Phosphorus Inventory for the Conterminous United States (2002-2012)

Abstract

Published reports suggest efforts designed to prevent the occurrence of harmful algal blooms and hypoxia by reducing non-point and point source phosphorus (P) pollution are not delivering water quality improvements in many areas. Part of the uncertainty in evaluating watershed responses to management practices is the lack of standardized estimates of phosphorus inputs and outputs. To assess P trends across the conterminous United States, we compiled an inventory using publicly available datasets of agricultural P fluxes, atmospheric P deposition, human P demand and waste, and point source discharges for 2002, 2007, and 2012 at the scale of the 8-digit Hydrologic Unit Code subbasin (~1,800 km²). Estimates of agricultural legacy P surplus accumulated from 1945 to 2001 were also developed. Fertilizer and manure inputs were found to exceed crop removal rates by up to 50% in many agricultural regions. This excess in inputs has led to the continued accumulation of legacy P in agricultural lands. Atmospheric P deposition increased throughout the Rockies, potentially contributing to reported increases in surface water P concentrations in undisturbed watersheds. In some urban areas, P fluxes associated with human waste and non-farm fertilizer use has declined despite population growth, likely due, in part, to various sales bans on P-containing detergents and fertilizers. Although regions and individual subbasins have different contemporary and legacy P sources, a standardized method of accounting for large and small fluxes and ready to use inventory numbers provide essential infromation to coordinate targeted interventions to reduce P concentrations in the nation's waters.

Figure 1.

Simplified conceptual diagram describing the majority of fluxes captured in the national P inventory for HUC-8 subbasins (n = 2115, ~1,800 km²) across the contiguous United States. **Note that livestock content is not included in the terrestrial surplus calculation and total non-hydrologic loss term since the net difference is captured with the inclusion of livestock waste input (i.e., livestock feed demand-livestock content = livestock waste).

Figure 2.

Largest source of P inputs for all subbasins of the CONUS in 2002 and 2012.

Figure 3.

Total human waste, non-farm fertilizer application, point source loading, and atmospheric P deposition rates for all subbasins of the CONUS in 2012 (left column). Right column displays the absolute difference of the P fluxes between the 2002 and 2012 inventories (i.e., 2012–2002) except for point source loads. The difference in point source loads was calculated between the years 2007 and 2012. Other fluxes not illustrated can be visualized using the supplementary database.

Figure 4.

Farm fertilizer application, livestock waste, and crop removal rates for all subbasins of the CONUS in 2012 (left column). Right column displays the absolute difference of the select P fluxes between the 2002 and 2012 inventories (i.e., 2012–2002). Other fluxes not illustrated can be visualized using the supplementary database.

Figure 5.

Total P inputs along with calculated terrestrial P surpluses (i.e., total P inputs-non-hydrologic P outputs) for all subbasins of the CONUS in 2012 (left column). Right column displays the absolute difference of total P inputs, non-hydrologic P outputs, and terrestrial P surpluses between the 2002 and 2012 inventories (i.e., 2012–2002). Please note that non-hydrologic outputs are not shown because they are the same as crop removal. The same legend is used for panels b, c, and d.

Figure 6.

For the 1945–2001 period, accumulated agricultural inputs (i.e., livestock waste + fertilizer) and agricultural legacy P surplus (i.e., amount of P remaining after accounting for crop removal for the 1945–2001 period) along with upper and lower bound confident estimates using variable crop removal efficiencies (c, d) for all subbasins of the CONUS as of 2001.