Climate, Growth and Drought Threat to Colorado River Water Supply Balaji Rajagopalan Department of Civil, Environmental and Architectural Engineering And Cooperative Institute for Research in Environmental Sciences (CIRES) University of Colorado Boulder, CO Presentation to KOWACO February 3, 2009

Collaborators Kenneth Nowak - CEAE / CADSWES Edith Zagona - CADSWES James Prairie - USBR, Boulder Ben Harding - AMEC, Boulder Marty Hoerling - NOAA Joe Barsugli - CIRES/WWA/NOAA Brad Udall - CIRES/WWA/NOAA Andrea Ray - NOAA

A Water Resources Management Perspective Time HorizonTime Horizon Inter-decadal Hours Weather Climate Decision Analysis: Risk + Values Data: Historical, Paleo, Scale, Models Facility Planning – Reservoir, Treatment Plant Size Policy + Regulatory Framework – Flood Frequency, Water Rights, 7Q10 flow Operational Analysis – Reservoir Operation, Flood/Drought Preparation Emergency Management – Flood Warning, Drought Response

What Drives Year to Year Variability in regional Hydrology? (Floods, Droughts etc.) Hydroclimate Predictions – Scenario Generation (Nonlinear Time Series Tools, Watershed Modeling) Decision Support System (Evaluate decision strategies Under uncertainty) Modeling Framework Forecast Diagnosis Application

Resources (PhD thesis) Regonda, 2006 Prairie, 2006 Grantz, 2006 Stochastic Streamflow Simulation

Colorado River Basin Overview 7 States, 2 Nations Upper Basin: CO, UT, WY, NM Lower Basin: AZ, CA, NV Fastest Growing Part of the U.S. Over 1,450 miles in length Basin makes up about 8% of total U.S. lands Highly variable Natural Flow which averages 15 MAF 60 MAF of total storage 4x Annual Flow 50 MAF in Powell + Mead Irrigates 3.5 million acres Serves 30 million people Very Complicated Legal Environment Denver, Albuquerque, Phoenix, Tucson, Las Vegas, Los Angeles, San Diego all use CRB water DOI Reclamation Operates Mead/Powell Source:Reclamation 1 acre-foot = 325,000 gals, 1 maf = 325 * 10 9 gals 1 maf = 1.23 km 3 = 1.23*10 9 m 3

When Will Lake Mead Go Dry? Barnett & Pierce, Water Resources Research, 2008 Water Budget Analysis One 50 maf reservoir, increasing UB demands (13.5 in >14.1 Maf/yr in 2030, 15.1 maf /yr inflows, current starting contents Linear Climate Change Reduction in Flows w/ some natural variability Results With Linear 20% Reduction in mean flows Over 50 years 10% Chance Live Storage Gone by % Chance Live Storage Gone by % Chance Loss of Power by 2017 Is that so?

Colorado River Demand - Supply

Declining Lakes Mead and Powell 75 Foot Drop (Max 140) 10.5 maf lost Current: ~56%, 14.5 maf 120 Foot drop 13 maf lost Current: ~48%, 12 maf 5 Years of 10 maf/yr 66% of average flows Worst drought in historic record

New York Times Sunday Magazine, October 21, 2007

Dropping Lake Mead ~1999 ~2004 Lake Powell – June 29, 2002 Lake Powell – December 23, 2003 Lake Mead’s Delta Circa 1999 Source: USGS, Reclamation

Colorado River at Lees Ferry, AZ Recent conditions in the Colorado River Basin Paleo Context Below normal flows into Lake Powell %, 59%, 25%, 51%, 51%, respectively 2002 at 25% lowest inflow recorded since completion of Glen Canyon Dam Some relief in % of normal inflows Not in 2006 ! 73% of normal inflows 2007 at 68% of Normal inflows 2008 at 111% of Normal inflows 5 year running average

observed record Woodhouse et al Stockton and Jacoby, 1976 Hirschboeck and Meko, 2005 Hildalgo et al. 2002

Past Flow Summary Paleo reconstructions indicate 20 th century one of the most wettest Long dry spells are not uncommon 20-25% changes in the mean flow Significant interannual/interdecadal variability Rich variety of wet/dry spell sequences All the reconstructions agree greatly on the ‘state’ (wet or dry) information How will the future differ? More important, What is the water supply risk under changing climate?

