About 60% of respondents are aged 45 and above, and only about 17% are aged 25-35. While this might be the result of the age distribution in the major fields of occupation which are potential CIMIS users, it could also be that the current interface of CIMIS caters less to younger potential users who might seek the data elsewhere. About a quarter of respondents are women, and their share decreases at higher age groups. This probably reflects the changing labor force characteristics in CIMIS related professions over the past few decades. In terms of geographic location, most respondents report only one area of activity, with the San Joaquin Valley leading the count. Figure 2.1 below shows the shares of respondents in each region. Note that we allowed more than one response for location. We ask all respondents to rank each type of data, offered by CIMIS, according to the frequency they search for it. Figure 2.2 shows the breakdown of answers for each of the frequency choices. ET and precipitation are large shares of the “often” column. These shares decrease when moving in the “never” direction. On the other hand, one can observe an opposite trend for insolation , soil temperature, and relative humidity, which seem to be of less interest for respondents. Interestingly, air temperature seems less correlated with the frequency response, with response rate for “often” lower than “sometimes”. This could stem from the use of air temperature data: while irrigation requires using ET data often,flower harvest buckets air temperature data applications might require less frequent data pulls. Respondents seem satisfied with CIMIS services. About 72% of respondents reported using CIMIS at least occasionally.
The user types reporting “often” using CIMIS the most were Agriculture, followed by Golf Course Management and Water Districts. These user types are indeed likely to use CIMIS on a day to day basis, at least for some part of the year. In research and planning, on the other hand, one might use CIMIS to draw data only at an initial stage of a given task. In general terms, of the respondents who report using CIMIS to some extent, 77% say it is at least “moderately important” for their operations, with 22% reporting CIMIS as “extremely important”. The frequency of use and importance scores are positively correlated: frequent users also report high importance of CIMIS to their operations, which makes sense. The correlations between frequency and satisfaction, and between importance and satisfaction, seem less pronounced. There might be users who use CIMIS infrequently, perhaps because only a smaller part of their tasks involve the weather or climate information provided. Nevertheless, they seem satisfied with CIMIS services, as the satisfaction scores are relatively high. We also asked respondents to rank factors which hinder further use of CIMIS. Various answers were provided, given the results of initial surveys, and there was also room to specify other answers. Two main concerns exist, especially for users in agriculture: how reliable is the data and how to integrate it into existing systems and practices. Many growers and consultants in agriculture complement CIMIS with other data sources, such as soil moisture sensors, irrigation logs, and flow meters. Integrating information from multiple sources into decision making is a challenge faced by virtually all growers. 599 respondents, about a quarter of our survey, reported agriculture to be their primary business. Out of these, about half work on one farm, and the rest are consultants of sorts . 89% of respondents in agriculture report using CIMIS to some extent. Growers and consultants were asked to report their total acreage, selecting into pre-determined ranges. Summing these, we have 318,156 acres covered by growers, and almost 3 million acres covered by consultants. Many of the questions for growers and consultants were similar. One notable exception is regarding water use. The team decided not to ask growers how much water they use, fearing that growers would not want to share this information and would not finish the survey. However, consultants were asked how much water their clients use on average. This question was presented in the online survey as a slider bar, with a default at the lower bar , and an option to check a “Not applicable” box.
This box was not checked very often. Instead, it seems like many consultants who did not want to answer this questions left the slider bar at the default value of 0.5 AF/acre. This is a very low value for irrigated crops, and we assume that all these responses are basically non-answers. Ignoring them, the average reported water use is 2.96 AF/acre per year . This seems like a very reasonable distribution for water use in irrigated crops. Indeed, the USDA’s most recent Farm and Ranch Irrigation Survey reports a total of 7,543,928 irrigated acres in California, with a total of 23,488,939 AF of water applied, and a resulting average water use of 3.11 AF/acre, only a minor deviation of the reported average. Given the responses from agricultural consultants, we seem to have captured a very large portion of the drip irrigated acres in California. As a baseline for valuation, we will use the total 2013 drip irrigated acreage from the USDA survey, 2.8 million acres. While some growers might use CIMIS with gravitational or sprinkler systems as well, our understanding of the qualitative and quantitative responses is that CIMIS is mostly important for drip. We exclude the potential of CIMIS values on non-drip acreage, noting that our estimates would therefore be conservative in that sense.Growers in our survey reported an average CIMIS water saving effect of 24.2%. The reported saving rates seem to be distributed evenly among crops and grower acreage. The average water saving rates reported for consultants is 21.5%, a slightly lower rate than the growers, but this difference is not meaningful in an economic or statistical way. Figure 2.3 plots the distributions of reported savings by growers and consultants, with very similar means and medians. Regressing the reported savings rate on all user types, one cannot reject the null hypothesis that the mean water saving effect is equal between growers and consultants with 95% confidence . Since each group deals with different acreages, we interpret this result as lack of substantial economies of scale in water saving by CIMIS. The monetary cost of water saved can be viewed as savings on the intensive margin. One can also consider gains on an extensive margin. The water saved by use of CIMIS is likely to be used in agriculture as well. This means more acres can be grown with the same initial amount of water. The “full” economic value of the water saved by CIMIS in agriculture is the value of agricultural output that can be produced with it on acres not irrigated before. This following analysis includes the economic value of growing alone, without the added values of post-harvest and economic multiplier effects, and probably a safe lower bound. We do not, however, include a counter-factual productivity of non-irrigated land. In California,round flower buckets this is probably range land or acreage that is too sloped for traditional irrigation methods, and therefore of very low economic productivity.
