 My name is Ted Grantham. I'm a Cooperative Extension Specialist in the Department of Environmental Science Policy and Management at UC Berkeley. I'm also one of the co-directors of the Cannabis Research Center at UC Berkeley. Today I'm going to be talking about cannabis water use in Northern California. And before I get started with this brief presentation, I'd like to acknowledge my co-author on this work, Chris Dillis, who led much of the analysis that I'll be presenting today. One of the major concerns over the environmental impacts of cannabis is how water use by cannabis is going to affect in-stream flows for sensitive ecological resources and species that we're trying to protect like salmon and other sensitive species. We know from enforcement efforts in the past that cannabis water users rely on in-stream diversion, such as depicted here in this slide. Here we see on the left a legal water diversion that's pumping water directly from a stream and this diesel generator is perched precariously above the stream bank. And in the slide to the right we see a stream that's been ponded in line with plastic to create a pond for diverting water. And these types of impacts really raise concerns over how cannabis water users are taking water from the environment and what this means for our sensitive species and resources. We also know that water storage is proliferating across the landscape in cannabis growing regions. We're seeing these external water storage tanks and upland storage systems, such as shown on the right slide here. But to date our information about cannabis water use has largely been anecdotal, obtained through enforcement efforts and site visits. And for the first time we're now able to answer some key questions with new data that's emerging from some of the state permitting programs. These key questions include where do cannabis farms actually source their water? Where are they obtaining water from the environment? How much water is actually being used? And we're really getting new insights on what the potential impacts to in-stream flows are from cannabis water use. So today I'm going to review the insights that we've gained on these really fundamental questions that to this date have not been easily answered owing to the historical illicit nature of this crop. And these data that we rely upon are self-reporting water use data that's emerged from the Cannabis Water Discharge Regulatory Program in the North Coast region of California. These data are from the first full cultivation season of the program. This was in the year 2017. So this is actually before most of the new state permitting regulations went into effect. The data that we were able to obtain from these reports include the size of the cultivation area, the type of the cultivation, whether that's outdoor, indoor, and hoop houses, for example. We also gained information on monthly volumes of water that are either input to storage or applied directly to plants. And finally we get some information about the storage types and capacity on site. So the results that I'll be showing today rely really on an analysis of about a thousand of these reports obtained for the 2017 growing season. We actually obtained a lot more reports than that, but we screened them for quality and ended up with this restricted set of 901 sites, mostly concentrated in Humboldt, Trinity, and Mendocino County with a few records in Sonoma. So our first key question, where do cannabis farms source their water? Now cannabis water users are largely located in these very rural remote watersheds and might be obtaining water from a variety of sources, including surface water sources such as streams and springs, which may sort of vary seasonally in the amount of water available. Or they may rely on more perennial sources like groundwater that could be extracted either from inland wells, such as shown on the right, or from wells right along the riparian or stream corridor, such as shown on the left. Now when we look at these reports, what the results have shown is that the vast majority of water users in the region, over 50%, nearly 60%, are relying on well water. And this was information that before these reports came out was not well understood. It was largely assumed that these water users were largely relying on surface water sources. So let me just take a minute to walk you through this graph. Basically what we see here is that in the gray bars, it's all of the records combined. And what we see is that overall nearly 60% of cannabis users in the region rely on wells. Somewhere around 20%, 15% to 20% of those users also rely on either surface water or spring water. Slightly lower fraction rely on rainwater catchments. And a very small number of users actually rely exclusively or in part on offsite water sources. And these could include municipal water deliveries or water deliveries from offsite sources or potentially direct tie-ins to municipal sources. So our key finding really from this initial analysis was that first, we observed a widespread use of subsurface water throughout the North Coast. But we also noticed that nearly 40% overall of users continue to rely on surface and spring water. And this is important because these types of water uses, water being taken from these sources are subject to new forbearance requirements that were established in 2019. And what I mean by that is that any cannabis cultivator who's interested in entering the legal permitted market that relies on these spring or surface water sources are now required to forbear or are prevented from taking any water during the summer dry season. So basically they have to provide enough storage on site to collect water in the winter to meet their irrigation needs or they need to drill a well. So I'll come back to that point a little bit later, but just wanted to highlight this finding here. And if you want to dig a little bit deeper into these specific results on water sourcing, I