Model input

This page describes all the possible model inputs. Both the model input for the basic version of WaTEM/SEDEM as the input for the model extensions is described.

WaTEM/SEDEM accepts input files and parameters. Below, all possible input layers and parameters are described. Somer parameters and input files are dependent on the choices that were made by the user.

Folders

input directory

Path to the directory where all the input files, used by the model, can be found. If the path does not exist, an exception is raised and the model run will stop.

output directory

Path to the directory where all the model output files will be written. If the directory does not exist, it is created by the model.

Files

All input rasters must have the same amount of columns, rows and cell size. If one of the input rasters has a different spatial extent, the model will raise an exception and stop the execution. See the section on the format for more information.

DTM filename

The filename of the raster with a digital terrain model (DTM). This raster must contain an elevation value (in meters)for every pixel inside the model domain or catchment. The Idirisi raster must be formatted as float32.

Note

WaTEM/SEDEM does not take nodata values (e.g. -9999) into account. When a nodata value in the DTM raster is encountered, it is considered as an elevation value. Consequently, slopes will be calculated wrongly. Thus, the user must ensure that all pixels in the model domain have an elevation value, and that at least two pixels outside the model domain have a valid elevation value.

Parcel filename

The filename of the Parcel or Land cover map. WaTEM/SEDEM requires information about land cover and parcel boundaries in the routing algorithm, but also when distributing the sediment through the model domain. Every pixel in the model domain must therefore contain a land cover value. Every value > 0 indicates a unique agricultural field. Meaning that all pixels withing one agricultural field should have the same, unique value, while pixels belonging to another parcel should have a different value. These unique parcel values are important to define the routing within a parcel.

Pixels with a value < 0 indicate land cover that is not an agricultural field. In the table below the different land cover classes are shown.

Land cover/use class	pixel id
agricultural fields	> 0
outside model domain	0
river	-1
infrastructure	-2
forest	-3
pasture	-4
open water	-5
grass strips	-6

The definition of these unique parcel values are important to define the routing within a parcel. Note that the data type of this raster is integer 16.

Note

1. The Parcel raster can contain only values between -32757 and 32757. Therefore, only 32757 unique agricultural field id’s can be used in the parcel map. When more parcel id’s are necessary (e.g. in very large catchments), two or more agricultural fields can be assigned the same id. Theoretically, the model will consider these two parcels as a single parcel. In practice, these two parcels will never be treated as one because chance that these parcels are adjacent parcels is negligibly small.

2. The concept of land use (agriculture, grass strips) and land cover (river, infrastructure, forest, pasture) are used interchangeably in this definition of the parcel raster. In this manual, we aim to define this as land cover rather than land use.

P factor map filename

The filename of the P-factor map.

The datatype of the raster is float32.

K factor filename

The filename of the K-factor map. The soil erodibility factor or K-factor of the RUSLE-equation for every pixel in the model domain is stored in the K-factor map (in \(kg.h.MJ^{-1}.mm^{-1}\)).

The datatype of the K-factor raster map is integer16.

C factor map filename

The filename of the C-factor map. This raster contains values between 0 and 1 and represent the dimensionless C-factor in the RUSLE equation. Pixels outside the model domain are set to zero.

The datatype of the outlet map is float32.

ktc map filename

The filename of the ktc map, a raster with transport capacity coefficients. This raster is only mandatory if the Model extension: Create ktc map is set to 0.

The dataype of the ktc map is float32.

Outlet map filename

The filename of the outlet map. This raster is only mandatory if the Model extension: Manual outlet selection is set to 1.

Every user defined river outlet needs a unique id (integers). The outlet pixels are given the value of their respective id’s in the outlet map. All other pixels have a value equal to zero.

The datatype of the outlet map is integer16.

ktil map filename

The filename of the ktil map. The ktil map contains values for ktil, the transport capacity coefficient for tillage erosion (see here: for more information about ktil an the tillage erosion model). This raster is only mandatory when Create ktil map = 0.

The datatype of the ktil map is integer16.

Sewer map filename

The filename of the sewer map. This raster is only mandatory if the Model extension: Include sewers is set to 1.

