Simulating overland flow following wildfire: mapping vulnerability to landscape disturbance

The probability of landscape-scale disturbances such as fire are expected t o increase in the future due to anticipated climate changes and past land m anagement practices. These disturbances can produce dramatic changes in hyd rologic responses (e,g. overland flow) that can pose risks to human life, i nfrastructure, and the environment. Assessing these risks and associated re mediation strategies requires spatially explicit evaluation of upland hydro logy. However, most current evaluation methods focus on a specified locatio n within a watershed, precluding estimation of spatially distributed, uplan d, hydrological response; and those that do consider spatial variability us ually do not account for redistribution of overland flow among adjacent sub units. Here we highlight the use of a spatially distributed model for asses sing spatial changes in upland hydrologic response following landscape-scal e disturbance. Using a distributed model called SPLASH (Simulator for Proce sses of Landscapes: Surface/Subsurface Hydrology), we simulated pre- and po st-fire scenarios based on the Cerro Grande fire (Los Alamos, NM, USA; May 2000) over 17 300 ha (resolution of 30 m x 30 m) for 2 year and 100 year de sign storms. For the 2 year storm, maximum overland flow rates for burned c ells in the post-fire scenario greatly exceeded those for pre-fire conditio ns (modes: pre-fire, 3.25 x 10(-10) m(3) s(-1); post-fire, 7.0 x 10(-10) m( 3) s(-1)). For the 100 year storm, maximum overland flow was much greater t han for the 2 year storm (modal pre-fire: 31.8 x 10(10) m(3) s(-1)), with t he difference between pre- and post-fire simulations being less dramatic (m odal post-fire: 48.6 x 10(-10) m(3) s(-1)). Mapped differences between pre- and post-fire provide a means for prioritizing upland areas for remediatio n using an approach that accounts not only for topography. soils, and plant cover, but also for the redistribution of overland flow. More generally, o ur results highlight the potential utility of spatially distributed models to focus and prioritize rehabilitation efforts for future assessments of ri sk following landscape-scale disturbance. Copyright (C) 2001 John Wiley & S ons, Ltd.