Estimating and modelling the risk of redox-sensitive phosphorus loss from saturated soils using different soil tests

Abstract

Phosphorus (P) loss from agricultural soils can negatively affect water quality. Models and management to decrease losses increasingly focus on P that is available and transported from areas in the landscape that are regularly saturated – and periodically anaerobic. Current models use soil tests conducted in oxic conditions, which do not represent anaerobic conditions. This could lead to inaccurate P loss predictions in wetter areas, where runoff and P loss would be generated. In the present study, anoxic water extractable P test (anoxic WEP) and sodium bicarbonate-dithionite extractable P (Dithionite-P) tests were first developed using samples from New Zealand and Ireland, and were used to predict P that is available under anoxic conditions in the short- and long-term. Using archived soils from New Zealand and Ireland, it was confirmed that present testing was under-predicting losses, which could have implications for modelling and correlations with agriculture and water quality. Anoxic WEP and Dithionite-P varied by soil order and land use and that anoxic WEP was greater than oxic WEP, which showed the short-term impact of soil anoxia on P release. Fluvisols, Gleysols and Luvisols were found to be particularly enriched in anoxic WEP and dithionite-P, owing to their periodic saturation. Therefore, P loss prediction from these soils could be significantly under-estimated during periods of saturation. Models predicting anoxic WEP and Dithionite-P (R² = 54%) at the 1:50,000 scale in New Zealand found relatively small proportions of agricultural land were enriched in Dithionite-P (31% >85 mg/kg) or anoxic WEP (3% >0–0.291 mg/L). The data shows that a uniform test under-predicts P losses in anoxic conditions. Although poorly-drained and saturated areas make up a small proportion of agricultural landscapes, these areas are important in terms of overall load and the P being lost from these parts is under-predicted. The new tests should be deployed in these areas and the other tests can still be used for other areas. This finer-scale data will help improve the isolation of critical source areas (CSAs) of P loss. The present study shows the importance of creating bespoke techniques for saturated areas, which will become important in a changing climate where the extent and duration of saturation in land is changing.