Step 1: Naturalized streamflow Reference runoff for each grid cell Step 2: VIC parameter calibration to reference runoff fields Results: calibration improves streamflow Simple reservoir implementation for flow simulation • Simulate 25 major reservoirs in the basin • At each reservoir, simulate reservoir operation based on historical guide curve, reservoir capacity, minimum release and maximum release • Add Δflow (= outflow – inflow) at each reservoir to all downstream grid cells with time lag (assume linear superposition) Results: the simple reservoir model captures main features of flow regulation • Weaker seasonal cycle; high flows are lowered by regulation Climate change is expected to alter streamflow and stream temperatures in the Southeast U.S. These climate-induced hydrological changes may affect human water use and in turn the reliability of power systems (either for cooling or for hydropower generation) as well as crop production Water, energy and food form an interdependent system The UW team uses the Tennessee River basin as an initial case study: • Implement a hydrologic modeling framework in the Tennessee River basin • Evaluate model ability to simulate water availability (both streamflow and stream temperature) • Demonstrate future climate impacts on water availability Future Climate Impacts on Streamflow and Stream Temperature in the Tennessee River Basin Yixin Mao 1 , Tian Zhou 1,2 , John Yearsley 1 and Bart Nijssen 1 1 Department of Civil and Environmental Engineering, University of Washington, Seattle, WA 2 Now at: Pacific Northwest National Laboratory, Richland, WA Modeling framework Background and Objectives Data and Methodology Stream Temperature Results - Naturalized Steps 1, 2 and 4: Simulate naturalized streamflow and stream temperature (i.e., unregulated) Steps 1-4: Simulate regulated streamflow and stream temperature VIC Hydrologic Model Calibration Reservoir Implementation – Initial Stage RBM stream temperature model [Yearsley, 2009 & 2012] • One-dimensional, time-dependent equation for thermal energy transfer • Semi-Lagrangian particle tracking numerical scheme Results: simulated naturalized stream temperature matches USGS observation • Main temporal patterns captured on big rivers • Some anthropogenic features in observed stream temperature on smaller tributaries are not captured by simulated naturalized temperature (not shown); need to consider reservoir effect in temperature modeling (work in progress) Future climate projection (right plot) • Projections from 5 Global Climate Models (GCMs) are used as demonstration (highlighted) • Light-colored crosses: 36 GCMs, 1 to 10 ensembles each Work in progress: improving reservoir modeling; implementing reservoir regulation in stream temperature modeling; improving stream temperature model parameterization More GCMs will be involved in future impacts analysis This modeling framework will be coupled with power system modeling and crop modeling This case study for the Tennessee River basin will be scaled up to the Southeast U.S. Discussion Arthur Ramsey and Carrie Williamson, TVA; Yifan Cheng, University of Washington; Domenico Amodeo, George Washington University This work is funded in part by the National Science Foundation under grant 1440852 to the University of Washington Acknowledgements The overarching goal of this study is to evaluate potential climate impacts on water-energy-food nexus over the southeastern U.S. (collaborators: University of Washington, Carnegie Mellon University, George Washington University and the University of Missouri) Step 3 – Reservoir Model Step 1 – Hydrologic Model (VIC) (Variable Infiltration Capacity model) Precipitation, temperature & wind speed at each grid cell Energy fluxes & runoff at each grid cell Step 2 – Routing Model (RVIC) Naturalized streamflow Step 4 – Stream Temperature Model (RBM) Stream temperature Regulated streamflow Study domain • Tennessee River basin • 1/8 o latitude-longitude grid cells (~12km) Study period • Historical: 1949-2010 • Projected: 2006-2099 Control period: 1950-2005 • Time step: daily Climate data • Historical: 1/8 o gridded meteorological product, 1949-2010 [Maurer et al., 2002] • Control and projected: 1/8 o downscaled CMIP5 climate projections [Reclamation, 2013] Tennessee River Basin Daily naturalized flow Reference runoff for each upstream grid cell • Inverse routing method developed at Princeton University [Pan and Wood, 2013] • Flow information needed: naturalized flow (i.e., as if no human regulation) at some stream locations; flow network and travel time • By inversely routing the naturalized flow to upstream, we disaggregate the flow to all its upstream grid cells (i.e., reference runoff for each grid cell) • Naturalized flow data: 21 dam locations, weekly pass- through data from Tennessee Valley Authority (TVA), downscaled to daily VIC hydrologic model Meteorological forcing Soil & vegetation parameters Output runoff SCE auto- calibration Reference runoff compare • For each grid cell, compare reference runoff with model output; adjust model parameters and run model again until reaching optimized output runoff • Shuffled Complex Evolution (SCE) autocalibration method [Duan et al., 1993] • Objective function: Kling-Gupta Efficiency (KGE) of runoff for each grid cell [Gupta et al., 2009] • Calibration period: water years 1961-1970 Validation period: water years 1971-2010 • Calibration improves streamflow seasonality , time series and flow duration curve (the latter two not shown) over most of the basin VIC model structure RBM model structure Daily stream temperature, Tennessee River at South Pittsburgh Results: future hydrologic projection (lower plots) • Future streamflow: slightly wetter in winter; no significant change in summer • Future stream temperature: warmer in all seasons Wetter Warmer Climate change – basin average Water years 1970-1999 to 2040-2069 Highlighted: access1-0 bcc-csm1-1-m canesm2 ccsm4 cesm1-bgc Future hydrologic projections, Tennessee River at South Pittsburgh Streamflow Stream temperature Duan, Q. Y., V. K. Gupta and S. Sorooshian (1993): Shuffled complex evolution approach for effective and efficient global minimization. J Optimiz. Theory App., 76, 501-521. Gupta H. V., H. Kling, K. K. Yilmaz and G. F. Martinez (2009): Decomposition of the mean squared error and NSE performance criteria: Implications for improving hydrological modelling, J. Hydrol., 377, 80-91. Maurer, E.P., A.W. Wood, J.C. Adam, D.P. Lettenmaier , and B. Nijssen (2002): A long-term hydrologically-based data set of land surface fluxes and states for the conterminous United States, J. Climate, 15(22), 3237-3251. Pan, M. and E. F. Wood (2013): Inverse streamflow routing: Hydrol. Earth Syst. Sci., 17, 4577-4588. Reclamation (2013): 'Downscaled CMIP3 and CMIP5 climate and hydrology projections: Release of downscaled CMIP5 climate projections, comparison with preceding information, and summary of user needs', Denver, Colorado. 47pp. Yearsley J. (2009): A semi-Lagrangian water temperature model for advection-dominated river systems, Water Resour. Res., 45, W12405. Yearsley J. (2012): A grid-based approach for simulating stream temperature, Water Resour. Res., 48, W03506. References Tennessee River Basin Future Climate Impacts – Naturalized Mean monthly flow, WY1970-2010, Tennessee River at Wilson Dam Thousand CFS French Broad River downstream of Douglas Dam Mean monthly flow (left) and weekly flow duration curve (right), WY 1950-1973 Thousand CFS Thousand CFS Thousand CFS