ASAnurag Srivastava

Anurag Srivastava

*This is an algorithmically generated author record based on publicly available information.

Anurag Srivastava has a total of 13 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    An interpretable remote sensing–machine learning framework for wildfire soil burn severity prediction in the northwestern United States

    by Subhankar Das, Mariana Dobre, Roger Lew, Sarah A. Lewis, Erin S. Brooks, Peter R. Robichaud, Anurag Srivastava, Mary Ellen Miller, Alex M. Watanabe, Marta Basso

    Background: Although soil burn severity (SBS) is traditionally mapped post-fire, pre-fire predictive approaches remain limited despite their potential to identify areas at risk of soil degradation, runoff and erosion. Aim: To develop a large-scale model for predicting post-fire SBS from pre-fire conditions at ~30 m spatial resolution. Methods: A harmonized dataset of 62 wildfires (2018–2024) was used to build machine learning (ML) frameworks based on Random Forest, eXtreme Gradient Boosting (XGBoost), and Light Gradient Boosting Machine (LightGBM), using pre-fire vegetation, climate, soil–terrain and anthropogenic predictors. Performance was evaluated using five-fold cross-validation (CV), leave-one-fire-out (LOFO) validation and independent testing on 2025 fires. Key Results: All models showed similar CV overall accuracy (OA ~0.58), but LOFO performance varied. LightGBM maintained more consistent OA (0.51–0.58) across fire size classes, whereas XGBoost and Random Forest exhibited comparable performance. Key predictors included vegetation biomass (Normalized Difference Moisture Index, Enhanced Vegetation Index), fuel stress (Evaporative Stress Index, Evapotranspiration), elevation and soil moisture. High SBS was associated with dry conditions and high biomass availability, with region-specific controls modulating these patterns. Conclusion: ML models using pre-fire environmental data enable proactive SBS prediction. Implications: This framework supports fuel management planning, potential post-fire hydrological impact assessment and the prediction of SBS from pre-fire conditions.


  • Open Access

    Evaluating Watershed Response Using WEPPcloud—EU for Wildfire Burned Areas in Portugal

    by Marta Basso, Erin Brooks, Anurag Srivastava, Jacob Keizer, R. Peter Robichaud

    WEPPcloud—European Union (EU) is an online interface of the Water Erosion Prediction Project (WEPP) model designed to predict streamflow and sediments in European forested watersheds. Although designed to be run in any location in Europe, its performance has not yet been evaluated against observed data. This study evaluates WEPPcloud—EU's performance in simulating streamflow and suspended sediment yield for two burned catchments, Ermida and Serra de Cima, in continental Portugal during the first year after the fire. The default parameters significantly underestimated streamflow and suspended sediment yield at the outlet of both catchments, suggesting the need for catchment‐specific parameterization. Selected model input parameters (bulk density, sand, clay, organic matter and rock content) were changed based on existing information on local conditions at the beginning of an autocalibration procedure. This combined re‐parameterization and autocalibration procedure (based on saturated hydraulic conductivity, anisotropy and baseflow coefficient) improved the performance of WEPPcloud—EU model, yielding satisfactory results for both streamflow and suspended sediment yield for both catchments. The autocalibration improved model accuracy, with predicted average streamflow from the default 3.3 to 6.3 mm day−1 (observed 6.9 mm day−1) and suspended sediment yield from 0.5 to 8.2 kg ha−1 day−1 (observed 11.7 kg ha−1 day−1) in the Ermida catchment. In Serra de Cima, predicted average streamflow improved from 1.8 to 1.9 mm day−1 (observed 1.8 mm day−1) and suspended sediment yield from 1.4 to 35.4 kg ha−1 day−1 (observed 28.5 kg ha−1 day−1). This study demonstrated WEPPcloud—EU potential as a decision‐support tool for post‐fire emergency stabilization in catchments in Europe. Future studies in different eco‐regions in Europe are needed for a broader assessment of the model's potential, including for emergency stabilization scenarios and with an emphasis on the importance of existing local information for achieving satisfactory predictions.


  • A Simulation Study to Estimate Effects of Wildfire and Forest Management on Hydrology and Sediment in a Forested Watershed, Northwestern U.S.

    by Anurag Srivastava, Joan Q. Wu, William J. Elliot, Erin S. Brooks, Dennis C. Flanagan

    Suitable fuel reduction treatments are needed in the Colville National Forest, Washington, to reduce the risk of severe wildfire. This study aimed to identify high-risk erosion hillslopes following wildfire to aid in forest fuel reduction planning and to evaluate the effects of fuel treatments on the watershed hydrological response. The specific objectives were (1) to assess the soil burn severity associated with wildfires and use that information to identify critical hillslopes for forest fuel treatments, and (2) to evaluate the potential changes in water yield and peak flows from pre-treatment (undisturbed forest) to post-treatment (thinning and prescribed burn) conditions, in the East Deer Creek Watershed (EDCW), a subwatershed of the Colville National Forest. Assessments were made using a modeling approach for hypothetical wildfire and fuel treatment scenarios. FlamMap, a fire behavior model, was used to predict the spatial distribution of wildfire intensity for a hypothetical event under current vegetation conditions. WEPP simulations were subsequently completed to obtain sediment and water yields based on fire intensity and topography. WEPP erosion estimations following a simulated wildfire showed hillslope sediment yield varying from 0 to 49.4 Mg ha-1 year-1 from the 777 hillslopes, which were ranked in descending order of sediment yield to identify critical hillslopes for fuel treatments. The WEPP model calibrated for a nearby gauged watershed was then applied to the EDCW for pre-treatment and post-treatment conditions. At the watershed scale, the increase in water yield from pre-treatment to post-treatment conditions ranged from 0.7% to 5.6% on hillslopes delivering 10% to 50% of the predicted post-fire sediment. Simulated water balance components at the treated hillslopes showed substantial changes. Surface runoff, subsurface lateral flow, and deep percolation increased 150% (5 mm), 50% (9 mm), and 40% (41 mm), respectively, whereas evapotranspiration (ET) decreased 23% (124 mm). The relative differences between pre- and post-harvest peak flows showed no clear trends as treatment area increased. The results suggest that thinning and prescribed burns to treated hillslopes in the EDCW may lead to an increase in water yield and significant alterations in hydrological processes. Keywords: Fuel treatments, Modeling, Peak flows, Sediment, Water yield, Wildfire.