A Protocol for Surveillance of Soil Degradation

We have developed a protocol for surveillance of soil degradation in the Lake Victoria Basin. 
We first delineated the river basins of the Lake Victoria Basin using spatial databases of digital elevation (1 km resolution) and stream networks (1:1,000,000) for Kenya, Uganda, Tanzania, Rwanda and Burundi.

Data derived from these spatial databases on the physical characteristics of the river basins indicated that Kenya river basins have larger average annual rainfall, average slope and sediment transport capacity than the other basins. The largest river basin, the Kagera, also showed large erosion potential. We conducted our first studies in the River Nyando Basin in Kenya as this basin had the largest erosion potential.

 
We confirmed the large erosion potential of the River Nyando when we uncovered a massive sediment plume flowing into the Winam Gulf of Lake Victoria from the river mouth . In this image the red, green and blue colours indicate different types of soil suspended in the lake water (yellow) [Follow this link for details and original image].

 

Press coverage:

Portait of an Invader. Science Volume 286, Number 5445, Issue of 26 Nov 1999, p. 1675.

Scientists Move to Save Lake Victoria from Dying. Panafrica News Agency, December 11, 2000.

 

Ground survey

 
Satellite imagery (Landsat) was used to stratify the River Nyando Basin for ground sampling (river basin is delineated by yellow line). [Description and photos of soil degradation in the Nyando River Basin]. Sampling locations were chosen to span the range of dominant parent materials and elevation range in the study area using 1:125,000 geological and 1:50,000 topographical maps.

During mid 1999 through early 2001, 522, 30 x 30 meter sized plots were surveyed in the Nyando River Basin. Total sampling effort by the involved five-person team was 29 days distributed over three separate dry seasons i.e., periods of minimal vegetation cover. All plots were georeferenced at their center-points using post-processed, differentially corrected GPS. Each plot was visually classified as apparently intact, sheet, rill, or gully eroded, or hardset using field criteria described by Schoeneberger et al. (1998) for soil erosion and criteria after Mullins et al. (1990) for soil hardsetting.

 

Detailed field observations and measurements were recorded for vegetation cover and composition, FAO land cover classification system classes, terrain, slope and soil surface conditions.

 

Three topsoil (0-20 cm) and 3 subsoil (20-50 cm) were recovered at the 5, 15, and 25 metre positions of the centre line of the plot and in direction of the dominant slope gradient. In some instances subsurface restrictions prevented recovery of subsoil samples. In these cases, the restriction type and depth were recorded.

The soil samples were air-dried and crushed to pass through a 2-mm sieve. A library of soil diffuse reflectance spectra was built from these samples.

 

Three sediment cores were collected from 3 meters water depth, circa 750 meters offshore of the outlet of the River Nyando in Lake Victoria using an 8.5 cm diameter gravity piston corer. Approximately 1-cm-thick sub-samples were extruded from each core (50 to 150 cm length). All samples were dried at 60ºC and crushed through a 2 mm sieve after removing macro-fossils and particulate organic matter. Preliminary Lead-210 and Cesium-137 dating of the master core, conducted at the University of Colorado at Boulder, indicated the bottom-most section of the core (at cm 145) could be dated to 1852 (95% CI = ± 3.3 years; D. Rowan, personal communication) using a constant sedimentation rate assumption.

The data generated by the protocol is being used to map prevalence of soil physical degradation and soil fertility constraints in the River Nyando basin at a spatial resolution of 30 m, at acceptable odds ratios for diagnostic surveillance purposes. We have developed empirical models that identify tentative risk factors for soil degradation and fertility constraints. The models are being used to provide practical recommendations for management of soil degradation that are quantitative and spatially explicit. For example, areas where preventative strategies as opposed to rehabilitation strategies can be located and critical levels of herbaceous ground cover for minimizing soil physical degradation can be specified in relation to slope and soil depth. Sequential diagnostic tests using imagery, ground observations, and soil spectral tests provide increasing levels of precision that are appropriate for different uses and audiences (e.g. broad mapping for policy makers, site specific diagnosis for field advisers and land users). The protocol is also providing a sampling frame for more detailed studies, such as prospective studies and spatial analogue surveys.