Because of this diversity, aggregation is likely to be misleading and to introduce error into the analysis. Second, unlike previous studies, this study focuses explicitly on supply reliability and the uncertainty over supply that confronts water users around the state at the time they make their important water use decisions. Researchers seek to measure this explicitly, both in the baseline situation and in climate change scenarios, because they believe that climate change in California is likely to affect water users primarily through its impact on supply reliability and uncertainty. This has not been analyzed in the existing work on climate change in California. In this context, it is important to note the uneven temporal distribution of water supply and water use in California: roughly 80% of the state’s precipitation falls between October and March, but about three quarters of all the water use in California occurs in the spring and summer, between April and September. What happens—or does not happen—during that period is the key to whether the state’s economy is benefited or harmed by water supply that year. Moreover, many important decisions that determine water use during this period are made at the beginning of the period. Farmers decide which crops to plant in the early spring, around March or early April. Once they have made that decision, they are limited in the degree to which they can vary their use of water during the growing season—they can under‐water their crops,nft channel or even abandon them, if it subsequently happens that they receive less water than they had anticipated at the time of planting, but they cannot switch to a different crop nor is it practical for them to make a major change in irrigation technology during the growing season.
With urban water use the context is somewhat different, but there is still a critical window for decision around April in that, if urban water managers think there is a fair chance that they will experience some degree of water shortage during the coming warm season, they generally need to put out a call for voluntary conservation no later than the end of spring. This sets up a pattern of water demand in their service area over the summer that is likely to be, at best, only partially reversible if water supplies turn out to be more abundant than originally anticipated. For somewhat similar reasons, environmental water managers in California, too, face a key decision point around April: because of the time lags in securing water supplies and arranging for their transfer, if managers are to meet critical in‐stream needs during the warm season they will need to take action by the end of spring. For these reasons, much of the water use that occurs in California between April and September is likely to be determined by water agencies’ expectations, as of the beginning of this period, regarding the amount of water that will become available to them during the coming summer. Supply reliability needs to be assessed by reference to these expectations. Most of the existing hydrologic/economic models—both in California and elsewhere—deal with supply uncertainty by ignoring it. They represent water supply using the actual, historical monthly deliveries. This approach amounts to characterizing uncertainty by the ex post realization of the random variable, which effectively eliminates the uncertainty. However, as explained above, given the timing of water use decisions in California it is clearly the ex ante probability of obtaining water during the warm season , as assessed some time around March or April, that has the most powerful influence on water users’ decisions in California.
Furthermore, it is reasonable to expect that these decisions will typically exhibit a significant degree of risk aversion. The important implication is that water use decisions are likely to depend not just on the mean of the ex ante probability distribution of warm‐season water supply, but also on other parameters of the distribution, such as the semi‐variance or the tail probabilities. In order to develop a linkage between changes in supply reliability and consequent economic impacts, one has to characterize supply reliability in terms of relevant parameters of the ex ante probability distribution of warm‐season water supply. Given the observations above about the heterogeneity among water districts with regard to their water supply, these distributions generally need to be assessed for each district separately. Implementing this approach, with its novel focus on measuring supply reliability at the level of individual water districts, is a major challenge, because of the limitations in the data that are readily available in California. It is easy to obtain data on historical water deliveries for the two big projects and State Water Project and for groups of irrigation districts combined into depletion study areas . Obtaining historical delivery data for individual districts not served by the two projects is often difficult. Obtaining a representation of the likely expectations of district managers in the form of an ex ante probability distribution is a major research task that has not previously been undertaken in California. To deal with problems caused by the limited availability of data, this study is pursuing a flexible and iterative strategy that iterates between data collection and data analysis. In the first year of research, researchers started by collecting the most readily available data and then pushed on to conduct a preliminary analysis of these data—recognizing that, although the data are still incomplete, many methodological issues arise during the course of data analysis, and it is useful to start confronting them as early in the research as possible.
