Premium
Assessment of wildland fire impacts on watershed annual water yield: Analytical framework and case studies in the United States
Author(s) -
Hallema Dennis W.,
Sun Ge,
Caldwell Peter V.,
Norman Steven P.,
Cohen Erika C.,
Liu Yongqiang,
Ward Eric J.,
McNulty Steven G.
Publication year - 2017
Publication title -
ecohydrology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.982
H-Index - 54
eISSN - 1936-0592
pISSN - 1936-0584
DOI - 10.1002/eco.1794
Subject(s) - environmental science , watershed , streamflow , rangeland , soil and water assessment tool , precipitation , context (archaeology) , climate change , hydrology (agriculture) , storm , water resource management , meteorology , drainage basin , agroforestry , geography , ecology , geology , cartography , geotechnical engineering , archaeology , machine learning , computer science , biology
More than 50% of water supplies in the conterminous United States originate on forestland or rangeland and are potentially under increasing stress as a result of larger and more severe wildfires. Little is known, however, about the long‐term impacts of fire on annual water yield and the role of climate variability within this context. We here propose a framework for evaluating wildland fire impacts on streamflow that combines double‐mass analysis with new methods (change point analysis, climate elasticity modeling, and process‐based modeling) to distinguish between multiyear fire and climate impacts. The framework captures a wide range of fire types, watersheds characteristics, and climate conditions using streamflow data, as opposed to other approaches requiring paired watersheds. The process is illustrated with three case studies. A watershed in Arizona experienced a +266% increase in annual water yield in the 5 years after a wildfire, where +219% was attributed to wildfire and +24% to precipitation trends. In contrast, a California watershed had a lower (−64%) post‐fire net water yield, comprised of enhanced flow (+38%) attributed to wildfire offset (−102%) by lower precipitation in the post‐fire period. Changes in streamflow within a watershed in South Carolina had no apparent link to periods of prescribed burning but matched a very wet winter and reports of storm damage. The presented framework is unique in its ability to detect and quantify fire or other disturbances, even if the date or nature of the disturbance event is uncertain, and regardless of precipitation trends.