Surveys of Community Garden Affiliates and Soils in Houston, Texas

Abstract

Although urban community food gardens have the capacity to strengthen and support neighborhoods in need, the benefits of such operations must be considered in tandem with the potential risks associated with urban environmental contamination. Therefore, research is needed to characterize existing community gardens in urban areas. In the present study, a survey of Houston, Texas community gardeners (N=20) was conducted to better understand their risk-based knowledge and perceptions, current gardening practices, and willingness to implement risk mitigation measures. Soil samples collected from the beds (N=22) and surrounding grounds (N=24) of existing community garden sites in Houston, Texas were screened for trace and heavy metals using X-ray fluorescence spectrometry. The survey indicated that community gardeners had few concerns with regard to potential soilborne hazards and were generally willing to use diverse strategies to reduce potential hazards related to garden soil contamination. Ground and garden bed soil collected from community gardens were found to have excess concentrations of arsenic compared to federal health screening limits. The information provided here provides insight into possible discordance between community gardening risk perception and contamination risk that could be addressed through outreach, engagement, and remediation approaches.

Introduction

Approximately 8.9% of the global population was affected by hunger and undernourishment in 2020 (Food and Agriculture Organization of the United Nations (FAO) et al., 2020). Although the availability of affordable foods has been increased by industrialization in the food supply chain, the sustainability of large-scale food production operations and practices remains a significant challenge. For example, food production capacity is impaired by freshwater, land, and soil degradation (FAO et al., 2020), while agricultural diversity loss due to practices such as monocropping render long-term food security vulnerable to plant disease outbreaks (National Academies of Sciences, Engineering, and Medicine, 2019). Drought, flooding, and other natural disasters present both present and emerging challenges to food production that are expected to become increasingly disruptive (National Academies of Sciences, Engineering, and Medicine, 2019). Food loss at subsequent stages of the food supply chain also hinders the availability of safe, nutritious, and affordable food. For example, around 14% of all harvested food is lost prior to entering retail and is therefore unavailable to consumers (FAO, 2020). In response to these challenges and others, local food production has garnered increased attention as a more sustainable approach to strengthening local food supply chains, including those located in urban areas.

The National Center for Environmental Health defines community gardens as “collaborative projects on shared open spaces where participants share in the maintenance and products of the garden, including healthful and affordable fresh fruits and vegetables” (National Center for Environmental Health, 2020). Community-level agricultural production is supported by government agencies at all levels, as reflected by an increasing number of programs and policies developed to strengthen production in local food systems (Low et al., 2015). For example, at the Federal level, the U.S. Department of Agriculture’s Natural Resources Conservation Service provides financial support through the Texas Urban and Rural Conservation Project for local-level agricultural efforts that aim to establish community gardens and high tunnels for food production, rainwater harvesting systems, and pollinator habitats (USDA-NRCS, 2019). A 2011-2012 survey of community garden organizations in the U.S. (N=445) indicated that around half of respondents received support from local governments in the form of access to gardening materials and equipment (50%) and expedited authorization to use public land (47%) (Lawson & Drake, 2013). Although gardening for food production has become more prevalent, as indicated by the 17% increase in food garden participation among U.S. households from 2008 to 2013 (National Gardening Association, 2014), this potential solution to food insecurity is challenged by important gaps in knowledge and practice related to soilborne contaminant risks (Kim et al., 2014).

Community gardeners tend to underestimate the risk of hazardous exposures related to gardening (Hunter et al., 2019; Kim et al., 2014; Wong et al., 2018). Community garden soils are rarely tested for potential hazards due to concerns about cost and the complexity of collecting, submitting, and processing samples (Hunter et al., 2019; Ramirez-Andreotta et al., 2019), and gardener knowledge of exposure reduction strategies and interventions is frequently limited (Wong et al., 2018). For example, in a study of Georgia gardens by Hunter et al., 44% of gardens (214 of 483) were tested only once over a 1 to 3 year period. Heavy metals are among the most common soil contaminants of concern among community gardeners (Hunter et al., 2019; Kim et al., 2014; Wong et al., 2018). Ingestion of contaminated crops is largely a known route of exposure to soilborne contaminants among community gardeners (Kim et al., 2014). The potential for soilborne contaminants to be transferred to produce has been documented, as has the internalization and uptake of certain metals into produce items (Natural Resources Canada et al., 1995; ASTDR, 2016; ASTDR, 2012). Incidental soil ingestion can also result from not washing hands or produce (Ramirez-Andreotta et al., 2019). Dermal contact with contaminated dust or soil is the most widely recognized exposure route among community gardeners, while inhalation of particulate-bound contaminants tends to be overlooked (Kim et al., 2014). Inhalation of particulate-bound contaminants such as arsenic, chromium, and lead may result in deposition in the interior lining of the lungs leading to absorption into the body (ASTDR, 2016; ASTDR, 2012; ASTDR, 2007).

