Extreme outliers were defined as those values outside 1.5 times the bounds of the interquartile range . Farms were not required to use water meters, and those without meters often estimated usage based on how frequently they filled and emptied small, temporary storage tanks otherwise used for gravity feed systems or nutrient mixing. The final dataset included 901 reports. Parcels of land where cannabis was cultivated — including multiple contiguous parcels under single ownership — constituted a site, and this is the scale on which reporting was conducted. The spatial extent of the cannabis program included all of California’s North Coast region ; however, only a subset of the counties in this region allow cannabis cultivation and therefore reports were only received from the following counties: Humboldt , Trinity , Mendocino and Sonoma . Because Sonoma County contributed relatively little data, we combined Sonoma County’s enrollments with those from Mendocino County when making county-level comparisons. The data used for this analysis included the source and amount of water that cultivators added to storage each month as well as the source and amount of water applied to plants each month. We did not analyze absolute water extraction rates. Rather, we used the amount of water extracted each month — whether water was added to storage or applied to plants directly from the source — to analyze seasonal variation in each water source’s share of total water extraction. Water sources included: surface , spring , rain , well ,grow tray delivery and municipal . The two external sources — delivery and municipal — were consolidated into a single category .
Because staff from the water quality control board were not able to corroborate the accuracy of reported data, enrollees may have classified water sources erroneously. A well placed in proximity to a stream, for example, might properly qualify as a diversion of surface water; so might rainwater catchment ponds or spring diversions that are hydrologically connected to a watercourse. We attempted to minimize these potential errors by restricting the dataset to reports prepared by professional consultants. As mentioned, enrollees were required to assess several standard conditions in their site reports, including water storage and use requirements. To encourage cultivators to join the regulated industry, and because many cultivation sites existed prior to adoption of the cannabis program, existing sites were not required to comply with standard conditions as a prerequisite for enrollment. Rather, cultivators unable to comply with the standards when they enrolled were required to indicate their lack of compliance and develop a plan for achieving compliance. Such sites were not held in violation of regulations, thus removing a potential motivation to falsely report site conditions. More than one-quarter of enrollees in the dataset reported noncompliance with the Water Storage and Use Standard Condition. To address question 1 — from which sources cannabis cultivators most frequently extract water across the North Coast region, and if extraction patterns differ across the region — we calculated the percentage of sites that reported use of each water source . We also calculated, for sites using each source, the percentage of sites that also used at least one other source category.
Directly applying water to plants and also placing water in storage did not constitute use of multiple extraction sources if the water was drawn from the same source category. Additionally, sites that used multiple inputs from the same category — for example, multiple wells — were not considered users of multiple sources, as this classification was reserved for extraction from multiple categories of sources. We performed all elements of our analysis for the entire dataset and for each county individually. To address question 2 — how reliance on each water source differed from one month to another — we divided each site’s monthly water extraction total by its annual extraction total to calculate the relative percentage of water extracted in each month, and performed similar calculations for each source category. The median amount of water extracted and interquartile range were calculated for each month — both for overall extractions and for each source category individually. To address question 3 — whether sites reporting compliance with the Water Storage and Use Standard Condition relied on different water sources than those reporting noncompliance — we compared water source extraction patterns for sites of both types. Specifically, we calculated for each compliance status the percentage of sites that extracted water from each source category and made comparisons accordingly; and did likewise for monthly extraction patterns, following procedures similar to those described in regard to question 2. The purpose of this comparison was strictly qualitative, and no inferential statistics were performed to determine statistically significant differences. Instead, this element of our analysis was performed for exploratory purposes, with the intention of identifying broad trends that warrant future attention.The most commonly reported water source was wells . Over half the sites reported at least some reliance on wells for their irrigation water. Surface water and springs were the next–most common sources.
