|
PREDICTING EVENT EROSION BETTER The USLE-M uses the product of the runoff ratio (QR) and the EI30 index as the R (erosivity) factor. This results in a model that predicts event erosion better than the USLE. |
|
|
|
|
|
Event soil loss from a bare plot at Arnot (Ithaca), NY in the USA in relation to the EI30 and QREI30 indices. The line in (A) represents the values predicted using the USLE. The line in (B) represents the prediction with the USLE-M |
|
|
In this example from Ithaca, NY, the efficiency of the USLE-M as
measured by the Z(log), the logarithmic form of the Nash-Sutcliffe model
efficiency statistic (similar to R2), is about 77% and 54% for the
USLE. For a bare plot at Morris, MN, the total loss predicted by the USLE-M
for the 10 events that produced the least erosion was 1.12 t/ha, an
error of just 0.3 t/ha. The USLE predicted the total loss to be
25 t/ha. For the 5 events that produced the most erosion, the USLE-M
predicted the total within –7%, compared with – 31% for the USLE.
|
|