All pixels in the sewer map should contain values between 0 and 1. These values represent the fraction of the outgoing sediment in this pixel that is entering the sewer system. A pixel with value 0 can be interpreted as a pixel where no sewer is present.

The datatype of the sewer map is float32.

Tillage direction filename

The filename of a raster with the tillage direction in degrees to the North. This raster is only mandatory if the :ref:Model extension: Include tillage direction is set to 1.

The datatype of the tillage direction raster is float32.

Oriented roughness filename

The filename of a raster with the oriented roughness. The oriented roughness is the height of the microrelief (in cm) due to ploughing. This raster is only mandatory if the Model extension: Include tillage direction is set to 1.

The datatype of the oriented roughness raster is float32.

Buffer map filename

The filename of the buffer map. This raster is only mandatory if the Model extension: Include buffers is set to 1.

The figure below shows an example of a buffermap with three buffer basins. The outlet of every buffer is marked with a buffer id (1, 2 and 3 in this example). The other pixels belonging to the buffer get the extension id. All other pixels in the raster are set to zero.

Example of a buffermap with three buffer basins.

The datatype of the buffermap is integer16.

Ditch map filename

The filename of the conductive ditch map. This raster is only mandatory if the Model extension: Include ditches is set to 1. See further for more information on how to create these routing maps.

Dam map filename

The filename of the conductive dam map. This raster is only mandatory if the Model extension: Include dams is set to 1. See further for more information on how to create these routing map.

River segment filename

The filename of the river segment map. This raster is only mandatory if the Model extension: Output per river segment is set to 1.

A river segment is a part of the river (usually a part between two confluences of the river with its tributaries). If detailed information about the sediment entering every river segment is requested, the user can make use of the river segment map option.

The river segment map is a raster where every river pixel (i.e. every pixel with value -1 in the parcel map) gets the id of the segment where it belongs to. Every segment has a unique (integer) id.

In the figure below, an example of a river segment map with seven segments is given. All pixels which are no river pixels get the value 0.

Example of a river segment map with seven segments.

The datatype of the river segment map is integer16.

adjectant segments

The filename of the Tab separated table with adjectant river segments. T his table is only mandatory if the Model extension: River routing is set to 1. The table consists of two columns: ‘from’ and ‘to’. Every row indicates a connection between two segments: segment from flows into segment to. The values in the table represent the segment-ids of the river segment map.

Based on the example river segment map, an example table with adjectant river segments is displayed below:

example adjectant segment file
from	to
1	3
2	3
3	5
4	5
6	2
7	5

upstream segments

The filename of the tab separated table with upstream segments. This table is only mandatory if the Model extension: River routing is set to 1. In the table three columns are present, namely:

edge (integer): segment id of the receiving segment
upstream edge (integer): segment id of one of the upstream segments of edge
proportion (float, between 0 and 1): the fraction of the upstream segment that flows into the considered downstream segment. If the fraction is < 1, the upstream segment should flow into two downstream segments adding up to 1.

Based on the example river segment map, an example table with adjectant upstream segments is displayed below:

example upstream segment file
edge	upstream edge	proportion
2	6	1.0
3	1	1.0
3	2	1.0
3	6	1.0
5	1	1.0
5	2	1.0
5	3	1.0
5	4	1.0
5	6	1.0
5	7	1.0

river routing filename

The filename of the river routing map. This raster is only mandatory if the Model extension: River routing = 1 is set to 1. See further for more information on how to create these routing maps.

Routing maps

The routing algorithm of WaTEM/SEDEM can take into account rasters that impose a single-flow routing along a line element in the landscape as defined by the user. The river routing map, ditchmap and dam map are made according to the principles described below.

A routing map contains integer values between 0 and 8. Every value indicates the direction which the routing should follow. A pixel set to zero has no imposed routing.

Consider pixel X in the figure below. If the routing must flow from X to the upper cardinal cell, pixel X will get value 1 in the routing map. If the routing must flow from X to the lower left pixel, X will get value 6. All other directions are set in the same way, according to the numbers in the figure.

Definition of flow routing.

An example of a routing map with two imposed routings is given here:

Example of a routing map

The datatype of a routing raster is integer16.

CN map filename

The filename of the CN map. This raster is only mandatory if the Model extension: Curve number is set to 1.

