Comparing modeling results with other model predictions
The original post-fire erosion analysis was carried out by a Forest Service Burned Area Emergency Response (BAER) specialist using the Erosion Risk Management Tool (ERMiT; Robichaud et al., 2007) to quickly estimate upland erosion following the Emerald Fire (Table 6; Young, 2016). Young (2016) followed recommended methods and only presented potential erosion from 2-, 5- and 10-year events, as the probability of a more severe event was unlikely. The three recorded precipitation events that occurred immediately following the wildfire (81, 56, and 114 mm) were similar to the 5 to 10-year return period range of the ERMiT predictions for single storms (Table 6), if a rainfall event had been the cause of ERMiT’s 10-y runoff event rather than rain-on-snow. Thus Young’s (2016) selection of events was close to what actually occurred. We estimated a spatially averaged erosion rate of 16-25 Mg ha-1, depending on hillslope profile for the three observed storms. The ERMiT predictions for these three storms would be approximately one 2-y event (0.22 Mg ha-1), one 5-y event (6.15 Mg ha-1) and one 10-y event (14.76 Mg ha-1) for a total of about 21 Mg ha-1. Young’s rapidly completed predictions, with a stated accuracy of plus or minus 50 percent, are similar to our more detailed analyses, confirming that the coarse-scale ERMiT predictions are reasonable for rapid post wildfire erosion analysis. The current BAER erosion prediction technologies as exemplified by Young (2016), however, did not consider the effects of roads in his analyses, which we have shown can be significant.