Phosphorus availability in selected soils of Rwanda: Comparison of four test procedures

Low levels of plant-available phosphorus (P) are a major factor limiting crop production in sub-Saharan Africa. Soil tests for phosphorus in this region have been applied mainly to highly weathered tropical lowland soils, while the number of investigations on the P status of less weathered tropical highland soils is limited. In this study, we evaluated four soil P tests with regard to their suitability for typical tropical highland soils and Technosols on mine spoils from pegmatite dump material in the Gatumba Mining District (GMD) of Rwanda. The tests comprised Bray 1, Mehlich 3, Olsen P and Pi test. Twelve soils (eleven from Rwanda and a reference soil from Germany) were analyzed for available P. The results reflected extremely low levels of available P in the Rwandan soils compared to the soil from Germany. A ranking from highest to lowest measured P followed the order Mehlich 3 > Bray 1 > Olsen > Pi. High correlations were found between Bray 1, Mehlich 3 and Pi. Extremely low levels of P extracted with Pi compared to Bray 1 and Mehlich 3, however, raised the question of its suitability for the investigated soils. Values of Olsen P generally showed low reproducibility and low correlation coefficients. Results from regression analyses with pH, clay content and soil organic matter also confirmed limited suitability of Olsen. A regional calibration for the detection of plant-available P in the GMD was successful for Bray 1 and Mehlich 3 demonstrating that these were suitable for all soils investigated.