Snow persistence and hydrologic response across the intermittent-persistent snow transition
Date
2018
Authors
Hammond, John Christopher, author
Kampf, Stephanie, advisor
Covino, Tim, committee member
Denning, Scott, committee member
Fassnacht, Steven, committee member
Journal Title
Journal ISSN
Volume Title
Abstract
In mountainous regions and high latitudes, seasonal snow is a critical component of the surface energy balance and hydrologic cycle. Snowpacks have been declining in many mountain regions, but the hydrologic responses to snow loss have varied due to interactions of climatic, vegetative, topographic and edaphic factors. With continued climatic change, it remains uncertain whether the southwestern U.S. and other subtropical and mid-latitude dry areas may experience significant reductions in water yield. In this dissertation snow persistence and trends are mapped globally; relationships between snow persistence and annual water yield are examined in different climates, and snowmelt and rain partitioning in the critical zone are modelled to examine potential effects of snow loss on hydrologic response. Chapter 2 involves mapping the distribution of snow persistence (SP), the fraction of time that snow is present on the ground for a specific period, using MODIS snow cover data, classifying similar areas into snow zones, assessing how snow persistence relates to climatic variables and elevation, and testing for trends in annual SP. SP is most variable from year to year near the snow line, which has a relatively consistent decrease in elevation with increasing latitude across all continents. At lower elevations, SP is typically best correlated with temperature, whereas precipitation has greater relative importance for SP at high elevations. The largest areas of declining SP are in the seasonal snow zones of the Northern Hemisphere. Trend patterns vary within individual regions, with elevation, and on windward-leeward sides of mountain ranges. This analysis provides a framework for comparing snow between regions, highlights areas with snow changes, and can facilitate analyses of why snow changes vary within and between regions. In Chapter 3, SP is used to evaluate how water yield relates to snow patterns at the annual time scale across the western U.S. in different climates. I first compare snow cover variables derived from MODIS to more commonly utilized metrics (snow fraction and peak snow water equivalent (SWE)). I then evaluate how SP and SWE relate to annual streamflow (Q) for 119 USGS reference watersheds and examine whether these relationships vary for wet/warm (precipitation surplus) and dry/cold (precipitation deficit) watersheds. Results show high correlations between all snow variables, but the slopes of these relationships differ between climates. In dry/cold watersheds, both SP and SNODAS SWE correlate with Q spatially across all watersheds and over time within individual watersheds. I conclude that SP can be used to map spatial patterns of annual streamflow generation in dry/cold parts of the study region. In Chapter 4 of the dissertation, I use a series of one-dimensional simulations to study how snow loss may impact hydrologic response in mountain areas at event to annual time scales. I use Hydrus 1-D simulations with historical inputs from fifteen SNOTEL snow monitoring sites to investigate how inter-annual variability of water input type (snowmelt, rainfall) and timing affect soil saturation and deep drainage in different soil types and depths. Greater input rate and antecedent moisture are observed for snowmelt compared to rain events, resulting in greater runoff efficiencies. At the annual scale runoff efficiencies increase with snowmelt fraction and decrease when all input is rainfall. In contrast, deep drainage has no clear correlation to snowmelt fraction. Input that is concentrated in time leads to greater surface runoff and deep drainage. Soil texture and depth modify partitioning, but these effects are small compared to those caused by variability in climate. This dissertation's findings have direct implications for climate change impacts in cold dry areas globally. Through the synthesis of the chapters described above I highlight areas where hydrologic response to snow loss may be most sensitive, provide methods for comparing regional snow patterns, demonstrate how snow persistence can help estimate annual streamflow generation, and improve process-based knowledge of hydrologic response to rainfall and snowmelt in the western U.S. Collectively these findings indicate that annual water yield is not directly sensitive to whether input is snowmelt vs. rainfall; instead it is more dependent on the effect that snowpack accumulation has on input timing and rate. Loss of concentrated melt from persistent snowpacks may lead to lower streamflow and compromise deep drainage, and thus aquifer recharge, in semi-arid cold regions. The consequences of streamflow and groundwater recharge loss could be severe in regions already water-stressed, and this needs to be addressed in long-term water supply planning.