Climate change and California's mountain snow pack how much will we lose?

Transcription

1 Water UCI Seminar 31 April, 2016 Irvine Climate change and California's mountain snow pack how much will we lose? Dan Cayan(1,2), 1. Scripps Ins4tu4on of Oceanography, UC San Diego 2. US Geological Survey thanks: David Pierce SIO Mary Tyree SIO Mike DeFnger USGS Phil Mote Oregon State Univ. Tapash Das CH2MHill U. Iris Stewart Santa Clara Observed Variability and Changes Projected Change-Pace and Detectability Sponsors: California Energy Commission NOAA RISA program DOE Calif Dept of Water Resources

2 Spring snowpack water varies greatly from year to year but on average contains about 70% of the water stored in California s Reservoirs Million Acre-Feet TOTAL WATER STORED (monthly) in 12 major & 148 other California Reservoirs with Statewide April 1 Snow-water contents (stacked atop each other) Calendar Years Total Water in Apr 1 Snowpack 148 other reservoirs extraordinarily low snowpack in 2014 and 2015 Total storage in: Shasta Oroville Trinity New Melones San Luis Don Pedro Exchequer Pine Flat Folsom Millerton Castaic Perris Figure 1 Monthly totals of water stored in (dark blue) 12 major reservoirs and (light blue) 148 other, mostly smaller reservoirs, stacked on top of each other, and (green bars) estimated statewide-total of water stored in April 1 snowpacks each year, January 1970 through April 2015

3 dry spell is characteristic of California s volatile precipitation climate Sacramento Delta Drainage Annual Precipitation coef of Variation 33% mean 19.7 inches std dev 6.6 inches California has a narrow seasonal window to generate its annual water supply. If atmospheric conditions are unfavorable during that period, a dry year results 2014 ~50% of long term average California Climate Tracker Western Regional Climate Center

4 last 4 years in California preponderance of warm daytime and night-time temperatures Tmax dark black line long term average Tmin

5 Extraordinarily low western U.S. Snow Pack Spring 2015 Spring 2010 a more normal snow year

6 Temperature is only moderately correlated with California Snowpack but lowest snow years tend to be quite warm Temperature DJF April 1st Snowpack (percent of normal) 200% 150% 100% 50% 0 SWE Apr

7 Annual Temperature global Western conterminous U.S. Freezing Level N American Freezing Level Tracker Western Regional Climate Center

8 Western Snowpack declines continue TRENDS in! April 1 snow-water content at! western snow courses! DIFFERENCES in! April 1 snow-water content at! western snow courses! s minus (past 61 yr) Mike Dettinger, after Phil Mote et al

9 Snow losses (Apr 1) have occurred in lower (warmer) elevations as shown directly from snow course observations (blue) and VIC model reanalysis (red) by Phil Mote and colleagues (2005) 9

10 994 JOURNAL OF HYDROMETEOROLOGY 996 VOLUME 12 JOURNAL OF HYDROMETEOROLOGY 996 Advance in snowmelt streamflow by few-several days VOLUME 12 JOURNAL OF HYDROM change in center of timing shift toward earlier flows in snow-dominated streams shown by changes in monthly fraction of annual flow higher winter/spring flows Lower spring/summer flows FIG. 5. Trends in fraction of annual streamflow for the period, categorized by SDCs. (1 3). FIG. 3. Trends in CT for each SDC. Trend values are given in days over the 61-yr period. Iris T. andof Edzer Pebesma, 2011: FHolger IG. 5.Fritze, Trends instewart, fraction annual streamflow for the 19 Shifts in Western North American Snowmelt Runoff. J. Hydrometeor, 12, ~ In lines with change in perslopeobtained after from a particular time Y. magnitude of the CT shifts and gauge elevation are shown category shifts. Sufficiently largeachanges in the the Parameter-elevation Regressions on centage of snowmelt runoff pulses caused basins to re- Independent Slopes Model (PRISM), ~ a high-quality, by Yi 2 Y, addition to y, we define a secondary predictor2.58 in Fig. 3. Many of the higher-elevation, snowmelt~ In ceive different classifications for thei first and second topographically sensitive, climate in slope after a particular time Y. lines minutes with (4akm) change mag set all the negative values of this vector to zero, and call it dominated streams are now flowing earlier by 6 12 subperiod and be identified as having experienced a data grid that uses a combined statistical and geographic ~ days

