An investigation into the impact of nine catchment characteristics on the accuracy of two phosphorus load apportionment models

Phosphorus (P) load apportionment models (LAMs), requiring only spatially and temporally paired P flow (Q) measurements, provide outputs of variable accuracy using long-term monthly datasets. Using a novel approach to investigate the impact catchment characteristics on variation, 91 watercourses’ Q-P datasets were applied two LAMs, BM GM, bootstrapped ascertain standard errors (SEs). Random forest regression analysis data pertaining catchments’ land use, steepness, size, base sinuosity used identify individual relative importance SE. For BM, increasing urban cover was influential raising SEs, accounting for c.19% observed whilst GM found no individually important characteristic. Assessment model fit evidenced consistently outperformed modelling values ±10% actual in 85.7% datasets, as opposed 17.6% by GM. Further are needed account SE variation within both models, interaction between variables may also be present. Future research should focus quantifying these possible interactions expand included random forest. Both LAMs must tested wide range high temporal resolution if they can adequately storm events catchments with diverse characteristics.