Spatial and temporal variations of loads and sources of total and dissolved Phosphorus in a set of rivers (Western France).

International audienceIn intensive agricultural regions with important livestock farming, long-term land application of Phosphorus (P)both as chemical fertilizer and animal wastes, have resulted in elevated P contents in soils. Since we know thathigh P concentrations in rivers is of major concern, few studies have been done at to assess the spatiotemporalvariability of P loads in rivers and apportionment of point and nonpoint source in total loads. Here we focuson Brittany (Western France) where even though P is a great issue in terms of human and drinking water safety(cyano-toxins), environmental protection and economic costs for Brittany with regards to the periodic proliferationsof cyanobacteria that occur every year in this region, no regional-scale systematic study has been carried out so far.We selected a set of small rivers (stream order 3-5) with homogeneous agriculture and granitic catchment. Bygathering data from three water quality monitoring networks, covering more than 100 measurements stations,we provide a regional-scale quantification of the spatiotemporal variability of dissolved P (DP) and total P (TP)interannual loads from 1992 to 2012. Build on mean P load in low flows and statistical significance tests, wedeveloped a new indicator, called “low flow P load” (LFP-load), which allows us to determine the importanceof domestic and industrial P sources in total P load and to assess their spatiotemporal variability compared toagricultural sources.The calculation and the map representation of DP and TP interannual load variations allow identification of thegreatest and lowest P contributory catchments over the study period and the way P loads of Brittany rivers haveevolved through time. Both mean DP and TP loads have been divided by more than two over the last 20 years.Mean LFDP-load decreased by more than 60% and mean LFTP-load by more than 45% on average over the sameperiod showing that this marked temporal decrease in total load is largely due to the decrease of domestic andindustrial P effluents. A global shift in P inputs apportionment to freshwaters thus occurred in Brittany since 20years as agricultural nonpoint sources now contribute a greater portion of inputs showing the efficiency of therecent control of point sources by enhancement of water treatment plant and removal of phosphates in detergents.The spatialized P loads provided by this study could give a basis for a better understanding of the factors that drivesthe P transfers in Brittany soils and hotspots of P emissions while the LFP-load indicator can be a tool to assesseffects of point-source P mitigation plan