11 Annual Temperature global Western conterminous U.S. Freezing Level N American Freezing Level Tracker Western Regional Climate Center

12 To estimate water balance, Including snow water we use the Variable Infiltration Capacity (VIC) model,a land surface hydrologic water/energy accounting model. VIC is run offline from GCMs or RCMs, using downscaled precipitation,temperature and winds as input variables. Most of the results here are from VIC calculations run at 1/8 (12km), although more recently we are using a finer (1/16 ) downscaling and attendant VIC hydrology. The GCMs, downscaling and VIC simulations cover They have been run using 2 scenarios of future GHG concentration, SRES B1 and A2 or RCP 4.5 and 8.5. Liang, X., Lettenmaier, D.P., Wood, E.P., Burges, S.J., A simple hydrologically based model of land surface water and energy fluxes for GSMs. J. Geophys. Res 99 (D7),

13 We considered 32 simulations 16 AR4 GCM s 16 A2 and 16B1 BCSD downscaled to 12 km Map depicts elevation >800m Sierra Nevada+ high terrain Hydrology translated using VIC (Variable Infiltration Capacity) driven by downscaled precipitation and temperature

14 VIC modeled vs snow course (35 sites) observed April 1 SWE correla4on=0.97 observed modeled VIC snow accumulation agrees closely with observed snow course observations variability of Apr 1 Snow Water Equivalent (SWE), VIC modeled vs. observations avg of 35 snow courses

15 end of 21 st Century median ΔT ~3 C ΔT= 2 C historical mean T 2.5 C

16 change in snow water projected for Sierra Nevada+ is substantial 16 GCMs, A2 and B1 emissions scenarios declining Apr 1 SWE: 2050 median SWE ~ 2/3 historical median 2100 median SWE ~ 1/3 historical median

17 90 th %ile Precipitation precipita4on considerable variability but li]le overall change Snow Water Equivalent 10 th %ile but spring snow declines steadily

18 precipitation fluctuations drive year-to-year SWE fluctuations but progressive climate warming diminishes California spring SWE Precipitation (ONDJFM) fluctuations: ±10% Δprecip! ±16% ΔSWE but precip fluctuations are not trending Temperature (ONDJFM) fluctuations: +1 C ΔT! -23% ΔSWE and ttemperature change is trending strongly Precipita4on-driven SWE change--do]ed Temperature-driven SWE change--solid

19 Loss of California Spring Snowpack from 21 st Century warming Under this scenario, California loses half of its spring (April 1) snow pack due to climate warming. Less snow, more rain, par4cularly at lower eleva4ons. The result is earlier run-off, more floods, Less stored water. This simulation by Noah Knowles is guided by temperature changes from PCM s Business-as-usual coupled climate simulation. (this is a low-middle of the road emissions and warming scenario) Knowles, N., and D.R. Cayan, 2002: Poten4al effects of global warming on the Sacramento/San Joaquin watershed and the San Francisco estuary. Geophysical Research LeDers, 29(18), 1891.

20 regional snow and hydrology a sensitive index of climate variation and change Douglas Alden Scripps Institution of Oceanography Installing met station Lee Vining, CA

21 As climate warms in this RCP 8.5 ensemble averaged over all all eight regions, all snow related measures change, but those that are most strongly dominated by temperature change more rapidly. David Pierce, D. Cayan : The uneven response of different snow measures to human-induced climate warming. Journal of Climate

22 Projected change temperature and precipita4on 31 Global Climate Models RCP 8.5 Los Angeles region S C a wetter a warmer future but uncertain changes in precipita4on warmer

23 Projected change temperature and precipita4on 31 Global Climate Models RCP 8.5 Los Angeles region wetter a warmer future but uncertain changes in precipita4on warmer

24 Projected change temperature and precipita4on 31 Global Climate Models RCP 8.5 Los Angeles region wetter a warmer future but uncertain changes in precipita4on warmer