Arsenic, chromium, and lead are contaminants of interest in garden soils because these elements are both naturally-occurring in the environment and produced from anthropogenic activities (ASTDR, 2016; ASTDR, 2012; ASTDR, 2007). The metalloid arsenic is both ubiquitous and may result from both natural (e.g., volcanic eruptions, wildfires, and low-temperature volatilization) and anthropogenic processes and activities (e.g., application of arsenic-containing pesticides in agricultural production, chromated copper arsenate-treated wood, and copper or lead smelting operations) (ASTDR, 2016). Chromium is a trace element that is native component of rocks and soils but may be introduced into the environment from industrial processes (e.g., electroplating, textile production, and leather tanning) or oil and gas combustion (ASTDR, 2012). The vast majority of lead in soil and dust is attributed to anthropogenic activities with leaded gasoline and paint being the greatest contributing sources, although these uses have been effectively banned in the U.S. since 1976 and 1988, respectively ) (ASTDR, 2007). However, older, painted homes and those locate near traffic congestion have still been shown to have high lead concentrations in soils (Taylor et al, 2021). Arsenic, inorganic arsenic compounds (International Agency for Research on Cancer (IARC, 2012)), and chromium (VI) compounds (IARC, 1990) are classified by as carcinogenic to humans (Group 1), while inorganic lead compounds are classified as being probable human carcinogens (Group 2A) (IARC, 2006).

Community gardens provide a multitude of benefits to communities, including expanding access to fresh, nutritious foods and promoting sustainability in local food supply chains. However, available evidence suggests that community gardeners and garden organizers tend to underestimate the risk of potential garden contamination and are often unsure of how to access soil testing services. These knowledge gaps are of particular concern in urban areas where community gardens are frequently in close proximity with potential sources of hazardous contaminants such as transportation infrastructure and industrial facilities. Community gardens located in the City of Houston, Texas are further challenged by a lack of zoning, which means transportation infrastructure and industrial land uses may be intermingled with residential and other types of land uses where gardens are later sited, such as parks, and that garden site history may include legacy contamination from prior uses (Hunter et al., 2018; Turner, 2020). Little is known about the knowledge, practices, and perceptions of risk of community gardeners in this area. Therefore, research is needed to characterize this population, identify risk-related knowledge gaps, and develop strategies and interventions that are both effective in reducing risk and perceived as being acceptable for implementation in community garden settings. To support and inform evidence-based strategies to mitigate potential risks associated with urban community gardening, a cross-sectional survey of community gardeners in Houston, Texas was conducted and ground and garden bed soils collected from community garden sites located in Houston, Texas were screened for trace and heavy metals.

Methods

Study Area

Houston, Texas is a large metropolitan city with a diverse population of over 2,325,500 individuals (U.S. Census Bureau, 2020). From 2010 to 2018, the population living in Houston grew by 11.1%, which compares to a national population growth rate over the same period of only 6.0% (U.S. Census Bureau, 2020). The population living in Houston is rich in diversity, with 44.8% of all people being of Hispanic or Latino ethnicity and a racial distribution of 57.6% White, 22.5% Black or African American, and 6.9% Asian (U.S. Census Bureau, 2020). The poverty rate in Houston, Texas is 20.4%, which is greater than that of both Texas (14.6%) and the United States (11.8%) (U.S. Census Bureau, 2020). In Harris County, Texas, approximately 16.3% of all people and 23.2% of the population under 18 years of age were challenged by food insecurity in 2017 (Houston Health Department, 2019). The City of Houston’s support for local agricultural initiatives includes an Urban Garden Program, which currently facilitates 11 community gardens located in city parks (Houston Parks and Recreation Department, 2019). The non-profit gardening association Urban Harvest, which provides community garden classes, youth education, and other educational programming, has registered 135 garden sites in its Community Gardens Program membership database, 91 of which are in the City of Houston (Urban Harvest, 2020).

