Calculation of input grid with CN-values for surface runoff

The model builder is used to modify Curve Numbers (CN-values) assigned to individual Grid cells of the raster based rainfall/runoff model.

After the following parameters concerning surface runoff were set ...
  1. Principle assignment of curve numbers according to the "hydrological soil type"
  2. Assignment of a correction factor in order to consider the landuse (mainly vegetation and urbanisation
  3. Assignment of another correction factor in order to consider the principle climatic zone (which is assumed to effect the soil depth).

...the grid file with the CN-values for surface runoff will be generated (or recalculated).

This calculation is performed using Surfer commands. This is invoked pressing the "Surfer calc key" The calculation lasts about one minute and is indicated by the hour-glass cursor.

The original grids as obtained from ArcView are usually in JTM with a 1000 m spacing. They are located in the "System Data" path as indicated at the bottom of the Core Runoff model form. The following grid files are required:

Note:The creation of this directory and the generation of the original grids from ArcView are presently still managed outside of this model. ArcView is able to generate ASCII grids in ESRI format. The conversion from ESRI to Surfer format is performed within this model using the ESRI2GRD utility of the Core Runoff model.


Processing of the grid for surface runoff is more complex than for percolation. The following steps will be automatically executed by the program:

  1. The original grid hysoil.grd is loaded and reclassified according to the table earlier edited. Because Surfer does not have a built-in reclassification command, this is done with a loop through all the classes resp. table records.
  2. In a next step, the CN-values are converted into S-values using the formula
    S=(1000/CN-10)*25.4
  3. Now, the landuse and urbanisation grids are combined using the formula
    C=0.1+A+1000*B
    with A being the code for landuse (160, 161, 162, 163) and B the indicator for urbanisation (1 = urbanisation, 0 = no urbanisation). The reason behind this algorithm is to permit to filter all urbanised areas independent of any other previous land use. The term 0.1 is used to prevent any zero grid cells.
  4. The combined landuse table is reclassified to relate it to the classification table as previously edited. Note, that this table is not using discrete values but intervals instead. All cells with a value below 150 are unclassified (correction factor =1) while all cells with a value above 163.5 are assumed to be urbanised. The result is a grid with the correction factors.
    Again, as Surfer does not have a built-in reclassification command, this is done with a loop through all the classes resp. table records.
  5. Now, the grid with S-values is multiplied with the correction factors for landuse as stored in the grid calculated in the previous step.
  6. In the next step, the previous grid multiplication is repeated, but this time with the correction factors for climate zones. By this procedure, runoff in the wetter zones is decreased and in the drier zones increased.
  7. After the S-values have been corrected, they are re-converted into CN-values:
    CN=1000/(S/25.4+10)
  8. Finally, the grid of 1000 m is recalculated to 2000 m to enhance modelling speed. The previous grid size of 2500 m is not used any more - the cubic spline algorithm Surfer applies will increase the extreme values in that case of interpolation.
  9. After the grid has been recalculated, the range of CN-values is shown in the form. This may serve as a control.