Modelling of phosphorus inputs to rivers from diffuse and point sources

Reconstructing historical changes in phosphorus inputs to rivers from point and nonpoint sources in a rapidly developing watershed in eastern China, 1980–2010

“…Due to the fundamental hydrological differences between nutrient inputs from point (PS) and nonpoint sources (NPS), riverine TP load on the ith day (L i , kg P d − 1 ) can be expressed as (Bowes et al, 2008(Bowes et al, , 2009a(Bowes et al, , 2009b:…”

“…First, it was assumed that 0 ≤ B ≤ 1, which implies that the PS-derived TP load will increase and concentration decreases with increasing river discharge (Chen et al, 2013). Second, the model was constrained to only consider D N 1, as NPS-derived TP load and concentration both increase with increasing river discharge (Bowes et al, 2008). Furthermore, parameters A and B were both constrained as N 0.…”

“…Based on the fundamental hydrological differences between nutrient inputs from PS and NPS to rivers, a load apportionment model (LAM) may be developed for statistically quantifying point and nonpoint source nutrient inputs as a power-law function of river discharge (Bowes et al, 2008(Bowes et al, , 2009a(Bowes et al, , 2009b(Bowes et al, , 2010(Bowes et al, , 2014Greene et al, 2011;Chen et al, 2013). The LAM does not require detailed watershed attribute information and can potentially be applied to any dataset comprised of paired P concentrations and river discharge measurements.…”

“…The LAM does not require detailed watershed attribute information and can potentially be applied to any dataset comprised of paired P concentrations and river discharge measurements. This approach has been successfully applied to a wide range of watersheds in England, Ireland, U.S.A. and China (Bowes et al, 2008(Bowes et al, , 2009a(Bowes et al, , 2009b(Bowes et al, , 2010(Bowes et al, , 2014Greene et al, 2011;Jarvie et al, 2012;Chen et al, 2013). In terms of long-term riverine P source apportionment, however, existing LAM applications encounter some important limitations.…”

“…In conventional applications of LAMs, the remobilization of P from the river channel is incorrectly assigned to the nonpoint source input load, which leads to overestimation of P from NPS and underestimation of PS. A second limitation for conventional LAM application is that they are usually validated from the results of other models, commonly export coefficient models (Bowes et al, 2008(Bowes et al, , 2009a(Bowes et al, , 2009b(Bowes et al, , 2010. However, it remains a challenge to verify LAM results using export coefficient based models for many watersheds since long-term specific P export coefficients are not available or rapidly changing over time.…”