Risk-assessment of post-wildfire hydrological response in semi-arid basins: The effects of varying rainfall representations in the KINEROS2/AGWA model

Authors: SIDMAN, G. - University Of Arizona; GUERTIN, D.P. - University Of Arizona; Goodrich, David - Dave; Unkrich, Carl; BURNS, I.S. - University Of Arizona

Submitted to: International Journal of Wildland Fire
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/7/2015
Publication Date: 7/7/2015
Citation: Sidman, G., Guertin, D., Goodrich, D.C., Unkrich, C.L., Burns, I. 2015. Risk-assessment of post-wildfire hydrological response in semi-arid basins: The effects of varying rainfall representations in the KINEROS2/AGWA model. International Journal of Wildland Fire. 25(3):268-278. https://doi.org/10.1071/WF14071.
DOI: https://doi.org/10.1071/WF14071

Interpretive Summary: Wildfires are increasing in both their intensity and area burned throughout the Western United States. After a wildfire is contained an assessment of damage and hazards, both on-site and downstream of the fire, is conducted by Burn Area Emergency Response (BAER) teams. BAER assessments with local emergency officials are used to plan and target post-fire mitigation efforts. Post-fire flooding and erosion can be severe and computer-based watershed models are one of the tools BAER teams use indicate where flooding and severe erosion might occur. However, the BAER teams typically must do the post-fire watershed modeling before an actual rainfall event has occurred. Therefore a design storm that has some of the statistical characteristics of historically observed storms in the areas is often used as input to the watershed models used by the BAER teams. This study used the KINEROS2/AGWA model to compare several spatial and temporal rainfall representations of post-fire rainfall-runoff events to determine the effect of differing representations on modeled peak flow in determining at-risk locations within a watershed. Post-fire rainfall-runoff events at Zion National Park in Utah and Bandelier National Monument in New Mexico were modeled. Representations considered included both uniform and design storms over the entire watershed and applying rain only on the burned area, and varying rainfall both temporally and spatially according to radar derived rainfall data. Results showed that rainfall representation greatly affected modeled peak flow, but did not significantly alter the model’s predictions for high-risk locations. This has important implications for post-fire assessments prior to a flood-inducing rainfall event, or for post-storm assessments in areas with low-gage density or lack of radar data due to mountain beam blockage.

Technical Abstract: Representation of precipitation is one of the most difficult aspects of modeling post-fire runoff and erosion and also one of the most sensitive input parameters to rainfall-runoff models. The impact of post-fire convective rainstorms, especially in semi-arid watersheds, depends on the overlap between locations of high intensity rainfall and areas of high severity burns. One of the most useful applications of models in post-fire situations is risk-assessment to quantify peak flow and identify areas at high risk to flooding and erosion. This study used the KINEROS2/AGWA model to compare several spatial and temporal rainfall representations of post-fire rainfall-runoff events to determine the effect of differing representations on modeled peak flow and determining at-risk locations within a watershed. Post-fire rainfall-runoff events at Zion National Park in Utah and Bandelier National Monument in New Mexico were modeled. Representations considered included both uniform and SCS Type II hyetographs, applying rain over the entire watershed and applying rain only on the burned area, and varying rainfall both temporally and spatially according to radar data. Results showed that rainfall representation greatly affected modeled peak flow, but did not significantly alter the model’s predictions for high-risk locations. This has important implications for post-fire assessments prior to a flood-inducing rainfall event, or for post-storm assessments in areas with low-gage density or lack of radar data due to mountain beam blockage.