As a land-grant institution, UC receives federal support; were UC to engage in work that directly supports or enhances marijuana production or profitability, it would be inviolation of federal law and risk losing federal support. As a result, UC research on California cannabis production has been limited and focused on the geography of production and its environmental impacts . These studies have documented the negative effects of production on waterways, natural habitats and wildlife. While such effects are not unique to cannabis agriculture perse, they do present a significant threat to environmental quality and sensitive species in the watersheds where cannabis is grown . Science-based best management practices to mitigate or avoid impacts have not been developed for cannabis. Because information on cannabis production practices is so limited, it is currently not possible to identify key points of intervention to address the potential negative impacts of production. As a first step toward understanding cannabis production practices, we developed a statewide survey on cultivation techniques, pest and disease management, water use, labor and regulatory compliance. The objective was to provide a starting point from which UC scientists could build research and extension programs that promote best management practices — which are allowable as long as their intended purpose is not to improve yields, quality or profitability. Survey results also establish a baseline for documenting changes in cultivation practices over time as legal cannabis production evolves in California. To characterize key aspects of cannabis production in California, we developed an anonymous online survey using Qualtrics survey software . A web-based survey that masked participants’ identity was determined to be the most suitable approach given that in-person interviews were limited by legal restrictions on UC researchers visiting cannabis farms,vertical racking system and mail or telephone surveys were constrained by the lack of any readily available mailing address or telephone contact information for most cannabis growers, who are understandably discrete with this information.

Survey questions focused on operational features , pest and water management, labor, farm revenue and grower demographics. Two draft surveys were reviewed by a subset of cannabis growers to improve the relevance of the questions and terminology. A consistent critique was that the survey was too long and asked for too much detail, taking up to 2 hours to complete, and that such a large time commitment would significantly reduce the response. We therefore made the survey more concise by eliminating or rephrasing many detailed questions across various aspects of cannabis production. The final survey included 37 questions: 12 openended and 25 structured . Structured questions presented either a list of answer choices or a text box to fill in with a number. Each list of answer choices included an “Other” option with a box for growers to enter text. Open-ended questions had a text entry box with no character limit. Condensing the survey to capture more respondents resulted in less detailed data, but the overall nature of the survey remained the same — a survey to broadly characterize multiple aspects of cannabis production in California. Data from the survey has supported and contextualized research by other scientists on specific aspects of cannabis production, such as water use , permitting , law enforcement , testing requirements , crop prices and perceptions of cannabis cultivation in the broader community . Recruitment of survey participants leveraged networks of California cannabis growers who had organized themselves for various economic and political purposes . These were a combination of county, regional and large statewide organizations, with many growers affiliating with multiple groups. We identified the organizations through online searches and social media and sent recruitment emails to their membership list-serves. The emails contained an explanation of the survey goals, a link to the survey website and a message from the grower organization that endorsed the survey and encouraged members to participate. The emails were sent in July 2018 to approximately 17,500 email addresses, although not all members of these organizations necessarily cultivated cannabis, and the organizations noted that their mailing lists somewhat overlapped the lists of other groups that we contacted.

For these reasons, the survey population was certainly less than 17,500 individual cannabis growers, but because we were not able to view mailing lists nor contact growers directly, and because there are no comprehensive surveys of the number of cannabis farms in California, we could not calculate a response rate or evaluate the representativeness of the sample. Respondents were given until Aug. 15, 2018, to complete the survey. All survey participants remained anonymous, and response data did not include any specific participant identifiers. In total, 101 surveys were either partially or fully completed. Responses to open-ended questions were coded before summary. Since incomplete surveys were included in this summary, the number of responses varied between questions. Each response was considered a unique grower and farm operation. As noted, survey response rate was difficult to quantify, and participants were self-selecting, which introduces bias. The survey data should be taken only as a starting point to guide more detailed evaluations of specific practices in the future, not as a basis for developing recommendations for production practices or policies.Survey respondents operated farms primarily in Humboldt , Mendocino and Nevada counties, but survey responses also came from Trinity , Santa Cruz , Sonoma , San Luis Obispo , Sacramento , Butte , Calaveras , Fresno , Los Angeles , San Diego , San Mateo and Siskiyou counties and Josephine County, Oregon . In line with California regulatory guidelines, small farms were defined as those of 10,000 sq ft or less, medium farms 10,001 to 22,000 sq ft and large farms 22,001 sq ft or more. Accordingly, 74% of farms were small, 16% were medium and 8% were large . For those growers who reported on their land use in 2013 , most farmed on land that was previously used entirely or in part for cannabis production . The other 22% indicated that the land was used in 2013 for agricultural crops, ranching, open space or “other” land uses.For this survey, we differentiated between outdoor , greenhouse and indoor farming . The most common ways to farm were all outdoors , combined outdoor and greenhouse and greenhouse only . This was followed by various combinations of greenhouse and indoor , greenhouse and other , outdoor and other , outdoor and indoor , all indoor and other .

