4.3 Projections in a changing climate
Over the past decades, the western United States has experienced declining snowfall (Kunkel et al., 2009; Harpold et al., 2012) and earlier melt pulses (Clow, 2010), a trend that is projected to continue (e.g., Siirila-Woodburn et al., 2021). Based on observations that solute fluxes are strongly tied to the magnitude and timing of the melt pulse, this will potentially result in decreased carbon and nitrogen export and greater retention within the catchment. While regionally, the western United States is expected to become dryer, some areas may actually see increased precipitation as a result of local conditions. The leeward side of the central Rocky Mountains appears to potentially be one such case (Kunkel et al., 2009). Indeed, from 1951 to 1994, Niwot Ridge LTER experienced a statistically significant (p < 0.01) increase in annual precipitation of 7.5 mm yr-1 (Williams et al., 1996b). However, this trend appears to have reversed, or at the very least stalled in the more recent SNOTEL data from 1995 to 2021 (slope = -3.9 mm yr-1; p = 0.19).
In addition to the overall volume, the form of precipitation (i.e. snow versus rain) may also change. With warmer temperatures, a greater fraction of precipitation would be expected to fall as rain rather than snow. As illustrated by the sensor data, rainfall events are clearly capable of rapidly mobilizing and transporting solutes, perhaps even more effectively than snowmelt. This suggests that a shift from snow to rain may not result in as large a decline in solute export as otherwise expected. Warmer temperature could also result in more frequent “mini-melt” periods. Modeling studies (Jennings and Molotch, 2020) suggest only a 3° C increase is required for much of Como Creek watershed to experience significant melting throughout the winter. More precipitation falling as rain rather than snow, combined with more frequent periods of winter melt would alter the timing of export such that it is more evenly distributed and less condensed into a singular seasonal pulse
Several additional sources of very large uncertainty also remain. In addition to changing precipitation, changing catchment characteristics may also influence stream hydrology. Warmer temperatures will likely result in a general upward migration of the tree line, though temperature is far from the only controlling factor (Bueno de Mesquita et al., 2018). The fraction which is forested has been demonstrated to exert substantial control on rates of evapotranspiration and water yield in nearby catchments (Sueker et al., 2001). It is possible that a shift to an earlier melt pulse, when rates of evapotranspiration are lower, could at least partially offset these losses (Barnhardt et al., 2020). Recent modeling work in Como Creek (Barnhart et al., 2021) has also suggested that expansion of forested area in response to warming temperatures may actually increase streamflow by decreasing snow wind-scour. In short, there is large uncertainty in how temperature driven changes to catchment vegetation will interact to affect catchment hydrology, including runoff.
Another large unknown is how biogeochemical processing of carbon and nitrogen within the catchment will respond to changing temperatures and snow cover. Biogeochemical cycles consist of multiple interconnected and sometimes reciprocal pathways. With the N cycle for example, atmospheric fixation, mineralization of organic nitrogen, nitrification and denitrification are all sensitive to temperature and soil moisture (e.g., Fisk and Schmidt, 1995; Osborne et al., 2016; Chen et al., 2020; Maslov and Maslova, 2021). Somewhat counterintuitively, reduction in the depth and duration of the insulating snowpack may actually result in colder subnivean temperatures and reduced biological activity (Williams et al., 1998). Snow cover has been identified as a critical control on subnivean microbial processing (Brooks et al., 1996), and by extension, soil water chemistry (Lewis and Grant, 1980). Williams (et al., 1998) projects that increased snow cover will enhance net soil nitrification and result in greater NO3-N availability, while a reduction in snowfall will have the opposite effect and result in greater retention. This is strongly supported by the observation that solute flux appears most closely associated with persistence of the snowpack later into the spring (which results in higher discharge but also higher concentrations). While the total snowfall, maximum snowpack depth, and its duration are certainly linked, they are not perfectly correlated. An unseasonably warm spring can quickly melt a deeper than average snowpack, as we observed in 2020. This provides a third example of how changing temperatures over the coming decades can alter the solute export, even if total annual precipitation remains unchanged.