Although community gardening is one approach to address food insecurity at local levels, variable environmental conditions and contaminant loads may inequitably affect the safety of produce grown locally since there may be risks with exposure over time (Santo et al., 2021). Development within the City of Houston is not regulated by zoning restrictions (Turner, 2020) and, as a result, residential properties may be located adjacent to industrial facilities and potential point sources of hazardous releases or legacy contamination. Prior research indicates that soils collected from predominantly minority and low-income areas in Houston have significantly elevated concentrations of barium, chromium, copper, and nickel compared to areas with lower proportions of minority and low-income residents (Chu, 2019). Thus, in order to protect vulnerable populations against further exposure to soilborne hazards, it is imperative that environmental conditions be considered in the development of community gardens.

Community Gardener Survey

Community Gardener Survey Recruitment

The study protocol, questionnaire instrument, information sheet, and recruitment e-mail were reviewed and approved by the Texas A&M University Institutional Review Board prior to data collection and determined to be exempt (IRB# IRB2019-1204M). Contact information for 38 of 91 community gardens in the City of Houston was obtained from publicly-available online membership listings and community garden websites. The recruitment email was distributed with the information sheet attached to each garden contact once per week for three weeks, for a total of three separate recruitment attempts. The recruitment email included a link to the online survey, which was disseminated on the Qualtrics (Provo, UT) survey platform. Potential participants were asked to confirm that they are 18 years of age or older and affiliated with a community garden in Houston, Texas. Participants who confirmed eligibility to participate were presented a final question that requested their consent to participate. Individuals who indicated that they are not 18 years of age or older, who were not affiliated with a community garden in Houston, Texas, and/or declined to consent to participate were be directed to a thank you page and did not advance to the survey.

The questionnaire instrument was comprised of 17 questions within the domains of participant demographics, community garden characteristics, and garden risk knowledge and perceptions (Supplemental Document 1). Specifically, there were 6 questions about respondent demographic and household attributes, 4 questions about community garden characteristics, 2 questions designed to ascertain knowledge and perceptions about garden-associated risk, and 5 questions to discern willingness to use risk reduction measures such as garden soil testing, remediating soil, changing gardening practices, and produce washing.

Survey Data Analysis

Following the four-week data collection period, the web-based survey response period was ended. Data was exported from Qualtrics (Provo, UT) and imported into Microsoft Excel (Redmond, WA) for descriptive analysis including counts and frequencies. All open-ended response were recorded in a Microsoft Excel (Redmond, WA) spreadsheet and reported verbatim.

Community Garden Soil Screening

Sample Collection and Preparation

Community garden sites in Houston, Texas were identified from publicly-available membership listing of Urban Harvest (Urban Harvest, 2020). Because of the restrictions associated with accessing community gardens located on the premises of primary or secondary schools (n=62), which require completing a background check and other documentation, only non-school based gardens were approached in the spring of 2020 for soil sampling. A composite sample of ground soil was collected from the perimeter adjacent to and surrounding garden bed locations. When accessible, a composite sample of soil was also collected from the garden beds. For each composite sample, six surface soil plugs were collected using a new, single-use Terra Core Soil Sampler (En Novative Technologies, Inc.) and inserted into a sterile polyethylene bag. Once complete, the plastic bag was sealed, double-bagged, and then transferred into a cooler containing frozen ice packs for transport to the Texas A&M University School of Public Health laboratory.

Upon arrival at the laboratory, samples were immediately placed in −80 °C storage until frozen and then lyophilized for a period of 36 to 48 hours. Subsequent to freeze-drying, particle size uniformity was achieved by grinding each sample with a clean, dry mortar and pestle and then passing each sample through a clean, dry 60-mesh sieve in accordance with U.S. Environmental Protection Agency (EPA) Method 6200 (U.S. EPA, 2007). To achieve homogenization, each soil sample was transferred onto a clean sheet of butcher paper and rolled as described in EPA (U.S. EPA, 2007). Single-use, double open-ended XRF sample cups (polyethylene, 31.0 mm internal diameter) were labeled and sealed on one end with Mylar film (2.5 μm). Each sample cup was then filled with a prepared soil sample, compacted and flattened by tamping, and sealed with Mylar film in preparation for analysis.

