Can prediction and understanding of water quality variation be improved by combining phosphorus source and waterbody condition parameters?

Phosphorus (P) pollution impairs river systems globally. There is long-standing interest in understanding catchment source P loads to inform mitigation to improve water quality. However, P sources to the hydrosphere differ individually in discharge behaviour, P intensity, bioavailability, and cumulative impacts. River condition also varies (e.g., riparian disturbance, climate change impacts) such that source and river resilience are likely synergistic but poorly studied controls on water quality variation. To challenge the use of overly-simplistic factors (e.g., basic soils and landcover) in empirical catchment pollution source-impact assessments, we pooled spatial data according to conceptual aspects of P source mechanisms and waterbody riparian condition. These were related empirically to P concentrations and loads, and trophic diatom indices, for 19 Scottish catchments (~10–250 km2) representing some mechanistic aspects of pollution loading and river impacts. Sources of P from septic tanks and farmyards influenced loads and ecological impacts. Some secondary calculations pooling spatial data such as septic tank source-delivery methods were novel, involving complex, but available, soil water flowpath data. In contrast, inclusion of channel condition and farmyard P loads used simple aerial imagery. Multiple Factor Analysis combined with Redundancy Analysis showed that source P loads expressed as bioavailable forms of P were better explanatory factors of diatom classification groups than stream soluble reactive P concentrations, although used together they improved explanation further. Riparian quality metrics were less powerful predictors than expected, likely with more scale-dependant effects on ecological functions than can be quantified by visual condition assessment on isolated short reaches. There was strong justification for examining separate P fractions (total, dissolved, particulate and bioavailable forms) by distinct catchment source types to understand better nutrient dynamics across land to waters, ecosystem degradation and waterbody impacts in the contemporary hydrosphere