This raster contains a CN-value (between 0 and 100) for every pixel in the model domain.

The datatype of the CN raster is float32.

Rainfall filename

Filename of a textfile with rainfall values. The text file contains a table (tab-delimited) with two columns without header. The first column contains the time in minutes (starting from 0), the second column contains the rainfall in mm. The rainfall of the first timestamp must be zero.

Parameters

This section describes all parameters that can be used to make a basic WaTEM/SEDEM run, without the use of any extensions.

max kernel

If the routing algorithm of WaTEM/SEDEM encounters a local minimum in the digital elevation model, it will not find a lower, neighbouring pixel. Therefore, the algorithm is set to search for a lower pixel within a search radius around the local minimum (see routing algorithm. The variable ‘max kernel’ defines this search radius expressed in pixels.

max kernel river

If the routing algorithm of WaTEM/SEDEM encounters a local minimum in the digital elevation model it will not find a lower, neighbouring pixel. If this pixel is a river pixel, the routing will remain in the river and the routing will look within a search radius around the local minimum with the same landuse (river). The variable ‘max kernel river’ defines the search radius expressed in pixels.

bulk density

The average bulk density (in \(kg.m{-3}\)) of the soil in the catchment (as integer). This value is used to convert the mass of the transported sediment to volumes. A good default value for Flanders is 1350 kg/m³.

R factor

The R-factor or rainfall erosivity factor in the RUSLE equation (float, in \(MJ.mm.ha{-1}.h{-1}.year{-1}\)). This input is mandatory, except except if the Model Choice: Only routing is set to 1.

Note

1. the user must make sure that the R and C-factor are calculated for the same time span (year, month, week,…).

2. R-factor values can be computed with the R-factor Python package.

Parcel connectivity cropland

The ‘parcel connectivity cropland’ expresses the reduction of the upstream area (\(A_{pixel}\)) at a parcel boundary. It is an integer value between 0 and 100. The reduction on the upstream area is applied when the target pixel is of the land cover ‘cropland’ (Parcel map value: >0).

\[A_{pixel} = A_{pixel}\frac{connectivity_{cropland}}{100}\]

Parcel connectivity grasstrips

The ‘parcel connectivity grasstrips’ expresses the reduction of the upstream area (\(A_{pixel}\)) at the boundary between a parcel and a grasstrip. It is an integer value between 0 and 100. The reduction on the upstream area is applied when the target pixel is of the land cover ‘grasstrip’ (Parcel map value: -6). The default value for this parameter is 100.

\[A_{pixel} = A_{pixel}\frac{connectivity_{grasstrip}}{100}\]

Parcel connectivity forest

The ‘parcel connectivity forest’ expresses the reduction of the upstream area (\(A_{pixel}\)) at a boundary of a forest. It is an integer value between 0 and 100. The reduction on the upstream area is applied when the target pixel is of the land cover ‘forest’ (Parcel map value: -3).

\[A_{pixel} = A_{pixel}\frac{connectivity_{forest}}{100}\]

Parcel trapping efficiency cropland

The parcel trapping efficiency (PTEF) is used to compute the upstream area for every raster pixel (\(A_{pixel}\)) (see also L-model). The PTEF also takes the land-use, defined by the parcels raster, into account. This then, contributes to the upstream area by a given percentage (100-PTEF).

The parcel trapping efficiency for cropland is defined by the ‘Parcel trapping efficiency cropland’ (in % as integer; e.g. PTEF = 87).

\[A_{pixel} = res^2(1-\frac{PTEF_{cropland}}{100})\]

Parcel trapping efficiency pasture

The parcel trapping efficiency for pasture is defined by the ‘Parcel trapping efficiency pasture’ (in % as integer e.g. PTEF = 25). For a definition of the Parcel trapping efficiency, see Parcel trapping efficiency cropland

Parcel trapping efficiency forest

The parcel trapping efficiency for forest is defined by the ‘Parcel trapping efficiency forest’ (in % as integer e.g. PTEF = 25). For a definition of the Parcel trapping efficiency, see Parcel trapping efficiency cropland

Parameters extensions

This section contains all parameters that are used within model extensions. For every parameter in this section a description is given about the meaning of the parameter and when to use it (i.e. in combination with which extension).

