Estimation of phosphorus loads with sparse data for agricultural watersheds in the Czech Republic

Agricultural watersheds in the Czech Republic are one of the primary sources of non-point-source phosphorus (P) loads in receiving waters. Since such non-point sources are generally located in headwater catchments, streamflow and P concentration data are sparse. We show how very short daily streamflow and P concentration records can be combined with nearby longer existing daily streamflow records to result in reliable estimates of daily and annual P concentrations and loads. Maintenance of variance streamflow record ex tension methods (MOVE) can be employed to extend short streamflow records. Constituent load regressions are used to predict daily P constituent loads from streamflow and other time varying characteristics. Annual P loads are then estimated for individual watersheds. Resulting annual P load estimates ranged from 0.21 to 95.4 kg year(-1) with a mean value of 11.77 kg year(-1). Similarly annual P yield estimates ranged from 0.01 to 0.3 kg ha(-1) year(-1) with an average, yield of 0.07 kg ha(-1) year(-1). We document how short records of daily streamflow and P concentrations can be combined with a national network of daily streamflow records in the Czech Republic to arrive at meaningful and reliable estimates of annual P loads for small agricultural watersheds.