Future Climate

The Fundamental Problem with Climate Change For Water Management All water resource planning based on the idea of “climate stationarity” – climate of the future will look like the climate of the past. Reservoir sizing Flood Control Curves System Yields Water Demands Urban Runoff Amounts This will be less and less true as we move forward. Existing Yields now not as certain given both supply and demand changes New water projects have an additional and new element of uncertainty. Stuff and m Science, February 1, 2008

IPCC 2007 AR4 Projections Wet get wetter and dry get drier… Southwest Likely to get drier

Models Precip and Temp Biases Models show consistent errors (biases) Western North America is too cold and too wet Weather models show biases, too Can be corrected

A Large Number of Studies Point to a Drying American Southwest Milly et al., 2005 Seager et a.l, 2007 IPCC WG1, IPCC WG2, 2007 National Academy Study, 2007 IPCC Water Report, 2008 CCSP SAP 4.3, 2008 “From 2040 to 2060, anticipated water flows from rainfall in much of the West are likely to approach a 20 percent decrease in the average from 1901 to 1970, and are likely to be much lower in places like the fast-growing Southwest.” ~ May 28, 2008, New York Times

Temperature Precipitation General CirculationModel Hypothetical Scenarios Regression Hydrology Models: NWSRFS VIC PRMS CRSS CRMM Reservoir storage Hydroelectric power UB Releases 1. Climate Change Data Source 2. Flow Generation Technique 3. Water Supply Operations Model OR Progression of Data and Models in studies about the influence of climate change on streamflows in the Colorado River Basin Streamflow Stuff and m

StudyClimate Change Technique (Scenario/GC M) Flow Generation Technique (Regression equation/Hydrologic model) Runoff ResultsOperations Model Used [results?] Notes Stockton and Boggess, 1979 ScenarioRegression: Langbein's 1949 US Historical Runoff- Temperature- Precipitation Relationships +2C and -10% Precip = ~ -33% reduction in Lees Ferry Flow Results are for the warmer/drier and warmer/wetter scenarios. Revelle and Waggoner, 1983 ScenarioRegression on Upper Basin Historical Temperature and Precipitation +2C and -10% Precip= -40% reduction in Lee Ferry Flow +2C only = -29% runoff, -10% Precip only = -11% runoff. Nash and Gleick, 1991 and 1993 Scenario and GCM NWSRFS Hydrology model runoff derived from 5 temperature & precipitation Scenarios and 3 GCMs using doubled CO2 equilibrium runs. +2C and -10% Precip = ~ -20% reduction in Lee Ferry Flow Used USBR CRSS Model for operations impacts. Many runoff results from different scenarios and sub- basins ranging from decreases of 33% to increases of 19%. Christensen et al., 2004 GCMUW VIC Hydrology model runoff derived from temperature & precipitation from NCAR GCM using Business as Usual Emissions. +2C and -3% Precip at 2100 = -17% reduction in total basin runoff Created and used operations model, CRMM. Used single GCM known not to be very temperature sensitive to CO2 increases. Hoerling and Eischeid, 2006 GCMRegression on PDSI developed from 18 AR4 GCMs and 42 runs using Business as Usual Emissions. +2.8C and ~0% Precip at = -45% reduction in Lee Fee Flow Christensen and Lettenmaier, 2006 GCMUW VIC Hydrology Model runoff using temperature & precipitation from 11 AR4 GCMs with 2 emissions scenarios. +4.4C and -2% Precip at = -11% reduction in total basin runoff Also used CRMM operations model. Other results available, increased winter precipitation buffers reduction in runoff.

