Analysing soil organic C gradients in a smallholder farming village of East Zimbabwe

• Digital soil mapping with Landsat imagery explained 50% of observed SOC variance.
• SOC variance could not be linked to farm or field management.
• Clay content was dominant for predicting variance in SOC at the village scale.
• Our inability to detect soil fertility gradients is likely due to a scale-mismatch.

We set out to map the soil organic carbon (SOC) content, as an indicator of soil fertility, at the village scale, and to relate the SOC content to farm scale management and landscape scale characteristics. Topsoil samples were taken at 100 random locations in the Murewa smallholder farming area in Zimbabwe and analysed for organic carbon. Using digital soil mapping techniques and Landsat TM images we could explain 50% of the observed SOC variance. The average SOC content was estimated to be 1.5%, although the sandy cropping area had a much lower average of 0.8% and the red clays and valleys had higher average of 1.8%. The SOC variability could not be linked to farm management. No fertility gradients were observed, mostly due to a strong dominance of clay content on the spatial distribution of SOC. Clay content was able to explain 57% of the SOC variance, while farm area and labour size, typically used for farmer typology, were able to explain only an additional minor part of the SOC variance. This strong landscape scale effect needs to be included in future village-scale studies. We conclude that digital soil mapping of soil fertility gradients at the village scale has several scale issues that need to be addressed if the envisioned global digital soil map is to be relevant for smallholder farmers in sub-Saharan Africa.