You are on page 1of 8

1

Environmental Impacts of
Marijuana Cultivation in Humboldt & Mendocino Counties, CA
Overview of a Pilot Study (Draft)
California Department of Fish and Wildlife,
Coastal Conservation Planning, Eureka Field Office
SUMMARY
1. Large-scale marijuana cultivation has been documented in many watersheds throughout
northern California, and appears to be having substantial cumulative regional effects on
watershed health and aquatic species. However, due to the inaccessibility of private forested
lands, the impacts of marijuana cultivation have been difficult to quantify.

2. In order to estimate marijuana cultivation water demands, we interpreted high resolution aerial
images to map marijuana cultivation sites in four watersheds. We then used USGS stream flow
records to estimate summer low flow within the watersheds, and compared these low flow
estimates with marijuana water demands.

3. Our results show that direct surface water diversions for marijuana cultivation have the
potential to completely dewater steams in three of the four study watersheds.

4. Greater efforts are needed by law enforcement to better enforce existing environmental and
land use laws now being largely disregarded by the marijuana cultivation community.

INTRODUCTION
Marijuana (Cannabis sp.) has been cultivated in the backwoods of northern California at least since
the countercultural revolution of the 1960s with few documented environmental impacts. Landscape-
level marijuana cultivation however is a more recent phenomenon and appears, in part, a response to
Proposition 215, the Compassionate Use Act of 1996, a California law which provides for the legal
use of medical marijuana. As a result of Proposition 215 and other factors, large-scale, widespread,
and largely unregulated cultivation of marijuana has increased rapidly since the mid-1990s in
Humboldt County and in many counties throughout California. In Northwestern California, large-scale
marijuana cultivation often entails bulldozing and grading sites of a half-acre or more to grow outdoor
plants, install one or more large greenhouses (often >400 ft
2
), marijuana drying and processing
facilities and related structures, and often, residences for the cultivators.
Humboldt County has over 1.9 million acres of forestland comprising more than 80% of its land base.
Because of the quasi-legal status of marijuana cultivation in California (cultivation remains illegal
under federal law) and the crops high market value, making it vulnerable to theft, large outdoor
marijuana cultivation sites (MCSs) are most commonly located in remote forested and mountainous
locations where detection by the public and law enforcement is minimized.
Marijuana is a water loving plant (Cervantes, 2006) and California has a dry summer Mediterranean
climate with >90% of its annual rainfall occurring between October and April (Western Regional
Climate Center 2013). Consequently, MCS are typically situated on land with reliable year-round
water sources to provide for irrigation throughout the hot dry summer growing season. Diverting

2
mountain springs, seeps, and perennial streams is the most common means for MCS to acquire
irrigation water, though California Department of Fish and Wildlife (CDFW) field staff have
documented the use of groundwater wells and importing water by truck.
Humboldt Countys forests and freshwater aquatic habitats are home to numerous state and federally
listed and otherwise sensitive species. Almost every anadromous salmon and steelhead population
in Humboldt County is either state or federally listed as a result of a significant long-term decline in
their populations. The combination of forest conversion by MCSs, poor sediment control during and
after site construction, use of rodenticides and other poisons, release of contaminants such as diesel
fuel and fertilizers, and water diversions all impact terrestrial and aquatic species in the region.
Quantifying these impacts are exceedingly challenging given the clandestine nature of these
operations, cultivators mistrust of county, state and federal regulators and scientists attempting to
assess environmental impacts, and related legal and safety concerns. This study quantifies
marijuana cultivation surface water diversions and its potential environmental impacts in four
watersheds in Humboldt and Mendocino counties: Redwood Creek near Redway (RCS), Salmon
Creek near Miranda (SC), Upper Redwood Creek near Orick (URC), and Outlet Creek near Willits
(OC) watersheds.
Table 1. Study watershed locations and attributes.
Watershed County Area (acres)
Upper Redwood Creek (URC) Humboldt 43,328
Salmon Creek (SC) Humboldt 23,514
Redwood Creek South (RCS) Humboldt 16,655
Outlet Creek (OC) Mendocino 103,562

