Dynamically managing soil phosphorus could halve phosphorus losses from cropland in lake basins

Abstract

Excessive fertilization leads to the accumulation of phosphorus (P) in the soil, potentially leading to P pollution in water bodies. A key challenge is to provide long-term solutions for sustainable P management to meet environmentally safe levels, meanwhile maintaining high crop yields. Here, we developed a dynamic P management (DPM) strategy that utilizes a P cycling model integrated with machine learning to manage the long-term soil P status and reduce P losses from soil and fertilizers within the safe threshold. We used the Erhai Lake Basin as an example witnessed a substantial increase in P balance, from 20.8 kg ha⁻¹ to 43.6 kg ha⁻¹, with an increase in soil available P by 51 % and total P losses by 63 % between 2010 and 2020. We then estimated the soil P threshold as 34.0 mg kg⁻¹ and 18.9 mg kg⁻¹ on average for the water environmental and agronomic thresholds, respectively. Currently, 90 % of the townships exceed these soil P thresholds. Employing a DPM strategy could achieve an optimal steady-state level of soil P in the basin within 44 years, balancing agronomic and environmental needs and reducing P application and losses by 60 % and 56 %, respectively. This study provides long-term solutions for sustainable P management in lake basins.