ktc low

ktc low is the transport capacity coefficient (as float) for pixels with a low erosion potential (see ktc limit). The parameter is only mandatory if the Model extension: Create ktc map is set to 1.

ktc high

ktc high is the transport capacity coefficient (float) for pixels with a high erosion potential (see ktc limit). The parameter is only mandatory if the Model extension: Create ktc map is set to 1.

ktc limit

ktc limit is a threshold value (float). Pixels with a C-factor higher than ktc limit will get the value of ktc high in the ktc map, pixels with a C-factor below ktc limit, will get ktc low in the ktc map. This parameter is only mandatory if the Model extension Create ktc map is set to 1 or Calibrate = 1

ktil default

The transport capacity coefficient for tillage erosion on agricultural fields. This value (as integer) should be expressed in \(kg.m{-1}.year{-1}\). A recommended default value is \(600 kg.m{-1}.year{-1}\).

This parameter is only mandatory if the Model extension: Create ktil map is set to 1.

ktil threshold

ktil threshold is a float between 0 and 1. Pixels with a C-factor higher than ktil threshold will get the value of ktil default in the ktil map, pixels with a C-factor below ktil threshold, are set to 0. A typical value for ktil threshold is 0.01.

This parameter is only mandatory if the Model extension: Create ktil map is set to 1.

LS correction

Notebaert et al. (2005) describes that changes in spatial resolution have major scaling effects on topographic variables like the L and S-factor.

The LS-factor will decrease on a higher resolution (smaller pixels, more height information) and extreme LS values will occur more. To be able to compare the calculated RUSLE values on different spatial resolutions, a correction factor can be calculated. This correction factor \(LS_{cor}\) is calculated as:

\[LS_{cor} = \frac{LS_{avg,x}}{LS_{avg,y}}\]

with

\(LS_{avg,x}\): the average LS factor in a catchment on resolution x
\(LS_{avg,y}\): the average LS factor in a catchment on resolution y

The input variable is a float (default value 1, i.e. no correction). The LS-factor in the model is divided by this variable.

Sewer exit

An integer value between 0 and 100 that represents the fration of the discharge that enters the sewer system. It is only applied on pixels where the sewer map is not zero.

This variable is only mandatory if the Model extensions: Curve number and Include sewers are set to 1.

Note

1. The values stored in the sewer map are not used in the discharge calculations of the CN module. The sewer map is only used to check if a pixel is a sewer or not.

2. In the sediment calculations, a different trapping efficiency for every sewer pixel in the model can be defined, but this is not the case in the discharge calculations.

Clay content parent material

The average fraction of clay in the soil of the modelled catchment (in decimals; float32, between 0 and 1). This variable is only mandatory if the Model extension: estimate clay content is set to 1.

5-day antecedent rainfall

The total rainfall (in mm) during 5 days before the start of the rainfall event. This variable is only mandatory if the Model extension: Curve number is set to 1.

stream velocity

As float, only mandatory if the Model extension: Curve number is set to 1.

alpha

Alpha (as float) is a calibration parameter of the CN-model. It determines the relation between the runoff and the rainfall intensity. This parameter is only mandatory if the Model extension: Curve number is set to 1.

beta

Beta (as float) is a calibration parameter of the CN-model. It determines the relation between the runoff and the antecedent rainfall. This parameter is only mandatory if the Model extension: Curve number is set to 1.

Desired timestep for model

Runoff calculations are done using this timestep. The given timestep must comply with the Courant Criterium. This criterium limits the timestep as a function of the spatial resolution (m) and the stream velocity of water over land (m/s).

\[dt \leq \frac{spatial resolution}{stream velocity}\]

The parameter is an integer value expressed in minutes and is only mandatory if the Model extension: Curve number is set to 1.

Final timestep output

The user has the option to resample the time-dependent output (runoff, sediment concentration, sediment load) to a different timestep than the Desired timestep of the model. The parameter is an integer value expressed in minutes and is only mandatory if the Model extension: Curve number is set to 1.

Endtime model

Total timespan (in minutes) the model has to simulate. This parameter is an integer value and must be a multiple of the timestep of the model.