METHODS

The number, size, and location of MCSs were determined by interpreting aerial imagery using Esri
ArcGIS

software (Environmental Systems Research Institute, Inc., 2013) in conjunction with


Google Earth

(Google Inc., 2013). Using current high-resolution aerial imagery (taken in summer
2012) in Google Earth

as a reference, features were mapped using ArcGIS software. Watershed
boundaries were digitized using USGS 7.5 minute Digital Raster Graphic (DRG) images to correct for
hydrological inconsistencies. These watershed boundaries and a grid with 1- km
2
cells were
displayed in both Google Earth

and ArcMap enabling easy navigation between the two programs cell
by cell. The mapping and data collection were accomplished as follows:


3
1. Greenhouses, outdoor marijuana cultivation sites, water tanks, and residences were
identified using Google Earth and then mapped as points in ArcMap
TM
with additional
attribute data.
Greenhouses: Identified by color and transparency, elongated shape and often apparent PVC
framework. Greenhouses were measured using Google Ruler to obtain area, and these
measurements were noted in the corresponding attribute table in ArcMap.

Outdoor Marijuana Cultivation Sites: Plants often varied greatly in size but were easily identified by
their shape, coloration, and placement in rows or other regularly spaced configurations. In addition,
we examined previous years imagery to confirm plantings were not perennial crops such as fruit or
other tree orchards.

Outdoor plants that were not in close proximity to greenhouses, or easily visible in clearings, were
sometimes placed along and within forest edges. Thus, we consider our counts to be conservative
estimates because of the difficulty associated with identifying partially hidden plants in areas with
overstory cover. The total number of plants within each MCS was noted in the corresponding
attribute table.

Tanks: These features were identified by shape and color. We noted whether each was clearly
associated with an MCS or not, and measured either the diameter (for round tanks) or the length and
width (for square tanks or water bladders) and noted these measurements in the corresponding
attribute table.

2. Ponds and areas under cultivation were digitized as polygons, and areas were calculated.

Ponds: Water features were identified by their dark-green color (easily visible on both NAIP and
Google Earth imagery). Ponds were often associated with residences or MCS operations. The
outline of each pond was digitized as a polygon feature.

Area of land under cultivation: In order to quantify the area of land cultivated for growing marijuana,
we digitized the outlines of gardens and greenhouses including any additional land clearing
associated with the operation. Imagery from years 2009, 2010, and 2012 were compared to
determine if land area associated with MCSs increased during that period. Due to lack of
availability/access to high resolution imagery from prior years, we were unable to digitize areas of
land in years prior to 2012.

Instead, we calculated the area under marijuana cultivation in 2009 and 2010 by a process of
subtraction: If an MCS existed in the summer 2012 imagery in Google Earth, the area was digitized in
the 2012 NAIP imagery in ArcMap to obtain a total acreage of land area under cultivation in that
watershed. We then used the Google Earth historical imagery feature to view these same MCSs in
prior years (2009 and 2010) and were thus able to determine whether or not each MCS existed in
those prior years. We then subtracted the acreage of MCSs which existed in 2012 but not in 2010
and/or 2009 from the total acreage in order to graph the change of marijuana cultivation over time.



