Jump to Main Content
Estimation of event‐based rainfall erosivity from radar after wildfire
- Zhu, Qinggaozi, Yang, Xihua, Yu, Bofu, Tulau, Mitch, McInnes‐Clarke, Sally, Nolan, Rachael H., Du, Zheyuan, Yu, Qiang
- Land degradation & development 2019 v.30 no.1 pp. 33-48
- Revised Universal Soil Loss Equation, energy, erosion control, meteorological data, monitoring, national parks, prediction, radar, rain, rain gauges, rain intensity, soil erosion, storms, time series analysis, topographic slope, wildfires, Australia
- Rainfall erosivity impacts all stages of hillslope erosion processes and is an important factor (the ‘R factor’) in the Revised Universal Soil Loss Equation. It is estimated as the average annual value of the sum of all erosive events (EI₃₀) over a period of many years. For each storm event, the EI₃₀ value is the product of storm energy, E in MJ ha⁻¹, and peak 30‐min rainfall intensity (I₃₀, mm hr⁻¹). Previous studies often focused on estimation of the R factor for prediction of mean annual or long‐term soil losses. However, many applications require EI₃₀ values at much higher temporal resolution, such as postfire soil erosion monitoring, which requires a time step at storm events or on a daily basis. In this study, we explored the use of radar rainfall data to estimate the storm event‐based EI₃₀ after a severe wildfire in Warrumbungle National Park in eastern Australia. The radar‐derived rainfall data were calibrated against 12 tipping bucket rain gauges across an area of 239 km² and subsequently used to produce a time series of rainfall erosivity maps at daily intervals since the wildfire in January 2013. The radar‐derived daily rainfall showed good agreement with the gauge measurements (R² > 0.70, Ec = 0.66). This study reveals great variation in EI₃₀ values ranging from near zero to 826.76 MJ·mm·ha⁻¹·hr⁻¹ for a single storm event. We conclude that weather radar rainfall data can be used to derive timely EI₃₀ and erosion information for fire incident management and erosion control. The methodology developed in this study is generic and thus readily applicable to other areas where weather radar data are available.