Specific Objectives of the Project

The specific objectives of the project were three-fold. First, we enhanced the panel (longitudinal) dataset of agricultural land‐use decisions of 52 irrigation and water districts from the San Joaquin Valley for the years 2007-2015. We did so by adding the land-use decisions for approximately 200 irrigation, reclamation, water storage, and water districts. Second, we enhanced our model by modelling the extensive margin of agriculture. To do this, we will match land-use decisions within districts, and reservations at the field level year-on-year. Comparing the same field year-on-year will allow us to identify and isolate drought-induced deviations from normal cropping rotations. This approach is novel, and will allow for the first estimation of how drought impacts the extensive margin of agriculture. Third, we are in the process of refining the empirical model which will be used to estimate land‐use and crop changes under different water allocations. The results are preliminary, and the project is ongoing. We will update the data with crop acreage data each year. This model will be used to conduct agricultural policy analysis for the San Joaquin Valley via simulation exercises as soon as the

Summary of Results

This project estimates the effect of changes in surface water availability on the probability of a crop experiencing a given crop transition in the Central Valley of California using detailed data on crop production acreage in the Sacramento and San Joaquin Valley and Central Valley and State Water Project surface water allocations 2007-2015. We estimate the effect of water availability on transition probabilities using linear probability models (LPM) for each transition. We use a Difference-in-Differences estimator by comparing fields in irrigation districts against a plausible control group of fields located outside of irrigation districts. We exploit residual variation in water availability from administrative water allocation decisions to estimate the probability of switching from one crop to another. We find that Northern irrigation districts increase (decrease) acreage from field-to-rice, fallow-to-rice, in response to an increase (decrease) in surface water availability. Southern irrigation districts decrease (increase) acreage from fallow-to-fallow and tomatoes-to-fallow with an increase (decrease) in surface water availability. Additionally, Southern irrigation districts increase (decrease) Alfalfa-to-alfalfa acreage in response to an increase (decrease) in water availability. All of our findings are consistent with the intuition that farmers substitute away from more water intensive crops to less water intensive crops following a decrease in water availability as the marginal cost of water has increased making it more expensive to irrigate.

Specific Objectives of the Project

The objective of this paper was to empirically estimate surface water elasticities for 52 irrigation districts south of the San Joaquin and Sacramento River Delta by observing districts' land-use responses to the reductions in contract allocations from the Central Valley Project (CVP) and State Water Project (SWP) during drought years.

Project Report/Summary of Results

California has experienced four major droughts over the course of the last forty years, and is currently in the fifth year of a severe drought. Emergency drought policies enacted during times of drought restrict access to cheap surface water supplies, which requires farming operations in the agricultural sector to change their cropping decisions, and utilize more expensive groundwater supplies to adapt to the lack of surface water availability. Understanding how surface water availability affects the intensive and extensive margin for agriculture, in other words cropping intensity and crop choice. This paper estimates the impact of drought on the intensive margin for agriculture.

This paper finds that fallowing supply response elasticities range from 0 to -5.9. These reductions in irrigated land could for lack of access to water or due to the decision to sell water through transfers to other districts. Supply elasticities for deciduous trees are smaller than those for grain, truck, and pasture land-uses. Supply elasticities for grain acreage range from -7.6 to 1.0, which is evidence of a heterogeneous grain acreage response to drought-induced delivery reductions.

The model presented in this paper will be used for policy simulation to inform and guide policymakers on how to design a more refined allocation mechanism, and to obtain necessary reductions in surface water consumption while minimizing the welfare impacts to farmers and environmental impacts from future droughts. Additionally, this work has potential to inform water transfer negotiations between irrigation districts to minimize welfare losses during future droughts. The main contribution of this work is that it is driven by the actual decisions of the landowner responding to drought policy.

Specific Objectives of the Project
The objective of the project was three-fold. First, we constructed a novel panel data set tracking urban water retailers’ price schedules, water demands, non-price conservation activities and other demand factors. Second, we estimated the price elasticity of demand; the wide coverage of retailers across the state allowed for estimating spatial heterogeneity in price elasticities. Third, we used these estimates to evaluate the potential reliability benefits of alternative water supply investments.

Summary of Results
We estimate water supply reliability benefits by evaluating welfare losses from annual water supply disruptions that are mediated through urban water utilities with substantial fixed costs. Our analysis considers spatially heterogeneous water demand functions and encompasses two features not typically addressed in the literature on water shortages: (i) welfare changes are measured from baseline water rates that are not allocatively efficient; and (ii) welfare losses are compounded by the need for water utilities to recover fixed costs. We exploit panel data on California urban water suppliers over the period 1996-2009 to estimate regional price elasticities of residential water demand. We use our estimates to examine the welfare consequences of an annual water supply disruption in California. The estimated welfare loss ranges from an average of $1,500 per acre-foot under a 10% shortage to over $5,000 per acre-foot under a 30% shortage in water supply. The spatial distribution of welfare impacts exhibits substantial variation across regions, which highlights the need to consider decentralized impacts of water supply disruptions in the design of efficient rationing rules. These results inform benefits analyses of water supply infrastructure/ conveyance projects that serve both urban and agricultural users, most notably the Bay Delta Conservation Plan.

The project lead to the development of a stochastic dynamic programming algorithm for operation of a groundwater storage facility. The features of the program allow for spatial and temporal arbitrage. That is, the groundwater storage facility connects two parallel canals,and has the capacity to store significant amounts of water for several years. The model is solved using Approximate Dynamic Programming (Powell, 2007).

Publications, Working Papers, or Reports Resulting from the Project:

“The Value of Groundwater Storage with Spatial and Temporal Arbitrage.” With Newsha Ajami and David Zilberman. “Economic Value of Improved Water Forecasts.” With Newsha Ajami and George Hornberger. Water Resources Research (2009): revise and resubmit.