While conducting the preliminary data analysis, this study continues to expand the data and fill in the gaps. After a second round of data collection efforts, researchers will take a second crack at the data analysis, while still continuing with efforts to complete the data collection and with a view to a subsequent final data analysis. Thus, rather than working in sequence, researchers are conducting the various components of this analysis in parallel.Water supply uncertainty is assessed at the district level, because the source and cost of water, the reliability of water supply, and the available quantity of water supply all vary primarily across districts. Water supply and price variation exists within districts as well, but it is comparatively limited. To implement a district‐based research approach, researchers have been creating a database with information on these variables, which permits them to better assess the uncertainty of existing water supplies and to project changes in future supply uncertainty as the result of climate change. To measure water supply reliability at the district level, researchers collected data on several variables, including deliveries for project districts spanning 20 years, supply forecasts for project districts spanning 20 years, some water rights information, water source information, and electricity use data related to groundwater pumping. To measure the economic value of reliability they collected land value, water price, water transfers for many years,hydroponic nft and cropping by districts. The climate data collected includes PRISM data, showing 100‐year run minimum and maximum monthly temperatures and precipitation for four grid points surrounding each farm observation. The population dataset for this study covers each of 7049 Census tracts in California; the soil data is derived from the STATSGO soil survey. Researchers have a very detailed database of groundwater, constructed from more than 16,000 well observations. The surface water rights data includes information about project entitlements to Districts from the Central Valley and State Water Projects. Water rights information also came from the ACWA database. Water price data was obtained from ACWA and the Irrigation Water Rates Manual. This section illustrates some of the data that has been collected, with a series of maps and charts, including data about water source, surface water supply, water transfer, cropping and land values in agricultural districts in the Central Valley. These results clearly indicate that there are large differences among districts in the Central Valley with respect to water supply reliability, water sources, water rights, land values, and cropping patterns—all of which greatly affect agricultural water use in California. The analysis was done for different types of users according to their geographic location, their source of water and different water rights status. These different users were also aggregated into different levels. The first level considered the whole Central Valley system.The second level compared reliability measures for broad geographic categories of users: North of the Delta Project and non‐project users; State Water Project South of the Delta users; Central Valley Project SOD users; and East San Joaquin users. Figure 2‐2 shows a map of these broad categories of users.
The third and final level went within some of these groups to assess the reliability for more specific type of users.An example of the analysis done at this step was the comparison of the reliability among different types of CVP users SOD , and Refugee Contractors. The reliability curves are presented in a series of figures. These curves should be read first looking at a delivery target along the x axis, and then at the percent of time this target is equaled or exceeded on the x axis.Researchers collected information showing surface water deliveries to water districts across the Central Valley. The information gathered includes data showing surface deliveries from project sources , local water deliveries to non‐project districts, riparian withdrawals, and local canal company deliveries. The information was provided from a variety of sources, including the regional offices of the California Department of Water Resources , individual water district offices, consultants, and officials from the State Water Resources Control Board . The data from these sources were combined with cropping data to show total surface deliveries per cropped acre. The information is stored in a series of access databases, Excel spreadsheets, and geographical information systems databases. The information about surface water deliveries contained in the GIS database is illustrated for a subset of districts in the San Joaquin Valley . It is apparent that districts within the same county have widely different water supplies. In particular, districts in the eastern and northern portions of the San Joaquin Valley have relatively large surface water supplies, compared to other districts in the Valley.Flexible water markets can help efficiently distribute water across the state during periods of drought. Climate change in California will likely decrease agricultural water supplies and increase water demands. Therefore, the impacts of climate change will vary depending on the extent that participants adopt flexible market structures to reallocate water supplies and regulate groundwater storage. For example, a recent drought study focused on the Delta Mendota region, suggests that water shortage may lower gross state product $175 million without water markets and by only $20 million with water markets . Water markets between agricultural and urban areas are particularly valuable, helping as they do to prevent urban drought shortages. Currently, the overall volume of trading in the different water markets is relatively small. In the researchers’ database, the most water traded in a single year was about 800,000 acre‐feet.4 Most of this trading was between agricultural districts. Currently only a handful of water districts have sold water to urban areas, despite continued efforts by urban areas to buy water . Of course, agriculture to urban transfers are limited by the level of urban demand for water in the state. However, rapid urban growth suggests the demand for low cost agricultural water supplies exceeds the supply in many parts of the state for reasons that have yet to be worked out.Many more districts have sold water to other agricultural districts, but the sales of water in this case are virtually all short‐term sales. Researchers suspect these short‐term sales are more in the category of trades between neighbors rather than sales that would support an open and well functioning water market. Figure 2‐6 shows the frequency of California water transfers between 1985 and 2000. The plan is to analyze the economic and political factors that influence a district’s decision to sell water. This analysis will test a number of hypotheses raised in the technology adoption and market adoption literature. These include rigid short‐run farm production technologies, lack of sufficient conveyance infrastructure, transactions costs, third‐party impacts, and political and legal risk.