Soil Sample Analysis

A Niton™ XL3t XRF Analyzer (Thermo Scientific™, Boston, MA) with a high-performance semiconductor detector, X-ray tube (Au anode 50 kV maximum, 200 μA maximum) excitation source, and 50 kV voltage was used in accordance with manufacturer instructions to evaluate the presence and approximate concentration of selected trace and heavy metals in prepared soil samples. To minimize potential interference from device positioning and inconsistent distance between the device probe window and sample (U.S. EPA, 2007), the XRF device was placed in a stationary test stand (Thermo Scientific) with the device probe window positioned in direct contact with the Mylar interface of the prepared, flattened soil sample. Internal device calibration was performed in advance of sample analysis in accordance with the instrument manufacturer’s instructions. Standard reference materials comprised of 2709a San Joaquin soil (National Institute of Standards and Technology, 2010) and TILL-4 soil (Natural Resources Canada, 1995) were used to evaluate the analytical device’s accuracy and the stability and consistency of analysis. Samples of these reference materials and a blank sample of silicon dioxide prepared in XRF sample cups were procured from Thermo Scientific and analyzed at the start and end of each working day and between each sample set of 20.

Data Analysis

The Standard Thermo Scientific™ Niton Data Transfer (NDT™) PC software suite was used to facilitate data transfer. Concentrations reported represent the average of triplicate readings of each sample and are appropriately interpreted as estimates. Concentrations reported as below the instrument’s limit of detection for each sample were assigned a value of zero before analysis. Results were compared to the EPA’s Generic Soil Screening Levels (SSLs) for residential scenario and ingestion-dermal exposure (U.S. EPA, 2019), the U.S. Geological Survey’s 2013 Geochemical and Mineralogical Data for Soils of the Conterminous U.S. (Smith et al., 2013), the Texas Commission on Environmental Quality’s (TCEQ) Texas-Specific Soil Background Levels (TCEQ, n.d.), and Houston background concentrations determined from 96 surface soil samples collected across the city (Chu, 2019). The datasets generated and analyzed during the current study are available from the corresponding author on reasonable request.

Results

Community Gardener Survey

Participant Demographics

In total, there were 20 eligible individuals who agreed to participate in and completed the survey (Table 1). Of the 18 individuals who entered a response for gender, half were male and half were female. Nearly half (45%) of all respondents were 65 years of age or older. Respondents most commonly identified as White (50%), followed by Black or African American (30%); none reported Hispanic ethnicity. The majority of respondents (N=18) had at least an associate’s degree-level educational attainment. Eighty percent of respondents reported a household income above the federal poverty limit for their household size.

Table 1.

Demographic characteristics of community gardeners (N=20).

Variable	n (%)
Sex
Male	9 (45.0)
Female	9 (45.0)
Response Missing*	2 (10.0)
Age
>65	9 (45.0)
55-64	7 (35.0)
26-54	3 (15.0)
Response Missing*	1 (5.0)
Race
White	10 (50.0)
Black or African American	6 (30.0)
Other race	1 (5.0)
Response Missing*	3 (15.0)
Ethnicity
Hispanic	0 (0.0)
Non-Hispanic	17 (85.0)
Response Missing*	3 (15.0)
Level of education
Associate’s Degree	7 (35.0)
Bachelor’s degree	9 (45.0)
Master’s degree	2 (10.0)
Response Missing*	2 (10.0)
Number of people in household
1	3 (15.0)
2	11 (55.0)
3	2 (10.0)
4	1 (5.0)
Response Missing*	3 (15.0)
Household income above poverty for household size
Yes	16 (80.0)
No	1 (5.0)
Response Missing*	3 (15.0)

^*Includes “prefer not to respond” and absent responses

Community Garden Characteristics

Respondents were asked about the types of produce grown in and soil amendments added to their respective community gardens. The most common produce types reported were root vegetables (100%), leafy vegetables (100%), herbs (95%), and fruits (70%). With respect to soil amendments, 90% used compost, 75% used top soil, and 40% used manure. Only 20% of respondents indicated using general pesticides, Miracle-Gro fertilizer, and/or MicorLife fertilizer, respectively.

