Effect of plot size on measured soil loss for two Italian experimental sites

The objective of this investigation was to determine empirically the plot width and length effects on runoff volume, Ve, soil loss, Ae, and sediment concentration, Ce, by using data collected, at the temporal scale of the erosive event, on bare plots differing in width (2 to 8 m) and length (11 to 22 m) for two Italian stations (Masse, Umbria; Sparacia, Sicily). Mean results differed by a maximum factor of 1.6 for Ve, 1.8 for Ae and 1.2 for Ce when plots differing in width were compared and by a maximum factor of 1.4 for Ve, 1.2 for Ae and 1.3 for Ce when comparison between plot lengths was conducted. Differences between two plot widths or two plot lengths were not statistically significant (P = 0.01). Soil loss was the variable most sensitive to plot size (maximum factor of discrepancy equal to 11.6 for plot width and 25.0 for plot length). Plot width and length effects varied with event erosivity, being low for the most erosive events. Relatively short and narrow plots may give long term results representative of longer and wider plots within the sampled sizes. An appropriate sampling scheme may be based on relatively narrow and short plots instead of wide and long plots, with a reduction of experimental efforts, at least when the events of interest are the most erosive ones. Reducing plot length from 22 m to 11 m is not expected to reduce soil loss per unit area at the two considered locations.