Although the east covers 1004 × 103 km2 and the west 711 × 103 km2, the number of catchments in the east is less than in the west (28 and 83 respectively). This is because smaller catchments are located in the west than in the east (catchment size in selleckchem the dataset differ from 302 km2 to 280 × 103 km2). It is this difference that motivates the primary use of specific loads in the study with total loads as complimentary data. For each group (east, west and east + west), aggregated

yearly time series were constructed for temperature, precipitation, discharge, TNC, TNL, TPC, TPL and the N:P ratio to characterize the interannual variability. The aggregated yearly averages for the time series (Fig. 1) and the aggregated averages of all years (Table 1) for the three groups were calculated by accounting for the catchment size. Furthermore, a paired t-test was applied Natural Product Library to test whether variables are significantly different for east and west. To detect significant trends in the monthly time series of temperature, precipitation, discharge, TNC, TNL, TPC, TPL and the N:P ratio, a seasonal Mann–Kendall trend test was carried out for each catchment in the BSDB (the significance level was set to 0.05). The seasonal Mann–Kendall trend test is a non-parametric test for the existence of a monotonic trend and has the advantage that the power and significance

of the test are not affected by the actual distribution of the data (Hamed, 2009 and Hipel and McLeod, 2005). For all significant trends, the slope was determined using an ordinary least square

regression to estimate the true slope of the linear trend present in the time series. The slopes were categorized using the Jenks natural optimization method. This statistical mapping method is a common way to determine optimal size classes by minimizing the squared deviations of the class means. The Mann–Kendall trend test was also carried out to investigate 6-phosphogluconolactonase the existence of trends in the aggregated annual temperature, precipitation, discharge, TNC, TNL, TPC, TPL and the N:P ratio time series. In addition to these straightforward trend investigations, the Kendall rank correlation coefficient τ was estimated to determine the statistical dependence between two time series of variables based on the slope of significant trends. Tau-values near zero indicate statistical independence of the compared quantities, while τ-values near 1 (or −1) indicate that the two variables tend to strongly move in the same (opposite) direction. TNL and TPL were excluded from this analysis because loads are composites of discharge and TNC or TPC and thus lead to spurious correlations. To analyze potential differences in processes impacting nutrient loads and concentrations by land cover and climate change, a classical factor analysis was carried out.