This paper by Nicholas Siler, Cristian Proistosescu, and Stephen Po‐Chedley was published in Geophysical Research Letters in December 2018: Natural Variability Has Slowed the Decline in Western U.S. Snowpack Since the 1980s. It is freely accessible. Click on the title to access the online version, which will have all the hot links activated,
Significant work! The lead author, Nick Siler, is my colleague at OSUCEOAS.
Here is an OSU press release on the paper.
Download Siler_et_al-2019-Geophysical_Research_Letters
Abstract
Spring snowpack in the mountains of the western United States has not declined substantially since the 1980s, despite significant global and regional warming. Here we show that this apparent insensitivity of snowpack to warming is a result of changes in the atmospheric circulation over the western United States, which have reduced snowpack losses due to warming. Climate model simulations indicate that the observed circulation changes have been driven in part by a shift in Pacific sea surface temperatures that is attributable to natural variability, and not part of the simulated response to anthropogenic forcing. Removing the influence of natural variability reveals a robust anthropogenic decline in western U.S. snowpack since the 1980s, particularly during the early months of the accumulation season (October–November). These results suggest that the recent stability of western U.S. snowpack will be followed by a period of accelerated decline once the current mode of natural variability subsides.Plain Language Summary
Melting snowpack is a vital source of water in the western United States during the summer, when rainfall is usually scarce. Although the amount of water contained in the snowpack has declined over the past century, it has been surprisingly stable since the 1980s, despite 1 °C of warming over the same period. At first glance, this result might appear to indicate that the snowpack is quite resilient to warming. However, here we show that the contribution of global warming to western U.S. snowpack loss has in reality been large and widespread since the 1980s, but mostly offset by natural variability in the climate system. This result points to a faster rate of snowpack loss in coming decades, when the impact of global warming is more likely to be amplified, rather than offset, by natural variability.
1) Introduction
Snowpack in the western United States acts as a natural reservoir, storing water during the cool wet season (October–March) and releasing it to the landscape during the warm dry season (Barnett et al., 2005; Cayan, 1996). Containing more than 150 cubic kilometers of water at its peak—usually reached around 1 April—the western U.S. winter snowpack holds about as much water as all the man‐made reservoirs in the western United States combined (Mote et al., 2018). Since 1950, the snowpack on 1 April has decreased by 15–30% over much of the western United States, as warmer temperatures have caused a shift from snow to rain, particularly at low elevations (Mote et al., 2018). Models indicate a further decrease in winter snowpack of perhaps 60% by 2050 (Ashfaq et al., 2013; Fyfe et al., 2017), leading to a dramatic reduction in summer stream flows in a region where water shortages are already common (Barnett et al., 2005; Christensen et al., 2004)
Despite this alarming forecast, snowpack has been surprisingly stable in recent decades. Figure 1a shows the trends in 1‐April snowpack at 329 automated Snowfall Telemetry (SNOTEL) stations where the average 1‐April snowpack contains at least 25 cm of water equivalent. Trends were calculated over 35 winters beginning in 1983–1984, when the network reached nearly its current level of coverage (Figure S1 in the supporting information), and 1 year after the record‐setting El Niño of 1982–1983. During this 35‐year period, only four sites experienced a statistically significant decline in 1‐April snowpack (95% confidence), while 208 sites (63%) experienced an insignificant decline. The other 117 sites (36%) experienced a positive (but statistically insignificant) trend. We find a similar result when we repeat the analysis using regionally averaged time series of 1‐April snowpack (Figure 1b), with no region exhibiting a statistically significant trend since 1983–1984. This result is consistent with Mudryk et al. (2018), who found similar stability in the spring snowpack of western Canada over roughly the same time period. Meanwhile, the average winter surface temperature over the western U.S. increased by more than 1 °C during this period—on par with the average warming trend across global land surfaces (Figure S2). Taken at face value, this result suggests that winter snowpack in the western United States may be remarkably insensitive to warming, seemingly contradicting model forecasts of rapid decline (Ashfaq et al., 2013; Fyfe et al., 2017).