Participants were asked to rate certain attributes about their community garden as being excellent, good, average, poor, or terrible. Garden location and quality of garden resources (e.g., soil and water) were rated as excellent or good by 90% of respondents. Seventy percent of respondents felt knowledge of community garden staff and volunteers was excellent and the social atmosphere of the garden was excellent, respectively. Most respondents agreed that working in a community garden strengthened their knowledge of, and care for the, environment. Eighty-five percent strongly agreed or agreed that they knew more about the environment as a result of working in the garden. Three-quarters of respondents strongly agreed or agreed that they cared more as a result of working in the garden.

Garden Risk Knowledge and Perceptions

Respondents were asked in open-ended questions to describe any health concerns related to gardening and any soil contaminants that may affect urban gardens. Only two respondents listed health concerns, which included stinging insects and cancer clusters reported in Houston, Texas. Similarly, only two respondents were aware of soil contaminants relevant to urban gardens but no specific types were identified. Although community garden affiliates were largely unaware of potential urban soil contaminants, 80% would be willing to test soil if contamination with metals and/or organic chemicals were suspected.

Respondents were asked about their willingness to implement certain risk reduction measures if community garden soil contamination was confirmed (Table 2). The majority of community garden affiliates would be willing to implement or consider implementing the proposed risk reduction measures, though behavioral and operational changes were favored over physical soil remediation. For example, most respondents would be willing to wash their hands immediately after gardening (85%) and wash produce prior to consumption (85%). With respect to soil remediation approaches, 60% of respondents would add “natural” soil amendments such as compost or minerals and 55% would add contaminant-binding sorbent materials, 55% would plant non-edible plants capable of phytoremediation, and 50% would add fungi or fungal metabolites capable of degrading, removing, and/or stabilizing contaminants in soil. Although 85% of respondents would be willing to grow produce in raised beds or containers if soil contamination was confirmed, only 35% would remove and replace contaminated soil and only 30% would be willing to install a barrier to contain contaminated soil.

Table 2.

Willingness to use testing and risk reduction strategies among Houston, Texas community gardeners (N=20).

Variable	Yes n (%)	Maybe n (%)	No n (%)	Response missing* n (%)
Install a barrier over contaminated soil such as a plastic cover	6 (30.0)	7 (35.0)	3 (15.0)	4 (20.0)
Add natural soil amendments such as compost or minerals	12 (60.0)	2 (10.0)	1 (5.)	5 (25.0)
Add sorbent materials that bind contaminants	11 (55.0)	5 (25.0)	0 (0.0)	4 (20.0)
Remove and replace contaminated soil	7 (35.0)	7 (35.0)	3 (15.0)	3 (15.0)
Plant certain types of plants that can degrade, remove, and/or stabilize contaminants in soil	11 (55.0)	5 (25.0)	0 (0.0)	4 (20.0)
Add fungus or fungal metabolites that can degrade, remove, and/or stabilize contaminants in soil	10 (50.0)	5 (25.0)	1 (5.0)	4 (20.0)
Stop eating produce grown in contaminated areas	15 (75.0)	0 (0.0)	1 (5.0)	4 (20.0)
Grow produce in raised beds or containers	17 (85.0)	1 (5.0)	1 (5.0)	2 (10.0)
Stop growing certain crops such as root vegetables	10 (50.0)	6 (30.0)	0 (0.0)	3 (15.0)
Wear gloves while gardening	14 (70.0)	3 (15.0)	1 (5.0)	2 (10.0)
Wash hands immediately after gardening	17 (85.0)	0 (0.0)	1 (5.0)	2 (10.0)
Wash produce with water before consuming	17 (85.0)	0 (0.0)	1 (5.0)	3 (15.0)
Wash produce with vinegar, detergent, or other treatment solution before consuming	11 (55.0)	7 (35.0)	0 (0.0)	2 (10.0)
Peel produce before consuming	13 (65.0)	3 (15.0)	1 (5.0)	3 (15.0)

