![](/upload_mm/c/1/1/dafe7460-7f8e-4e21-921f-f2d6b5eb6b4c_39%20Landdynamiek_shutterstock_26465362__dd3ece35_750x400.jpg)
Case Study Kenya
This MSc-thesis project in 2002 focussed on the impossibilities and sensitivities of landscape modeling using low resolution and medium quality data.
![KY_small.png KY_small.png](/upload_mm/8/f/3/9fab0aca-39e5-4461-b96a-81a8fc604dbf_KY_small_8c16da1a_239x126.png)
In an arid area in Eastern Kenya close to Embu, existing data had low resolution and accuracy. Data for running, calibrating and validating LAPSUS, including the DEM, were manually collected in a regular 125m grid for two research areas of approx. 7.5 km2 each. Methods for additional data collection were simple: infiltration was measured by doing fifteen minutes, 20 cm2 infiltration tests; soil erodibility was estimated with the aggregate stability test. For calibration and validation purposes, actual erosion and sedimentation were estimated as a percentage of the surface for each grid cell.
![Research_Area_small4.png Research_Area_small4.png](/upload_mm/e/a/f/9d8d549e-e8b7-4ef2-8903-b6d37f0487f1_Research_Area_small4_be88bee3_670x303.png)
LAPSUS Application
The research was successful in the sense that data for running LAPSUS were collected in challenging conditions. Hand-measuring a Digital Elevation Model with a GPS unit was particularly problematic.
Results
Calibration and validation efforts were not very successful, with the two most important reasons being a too large resolution of the DEM and large uncertainty in calibration data. R-square value (cell-by-cell) of the best calibration was 0.47.
![DEMs2.jpg DEMs2.jpg](/upload_mm/b/0/3/b1477f1a-f70d-4f50-8c38-79ef5d937233_DEMs2_a7da4b43_670x355.jpg)
Sensitivity analysis of the results focussed on both patterns and values of model outputs. Sensitivity to multiple flow parameter p was lowest, and sensitivity to transport capacity parameters m and n was highest.
![results1.jpg results1.jpg](/upload_mm/6/b/d/8d3caf02-7893-402e-9a51-d6a1c23fd927_results1_f8839c1c_670x596.jpg)
Different higher-resolution DEMs were interpolated from the field-measured DEM. LAPSUS runs with these DEMs were planned to find out which DEM interpolation method performed better, but LAPSUS (2002) could not handle the sinks created by interpolation.
Publications
- Temme, A.J.A.M., 2002, The impossibilities and sensitivities of using a landscape model with low resolution and medium-quality data: a case study in East-Kenya: Thesis report.
» more LAPSUS Publications.
Future research
This case study started the drive for a version of LAPSUS that could deal with sinks in a DEM. Sinks are not necessarily mistakes and can represent true landscape features.
» more about how LAPSUS currently deals with sinks.