EFFECT OF KINETIC ENERGY-INTENSITY RELATIONSHIPS AND 0.5-H MAXIMUM INTENSITY ESTIMATION METHODS ON RAINFALL EROSIVITY

摘要

This paper analyses and compares rainfall erosivity values computed with five different kinetic energy-intensity (KE-I) relationships using 0.5-h and 1-h maximum rainfall intensities (I_(30) and I_(60), respectively). Three exponential, a logarithmic and a linear KE-I relationships were used to compute rainfall erosivity from pluviographic records of 30 sites located in Central Chile. A total of 415 years of data were analyzed and more than 18,000 storms were identified. The results showed that, among the exponential equations, the McGregor relationship yielded statistically equal erosivity results with the Van Dijk and the Brown and Foster relationships. However, when comparing the Van Dijk and the Brown and Foster relationships, significant erosivity differences were found, showing that the exponential equation is highly sensitive to changes in its regression parameters and therefore site-specific. Among all the relationships, the Wischmeyer and Smith logarithmic equation yielded the largest erosivity estimates. This equation provided slightly larger, but statistically equal erosivity estimates than the McGregor and the Van Dijk relationships. However, the logarithmic relationship provided statistically different erosivity estimates than the Brown and Foster relationship. An average difference of 36% was observed between the erosivity estimates of these equations for every site. The linear relationship proposed by Hudson yielded statistically different erosivity values to the rest of the equations. This equation provided erosivity estimates that were, in average three times smaller than those predicted with the logarithmic equation. On the other hand, regardless of the KE-I relationship and the site, computing erosivity using the I_(60) provided erosivity estimates that were 10% smaller than those obtained using the I_(30). This is explained because the I_(60) was in average 10% smaller than I_(30) in the study sites. Because erosivity is defined as the kinetic energy of the storms multiplied by their respective I_(30), a 10% variation in the I_(30) accounts for a 10% change in erosivity. Thus, the differences in erosivity reported among the KE-I relationships are independent of the I_(30) estimation method. Finally, because the rainfall erosivity estimates in 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 accurate erosivity estimations. These differences were augmented because of the typically low rainfall intensities at the study sites. Under this condition the KE-I relationships show the largest differences among them. However, with higher rainfall intensities all the KE-I relationships provide similar kinetic energy estimates, making the KE-I relationship selection less important than the use of a reliable I_(30) value for computing rainfall erosivity.