Wildfire impacts to water quality investigated using a LANDIS-VELMA model linkage for water resource protection

Wildfire activity in western US forests has increased over the last two decades, resulting in increased solids and nutrient loadings to streams, and in some cases threatening drinking water supplies. These increases are attributed to rainfall events and the distribution of water throughout the watershed. Modeling the influence of riparian biomass on watershed dynamics is essential for the protection of aquatic ecosystem health and drinking water quality by understanding interactions between climate, land use change, and complex hydrodynamic-biogeochemical processes. To fully assess these tradeoffs across media requires an integrated modeling platform. The Hayman Fire of 2002, one of the largest and most devastating wildfires in Colorado's recorded history, served as the study area for the linked model application. We utilized the Visualizing Ecosystem and Land Management Assessment model (VELMA) eco-hydrologic model and the LANDscape DIsturbance and Succession (LANDIS-II) plant community model to simulate the Hayman Fire pre- and post-fire conditions (Venable et al., 2024).

Our research goal is to develop an experimental modeling platform to protect water quality while evaluating effective management options deemed by the USDA Forestry Services’ silvicultural treatments (2024) based on LANDIS-II’s biomass characterization. To protect water resources, our case study modeling tool will 1) improve baseline water quantity through model optimization using NSE evaluation, 2) specify riparian buffer location, age, and species, 3) evaluate effectiveness of thinning to reduce burn severity, and 4) provide modeled EPA water quality criteria conversion for interactive visualizations and reports to compare scenario differences associated with best management practices and mitigation strategies.

Here we present the result of a single catchment, Brush Creek, due to the availability of observed water quality data available pre- and post-fire, and demonstrate the development, deployment, and efficacy of the modeling workflow. The products developed are based on Python open-source software and provide greater efficiency, reducing manual processing errors, and facilitating reusable science for other land surface modeling studies.