^*Includes “prefer not to respond” and absent responses

In a final open-ended question, respondents were asked if there were other risk reduction measures that they would consider implementing if garden soil contamination were confirmed. The five responses received highlighted important factors and uncertainties that would need to be considered in decision-making. These included “I would want information on the contaminants and go from there,” and “We would have to work with the owner and see what she is willing to do to reduce the risk.” Others considered the level of investment that would be required to remediate contaminated soil, with one stating that they were “Willing to try other measures as long as it does not require strenuous labor,” and another that “depending on the health and safety risks, we would pursue all mitigation steps needed including completely starting over with new soil if necessary.” A final response brought emphasis to the dire reality that may result from such a discovery, responding that “If our soil was contaminated, we would not continue our business.”

Community Garden Soil Screening

Soilborne metal concentrations in samples collected from the garden beds (N=22) and grounds (N=24) of community garden sites in Houston, Texas are shown in Table 3. Notably, all sampled garden beds were found to be raised. Except for calcium, copper, molybdenum, strontium, scandium, mean concentrations of all elements in ground soil samples were higher than in garden soil samples. No garden bed nor ground soil samples were found to have cobalt, gold, or palladium. Arsenic was detected in 95.5% of garden bed soils and 91.7% of garden ground soils. The mean concentration of arsenic in ground soils was 4.1 mg/kg ± 2.6 SD, which was moderately higher than that of garden bed soils (2.9 mg/kg ± 2.1 SD). The maximum concentration of arsenic in garden bed soil (7.8 mg/kg) was moderately lower than that of ground soil (10.4 mg/kg). Cadmium was detected in a single garden bed sample at a concentration of 1.2 mg/kg and in 29.2% ground soil samples at an average concentration of 0.5 mg/kg ± 1.0 SD. The maximum concentration of cadmium across ground soil samples was 3.7 mg/kg. Chromium was present in the bed soils of 86.4% of gardens and 100% of garden ground soils. The average concentration of chromium in garden bed soils was 17.1 mg/kg ± 12.2 SD, while the mean concertation of chromium in ground soil samples was 32.6 mg/kg ± 12.8 SD. The highest levels of chromium were 56.7 mg/kg in garden bed soil and 62.1 mg/kg in ground soil. Soils collected from garden beds had an average lead concentration of 16.8 mg/kg ± 9.6 SD and a maximum concentration 40.6 mg/kg. The average concentration of lead in ground soil samples was 52.4 mg/kg ± 69.8 SD with the maximum level observed being 328.7 mg/kg. Mean and median concentrations of antimony, barium, calcium, copper, iron, manganese, molybdenum, nickel, potassium, rubidium, strontium, thorium, tin, titanium, vanadium, zinc, and zirconium in ground and garden soils are available in Table 2.

Table 3.

Elemental concentrations (mg/kg) in garden bed and ground soil collected from community gardens in Houston, TX

	Instrument Limit of Detection (LOD)*	Garden Bed Soil (N=22)				Ground Soil (N=24)
Element		N>LOD	Mean	SD	Median	N>LOD	Mean	SD	Median
Antimony	20	0	-	-	-	0	-	-	-
Arsenic	7	2	7.84	0.04	7.84	6	8.43	1.02	8.17
Barium	45	12	148	59.6	143	23	195	54.8	180
Cadmium	12	0	-	-	-	0	-	-	-
Cesium	35	0	-	-	-	0	-	-	-
Chromium	22	8	28.3	10.9	23.7	18	37.9	9.85	35.9
Cobalt	90	0	-	-	-	0	-	-	-
Copper	13	22	34.1	10.9	31.7	24	31.4	12.7	26.8
Gold	9	0	-	-	-	0	-	-	-
Lead	8	21	17.3	9.3	13.4	24	52.4	68.3	24.2
Manganese	50	22	233	50.9	223	24	237	87.9	225
Mercury	9	0	-	-	-	0	-	-	-
Molybdenum	3	4	4.44	1.87	3.49	5	3.70	0.55	3.60
Nickel	30	1	33.2	0.00	33.2	11	35.9	3.93	36.4
Palladium	12	0	-	-	-	0	-	-	-
Rubidium	3	22	21.6	5.36	21.3	24	35.0	14.9	33.7
Scandium	75	4	130	39.1	112	1	78.7	0	78.7
Selenium	4	0	-	-	-	0	-	-	-
Strontium	3	22	71.8	35.6	63.5	24	48.5	17.2	44.6
Tellurium	35	1	38.4	0.00	38.4	6	39.8	4.65	37.9
Thorium	4	7	4.71	0.39	4.51	22	6.19	1.72	5.85
Tin	20	0	-	-	-	1	20.5	0	20.5
Tungsten	30	0	-	-	-	0	-	-	-
Uranium	4	3	4.21	0.25	4.04	12	4.67	0.96	4.35
Vanadium	25	22	38.0	7.01	37.7	24	56.6	17.4	58.9
Zinc	10	22	105.8	43.6	94.8	24	187	187	104
Zirconium	4	22	225	99	211	24	307	74.8	295

