Include main channels=0
Channel Infiltration=0
[Additional options]
Grassstrip Mannings n=0.3
Splash Delivery Ratio=0.1
Include wheeltracks=0
Include grass strips=0
Include crusts=0
Ksat calibration=100
Impermeable sublayer=0
SWATRE internal minimum timestep=
Matric head files=0
Geometric mean Ksat=0
LISEM MANUAL version 2.x - January 2, 2002
Runoff maps in l/s/m=0
Timeseries as PCRaster=0
Regular runoff output=1
Output interval=4
User defined output=0
Output times=
[Texture classes]
[map names]
LISEM MANUAL version 2.x - January 2, 2002
LISEM MANUAL version 2.x - January 2, 2002
LISEM MANUAL version 2.x - January 2, 2002
==========end of file========================
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF