Bark beetle outbreaks have impacted over 58 million acres of coniferous forest in the Western US since 2000, an area slightly larger than the state of Utah. Most of these beetle-impacted forests are in semi-arid, snow-dominated headwater catchments that generate a disproportionate fraction of water supplies. Limited previous studies have shown severe beetle-kill can cause mixed increases and decreases in streamflow. This study is the first to empirically explore changes in streamflow following a recent spruce beetle outbreak in southwest Colorado using a paired catchment approach. The period following beetle kill (2014-2019) was 0.95° C warmer and 5.8 cm/year drier than the 21-year period prior to the disturbance’s peak (1993-2013). There was no change in streamflow in the control basins after beetle kill. In contrast, post-beetle kill had 34% higher peak flows on average and consistent predictions of >14% increases in streamflow in wetter basins and >20% in drier basins. Our results suggest that higher streamflows are primarily driven by 44% higher runoff efficiencies during the snowmelt period. The increased flows due to beetle kill are occurring at a time when control catchments have unchanged runoff efficiencies. These findings are the first to clearly show streamflow increases following extensive spruce beetle kill in watersheds that contribute water to millions of downstream residents. Moreover, our findings contrast with evidence of unchanged or decreased streamflow following mountain pine beetle kill in nearby parts of Colorado, highlighting the need for better post-disturbance hydrologic predictions in these important montane forests.

Snow disappearance date (SDD) has a substantial impact on the ecohydrological dynamics of montane forests, by affecting soil moisture, ecosystem water availability, and fire risk. The forest canopy modulates SDD through competing processes, such as intercepting snowfall and enhancing longwave radiation (LWR) versus reducing near surface shortwave radiation (SWR) and wind speed. Limited ground-based observations of snow presence and absence have restricted our ability to unravel the dominant processes affecting SDD over mountains with complex forest structure. We apply a lidar-derived method to estimate fractional snow cover area (fSCA) at two relatively warm sites in the Sierra Nevada and two colder sites in the Rocky Mountain. Our analyses show that warm sites and lower elevations are characterized by higher LWR and canopy snow interception leading to less snow retention under dense forest canopy. In contrast, snow retention in colder forests can be longer in open or under canopy depending on interactions between vegetation structure and topography. These colder climates have greater under canopy snow retention on north-facing slopes and under low vegetation density areas, but greater snow retention in open areas at lower elevations and south-facing slopes. We develop a new conceptual model to incorporate the role of topography and vegetation structure into existing climate-based frameworks. The inferences into the interacting energy and mass controls, derived from our lidar datasets give opportunities to improve hydrological modeling and provide targeted forest management recommendations.

Dissolved oxygen (DO) sustains aquatic life and is an essential water quality measure. Our capabilities of forecasting DO levels, however, remain elusive. Unlike the increasingly intensive earth surface and hydroclimatic data, water quality data often have large temporal gaps and sparse areal coverage. Here we ask the question: can a Long Short-Term Memory (LSTM) deep learning model learn the spatio-temporal dynamics of stream DO from intensive hydroclimatic and sparse DO observations at the continental scale? That is, can the model harvest the power of big hydroclimatic data and use them for water quality forecasting? Here we used data from CAMELS-chem, a new dataset that includes sparse DO concentrations from 236 minimally-disturbed watersheds. The trained model can generally learn the theory of DO solubility under specific temperature, pressure, and salinity conditions. It captures the bulk variability and seasonality of DO and exhibits the potential of forecasting water quality in ungauged basins without training data. It however often misses concentration peaks and troughs where DO level depends on complex biogeochemical processes. The model surprisingly does not perform better where data are more intensive. It performs better in basins with low streamflow variations, low DO variability, high runoff-ratio (> 0.45), and precipitation peaks in winter. This work suggests that more frequent data collection in anticipated DO peak and trough conditions are essential to help overcome the issue of sparse data, an outstanding challenge in the water quality community.