Rainwater catchment and off-site water were the least commonly used water sources . Sites using wells and off-site sources were the least likely to use additional sources . In contrast, sites using rain catchment systems most frequently reported using an additional source category , followed by sites reporting use of spring diversions and surface diversions . To determine if the observed high frequency of well use was due to bias associated with examining only reports prepared by consultants, we reincorporated sites without consultants and reran the analysis on this dataset . Reported well use was slightly more common among sites not using consultants than among sites using consultants . Counties displayed notable variation in the frequency with which cannabis cultivators used particular water sources . Compared to all sites in the dataset, sites in Humboldt County relied more on surface water and spring diversions , with fewer relying on wells . The pattern was reversed in Trinity County,hydroponic trays with a high percentage of sites there reporting well use and relatively few using surface and spring diversions. A large number of sites in Trinity County were located in a single watershed known for a high concentration of similar cultivation practices, so we recalculated the percentages with these sites excluded. The resulting totals for Trinity County were closer to the overall results: wells , surface , spring , rain and off-site . Mendocino and Sonoma counties reported a similar pattern of extraction sources per site: wells , surface , spring , rain and off-site . Patterns of using multiple sources varied among counties. Sites in Humboldt County using well water extraction much more commonly used additional sources of water than did similar sites in Trinity and Mendocino/Sonoma counties. Use of additional sources was also more common among Humboldt County sites extracting surface water and spring water than among sites using surface and spring water in Trinity County and Mendocino/Sonoma counties . Wells were a prominent water source for cannabis cultivators during the summer months . Extraction from wells generally peaked in August and declined in off-season months. The pattern was reversed for rainwater use, with most extraction occurring in off-season months. Spring water use was generally even across the year, with slightly higher use during the growing season. Surface diversions occurred throughout the year, but declined late in the growing season, likely reflecting declining availability of surface water. The pattern exhibited in off-site water use closely matched that of well water; the former, however, was a less substantial source of water in general. There appeared to be differences in the extraction sources reported by compliant and non-compliant sites . Although nearly one-third of non-compliant sites used well extraction, this source was more than twice as frequently reported among compliant sites . In contrast, non-compliant sites reported surface diversion and spring diversion more commonly than did compliant sites .Use of additional alternative sources was lower for compliant sites with wells than for non-compliant sites with wells . The seasonal extraction patterns of compliant and non-compliant sites were generally similar , following the overall pattern discussed above.We found that well water is the most commonly reported source of extracted water for cannabis cultivation in the North Coast region of California. Furthermore, among the source categories, wells are least frequently supplemented with alternative sources. Spring and surface water diversions together are also important water sources, with seasonal patterns of use that are distinct from well water extraction.
Reported timing of well water extraction closely tracks the water demand patterns of plants, indicating that cultivators are applying well water directly to plants, rather than storing it. In contrast, the timing of extractions of spring water and surface water remains relatively consistent throughout the year, suggesting that water from these sources may be diverted to storage in the winter, reducing the need for extraction in the summer months. These seasonal extraction patterns and the relative predominance of each source may inform assessments of cannabis cultivation’s impacts on water availability. The use of well water for cannabis cultivation, in comparison to other water sources, presents both potential threats and benefits for in stream flow. In upper reaches of small watersheds, streams are dependent throughout the summer months on subsurface water flows from the landscape into the stream. Well water extraction may reduce cold water inputs — limiting stream flow or, in extreme conditions, dewatering stream channels . The extent to which use of subsurface water affects stream flow and water temperature depends on the degree to which well water sources are hydrologically connected to streams. When wells are shallower and closer to streams, and when soil conductivity is greater, subsurface water pumping is more likely to directly capture stream flow. However, if wells are less hydrologically connected to streams, the effects of extraction will be attenuated, resulting in smaller-magnitude and temporally lagged stream flow depletions. With sufficient groundwater recharge in wet months, well water extractions may affect stream flow less than surface water diversions, which were previously assumed to be cannabis cultivators’ predominant means of obtaining water in the region . Further analysis is necessary to understand the potential impacts of well use on stream flow depletion. Such an analysis would incorporate information on well locations and depths and would consider the underlying geology and soil properties at cultivation sites . Meanwhile, the prevalence and distribution of wells relative to other water sources are influenced by broader geospatial characteristics such as topography and precipitation patterns. Understanding these issues will also be important for assessing the threats and benefits associated with subsurface water extraction. Variation between counties in well extraction patterns demonstrates that, although subsurface water may be the most common source of water in North Coast cannabis cultivation, the availability of alternative sources may play an important role. Humboldt County watersheds included in this study consistently receive more average annual precipitation than do those in Trinity , Mendocino and Sonoma counties . This difference translates into more available surface and spring water in Humboldt County over the course of the growing season. The observation that fewer sites in Humboldt County report well use, compared to other counties in the study, suggests that if surface or spring water is available, cultivators are likely to use it. Conversely, the potential necessity of groundwater use in counties that receive less rainfall holds particular importance in consideration of emerging areas of industry growth throughout California. Further analysis is needed to understand how likely cultivators are to rely on wells if other sources of water are available to them. The winter preceding the 2017 growing season was the wettest on record. It is important to understand how cultivators may source their water during years in which summer water availability is not as abundant. These findings suggest that cultivators may utilize wells both as insurance against surface water scarcity in the summer drought months and as a means of achieving regulatory compliance. The observation that nearly one-third of non-compliant sites reported well extraction indicates that use of subsurface water may be a common means to avoid water scarcity in the late growing season.