^*Limits of detection (LODs) of the Niton XL3t GOLDD+ XRF analyzer were obtained from the device manufacturer (Thermo Scientific 2010). Shown here are LODs of the Niton XL3t GOLDD+ XRF analyzer operated in Soil Analysis Mode with an analysis time of 60-seconds per filter for a sample consisting of SiO₂ with Ca/Fe, which is intended to represent a typical soil matrix.

Reference concentrations for soil metals are shown in Table 4. The mean concentrations of antimony in garden bed and ground soils were higher than the background level for the conterminous U.S. (0.8 mg/kg) and Texas (1.0 mg/kg) but below the respective EPA SSL (31.0 mg/kg). Soil arsenic concentrations were comparable to background levels for the U.S. (6.4 mg/kg), Texas (5.9 mg/kg), and Houston (2.9 mg/kg) but in excess of the corresponding EPA SSL (0.4 mg/kg). Barium concentrations in ground and garden bed soils were lower than the EPA SSL (5,500 mg/kg) and the background levels for the U.S. (518 mg/kg) and Texas (300 mg/kg), though garden bed soils had a lower mean barium concentration than Houston (120.9 mg/kg). Concentrations of cadmium were comparable to the U.S. background level (0.3 mg/kg) and well below the EPA SSL (78.0 mg/kg). Chromium concentrations were below the EPA SSL (390 mg/kg) and similar to background U.S. (36.0 mg/kg) and Texas (30.0 mg/kg) levels but higher than Houston (11.5 mg/kg). Mean lead levels in garden bed soils were comparable to the background level across Texas (15 mg/kg) and lower than that of the U.S. (25.8 mg/kg), while ground soils contained similar concentrations of lead to Houston (60.2 mg/kg). Ground and garden bed soils had mean lead concentrations that were lower than the EPA SSL of 400 mg/kg. Average nickel in garden bed soil was comparable to the background level for Houston (8.0 mg/kg) but lower than the U.S. (17.7 mg/kg) and state (10.0 mg/kg) background levels. Although the mean nickel concentration in garden ground soil exceeded city, state, and national background concentrations, it did not exceed the U.S. SSL of 1,600 mg/kg. The average vanadium levels in ground and garden bed soils were higher than the background concentration for Houston (11.7 mg/kg) and lower than that of the U.S. (60 mg/kg); however, mean vanadium concentrations in garden soils were below the corresponding EPA SSL of 550 mg/kg. With respect to zinc, mean concentrations in garden bed and ground soils were lower than the background level for Houston (279.8 mg/kg) but higher than the U.S. (66 mg/kg) and Texas (30 mg/kg) background levels. Average zinc levels in garden bed and ground soils did not exceed the EPA SSL for zinc (23,000 mg/kg).

Table 4.

Reference concentrations (mg/kg) for soil metals in Houston, TX, and the U.S.

