The frontier of wildfire-related risk assessment is moving into data science territory, and with good reason. Computational statistics, built on a foundation of high resolution remote sensing data, ground data, and theory, forms the basis of powerful risk assessment tools. The need for data based risk assessment has increased in past years, in view of longer wildfire seasons in the U.S., associated with more frequent droughts, more human ignitions and accumulating fuel loads. We present an application of machine learning (ML), which makes it possible to analyze complex data without a priori definition of interactions—this is a major advantage because these interactions are not known beforehand. Specifically, we build a stochastic gradient boosting machine (GBM) toolkit to assess the change in river flow after wildfire in the contiguous United States (CONUS) over a 5-year period. The GBM accounts for nonlinear relationships and interactions between wildland fire characteristics, watershed geometry, climate variability, topography and land cover. Building the GBM is a sequential process where a loss function is minimized at each fold, along a gradient defined by pseudo-residuals. This process allows the program to progressively learn more about how the variables in the large dataset interact to result in the response (i.e., river flow). Our results show that wildfires increase annual river flow in the CONUS when more than 20% of a gaged basin is burned. Data science tools like the GBM presented here, are essential in generating practical knowledge on how wildfire impacts on ecohydrology can ultimately affect hydrological services, socio-hydrosystems and water security in fire-affected regions.

Understanding drivers of thermal regimes in headwater streams is critical for a comprehensive understanding of freshwater ecological condition and habitat resilience to disturbance, and to inform sustainable forest management policies and decisions. However, stream temperatures may vary depending on characteristics of the stream, catchment, or region. To improve our knowledge of the key drivers of stream thermal regime, we collected stream and air temperature data along eight headwater streams in two regions with distinct lithology, climate, and riparian vegetation. Five streams were in the Northern California Coast Range at the Caspar Creek Experimental Watershed Study, which is characterized by permeable sandstone lithology. Three streams were in the Cascade Range at the LaTour Demonstration State Forest, which is characterized by fractured and resistant basalt lithology. We instrumented each stream with 12 stream temperature and four air temperature sensors during summer 2018. Our objectives were to compare stream thermal regimes and thermal sensitivity—slope of the linear regression relationship between daily stream and air temperature—within and between both study regions. Mean daily stream temperatures were ~4.7 °C warmer in the Coast Range but were less variable (SD = 0.7 °C) compared to the Cascade Range (SD = 2.3 °C). Median thermal sensitivity was 0.33 °C °C-1 in the Coast Range and 0.23 °C °C-1 in the Cascade Range. We posit that the volcanic lithology underlying the Cascade streams likely supported discrete groundwater discharge locations, which dampened thermal sensitivity. At locations of apparent groundwater discharge in these streams, median stream temperatures rapidly decreased by 2.0 °C, 3.6 °C, and 7.0 °C relative to adjacent locations, approximately 70–90 meters upstream. In contrast, thin friable soils in the Coast Range likely contributed baseflow from shallow subsurface sources, which was more sensitive to air temperature and generally warmed downstream (up to 2.1 °C km-1). Our study revealed distinct longitudinal thermal regimes in streams draining contrasting lithology, suggesting that streams in these different regions may respond differentially to forest disturbances or climate change.