Relationships between soil test phosphorus and county-level agricultural surplus phosphorus

National nutrient inventories provide surplus phosphorus (P) estimates derived from county-scale mass balance calculations using P inputs from manure and fertilizer sales and P outputs from crop yield data. Although bioavailable P and surplus P are often correlated at the field scale, few studies have investigated the relationship between measured soil P concentrations of large-scale soil testing programs and inventory-based surplus P estimates. In this study, we assessed the relationship between national surplus P data from the NuGIS dataset and laboratory-measured soil test phosphorus (STP) at the county scale for Arkansas, North Carolina, and Oklahoma. For optimal periods of surplus P aggregation, surplus P was positively correlated with STP based on both Pearson (Arkansas: r = 0.65, North Carolina: r = 0.45, Oklahoma: r = 0.52) and Spearman correlation coefficients (Arkansas: ρ = 0.57, North Carolina: ρ = 0.28, and Oklahoma: ρ = 0.66). Based on Pearson correlations, the optimal surplus P aggregation periods were 10, 30, and 4 years for AR, NC, and OK, respectively. On average, STP was more strongly correlated with surplus P than with individual P inventory components (fertilizer, manure, and crop removal), except in North Carolina. In Arkansas and North Carolina, manure P was positively correlated with STP, and fertilizer P was negatively correlated with STP. Altogether, results suggest that surplus P moderately correlates with STP concentrations, but aggregation period and location-specific factors influence the strength of the relationship.