Note

In a first model run for a catchment with a given rainfall event, a large enough endtime should be given. This, in order to ensure that the whole runoff peak is modelled. After the first simulation, the model user can shorten the endtime to optimise the calculation time of the model.

This parameter is only mandatory if the Model extension: Curve number is set to 1.

Number of buffers

The amount of buffers present in the buffer map is given in this parameter (as integer). The parameter is only mandatory if the Model extension: Include buffers is set to 1.

Number of forced routing

The amount of locations where the user wants to force the routing is given by this parameter (as integer). This variable defines the number of forced routing sections in the ini-file. This variable is only mandatory if the Model extension: Force Routing is set to 1.

Bufferdata

The inclusion of erosion control buffers is based on input rasters and buffer parameters. How these input rasters should be created, is described here. If the Model extension: include buffers is set to 1, the buffer parameters must be defined in the ini-file in the following manner:

[Buffer 1]
volume = 329.0
height dam = 0.37
height opening = 0
opening area = 0.03
discharge coefficient = 0.6
width dam = 7
trapping efficiency = 75
extension id = 16385

[Buffer 2]
volume = 1123.0
height dam = 1.5
height opening = 0
opening area = 0.03
discharge coefficient = 0.6
width dam = 7
trapping efficiency = 75
extension id = 16386

with:

volume: the maximum volume of water that can be trapped in the bufferbasin, \(V_{basin}\) (\(m^{3}\)).

height dam: the height of the dam of the buffer basin, \(H_{dam}\) (\(m\)).

height opening: the height of the opening of the discharge pipe of the basin, \(H_{opening}\) (m).

opening area: the area of the discharge opening \(A_0\) (\(m^{2}\)).

discharge coefficient: the discharge coefficient \(C_d\) (-) of the buffer basin.

width dam: the width of the overflow on the bufferbasin dam \(W_{dam}\) (m).

trapping efficiency: the trapping efficiency is the fraction of the incoming sediment that is trapped.

extension id of a buffer is calculated as the buffer id + 16384. It is an integer value. All pixels of the buffer in the buffer map are given the value of the extension id, except the outlet pixel.

The extension id and trapping efficiency are mandatory for every buffer. The other buffer parameters are only mandatory when the the CN-module seperately (i.e. the Model Choice: Curve number is set to 1).

Note that the amount of sections with buffer data has to be defined with the variable Number of buffers.

A full description of the CN calculation in buffers can be found here.

Note

The definition of the buffer extension id equal to buffer id + 16384, implies only 16384 buffer basins can be modelled.

Forced routing data

In the case that the analysis of the routing and field validation shows that the routing is defined incorrectly by the model, a forced routing from a specified source to target pixel can be defined by the user. Forced routing is defined by stating the column and row of both the source and target pixel in the ini-file, as shown here:

[Forced Routing 1]
from col = 10
from row = 10
target col = 11
target row = 11

[Forced Routing 2]
from col = 15
from row = 16
target col = 20
target row = 19

Note that the amount of sections with forced routing vectors has to be defined with the variable Number of forced routing

Calibration data

The following parameters are only mandatory if the Model extension: Calibrate is set to 1 . These parameters must be grouped in a seperate section in the ini-file with the header ‘Calibration’, as shown here:

[Calibration]
KTcHigh_lower=1
KTcHigh_upper=20
KTcLow_lower=1
KTcLow_upper=20
steps=20

KTcHigh_lower

The lower range of ktc-high values in the calibration mode. The value is a float and by default 5.

KTcHigh_upper

The upper range of ktc-high values in the calibration mode. The value is a float and by default 40.

KTcLow_lower

The lower range of ktc-low values in the calibration mode. The value is a float and by default 1.

KTcLow_upper

The upper range of ktc-low values in the calibration mode. The value is a float and by default 20.

steps

The amount of steps between the lower and upper values for ktc low and ktc high during a calibration run. This value is an integer and by default 12.

References

Notebaert, B,. Govers, G.n Verstraeten, G., Van Oost, K., Ruysschaert, G., Poesen, J., Van Rompay, A. (2005): Verfijnde ersoiekaart Vlaanderen: eindrapport, Departement Omgeving, Brussel, 53 pp.