Element	U.S. SSL (U.S. EPA 2019)	Mean Conterminous U.S. Soil (Smith et al. 2013)	Median TX Soil (TCEQ n.d.)	Mean Houston Soil (Chu 2020)
	mg/kg
Antimony	31	0.8	1	-
Arsenic	0.4	6.4	5.9	2.9
Barium	5,500	518	300	120.9
Cadmium	78	0.3	-	-
Calcium	-	-	-	-
Chromium	390	36	30	11.5
Copper	-	17.9	15	58.2
Iron	-	-	15,000	6,403.1
Lead	400	25.8	15	60.2
Manganese	-	612	300	220.6
Molybdenum	-	1.04	-	-
Nickel	1,600	17.7	10	8.0
Potassium	-	1.46 wt%	-	-
Rubidium	-	66.2	-	-
Strontium	-	159	100	-
Thorium	-	8	9.3	-
Tin	-	1.6	0.9	-
Titanium	-	0.27 wt%	2,000	-
Vanadium	550	60	50	11.7
Zinc	23,000	66	30	279.8
Zirconium	-	-	-	-

Discussion

Because soil testing is not routinely performed at community garden sites, it is often difficult to determine the extent to which the soil in community garden sites may be contaminated (Hunter et al., 2019; Ramirez-Andreotta et al., 2019). Researchers seeking to support local gardeners have reported a need for information and guidance pertaining to soil testing and remediation strategies since there can frequently be discordance between perceptions of risk and potential for contamination (Hunter et al., 2019; Kim et al., 2014; Wong et al., 2018). Citizen science may present opportunities to improve understanding and acceptability of remediation (Taylor et al., 2021). In the present study, concerns expressed by gardeners with respect to soil contamination were minimal, and no specific contaminant type (e.g., metals, pesticides, microbial pathogens) was stated by any respondent. Community gardeners would be willing to seek soil testing if they suspected their soil to be contaminated. Although willingness to use certain physical interventions and behavioral changes to reduce exposure to potential soilborne hazards differed, the survey respondents were largely willing to use common interventions that are known to be effective in reducing risk, indicating an opportunity to provide educational and outreach services to community gardeners in Houston Texas, through community/university partnerships.

With the exception of arsenic, the mean concentration of metals in ground and garden soils collected from Houston, Texas community gardens did not exceed the EPA’s SSLs for residential locations and ingestion-dermal exposure (U.S. EPA, 2019). Some of the soil samples collected from the grounds adjacent to raised community garden beds had elevated concentrations of soilborne hazards, including lead. The highest concentration of lead in garden ground soil was 328.7 mg/kg, which is comparable to prior reports of urban garden soils. For example, XRF analysis of soil collected from 23 raised garden beds in urban Massachusetts revealed an average lead concentration of 336 mg/kg (Clark et al. 2008). However, the variability in soil metal concentrations between existing community garden sites in Houston, Texas highlights the need for site-specific soil testing. Each of the garden beds sampled were raised beds, which have been shown to be effective in reducing contamination risk (Clark et al., 2008). Given that ground soils had consistently higher concentrations compared to soils collected from garden beds, future food garden developments should implement raised garden beds when possible, which have been shown to be effective at reducing the risk of contamination in other urban settings (Mitchell et al., 2021).

This study has several important limitations. The demographic distribution of the study population is not representative of the larger Houston community. For example, no respondents identified as Hispanic or Latino in spite of 44.8% of Houston’s population being Hispanic or Latino (U.S. Census Bureau, 2020). Although the Houston gardening community has not been well-characterized, this disparity could be attributed to the limited sample size or to the lack of representativeness of those engaged in community gardening in the City of Houston, White community gardeners are overrepresented in other studies of gardens in U.S. cities such as St. Louis, MO and Atlanta, GA (Wong et al. 2018; Hunter et al. 2019), which may provide opportunities for targeted engagement of minority communities with community gardening, such as in citizen science or community-extension partnerships (Mitchell et al., 2014; Taylor et al., 2021). In addition, the sampling design employed in this study does not capture concentration gradients across the garden site. In spite of these limitations, results obtained are adequate to begin community engagement and research translation efforts related to potential exposures to metals in garden soils. It is important to note that these values represent estimated concentrations are intended only for characterization and screening. Thus, confirmatory soil testing is advised even for the sites that were screened, and all proposed community garden sites should also be tested in advance of development to ensure that local soil conditions do not present excess risk of produce contamination.

Community gardens are an essential component of the local food system in Houston, Texas. Although limited in scope, the information from this study can be used to tailor research engagement programming with the gardening community of Houston, Texas. Findings from this survey may be used to direct the development of new practices and technologies to address potential garden contamination and to identify opportunities for outreach and engagement with community gardeners.