4
3. For certain features of the watershed (water tanks and ponds) we assessed likelihood of
association with an MCS by assigning a 1, 2, or 3 in decreasing degrees of certainty, 1 having
the highest likelihood of association.
We assigned these values in order to further qualify our data to permit more conservative analyses in
the future (for example, when counting tanks we will assign a value of 1 to those that are directly
adjacent to or inside an obvious MCS, and a value of 2 to those in close proximity to MCSs but not
directly adjacent to or inside one, etc).
4. Water demands were estimated using above data in conjunction with calculations provided
in an industry document (HGA 2010).
We totaled the number of plants counted from outdoor MCSs and combined this number with
estimates of marijuana plant canopy contained in greenhouses. These estimates were based field
visits to MCSs by CDFW Environmental Scientists during 2013. During these field visits, 32
greenhouses were measured and the spatial arrangement and canopy area of marijuana plants
inside were recorded. The resulting average ratio of greenhouse area per plant was 12 ft per plant.
Water estimates were derived from the Humboldt Growers Association 2010 Humboldt County
Outdoor Medical Cannabis Ordinance draft (Humboldt Growers Association, 2010). According to this
document, it is reasonable to estimate water use of 6 gallons per plant per day during the growing
season, which typically extends from June-October (150 days). This water use estimate was based
on a popular watering method (soaker hose or emitter line watering) and typical watering patterns.
Although water use estimates for marijuana cultivation vary widely, with the United States Department
of Justice estimating 5 gallons per plant per day (National Drug Intelligence Center, 2007) and some
industry estimates as high as 15 gallons per day (PRWeb, 2012), the 6 gallon per day figure is an
average put forth by the industry. The figure does not account for water delivery inefficiencies
observed by CDFW biologists at multiple MCSs (loss during transmission due to leaking water lines
and tank overflow).
5. Seven-day low flow minimum and maximum values for Redwood Creek and Salmon Creek
were estimated using records from three gaged streams within the South Fork Eel watershed.
Seven-day low flow values were estimated for our ungaged study watersheds (RCS and SC) using
data from nearby gaged streams and methods developed by Rantz (1968). Gaged streams from
within the South Fork Eel River Watershed were used to estimate the seven-day low flow in RCS and
SC (Bull Creek near Weott, Humboldt County USGS gage 1146600; and Elder Creek near
Branscomb, Mendocino County, USGS gage 11475560).
6. Seven-day low flow minimum and maximum values for Upper Redwood Creek and Outlet
Creek were calculated using gage data.
In our gaged watersheds (URC and OC), seven-day low flow minimum and maximum values were
calculated using flow data from USGS gages within the watershed. Seven-day low flow data from the
Okane Gage (USGS gage 11481500, Redwood Creek near Blue Lake) were used to compare to
water demands from marijuana cultivation to stream flow in RCN. OC was gaged from 1956 (USGS

5
gage 11472200, Outlet Creek near Longvale) through 1994, and these data were used to calculate
seven-day low flow data for OC.
7. Water demand estimates were compared to seven-day low flows and we calculated the
percentage of stream flow that could be diverted for marijuana cultivation demands.
We used minimum and maximum values for seven-day low flows in order to obtain a range of values
for the percentage of water that marijuana cultivation could potentially consume depending on
whether the year was above or below average in terms of stream flow. We then calculated the daily
water demand as a percentage of stream flow.
Table 2. Estimates of seven-day low for ungaged watersheds.
Ungaged Watershed
Estimated annual seven-day
low flow (minimum, cfs)
Estimated annual seven-day
low flow (maximum, cfs)
Redwood Creek (southern) 0.10 0.61
Salmon Creek 0.10 0.64

Table 3. Annual seven-day low flow range for period of record for gaged streams.

Calculated seven-day low flow for period of record
(cfs)
Gage Watershed Area (Acres) Minimum Average Maximum
SF Eel Miranda 343,680 12.4286 31.1776 63.4286
Bull 17,920 0.2100 1.0963 3.0143
Elder 4,160 0.2714 0.6363 1.3000
Outlet 103,040 0.0000 0.5734 1.7571
Redwood Creek (nr Blue Lake) 43,328 0.9357 3.7584 9.1857

