Effect of data availability and pedotransfer estimates on water flow modelling in wildfire-affected soils

Abstract

Understanding the impact of wildfires on soils exposed to fire is critical, especially in the current climate scenario, where an increase in the occurrence of wildfires is expected. Near-surface soil physical properties are affected by temperature increases caused by wildfires; therefore, changes in the soil water retention curve (SWRC) are expected. Parameters describing the SWRC can be obtained either by measuring or deriving using pedotransfer functions (PTF). However, PTFs have been developed using data from agricultural soils without major heating events; therefore, it is uncertain whether the estimation of parameters in fire-affected soils is reliable. This study evaluated changes in the hydraulic properties of near-surface soil due to fire during three wildfire events of different magnitudes. The objectives were: a) to identify changes in soil properties and SWRC due to wildfires, b) to assess the PTF performance (Rosetta versions 1, 2, and 3) of non-affected and fire-affected soils and (c) to evaluate changes in SWRC due to wildfires and water flow behavior changes through modelling using the HYDRUS-1D model. Decreases in organic matter (OM) and Ksat and increases in pH and bulk density (BD) were observed in fire-affected soils compared to non-affected soils. Based on sand, silt, clay, bulk density, and field capacity, Rosetta version 1 had the lowest values of root-mean-square error for the entire range of suctions, although it did not accurately estimate ?s or Ksat. Among Rosetta’s estimations, Ksat showed the highest variations, which were more marked in fire-affected soils, when measured values were 15.85 cm d-1 while those estimated were 79.14 cm d-1 on average. The implications for hydrologic modelling were translated into lower annual water content and higher infiltration when using Rosetta inputs compared to inputs based on the measured SWRC.