can direct you to this paper that was published in 2019 in California agriculture. So the next question, how much water is actually being used by cannabis? There has been some previous work trying to address this question and basically through interviews with grower associations, a number that was developed was this idea that plants are generally using about six gallons per plant per day over the growing season. And this plant-based estimate of water use has been applied in a few different contexts. And one was the study here shown on the right where they looked at cannabis farms in different watersheds in the North Coast region and used this plant-based estimate to generate basically a water budget to estimate total cannabis water demands in relation to available stream flow. However, there are a few limitations of this approach. First, we know that cannabis is grown in different ways, right? We have outdoor plants that tend to be quite large, such as we see in the upper right-hand corner. And we also have cannabis plants that are grown in these hoop houses or covered areas where the plant density is much higher and the plants themselves are much smaller. And so we really can't compare the plant-based use of these two cultivation types. Second, we can expect that there's going to be a seasonality of water demand. In current previous approaches, they've basically used gallons per plant estimate and applied it across the entire growing season from March until October. And we can expect that plant water demands are actually going to ramp up as they get bigger and as the season progresses and probably peak in the late summer. Finally, these plant-based approaches are limited in the way that we can interpret the data in relation to stream flow because this really assumes that the water that's being applied to plants are being taken on demand from the environment rather than using water storage. So a really important point here is that there's a difference between plant water demand or water use and water extraction. So in this case, water use is just the total amount of water applied from either natural sources or storage directly to the plants to grow these plants. And it really reflects the plant water demand in the growing season. And this has really been what's been used previously to think about potential environmental impacts of cannabis water use. In contrast, water extraction is probably a more reliable way of thinking about environmental impacts. And this really represents the sum of water taken from the environment and either directly applied to plants or diverted into storage. So it really reflects the total amount of water and the timing of water withdrawn from the environment. And this way of thinking about water use is probably more ecologically relevant when thinking about potential impacts to salmon and other sensitive species. So further analysis of this data has shown that the source of water, which I've shown can vary quite substantially depending on where they're located and every farm, has a huge influence on the timing of water extraction. And here we just see a really notable example where if cannabis farms are relying on rainwater catchment, their patterns of extraction are going to be entirely in the winter wet seasons. In other words, the timing of water use is going to be concentrated in November to April, whereas we see that farms that rely primarily on groundwater or well water pump primarily on demand. In other words, they're not pumping from the wells into storage, but rather they're taking that water from wells and applying it directly to plants as they need it. And so in this case, if they're relying on wells, the demands largely reflect really the plant water demand in the growing season. With farmers relying on both surface and spring water sources, we see a more steady demand across the growing season, where we see perhaps some peak in the growing season months, but the timing of withdrawals is actually much more consistent. And this is really explained by the role of storage that I mentioned earlier. And in particular, it appears that growers that rely on seasonal water sources like springs and surface water sources like streams often use storage, such as like the pond shown here or these above-ground water storage tanks to augment their irrigation supplies in the dry growing season. And so we took all these data and we developed a model to try to understand how the access to different water sources and or the presence of water storage actually affects these seasonal water extraction patterns. So let me take a minute to walk you through this slide. What this shows is the seasonal extraction profiles of different classes of water users. In the yellow-orange color, those are water users that rely on perennial water sources like groundwater. And as I showed earlier, this pattern of demand largely reflects cannabis, the plant water demand itself. And so it really tracks the growing season. In contrast, the blue line here represents cultivators that have access to a pond on site and are able to store large quantities of water in the wet season. And so here we basically see the exact opposite pattern where all of the vast majority of water is being extracted in the winter months and very little water, relatively little water is being extracted in the growing season. And we see that other users that rely on seasonal sources but don't have access to large storage like a pond have that sort of steady water use profile across the season. Just one note here is that these graphs represent the estimated use for the median farm size in the area, which is about a quarter of an acre. So those demand volumes that you see on the y-axis are relative to the sort of the median farm size. You also see in light gray here a line that shows the plant-based estimate of total demand. So that's basically six gallons per plant per day. And what we see is that for groundwater users, the