Where Does Missing Water Go? Constraining Hydrologic Anomalies in Snow-Dominated Mountain Catchments

Abstract

This dissertation explores uncertainties that can drive hydrologic anomalies, which occur when observed streamflow falls far from projected values. Each chapter targets a different point where uncertainty enters the mountain water balance: how much snow is lost before it can melt, how accurately we measure precipitation, and how much storage shapes streamflow over multiple years. In Chapter 2, we combine meteorological observations spanning site-to-synoptic scales from three field campaigns in a Colorado mountain valley to identify the events responsible for the majority of winter sublimation. We find that most sublimation occurs during a small number of discrete events coinciding with either short sunny, dry periods or long storm events with new snowfall and blowing snow. The number of these events varies considerably year to year with no significant long-term trend. Accurately representing seasonal sublimation requires capturing discrete events driven by identifiable synoptic and local meteorological conditions. In Chapter 3, we compare precipitation measurements from 11 co-located gauges, supplemented by radar-derived and gridded products, to identify the dominant drivers of measurement variability at a mountain site. We find large discrepancies in total accumulation across gauges, with biases concentrated in a small number of winter events driven by power failures, gauge burial, blowing snow, and high snowfall rates. For future deployments in snow-dominated mountain terrain, we recommend co-located storage-type gauges with low power requirements to ensure reliable, continuous measurements. In Chapter 4, we use the Structure for Unifying Multiple Modeling Alternatives (SUMMA) model to evaluate whether representing long-term water storage improves simulated streamflow in two contrasting snow-dominated basins. Observed late-season low flows carried a memory signature that models without an explicit storage compartment could not reproduce, while adding a basin-appropriate compartment recovered it. These stores contributed disproportionately to low flows in dry years but did little to improve water-year volume accuracy. As mountain snowpack and groundwater decline, a model's ability to account for missing water will depend on how well it represents the dominant storage mechanism in the basin in question. Together, these chapters show that much of this uncertainty traces back to conditions that average behavior obscures: discrete high-impact events in the case of sublimation and precipitation measurement, and slow storage dynamics that only become visible across years. As mountain systems experience increased variability, these results suggest that improving representations of the mountain water balance requires deliberately targeting what average conditions leave out.