Sensitivity analysis of kinetic energy-intensity relationships and maximum rainfall intensities on rainfall erosivity using a long-term precipitation dataset

Abstract

This paper analyses and compares rainfall erosivity values that were computed with five different kinetic energy-intensity (KE-I) relationships using the 0.5-h and 1-h maximum rainfall intensities (I30 and I60, respectively). The KE-I relationships included three exponential equations (Brown and Foster, 1987 (BF); McGregor et al., 1995 (MG); van Dijk et al., 2002 (VD)), a logarithmic equation (Wischmeier and Smith, 1958 (WS)) and a linear equation (Hudson, 1961 (HU)). The KE-I relationships were used to com- pute rainfall erosivity from pluviographic records of 30 sites that are located in central Chile. A total of 415 years of data were used, and more than 18,000 storms were identified. The results showed that among the exponential equations, the MG relationship yielded erosivity results that were statistically identical to the VD and the BF relationships. However, when comparing the VD and the BF relationships, significant differences in erosivity were found, which showed that the exponential equation is highly sen- sitive to changes in its regression parameters and is therefore site-specific. Among all of the relationships, the WS logarithmic equation yielded the largest erosivity estimates; however, they were statistically equal to the estimates made with the MG and the VD relationships but were not statistically equal to esti- mates predicted by the BF relationship. In contrast, in comparison to the other equations, the HU linear relationship yielded significantly smaller erosivity values. Conversely, regardless of the KE-I relationship and the site, computing erosivity using I60 provided erosivity estimates that are 10% smaller than those obtained using I30. The relative size of the erosivity estimates is due to the relative values of I60 and I30; on average I60 was 10% smaller than I30 at the study sites. Finally, because the rainfall erosivity estimates at the study sites were highly affected by the type of KE-I relationship, these results demonstrate that selecting an appropriate KE-I relationship is crucial for accurately estimating erosivity. These differences were augmented because of the typically low rainfall intensities at the study sites. Under this condition, the differences between the KE-I relationships were greatest. However, with higher rainfall intensities, all of the KE-I relationships provide similar kinetic energy estimates, which makes the selection of the KE-I relationship less important than the use of a reliable I30 value for computing rainfall erosivity.