Search for: All records

Forested watersheds are instrumental in providing purified and reliable water to millions of people worldwide. The changing climate has increased the frequency and severity of global fire events. Forested watersheds and their ecosystem functions are greatly disrupted during fire activity. Postfire concerns in forested watersheds include unpredictable and potentially simultaneous alterations in source water quality and hydro-biogeochemical processes. The degree of fire severity can complexly modify water quality through the production of fire-transformed constituents on the burned forest floor (i.e., nutrients, metal(loid)s, dissolved organic matter, and the formation of disinfection byproducts). Correspondingly, fire severity and postfire rainfall patterns can refine hydro-biogeochemical processes that influence the transport of the fire-transformed constituents (i.e., vegetation function, soil structure, hydrological pathways, and microbial communities). Postfire alterations to water quality and hydro-biogeochemical processes introduce further complexity with varying temporal influence, which ranges from months to decades. As postfire water quality and watershed response research progresses, it is essential to homogenize interdisciplinary expertise to bridge knowledge gaps between fields ranging from forest ecology, hydrology, microbiology, and geochemistry. A multidisciplinary approach in wildfire research will facilitate a comprehensive perception of the diverse water quality risks associated with fire activity and mitigate fire concerns on a global level. 

This study evaluates the performance of multiple machine learning (ML) algorithms and electrical resistivity (ER) arrays for inversion with comparison to a conventional Gauss-Newton numerical inversion method. Four different ML models and four arrays were used for the estimation of only six variables for locating and characterizing hypothetical subsurface targets. The combination of dipole-dipole with Multilayer Perceptron Neural Network (MLP-NN) had the highest accuracy. Evaluation showed that both MLP-NN and Gauss-Newton methods performed well for estimating the matrix resistivity while target resistivity accuracy was lower, and MLP-NN produced sharper contrast at target boundaries for the field and hypothetical data. Both methods exhibited comparable target characterization performance, whereas MLP-NN had increased accuracy compared to Gauss-Newton in prediction of target width and height, which was attributed to numerical smoothing present in the Gauss-Newton approach. MLP-NN was also applied to a field dataset acquired at U.S. DOE Hanford site.