2C to 6 C -40% to +30% Runoff changes in ~115% ~80% CRB Runoff From C&L Precipitation, Temperatures and Runoff in Triangle size proportional to runoff changes: Up = Increase Down = Decrease Green = Blue = Red =

Colorado River Climate Change Studies over the Years Early Studies – Scenarios, About 1980 Stockton and Boggess, 1979 Revelle and Waggoner, 1983* Mid Studies, First Global Climate Model Use, 1990s Nash and Gleick, 1991, 1993 McCabe and Wolock, 1999 (NAST) IPCC, 2001 More Recent Studies, Since 2004 Milly et al.,2005, “Global Patterns of trends in runoff” Christensen and Lettenmaier, 2004, 2006 Hoerling and Eischeid, 2006, “Past Peak Water?” Seager et al, 2007, “Imminent Transition to more arid climate state..” IPCC, 2007 (Regional Assessments) Barnett and Pierce, 2008, “When will Lake Mead Go Dry?” National Research Council Colorado River Report, 2007

Almost all the water is generated from a small region of the basin at very high altitude GCM projections for the high altitude regions are uncertain

Future Flow Summary Future projections of Climate/Hydrology in the basin based on current knowledge suggest Increase in temperature with less uncertainty Decrease in streamflow with large uncertainty Uncertain about the summer rainfall (which forms a reasonable amount of flow) Unreliable on the sequence of wet/dry (which is key for system risk/reliability) The best information that can be used is the projected mean flow

Streamflow Scenarios Conditioned on climate change projections Water Supply System Risk Estimation Water Supply Model Management + Demand growth alternatives System Risk Estimates For each year

Streamflow Simulation Paleo Observations Need to Combine

Colorado River System has enormous storage of approx 60MAF ~ 4 times the average annual flow - consequently, wet and dry sequences are crucial for system risk/reliability assessment Streamflow generation tool that can generate flow scenarios in the basin that are realistic in wet and dry spell sequences Magnitude Paleo reconstructions arePaleo reconstructions are Good at providing ‘state’ (wet or dry) informationGood at providing ‘state’ (wet or dry) information Poor with the magnitude informationPoor with the magnitude information Observations are reliable with the magnitudeObservations are reliable with the magnitude Need for combining all the available information Observed Annual average flow (15MAF) is used to define wet/dry state. Need to Combine Paleo and Observed flows for stochastic simulation

Generate flow conditionally (K-NN resampling of historical flow) Generate system state Nonhomogeneous Markov Chain Model on the observed & Paleo data Proposed Framework Prairie et al. (2008, WRR) Superimpose Climate Change trend (10% and 20%) Simulations Each 50-year long Natural Climate Variability Climate Change

h window = 2h +1 Discrete kernel function Source: Rajagopalan et al., 1996

Nonhomogenous Markov model with Kernel smoothing (Rajagopalan et al., 1996) Transition Probability (TP) for each year are obtained using a discrete Kernel Estimator h determined with LSCV 2 state, lag 1 model was chosen ‘wet (1)’ if flow above annual median of observed record; ‘dry (0)’ otherwise. AIC used for order selection (order 1 chosen)

Transition Probabilities

Re-sample a block of years (as desired for planning – say 50 year) Using the TP for each year generate a ‘state’ (S t ) Conditionally Re-sample a streamflow magnitude from the observed flow Identify K-nearest neighbors from the observations to the ‘feature vector’ ( S t, S t-1 and x t ) Re-sample one of the neighbor – i.e., one of the years, say year j Flow of year j+1 is the simulated flow, X t+1 Simulation Generate flow conditionally (K-NN resampling of historical flow)

Threshold (e.g., median) Drought Length Surplus Length time Drought Deficit Drought and Surplus Statistics Surplus volume flow

Drought/Surplus Statistics Paleo + Obs K-NN-1 bootstrap Of observed flow Red  Paleo stat Blue  Observed stat

Storage Statistics 60

System Risk Streamflow Simulation System Water Balance Model Management Alternatives (Reservoir Operation + Demand Growth)

UC CRSS stream gauges LC CRSS stream gauges Lees Ferry, AZ gauge Demarcates Upper and Lower Basin 90% of the entire basin flow passes through this gauge Well maintained gauge Annual Average flow is about 15MaF Sizeable flow occurs between Lake Powell and Mead ~ 750KaF/year Small but useful flow below Mead also comes in to the system ~ 250KaF/year

