Have extractable phosphorus and potassium contents evolved in French agricultural soils since 2004?

In the last 25 years, amounts of phosphorus and potassium brought to soils through mineral fertilisers have decreased by 70% in France, without organic compensation. This situation con- duced to a reduction of nutrients balances. The aim of this study was to estimate if the decreased fertilisation could be linked with a significant decrease of extractable P and K contents in soil. The French National Soil Test (FNST) database was used to study this hypothesis. It gathers results of physico-chemical analysis performed on agricultural topsoils samples, at farmers’ request. The spatial and temporal sampling strategy were not controlled. The sampling year is known and the localisation is given by the municipality of origin of the sample. Two periods were compared: 1990-2004 and 2005-2014. A total of 2,516,743 and 2,161,811 soil test results for K 2 O and P 2 O 5 respectively were available in the FNST database, equitably allocated between the two periods. Analytical results were grouped within administrative spatial entities (cantons - groups of munic- ipalities; small agricultural regions - 713 units in France), and statistical estimator of nutrients contents were calculated after a resampling procedure (Monte Carlo approach) aiming at limiting sampling biases. Two indicators were studied: the evolution of extracible nutrients contents and the evolution of the agronomic diagnosis performed with the software RegiFert. The agronomic diagnosis evaluates the necessity of fertilisation to reach the potential yield considering the crop, the soil nutrient status and other soil characteristics. Results are provided according to 3 diagnosis classes: low, intermediate and high fertility. Evolution of contents was assessed using the non- parametric Mann–Whitney test, and evolution of diagnosis was assessed using the adequation 2 test with a limit p-value of 0.1, combined with expert rules on the evolution of the count of results into each fertility class between the two periods. Extractable P contents exhibit a clear decrease: P content decreased by more than 10% in 42.2% of the total cultivated area (TCA), with contrasted regional situations according to intensity of the animal production. The evolution of extractable K content was very low: it decreased for only 4.5% of the TCA. Evolution of agronomic diagnosis is very narrow, meaning that evolution in nutrients contents was too limited to significantly im- pact fertilisation recommendations. However, the overall decline in nutrients availability in soils encourage vigilance and soil fertility monitoring through regular soil testing