FIELD WORK

In addition to the spatial analysis, CDFW staff has evaluated more than 40 MCSs in the Eel River
watershed and other watersheds in northwestern California. During these site visits, the size,
spacing, and spatial arrangement, and growing season of marijuana plants were recorded as were
the water conveyance and storage methods. Every MCS assessed in the field utilized a surface water
diversion system that diverted and stored water from on-site or adjacent springs, seeps, and streams.
Only two MCSs utilized a groundwater well in addition to surface water. On-site surface irrigation
water is often supplemented by water imported by truck from within or from an adjacent sub-basin.
While it is difficult to assess the scope, scale, and precise diversion location(s) of imported water, or if
water is imported at all, agency experience with water drafting from other industries, such as timber
and ranching, indicates when water is imported, it is typically diverted from surface water sites in
close proximity to the MCS (due to high transportation costs). This would have a similar detrimental
effect on aquatic species as would water diverted on-site. While CDFW staff have not conducted field
visits to the majority of MCS analyzed in this study, our site visits to MCSs across the region provide

6
high confidence that surface waters are the primary irrigation source for the majority of MCSs in this
study.
RESULTS
Plant totals and estimated water demand are tabulated in Table 4. MCSs were not analyzed in areas
that were known to be serviced by municipal water or other community service districts (these
consisted of the City of Willits and the Brooktrails Community Services District in the Outlet Creek
Watershed).

Table 4. Total number of marijuana plants found in study watersheds.
Watershed
Outdoor
Plants
Greenhouses
Est. Plants in
Greenhouses
Estimated Total
Plants in
Watershed
Upper Redwood
Creek (URC)
4,434 220 18,612 23,046
Salmon Creek (SC) 11,697 302 18,440 30,137
Redwood Creek South
(RCS)
10,475 324 16,777 27,252
Outlet Creek (OC)* 15,165 266 16,730 31,895

Table 5. Estimated water demand for marijuana expressed as a percentage of low flow in four study
watersheds. Water demand was compared to high flow regime and low flow regime estimates for seven-day
low flow.**The seven-day low flow minimum was less than 0.0 L/s at the gaged reference stream.
Watershed
Area
(mi)
Plants per mi
Demand as percent of seven-
day low flow
High est. Low est.
URC 67.7 340 2% 23%
SC 36.7 821 44% 254%
RCS 25.0 1090 42% 243%
OC 161.8 197 17% **

*The seven-day low flow minimum was less than 0.0000 cfs at the gage

Our data show that marijuana water demands in these watersheds have the potential to consume
significant portions of stream flow, and in some cases have the potential to divert all the flow in
several of our study watersheds.
DISCUSSION
The marijuana cultivation water demands presented here are best estimates based upon prudent and
conservative assumptions, marijuana cultivation industry data, and field visit assessments of MCSs in