amount of water being applied to farms pretty much matches that amount in the peak of the dry season. But at the beginning of the growing season and the end of the growing season, total water demand is actually much less. So this sort of plant-based approach that's been used previously is pretty reflective of peak season water demands but may not be as a good an indicator estimate of demands in the earlier late season. So coming back to this question of water storage and some of the policy implications of this work, as I mentioned earlier, farms that rely on seasonal water sources, surface water sources like streams and springs, are under the new permitting system, are no longer able to take any water from those sources between April and October. In other words, they either have to store the water they need in the wet season or drill wells. And based on our analysis of these data, we see that those farms that rely on these seasonal water sources, most of them or the majority of them do not have sufficient storage to meet that requirement. And so this suggests to me that in the future we can either expect to see increasing well drilling to allow them to continue to meet their water demands and meet the permitting requirements or potentially the expansion of water storage if they're able to on their site. So some key findings from our analysis of water use first, farms with perennial water sources, in other words, farms that rely primarily on groundwater or Iperian well water, do not store much water. And the extraction patterns, the timing in which water is taken from the environment, really follows plant demand during the growing season. In contrast, farms relying on seasonal water sources really show a steady demand across the year. And this sort of lack of variation really reflects sort of continued inputs of water from the environment into storage in the winter season and then continued applications during the growing season. And finally, from this analysis, we found that farms with ponds generally extract most of their water, if not all of their water in the growing season and have enough water stored really to meet their plant water demands throughout the dry season. So it's the farms with the ponds that are, for them to meet some of these new permitting requirements. So let's just review our key questions and the conclusions drawn from this analysis. First question, where do cannabis farms source their water? Our key takeaway point here is that wells are the most common water source in this region followed by surface water and springs. How much water use is really the wrong question to ask? What we should be asking is how much water is extracted and when? And this we basically find that the timing and amount of water extracted for cannabis really depends on the water source that they're relying upon or water sources and available storage capacity. So what does this all mean for impacts to in-stream flows? This is an area of research that we're continuing to explore, but these findings provide some insight that we can use to think about these potential impacts. First, we see that for the majority of sites are the extraction patterns really track plant irrigation demands. And so in this region of California stream flows are naturally quite low in the growing season. And so this additional demand for agricultural water use raises some concern over in-stream flow impacts. However, it's important to note that groundwater use, which I've noted is widespread, and off-stream storage perhaps might moderate some of these impacts. In other words, the impacts that we might see from direct diversions are probably greater than the impacts that we might see from groundwater use. However, recent studies have shown that groundwater use can actually cause both acute and chronic stream flow depletion. And so it's not like groundwater could not have an impact on surface flows. It's just that those impacts might be delayed or take a little more time to manifest in the system. And when thinking about stream flow impacts and our approach to quantifying stream flow impacts, we really need to think about both the quantity and timing of water extraction patterns, as well as the location of diversion and whether they're taking water from surface or groundwater. And so if they're taking water from groundwater that's directly adjacent to the stream in the riparian zone, which is quite common, the impacts are probably the same effectively as if they were taking water directly from the stream. In contrast, if the water is being taken far from the stream from surface water sources that have sort of limited connectivity to the stream, then those impacts are likely to be delayed or lagged from the timing of diversion. So basically this points to the need for further research to really get a handle on how cannabis water diversions are impacting stream flow. And this is an active area of research that we're continuing to explore here at the Cannabis Research Center. Our next step is really to link some of the information that we're gaining from aerial imagery mapping of distribution of cannabis farms and linking them with these types of water use models to predict total estimated water demands throughout the year and trying to understand what this might mean for stream flow. Another important topic that I haven't discussed in this presentation are the potential impacts of cannabis on water quality. So we know from, again, from enforcement actions and from interviews with cannabis farmers that the use of pesticides and fertilizers and fuels for generators and so forth is fairly, fairly widespread and represents a fairly significant threat if these chemicals are to enter our waterways. And so that's another area of research that we will continue to explore in the future. Thank you very much for your attention. If you'd like to reach out and have any questions, feel free to contact me at the email below. Thank you.