Abstract:
California has roughly the 5th largest economy in the world, in a global setting with a very dynamic population and economic structure, and a dry and highly-variable climate. It also has one of the world’s most productive agricultural sectors, responsible for about 3% of California’s economy and 80% of its human water use. Water demands are changing. Technology is improving. California’s already highly-variable climate is becoming warmer and still more variable.
This presentation reviews a long-term effort by university researchers to construct an integrated planning and policy optimization model of California’s extensive and complex water supply system, called CALVIN. The model integrates California’s surface and groundwater resources, and their variability, together with its diverse water demands, infrastructure, and water policies. As an optimization model, CALVIN suggests an integrated portfolio of water supply and demand management actions to minimize statewide water operation and scarcity costs, within environmental, capacity, and water balance constrains over a long hydrologic time-series.
Technical challenges include representing complex portfolios of diverse water management actions, data management and documentation, reconciling diverse sources of data, and improving modeling technology. Human challenges include an ever-changing group of modelers and a challenging and decentralized policy and modeling environment.
Recent advances in rebuilding as open-source Python-based software, on-line documented databases, and improved methods for representing imperfect hydrologic foresight and multiple near-optima are highlighted, as well are hydro-economic implications for long-term water and drought management.
Acknowledgement:
This work relies on the diligent and creative efforts of dozens of graduate students and professional researchers over the last 23 years, as well as the people of California and the US for funding this work.
Bio:
Jay Lund is a Distinguished Professor of Civil and Environmental Engineering and Co-Director of the Center for Watershed Sciences at the University of California – Davis. He specializes in the application of economic, engineering, and optimization ideas and methods to water and environmental problems, as a hydro-economic approach to water management and systems engineering. Jay is a member of the US National Academy of Engineering.
Summary:
CALVIN: Hydro-economic model
Water problems are complex/multi-objective
Water supplied primarily for economic purposes
A hydro-economic model combines water infrastructure with economic aspects of management decisions
Goal: find better ways to organize water management
Infrastructure sizing & operation
Identify portfolios of configurations for different goals, future scenarios
Improve public discussion about water management
Non-goal: predicting how a given decision will affect operations
There are predictive models of various portions of CA’s water system
Once CALVIN provides a near-optimal overall policy, can use those models to fine-tune operations
Coupled hydro-economic modeling
Water sources
Water demands (agricultural, urban, hydro power, environment)
Devices (pumps, wells)
Costs
Economic
Operating
Water portfolio
Supply: sources (snow, groundwater, reservoirs), conveyances, storage, treatment options
Demands: efficiency, crops sown, urban convenience, industrial, ecosystems/parks
Incentives: pricing, markets, taxes, education, insurance
Economics-driven engineering optimization model
Economic costs
SWAP: Model of the agricultural economic values of water
Estimated as marginal cost of x% water cut
Modeled in terms of crop decisions based on water availability
Constraints (mass, environment, capacities, policies)
Inter-tied California system
Prescribes monthly system operation over 82 year representative hydrology (records), can be modified for climate change
Scale:
155 major reservoirs
40 million people
3000 governments
Model solved using linear programming solver
Case Study: Paleo megadroughts
Average river flows were 40%-60% of what they are today for ~100 years at a time
Ran CALVIN on the 10 years that mimicked the paleo-droughts to create a 72 year synthetic drought
Significant water scarcity (25% agricultural and 6% urban reduction compared to current water requests)
Will probably mean dropping cotton, corn, rice and similar low value crops
Groundwater and surface water storage will drop as a result
Reservoirs will never fill regardless of how much storage you build
Cost to state: few $billion/year if drought is managed well
Salt water:
Water salinity: salty water less suitable for agriculture and more damaging to pipes
Desalination
Water prices ~$800 per acre foot, may go to $1200 per acre foot
Current price of desalination is $2500 per acre foot
Not cost-effective in CA but is cost-effective in Israel
Model currently assumes that we can foresee droughts (~5% improvement in costs), they are extending it to model limited foresight
Sequence of 1 year optimizations with cross-year carryover, rather than 82-year optimization
Weather modeling
Critical to understand water supply, evaporation
Short-term weather doesn’t matter much since reservoirs smooth supply and demand
HOBBES: Web-based database
Hydrology time series data
Graph of the water supply network
Visualization of the network
Limitations
Representation of the delta and its regulations
Groundwater model and its mass balance
Many limitations on the data they’re taking in to describe the system