When measured by total plants, farms with combined outdoor and greenhouse facilities were responsible for 41% of crop production, followed by outdoor and other , greenhouse only , outdoor only , greenhouse and other , outdoor and indoor , greenhouse and indoor and other . A majority of survey respondents grew their cannabis crop in raised beds , native soil and/or grow bags , followed by hydroponic systems and plastic pots . The average number of plants grown in outdoor farms was 166 , in greenhouses, 582 and indoors, 383 . When adjusted for total cropping area, this equates to 0.05 plant per sq ft for outdoor cultivation , 0.13 plant per sq ft for greenhouse cultivation and 0.64 plant per sq ft for indoor cultivation .Most growers reported groundwater as their primary water source for irrigation , with some growers reporting use of multiple water sources. Those using groundwater extracted 87% of annual volume between June and October. Of those storing water,indoor grow facility most stored exclusively well or spring water, though some stored municipal water or rainwater . Extraction to storage was greatest in summer but was relatively well distributed throughout the year. Many growers reported that adding storage was either cost prohibitive or limited by regulatory constraints. Half the respondents indicated that additional storage was not needed, 40% indicated that the high costs of building storage were limiting, and 5% reported there was insufficient water available and 5% that they were unable to obtain permits to store . Most growers reported using variable amounts of water across the growing season. Outdoor growers applied, on average, 5.5 gal per day per plant in August and 5.1 gal per day per plant in September. Greenhouse growers applied an average of 2.5 gal per day per plant in August and 2.8 gal per day per plant in September . When standardized by area, application rates were very similar between cultivation types . In our survey, growers reported using low maximum pumping rates : 53% indicated rates ranging 1 to 50 gal per minute, 7% did not know their pumping rate and the remaining 40%, who used groundwater or municipal water sources, indicated that this question did not apply to them. Growers reported 14 different arthropods, 13 diseases and nine vertebrates that had negative impacts on cannabis production . The most frequent arthropod pest was mites , followed by thrips , aphids and unknown larvae . The most common vertebrate pests were gophers, mice and rats , followed by deer and wild boars . Powdery mildew was by far the most commonly reported disease , followed by other fungal diseases such as molds and rots .

While these findings are in line with cannabis pests and diseases reported by others , survey data are self-reported data and grower identification of pests and diseases may not be entirely accurate. For instance, the complex of mites reported included russet mites, spider mites, broad mites and red mites. Growers were likely referring to hemp russet mite , two-spotted spider mite , broad mite and Carmine spider mite , respectively, but this remains unclear because there are many species of mite commonly referred to as russet mite, spider mite and red mite . This similarly applies to aphids, thrips, larvae, mildew, rots and molds. Accurate species identification of these pests and diseases will remain uncertain until they can be more systematically collected and identified by UC academics or other scientists. The most common approach to pest and disease control was to apply some type of solution or chemical to the crop , followed by augmentation of natural enemies and various cultural practices .Products specifically used for control of arthropod pests included azadirachtin , soap solution , pyrethrins and Bacillus thuringiensis . Many respondents indicated that certain products were effective against both pests and diseases, for instance microbial pesticides , oils and compost tea . Sulfur was the most commonly applied product specifically used for disease control. In addition, 29% of respondents claimed to use certified organic products for pest and disease management but did not name any product specifically. Finally, 2% of respondents reported that they did not spray for pests and diseases at all. Augmentation of natural enemies involved the introduction of predatory mites , lady beetles , predatory nematodes and other unnamed beneficial insects . Cultural practices included removal of infested plant material , insect trapping , intercropping , use of diatomaceous earth and selection of resistant cultivars .Our survey, although of limited sample size, is the first known survey of California cannabis growers and provided insights into common forms of cultivation, pest and disease management, water use and labor practices. Since completing this survey, we have discussed and/or presented the survey results with representatives from multiple cannabis grower organizations, and they confirmed that the data were generally in line with production trends. Evident in the survey results, however, was the need for more data on grower cultivation practices before best management practices or natural resource stewardship goals can be developed. All growers monitored crop health, and many reported using a preventative management strategy, but we have no information on treatment thresholds used or the efficacy of particular sprays on cannabis crops. Likewise, the details of species-level pest and disease identification, natural enemy augmentation and sanitation efforts remain unclear. Growers did not report using synthetic pesticides, which contrasts with findings from previous studies that documented a wide range of synthetic pesticide residues on cannabis . Product selection for cannabis is very limited due to a mixed regulatory environment that currently does not allow for the registration of any insecticide or fungicide for use specifically on cannabis , although growers are allowed to use products that are exempt from residue tolerance requirements, exempt from registration requirements or registered for a use that is broad enough to include cannabis .