25 Median Apr 1 SWE 11.9cm over 21 st Century occurs a marked decline of chances of reaching or exceeding historical median Snow Water Equivalent Sierra Nevada+ 10% 40% 10 th % Apr 1 SWE 3.6cm.and, chances of historical 10 th percentile or less SWE increases greatly

26 According to VIC--- end-of-century April 1 SWE losses in California occur mostly because of increased rainfall (and decreased snowfall), but are compounded by Increased snowmelt David Pierce, D. Cayan : The uneven response of different snow measures to human-induced climate warming. Journal of Climate

27 Projections indicate Increasing Flood Flows 50 year return period annual maximum 3-day floods for both Northern (shown below) and Southern Sierra Nevada from VIC simulations Northern Sierra Nevada % change in 50-yrs floods (2% exceedence) 5% significance 5% significance SRES A2 SRES B1 some, not all, of flood flow increase can be attributed to change in rain/snow and snowmelt Percentage change in 50-yrs (2% exceedence) flood discharge. The % change is computed with respect to the 50-yrs flood computed over the period The second, third and next on points are computed for the period with 10-yrs sliding period (e.g., the second point represents change of the flood magnitude computed for the period with respect to flood discharge computed in the period ). The plot shows 25th, 50th and 75th percentiles from sixteen climate models from SRESA2 (red color curves) and SRESB1 (blue color curves) simulations for Northern Sierra Nevada (left) and Southern Sierra Nevada (right). In the plot, black color curves show the % change in 50-yrs flood with respect to historical flood ( ) from VIC simulation as simulated by historical observed meteorologies (Hamlet and Lettenmaier, 2005). 5% significant level computed using a long control simulation (750-yrs) from PCM1 is shown as dotted gray lines. Numbers in the x-axis indicate the middle year of each 49-yrs time window used to compute the flood. Das, T., M.D. Dettinger, a D.R. Cayan and H.G. Hidalgo, 2011: Potential increase in floods in California's Sierra Nevada under future climate projections. Climatic Change

28 early 21st Drier Summer Landscapes increased warming and diminished snow causes successively greater soil drying throughout 21 st Century (this picture could change somewhat under more recent CMIP5 simulations) middle 21st late 21st Cayan et al. Ch 6 Southwest Climate Assessment

29 since 1985 the number of large wildfires in western U.S. increased four-fold relative to previous 15 years, mostly forest fires, not shrubland fires Anthony Westerling et al. Science August

30 large summer wildfires occur more ooen in years with early/warm springs Late Snowmelt Years Early Snowmelt Years , NPS, USFS & BIA Fires over 1000 acres Area burned is propor4onal to size of red dots The warming and earlier springs during last few decades have extended and intensified the fire season in mid-eleva4on forests Tony Westerling et al Science 2006

31 Summary Variability of seasonal snowpack in western U.S. will con4nue to be strongly Influenced by amount of winter/spring precipita4on, but warmer temperatures will play an increasing role. In California, VIC hydrological simula4ons exhibit loss of aggregate spring Snowpack that equates to- 23% Apr 1 SWE per +1 C of warming. Snow and snowmelt hydrology is already changing: Less snow, more rain Diminished spring snow pack in mid- and low eleva4ons Earlier run-off Temperature related measures (like snow accumula4on) have much stronger long term change signal to shorter term variability noise than precipita4on measures and thus are more easily detected These trends toward snow reduc4on will con4nue as climate warms. Consequences are many, but include Higher floods Ecosystem impacts Increased wildfire vulnerability Poten4ally, less stored water

32

33 Climate Variability and Change and California Water Lahontan Water Board Barstow, CA November 13, 2014 Dan Cayan Scripps Institution of Oceanography and USGS (with Mike Dettinger, David Pierce, Suraj Polade, Mary Tyree, Alexander Gershunov) key points Broad impacts of climate change will increase in future decades Average precipitation may not change much, but volatile precipitation climate in present and future produces periodic dry (and wet) spells. Importance of the presence or absence of few very large storms

34 virtually all climate simulations project warming, but with a wide envelope of temperature change July Temperature change Sacramento +3 C CMIP5 GCMs project C summer warming by 2060, under mid and high RCPs 14 GCMs X 3 RCP Emissions Scenarios IPCC 5th Assessment (CMIP5) models Some important ques4ons: Which emissions pathway will we take? How much summer amplificaion of warming? How will temperature change in near term?

