A Spatial Analysis of Phosphorus in the Mississippi River Basin

Abstract

Phosphorus (P) in rivers and streams in the Mississippi River Basin (MRB) is a contributing nutrient to hypoxia in the Gulf of Mexico and impacts local water quality. Although nitrogen has been extensively studied and attributed to be the main nutrient causing the Gulf hypoxic zone, P has increasingly been recognized as having more of an effect than was previously thought. The primary inputs of P to rivers and streams are through surface runoff or tile drainage from agricultural fields and from point sources, primarily sewage effluent. The objective of this study was to analyze the spatial pattern of P inputs and outputs in the MRB as they relate to riverine P yields to the Gulf of Mexico and to determine the counties in the MRB that have the highest yields of P and the most critical factors causing this P loss. A database of factors contributing P to the environment was constructed for each county in the MRB from 1997 through 2006 including data on fertilizer application, major crop acreage and harvest, animal numbers for manure inputs and human population. Landscape and climate characteristics of tile drainage, soil characteristics, slope and precipitation were also included. Riverine yields of total P (TP), dissolved reactive P (DRP) and particulate P (PP) for 113 watersheds within the MRB were related to these factors to create multiple regression and spatial error regression models to predict P yields throughout the basin. The fraction of the land planted in crops, sewage effluent P inputs and precipitation were found to best predict TP yields with a spatial error model (R 2 =0.47). Dissolved reactive P yields were predicted by fertilizer P inputs, sewage effluent P inputs and precipitation in a multiple regression model (R 2 =0.46), whereas PP yields were explained by crop fraction, sewage effluent P inputs and soil bulk density in a spatial error regression model (R 2 =0.499). Inputs of P from animal manure were not found to be significant in any model predicting P yields, which contrasts with the results of the USGS SPARROW iii model. Overall, the Cornbelt region of the upper Midwest, where there is a large amount of cropland and fertilizer application, had the counties with the greatest P yields. This analysis of P in the MRB helps to point out specific areas where P reductions are needed to limit the occurrence of Gulf hypoxia and improve local water quality.iv ACKNOWLEDGEMENTS

Feeding the Corn Belt: Opportunities for phosphorus recycling in U.S. agriculture

“…These trends have coincided with important ecological impacts, particularly freshwater eutrophication (Dodds et al, 2008) and seasonal hypoxia in the Gulf of Mexico downstream of the Mississippi River (Rabalais et al, 2002). Fertilizer use for corn and soybean production, concentrated application of manure, and urban sewage sources have all been identified as key sources of P loading to the Gulf of Mexico (Alexander et al, 2008;Jacobson et al, 2011).…”

Phosphorus source—sink relationships of stream sediments in the Rathbun Lake watershed in southern Iowa, USA

“…Jacobson et al (2011) found that the fraction of land in row crops and the use of fertilizer P were correlated with total and dissolved P loads, respectively, in the Mississippi River basin. Research by Alexander et al (2008), however, indicated that 37 % of the P delivered to the Gulf of Mexico originated from pasture and rangeland where manures from grazing animals were present and/or fertilizers had been applied.…”

The fate of phosphorus fertilizer in Amazon soya bean fields

“…In many soils, a threshold of soil solution P is eventually reached [12], where the accumulated P surplus is large enough that only enough P fertilizer to replace that removed by crops is required to maintain yields, though this threshold will vary by crop, soil type and farming practices [3,13]. Accumulated P in less-plant-available pools has allowed many regions, including the Midwestern USA, to greatly reduce their P inputs, in some cases such that croplands now have net negative P balances annually [14]. A recent global model suggested that P inputs required to support global agriculture in 2050 may be 10-40 per cent lower than previously estimated when residual soil P pools are taken into account [11].…”