On decadal timescales, however, snowpack is influenced not only by anthropogenic warming, but also by natural variability in the large‐scale atmospheric circulation (Deser et al., 2012). Depending on its phase, natural variability can either offset or accelerate the changes in snowpack due to anthropogenic warming (Cayan, 1996; McCabe & Wolock, 2009; Mote, 2006; Smoliak et al., 2010; Stoelinga et al., 2010). In order to quantify the anthropogenic trend in western U.S. snowpack, one must first remove the component of the trend that can be attributed to natural variability in the atmosphere.
One approach that previous studies have taken to minimize the noise of natural variability has been to focus on the ratio of snowpack to accumulated precipitation (S/P), which tends to be more sensitive than snowpack to changes in temperature (Barnett et al., 2008; Pierce et al., 2008; Pierce & Cayan, 2013). Across the SNOTEL network, we similarly find more robust declines in S/P than in snowpack alone (Figure S3 vs. Figures 1a and 1b). However, while S/Pmay exhibit a clearer warming signal, it provides little insight into the magnitude of snowpack change resulting from anthropogenic warming versus natural variability.
In this paper, we use a method called “dynamical adjustment” to quantify the influences of both anthropogenic warming and natural variability on 1‐April snowpack at each of the 329 SNOTEL stations shown in Figure 1a. We find that, since 1983–1984, changes in the atmospheric circulation have contributed to a ∼30% increase in 1‐April snowpack in the Cascade Mountains and northern Sierra Nevada, and an increase of ∼10% in Utah and the Northern Rocky Mountains, offsetting much of the decline due to global warming. Simulations performed with an atmospheric general circulation model (AGCM) and prescribed historical sea surface temperatures (SSTs) indicate that the observed circulation change has likely been driven in part by internal atmospheric variability, and in part by a shift in Pacific SSTs toward the cool phase of the Interdecadal Pacific Oscillation (IPO). We find that such SST/circulation trends are not part of the simulated response to anthropogenic forcing in coupled GCMs but are likely associated with low‐frequency natural variability that will eventually subside, ushering in a period of accelerated snowpack loss.
Cutting to the chase...
4) Future Implications
Looking forward, the implications of our results depend on whether the observed SST/circulation changes in recent decades have been part of the climate's response to anthropogenic forcing, or simply a result of low‐frequency natural variability. To answer this question, we compare the observed SST/circulation trends in Figures 3b and 3c with those simulated by a large ensemble (n = 86) of coupled oceanic‐atmospheric GCMs from CMIP5 (Taylor et al., 2012) over the same 35‐year time period (Figure 5; Table S1). In contrast to the observed trends, the ensemble mean of these simulations shows little evidence of an IPO‐like shift in SST patterns (Figure 5a), or of a substantial change in the atmospheric circulation over the western United States (Figure 5b), suggesting that observed trends are not primarily driven by anthropogenic forcing.
On the other hand, the observed trend in U500 over the western United States does fall within the range that one would expect from natural variability, based on the CMIP5 ensemble spread (Figure 4b). Furthermore, among the ensemble members that exhibit the largest U500trends over the western United States (Figure 5d), SST trends resemble the observed shift toward the cool phase of the IPO (Figure 5c), supporting our hypothesis that circulation trends have been driven in part by shifting SST patterns. Meanwhile, the same subset of GCMs produced a wide range of U500 trends in other identical simulations (Figure 4b, red shapes), confirming that the CMIP5 ensemble spread is primarily a reflection of natural variability, and not of differences in model physics.
Of course, it is possible that observed SST/circulation trends represent a component of the forced response that most GCMs do not capture (Kohyama & Hartmann, 2017). If so, then changes in the atmospheric circulation may continue to offset some of the decline in western U.S. snowpack due to warming. But if recent circulation changes have instead been a result of natural variability—the more likely outcome in our view, given the close correspondence between observed SST changes and the IPO—then the relative stability of western U.S. snowpack since the 1980s is unlikely to persist, portending an accelerated decline in coming decades as the phase of variability becomes less favorable for snowpack accumulation.
Gotta recharge that stuff!
Enjoy!
"‘The Lorax' is actually a cautionary tale about the need for data analytics and best business practices in Higher Ed. 'The Cat in the Hat' is about peer review." - @ass_deans
Recent Comments