Figure 2 : Time trends of DIN and TP in the Conangles catchment.

From: Atmospheric phosphorus deposition may cause lakes to revert from phosphorus limitation back to nitrogen limitation

Figure 2

Time series of DIN in atmospheric deposition (a), Lake Redon (b) and Conangles creek (c); TP in Lake Redon (d) and in deposition (e); values for the monthly NAO index during the spring and summer (April–September) months (f). Chemical data (total deposition and average concentrations) are grouped by water years (that is, from 1 September–31 August). Note that the y axis break in (a) and (e). The error bars in (a) and (e) indicate the range of deposition measurements at two sampling stations. Straight lines (ad) are the computed Kendall's overall trend lines, the solid ones indicating a high significance (P=0.004 for DIN deposition, P=0.001 for lake water DIN and P=0.029 for lake water TP) and the dashed line a nonsignificant trend (P=0.179 for the stream DIN). Curves (e,f) are local regression (LOESS) smoothing trend lines. The significance of linear trends was tested by mean of the nonparametric Mann–Kendall test, using the software provided by the US Geological Survey27. For nonlinear trends, trend lines were obtained as a smoothed LOESS curve (with sampling proportion=0.7, polynomial regression degree=2 and rejection of outliers).