Water Balance Model Storage in any year is computed as: Storage = Previous Storage + Inflow - ET- Demand Upper and Lower Colorado Basin demand = 13.5 MAF/yr Total Active Storage in the system 60 MAF reservoir Initial storage of 30 MAF (i.e., current reservoir content) Inflow values are natural flows at Lee’s Ferry, AZ + Intervening flows between Powell and Mead and below Mead ET computed using Lake Area – Lake volume relationship and an average ET coefficient of Transmission Losses ~6% of Releases

Combined Area-volume Relationship ET Calculation ET coefficients/month (Max and Min) 0.5 and 0.16 at Powell 0.85 and 0.33 at Mead Average ET coefficient : ET = Area * Average coefficient * Storage (MaF) ET (MaF)

Management and Demand Growth Combinations A.The interim EIS operational policies employed with demand growing based on the upper basin depletion schedule. B with the demand fixed at the 2008 level ~ 13.5MaF C. Same as A but with larger delivery shortages D.Same as C but with a 50% reduced upper basin depletion schedule. E.Same as A with full initial storage. F.Same as A but post 2026 policy that establishes new shortage action thresholds and volumes. G.Demand fixed at 2008 level and post 2026 new shortage action. All the reservoir operation policies take effect from 2026 INTERIM EISINTERIM PLUSNEW THRESHOLD Res. Storage (%) Shortage (kaf) Res. Storage (%) Shortage (% of current demand) Res. Storage (%) Shortage (% of current demand)

Flow and Demand Trends applied to the simulations Red – demand trend 13.5MAF – 14.1MAF by 2030 Blue – mean flow trend 15MAF – 12MAF By MAF/year Under 20% - reduction

Flow trend with sample simulation 37.2% of simulations > 15MAF 22.3% of simulations > 17MAF 34.7% of simulations > 15MAF 18.8% of simulations > 17MAF

PDF of generated streamflows (boxplots) PDF of observed flow (red) AR-1NHMM

Natural Climate Variability

Climate Change – 20% reduction Climate Change – 10% reduction

When Will Lake Mead Go Dry? Water Resources Research, 2008 Water Budget Analysis One 50 maf reservoir, increasing UB demands (13.5 in >14.1 maf/yr in 2030, 15.1 maf /yr inflows, current starting contents Linear Climate Change Reduction in Flows w/ some natural variability Results With Linear 20% Reduction in mean flows Over 50 years 10% Chance Live Storage Gone by % Chance Live Storage Gone by % Chance Loss of Power by 2017 Problems 1.7 maf/year fixed evaporation plus bank storage Missing 850 kaf/yr inflows Forgotten / Ignored Issues System is on a knife-edge, even with existing flows Normal climate variability can push us over the edge without climate change

Probability of at least one drying – Barnett and Pierce (2008) Yellow – AR-1 (Barnett and Pierce, 2008) Red – Scenario I Green – Scenario II Blue – Scenario II

Probability of drying in a given year

Shortage Volume (MaF) Shortage Frequency Climate Change – 20% reduction Shortage Statistics

Shortage Volume (MaF) Shortage Frequency Climate Change – 10% reduction Shortage Statistics

Initial Demand – 12.7MaF Actual Average Consumption In the recent decade Sensitivity to Initial Demand Climate Change – 20% reduction Initial Demand – 13.5MaF

Initial Demand – 12.7MaF Actual Average Consumption In the recent decade Sensitivity to Initial Demand Climate Change – 10% reduction Initial Demand – 13.5MaF

Shortage Volume (MaF) Shortage Frequency Climate Change – 20% reduction Shortage Statistics

Initial Demand ~12.7MaF Sensitivity to Initial Demand Climate Change – 20% reduction Shortage Volume Initial Demand ~13.5MaF

Summary Interim Guidelines (EIS) are pretty robust Until these guidelines are as good as any in reducing risk Water supply risk (i.e., risk of drying) is small (< 5%) in the near term ~2026, for any climate variability (good news) Risk increases dramatically by about 7 times in the three decades thereafter (bad news) Risk increase is highly nonlinear There is flexibility in the system that can be exploited to mitigate risk. Considered alternatives provide ideas Smart operating policies and demand growth strategies need to be instilled Demand profiles are not rigid Delayed action can be too little too late Risk of various subsystems need to be assessed via the basin wide decision model (CRSS)