Changing Hydrologic Conditions in the Rio Grande Headwaters and Implications for Water Resource Management

Abstract:

Seasonal streamflow forecasts in the Rio Grande Headwaters, Colorado, are beneficial for water resource management in an area often affected by drought. These forecasts rely on accurate representation of basin-wide snowpack and relationships between snowpack and streamflow. In this study, statistical and physical modeling applications were used to assess recent changes in both climate and land cover and their effects on the snowpack-streamflow relationship, with implications for forecast accuracy. Results of multiple linear regression indicate that shifts in the seasonality of precipitation and land-cover changes, namely wildfire, explain 2%-18% more variability in streamflow prediction than traditional statistical methods utilizing SNOTEL data. Basin-wide snow water equivalent (SWE) and snow-covered area (SCA) were simulated using SnowModel and evaluated for trends and possible changing relationships at SNOTEL stations through the melt period April­–June. SWE and SCA showed significant decreasing trends of -4.33 mm/yr and -0.05%/yr, respectively. Relationships between SWE and SCA with measured SWE at select SNOTEL stations varied by station but show potential shifts in these relationships when stations are affected by bark beetle tree mortality or wildfire. More resilience was shown in stations located outside critical zones experiencing snowpack changes related to climate change. The effects of land-cover change on streamflow were evaluated using the Precipitation-Runoff Modeling System in a series of numerical experiments. Results show bark beetle tree mortality had little effect on streamflow, while wildfire increased streamflow an average 35% for 1–4 years post-fire. Simulated storage terms, such as soil moisture and evapotranspiration fluxes, were also altered in post-wildfire years. A statistical model predicting streamflow from SnowModel SWE did not perform significantly better than a SNOTEL-based model, highlighting the potential importance of hydrologic conditions other than snowpack when predicting seasonal streamflow. A spreadsheet tool was developed to graphically show the difference in statistical model forecasts when SNOTEL, precipitation, and forest disturbance lags are input to the model. These results and tool can help guide forecast development in the region.