Surveillance of Soil Fertility Condition

Soil condition classes were assigned to 522 georeferenced plots from the River Nyando basin based on a soil reflectance screening test and then calibrated to Landsat reflectance. We used an ordinal logistic classification procedure to assess the spatial distribution of soil condition classes. Because of the change in geometric support in the measurements (i.e., Landsat reflectance values were measured at the level of 30 x 30 m pixels,  whereas soils were classified at the level of samples within pixels), the classifier was fit in MIXOR©, and the

resulting model equations were transferred to the IDRISI32 GIS system for display.  

 

Expected Available Phosphorus Distribution in the River Nyando Basin
Soil phosphorus test levels should be above 15 mg kg-1 to ensure good yields of most crops.

Maize does not generally respond much to P applications above 8.5 mg kg-1

Between 64% and 74% of the basin area is deficient for growing maize.

 

The expected values of any of the soil properties used to create the multivariate soil fertility index can be mapped in this way. 

 

We are proposing the use of the soil fertility index as a tool for controlling for soil fertility effects in various studies and as an index for monitoring change in soil condition.