35

36 California and much of western region has been more-or-less dry since 1999 observed precipitation departure (% of average), thru Inot including the present water year) % departure from average

37 California and much of western region has been more-or-less dry since 1999 observed precipitation departure (% of average), thru % departure from average

38 Loss of water during the recent drought has caused the earth s crust to rise GPS displacements rela4ve to average 4mm upward displacement (10cm water) average over western US maximum loss of 50cm has occurred in Sierra Nevada Ongoing drought-induced upli; in the western United States by A. A. Borsa (1), D. C. Agnew(1) and D. R. Cayan(1,2) published in the online version of Science on August 20, Scripps Ins4tu4on of Oceanography, UC San Diego 2 US Geological Survey La Jolla, CA Figure 2. Maps of ver4cal GPS displacements on March 1 of 2011 through Uplio is indicated by yellow-red colors and subsidence by shades of blue. Gray region is where sta4ons were excluded in the Central Valley of California due to strong agricultural pumping signal. GPS es4mate by Borsa et al. in present study showing deficit of ~240km**3 water in March 2014 rela4ve to baseline over western U.S., west of 109 W. This is in general agreement with GRACE satellite observa4ons, S. Castle et al 2014 GRL found ~65km**3 water water loss in Colorado Basin during

39 Summary Points California s climate is prone to year-to-year and longer term varia4on in precipita4on drought is an expected part of our climate present and future. California dry spells ooen build up over mul4ple years,. A more/less dry pa]ern has been in place since A variety of climate patterns may produce drought--there is not a unique atmospheric drought-circulation pattern. The absence of a few very large storms is ooen a key driver of dry years. And large storms are frequently involved in bus4ng drought. Climate change will broadly affect California hydroclimate and impact sectors and systems across-the-board. Expected impacts of climate change: longer warm season, loss of spring snow pack, increased wildfire threat, more winter floods. Climate changes in annual precipitation is not so clear in California. However, climate change may shift precipitation characteristics fewer overall wet days but more intense heavy events. Climate change projections warmer, fewer overall wet days but more intense heavy events. Implica4ons: Less snow, more rain Earlier run-off from traditionally snow-fed mountain watersheds Higher floods Potentially, less stored water Water quality implications: warmer surface water, warmer dry spells, occasional higher runoff, greater sediment influx from high runoff events and burned landscape

40 Now consider VIC RCP 8.5 simulations over 8 regions in western U.S. that have historically been snow-laden David Pierce, D. Cayan : The uneven response of different snow measures to human-induced climate warming. Journal of Climate

41 Annual Temperature global Western conterminous U.S. Freezing Level N American Freezing Level Tracker Western Regional Climate Center

42 a few large storms (or their absence) account for a disproportionate amount of California s precipitation variability 5yr smoother 95%ile LARGE STORM CONTRIBUTION Mike Dettinger Dettinger and Cayan Drought and the Delta A Matter of Extremes accepted, San Francisco Estuary and Watershed Science, April 2014

43

44 Mike Dettinger

45 April 1st Snowpack (percent of normal) 200% 150% 100% 50%

46 Total precip. (mm) Sierra Nevada +145 (50.6%) +63 (62.3%) 169 ( 30.3%) Melts SWE Rain VIC Sierra Nevada simula4ons of future decades has reduced Apr 1 SWE in response to: increased rain increased snowmelt Snow

47 (a) FRACTION OF ANNUAL PRECIPITATION FALLING IN THE DAILY TEMPERATURE RANGE: -2C < Tavg < 0C FRACTION (b) Fraction of Western Precipitation Falling on Days with Tmin < 0C under Various Imposed Warmings (c) Areas in the Western US subject to Rain-Snow Conversions of Various Percentages Fraction of Total Precipitation as Snow vs Rain Fraction of West with > this Precip-Form Change % increase in rain fraction 20% change 30% change 40% change 50% change Warming, in degrees C Warming, in degrees C

48