7
similar watersheds throughout the region. Gathering more precise water use estimates, is presently
difficult because marijuana cultivation is a highly profitable, yet quasi-legal endeavor, thus naturally
engendering suspicion and mistrust of scientists and regulators seeking to assess its environmental
effects. Furthermore, marijuana cultivation in California is a largely unregulated and cultivators have
no incentive for detailed record keeping. More precise water use estimates, seasonal variability,
variability between MCSs, and confidence intervals of flow rates and volumes would require far more
open access to marijuana cultivation activities, and coordination with cultivators. Assessing
measurable effects of marijuana cultivation water diversions on salmonid populations and other
aquatic species would likewise require unobstructed access to streams on private properties with
MCSs or adjacent to them, which in the current legal and cultural setting, is infeasible.
This study shows the estimated rate of water diversion by marijuana cultivation during the low flow
season is highly likely to negatively affect listed salmonids by reducing available over-summering
habitat, impairing benthic macroinvertebrate production and food availability, and could result in direct
mortality of salmon and steelhead juveniles if stream reaches are completely dewatered. In addition,
lower steam flows lead to warmer water temperatures and lower dissolved oxygen levels which
negatively affect fish and amphibian populations (Moore & Townsend, 1998; Stevens, Blewett, &
Casey, 2006). Diversion of headwater streams and springs can dewater southern torrent salamander
(Rhyacotriton variegatus) and coastal tailed frog (Ascaphus truei) habitat, and lead to mortality of
these desiccation intolerant species. Lastly, sedimentation from land-clearing activities and road
construction has negative effects on watershed health including impacts on fish and amphibian
reproduction (Klein, 2008; Poole & Berman, 2001).
There is compelling evidence that even if the surface water diversions presented here are
substantially over-estimated, the diversion of surface water for marijuana cultivation would remain a
significant negative impact on aquatic species during the summer low flow period. Furthermore, a
technique known as light deprivation was observed at most MCS in 2013. This technique allows for
multiple crops per year, and thus even greater water consumption during the growing season. This
study and subsequent field observations indicate marijuana cultivation in the four study watersheds
has potentially significant negative effects on watershed health and sensitive aquatic species.
Given large-scale marijuana cultivation is documented in many watersheds throughout northern
California, it appears highly likely marijuana cultivation is having substantial cumulative regional
effects on watershed health, aquatic species, and specifically listed salmonid stocks. The current
legal and regulatory structures used to assess, minimize, and monitor other land uses, such as
surface mining, timber harvesting, and ranching appear inadequate to address the impacts of large-
scale marijuana cultivation. The fact that marijuana cultivation is illegal under federal law, and
uncertainty over how federal law and policies interact with state law under the current federal
administration, appears to make state and local authorities hesitant to more deliberately regulate
marijuana cultivation pursuant to Proposition 215. Resolution of the legal status of marijuana
cultivation under Proposition 215 under federal law could help provide the state and counties with a
greater ability and incentives, to regulate the harmful environmental effects of this industry. Short of
that, greater efforts are needed by law enforcement to better enforce existing environmental and land
use laws now being largely disregarded by the marijuana cultivation community.

8
REFERENCES

Cervantes, J. (2006). Marijuana horticulture: the indoor/outdoor medical growers bible. Sacramento,
CA: Van Patten Pub.
Environmental Systems Research Institute, Inc. (2013). ArcMap
TM
(Version 10.0). Retrieved from
http://www.esri.com
Google Inc. (2012). Google Earth (Version 6.1.0.5001). Retrieved from
http://www.google.com/earth/index.html
Humboldt Growers Association. (2010). Humboldt County Outdoor Medical Cannabis Ordinance
Draft. Retrieved from http://library.humboldt.edu/humco/holdings/HGA2.pdf
Klein, R. D. (2008). Timber Harvest and Turbidity in North Coastal California Watersheds. Presented
at the Third Interagency Conference on Research in the Watersheds, Estes Park, CO.
Moore, M. K., & Townsend, V. R. (1998). The Interaction of Temperature, Dissolved Oxygen and
Predation Pressure in an Aquatic Predator-Prey System. Oikos, 81(2), 329.
doi:10.2307/3547053
National Drug Intelligence Center. (2007). Domestic cannabis cultivation assessment 2007.
Johnstown, PA: U.S. Dept. of Justice, National Drug Intelligence Center.
Poole, G. C., & Berman, C. H. (2001). An Ecological Perspective on In-Stream Temperature: Natural
Heat Dynamics and Mechanisms of Human-CausedThermal Degradation. Environmental
Management, 27(6), 787802.
PRWeb. (2012, March 22). Leading California Marijuana Attorney Says Growers Must Focus on
Water Conservation. PRWeb. Retrieved January 28, 2014, from
http://www.prweb.com/releases/marijuana-attorney/california/prweb9316223.htm
Rantz, S. E. (1968). Average annual precipitation and runoff in north coastal California (No. HA -
298). United States Geological Survey. Retrieved from
http://pubs.er.usgs.gov/publication/ha298
Stevens, P. W., Blewett, D. A., & Casey, J. P. (2006). Short-term effects of a low dissolved oxygen
event on estuarine fish assemblages following the passage of hurricane Charley. Estuaries
and Coasts, 29(6), 9971003. doi:10.1007/BF02798661

You might also like