With 1.92 million AF of water saved by CIMIS, and an average use of 2.5 AF/acre by growers , the savings from CIMIS can water an extra 768,000 acres in California. To put this in context, this is about double the total walnut acreage in 2016. Because of economic and technical constraints of water transport, it is hard to determine which crops would be planted in these extra acres. A conservative approximation assumes that the water saved by CIMIS serves to replicate the existing distribution of crops , taking the average value of productivity of an acre as the benchmark. The weighted average of grower revenue per acre in 2016 was $3,757 per acre1 . Multiplying by 768,000 acres, a conservative approximation for the contribution from CIMIS to California’s GDP via agriculture is about $2.89 billion. This number may appear very high, yet this calculation took various conservative assumptions:in the calculation of the water saved, in assuming the value of extra acreage, and in not including post-harvest added value and multiplier effects. To be even more conservative, let us assume that the elasticity of demand for the products grown on these extra acres is -2. That is, an increase of 1% in quantity would drop the price by about 0.5%. This is a reasonable estimate for elasticities of high value crops . The resulting extra income for growers is then about $1.44 billion dollars. CIMIS allows for more precise irrigation, which means not only saving water but also increasing yields: water application can be adjusted to the plant requirements, which might depend on the weather and growing phase. We ask growers and consultants how does CIMIS contribute in increasing yields, ranking from 1 to 5 . How should we quantify these ranked contributions? Taylor, Parker, and Zilberman mention average yield effects of drip irrigation, ranging between 5% and 25% increase in output. This extra yield effect is explained by allowing for more consistent soil humidity and the precision of the irrigation. This aspect of drip depends on weather and ET information, such as the one provided by CIMIS, to assess the water intake by plants and the appropriate amount of water required. We calculate an average yield effect of CIMIS by reconciling the respondent rankings with a portion of the yield effects from drip irrigation. For a lower estimate, rankings between 1 and 3 are attributed 0% yield effect, and the rankings of 4 and 5 get 5%. For a higher estimate, ranking of 1 gets 0% yield increase, ranking of 2 and 3 get 5% yield increase, and the rankings of 4 and 5 get a 10% yield increase. These percent yield effects are then averaged among the respondents. We aggregate growers and consultants with equal weights. 41% of respondents rank the importance of CIMIS for yield effects at 4-5. The low estimate for yield contribution of CIMIS results in 2% output increase, and the higher estimate at 5.9% increase. At a conservative estimate of per-acre income of $3,757 for growers, this represents an extra yearly income of $76 – $222 per acre. For the 2.8 million acres using drip irrigation, this would account for $213 – $622 million yearly from the contribution of CIMIS to yields. Assuming again the demand is elastic with a coefficient of -2, these estimates would halve to $107 – $311 million. Weather data can have quality effects on crops. For example, using ET data and drip irrigation, the quality of tomatoes can be increased by controlled irrigation deficit in proper timing. For tomatoes grown under a contract, reaching threshold quality levels raises the price received by the grower . Another potential use of weather data is in pest control, avoiding not only yield loss but quality degradation as well. These two examples reflect a relationship between quality and price that has long been established in the literature . To assess the contribution of CIMIS to quality, we also asked respondents to rank it from 1 to 5 . We assume that a score of 4-5 represents a quality index resulting in a price increase of 5%. About 45% of all respondents report a score of 4-5. The average price increase due to quality is therefore 2.2%, or $83 per acre. For 2.8 million acres, this results in a total increased revenue of $231 million. Note that this price increase is due to quality improvement, and thus not accompanied by a quantity reduction in our analysis. These are gains from water saving in parks, golf courses, and gardens. They were assessed as a small portion of the total gains from CIMIS in the 1996 report by Parker et al., totaling about $2.3 million . Our current estimate for these gains is much higher. The discrepancy from the 1996 report is due to several factors. First, we believe to have reached out to more respondents in this sector. Second, water prices in California have gone up substantially. Third, there might be more use of CIMIS and smart irrigation planning in the sector compared to 20 years ago.