Hydrologic Partitioning Response to Severe Forest Disturbance Quantified by Eddy Covariance, Streamflow, Snow Surveys and Stable Isotope Fractionation

Forested montane watersheds are critical sources of fresh water, but forest disturbance of unprecedented rate and extent is challenging our ability to understand future water supplies. Disturbance from natural agents including fire, drought, and pathogen infestation may manifest early impacts of climate change, suggesting disturbance will persist and intensify. In Western North America, bark beetles are the most widespread natural disturbance agent, having impacted over 10 Mha since the mid-1990s. Hydrologic response to overstory die-off is generally expected to include reduced water vapor losses via interception and transpiration and increased streamflow, consistent with recent bark beetle disturbance studies employing land surface models or remote sensing evapotranspiration products (e.g. MODIS). Surprisingly, streamflow increases have not been reported after nearly two decades of the present epidemic. Recent studies of post-disturbance ecosystem water vapor flux or snowpack mass balance suggest that vapor losses may not decline as expected, but it remains unclear how ecosystem vapor flux response to forest disturbance regulates streamflow, and at what spatial and temporal scales.

We address this knowledge gap with simultaneous observations of climate, eddy covariance vapor flux, and streamflow in a subalpine forest ecosystem for three years following a severe (75% lodgepole pine mortality) bark beetle infestation in the central Rocky Mountains of Wyoming, and we compare the results to an undisturbed control ecosystem. Effective precipitation was quantified using multiple precipitation gauges and distributed surveys of peak snowpack to assess winter sublimation vapor losses prior to snowmelt. Effective precipitation was unaltered by the forest mortality, as reduced interception was compensated by increased snowpack sublimation. Annual vapor flux from the control ecosystem was relatively insensitive to annual climate differences, varying only from 573 to 623 mm, while the disturbed ecosystem was larger and more variable, ranging from 569 to 700 mm. Both direct streamflow observations and calculated residual streamflow from eddy covariance (Precipitation – Water Vapor Flux) agreed in showing a decline in annual streamflow from the disturbed ecosystem. Annual runoff ratios (streamflow: precipitation) at the control site varied from 14% to 40% and followed the pattern of annual precipitation, while disturbed site runoff ratios declined over the three years from 37% to 0%. Stable isotope fractionation in snowpack, soil water and stream water showed kinetic enrichment only at the disturbed site, demonstrating up to 40% abiotic evaporation of precipitation. Increased snowpack sublimation and soil water evaporation were most likely driven by increases in solar radiation reaching the subcanopy, which increased by ca. 70% in winter and 300 - 350% in the growing season and probable increases in subcanopy turbulence. Collectively, these results indicate that compensatory vapor fluxes counteracted the expected impacts of reduced canopy interception and transpiration. Evaporation and sublimation contributed to maintaining or increasing total annual ecosystem vapor flux, limiting streamflow over one to three years at the spatial scale of a small headwater stream.