Soil phosphorus retention can predict responses of phosphorus uptake and yield of rice plants to P fertilizer application in flooded weathered soils in the central highlands of Madagascar
Tomohiro Nishigaki, Yasuhiro Tsujimoto, Tovohery Rakotoson, Michel Rabenarivo, Andry Andriamananjara, Hidetoshi Asai, Haja Bruce Andrianary, Hobimiarantsoa Rakotonindrina, Tantely Razafimbelo
Abstract
Efficient phosphorus (P) fertilizer management is highly required for resource-limited farmers to achieve higher yields in typical P-deficient farmlands in sub-Saharan Africa. Although soil P retention has been conventionally used to estimate the possible responses to P fertilizer applied to soils, its applicability has not been clearly confirmed at the field level. In this study, the applicability of P retention to predict the response of rice plants to P application was investigated using pot experiments with soils collected from various locations (N = 62) and on-farm experiments at two nutrient-poor sites (N = 38) in the central highlands of Madagascar. Further, we explored a simple prediction method of P retention for multiple location assessment in local farmers’ fields using an alternative property of soils collected from rice fields in the same region (N = 213). P retention was negatively correlated with the increase in rice P uptake (ΔPuptake, r = − 0.550) in the pot experiment and with the increase in yield (ΔYield, r = − 0.697) in the on-farm field experiment as a response to P application. Path analysis revealed that oxalate-extractable aluminum (Alox) content was the most important factor of P retention across all the soils, indicating that rice plants grown on soils with higher Alox and P retention are less sensitive to P application. Given its high correlation with P retention (r = 0.642) and its simplicity in measurement, we proposed the use of moisture content of air-dried soils (ωair) as a parameter to predict soil P retention. It was further confirmed that ωair had a significant negative correlation with ΔPuptake based on the pot experiment across a wide range of soil statuses (r = − 0.518). However, ωair could not clearly explain ΔYield within a small range of soil properties in the field experiment. Overall, soil P retention can predict the responses of rice plants to P application in the typical P-deficient and low-yielding lowlands in the central highlands of Madagascar, and this simple evaluation technique using air-dried soil moisture content will be helpful for the assessment of multiple rice fields on a broad scale.