Evaluation of High Tunnel Systems for Spring Organic Lettuce Production in Georgia

Study 1: Evaluation of high tunnel systems for spring organic lettuce production in Georgia

Site Characteristics and High Tunnel Design:

A comparison of lettuce (Lactuca sativa var capitata L. and Lactuca sativa var. logifolia Lam.) production in high tunnel production compared to field production was conducted during 2015 and 2016 at the Durham Horticulture Farm located in Watkinsville, GA). The USDA plant hardiness zone for the site is 8a. The project site has been USDA certified organic since 2012 and all agricultural production methods were performed under these guidelines.

Two commercial-size gothic-shaped high tunnels (Atlas Greenhouse Inc., Alapaha, GA) (96 ft x 30 ft) and an adjacent field area were used for the study (150 ft x 30 ft). High tunnels were constructed in an east-west orientation to be perpendicular to the prevailing winter winds at the site. Bows were spaced every 6 ft. The high tunnels had inflated double polythene film roofs comprised of 6 mil plastic. The end walls were comprised of ~0.3 inch thick polycarbonate. Automated 6 ft tall z-lock drop-down side curtains constructed from 12 mil weave fabric were utilized.

Three planting dates (PD) (2015: 3/12, 4/2, 4/23 and 2016: 3/3, 3/24, 4/14) were selected by studying the last 20 years of last frost dates at the project site. Lettuce cultivars were selected after consulting with local organic farmers and several seed companies for recommendations for heat tolerant and/or top performing cultivars for the region. A few cultivars (‘Freckles’ and ‘Red Rosie’) were included for visual interest. The six butterhead cultivars included were: ‘Red Cross’, ‘Sylvesta’, ‘Adriana’, ‘Skyphos’ (Johnny's Selected Seeds), ‘Pirat’, and ‘Mirlo’ (High Mowing Seeds). The six romaine cultivars included were: ‘Salvius’, ‘Coastal Star’, ‘Green Forest’, ‘Red Rosie’ (Johnny's Selected Seed), ‘Super Jericho’ (Harris Seed), and ‘Freckles’ (High Mowing Seeds). In addition, farmer’s advice for fertilizer regimes, crop management, high tunnel management, etc. were also used to inform our decisions and management practices.

A total of eight raised beds were prepared in each high tunnel and field. Beds were 28 in wide x 8 in tall. Only six beds were used for the experiment. The two beds parallel to the side-walls in the high tunnel or the lateral edges of the field plots were designated as guard rows to minimize differential effects from environmental factors such as wind and light. Fertilizer applications were based on soil sample results taken one month before the lettuce crop was planted. A plot was harvested when more than 75% of the plants were judged to have firm, mature, marketable size heads or when more than 25% of plants demonstrated signs of bolting, tip burn, or other defects. Judgments about marketable size were made from comparisons to lettuce for sale at local farmers markets and grocery stores

The experiment was a split-split plot design with whole plot factor consisted of the growing system type (i.e., high tunnel or field), the split-plot was one of three planting dates, and the split-split plot was cultivar type. The experimental unit was 10 plants per plot. Two of the six experimental beds were randomly assigned to planting date 1 (PD1), planting date 2 (PD2) or planting date 3 (PD3). Planting date treatments were three weeks apart from each other (Table 1). The anticipated first planting date was delayed by one week in 2015, followed by planting date two and three. The field bed preparation was delayed for planting date one due to multiple heavy rain events. It was a drier spring in 2016 and transplanting was carried out according to the original target planting dates. Lettuce seedlings were planted in two rows per bed and in a staggered arrangement with 12 in between-rows and 10 in within-row. One guard row per bed was planted at the PD1 date and the other at PD3 date in order to have the continuous crop growth throughout the research period.

The majority of the season, irrigation was administered via drip tape every other day but this depended on the current weather and soil moisture conditions. Automatic drop-down side curtains were set to close at 50 ± 2°F and the end-walls were closed manually. Side curtains were closed when rain or winds (>15 mph) were predicted; however, to maintain the inside temperature end-walls were kept open. When the air temperature was predicted to be ≤32°F, intermediate weight row covers (0.6 oz yd^-1) were used for frost protection in both the high tunnel and field systems (Gro-Guard UV Row Cover #20). Row covers were draped over wire hoops spaced every 6 ft and edges held in place with weighted sand bags. Weeding was done several times throughout the growing period with hand tools.

Environmental monitoring stations were located in each high tunnel and field block (Em50 Digital/Analog Data Logger; Decagon Devices Inc). Each station included sensors to measure air temperature and relative humidity, photosynthetic photon flux (PPF), soil temperature and leaf wetness (LWS). Average values for each parameter were recorded at hourly intervals. Environmental monitoring stations were located within experimental beds and they were also covered with row covers on nights when temperatures were predicted to be ≤32^oF.

Integrated pest management (IPM) scouting was carried out, twice each week. In 2015, fire ants (Solenopsis spp.), aphids (Aphidoidea spp.), and armyworms (Spodoptera exigua) were the major pests along with the disease, lettuce drop (Sclerotinia sclerotiorum). Two spinosad-based products were used to manage fire ants including “Entrust” naturalyte insect control and “Come and get it” fire ant killer bait. Bacillus thuringiensis (DiPel DF) was used once to manage armyworms. Plants infected with lettuce drop (S. sclerotiorum) were removed along with the surrounding surficial soil and visible sclerotia were disposed in the trash. Dead or severely damaged plants, generally from fire ants or S. sclerotiorum infection, were recorded and replaced up to 2 weeks from their original transplanting date.

The main factors of interest in this study were a comparison between two production systems (i.e., high tunnel vs. field) and three planting dates (i.e., early PD1, mid-PD2, late PD3). A set of analyses to assess micro-environment differences were also completed. Continuous response variables included: air temperature, soil temperature, relative humidity, LWS, and PPF. The aforementioned variables were all approximately normally distributed so the statistical analysis was carried out using a mixed effects ANOVA model (SAS Institute, Cary, NC). These variables were not transformed. Tukey’s mean separation method with 95% confidence level was used to determine if they were significantly different between the two production system types or three planting dates. Each year was analyzed separately in order to account for different weather patterns. Statistical analysis was not conducted on the average monthly micro-environment data. These averages were presented only in order to generalize comparisons between years and between the early to late spring at our research site and other regions.

Study 2: Consumer lettuce preferences and taste evaluation

The goal of this survey was to identify the purchasing habits, visual preferences and taste ratings of organic head lettuce among shoppers at a local farmers market. Multiple lettuce cultivars were represented in the visual and taste evaluation portion of the study. These cultivars were selected based on their strong performance (i.e., highest marketable yields) in the concurrent production research trial or their unique visual appeal. ‘The Athens Farmers Market’ located in Athens, GA was the location of the study. Approximately, 1,000 to 1,500 people visit the Athens Farmers Market every Saturday morning (open 8am- noon) from early April to mid-December (personal communication). One-hundred and twenty-six surveys were completed on 7 May 2016.

The top three performing cultivars (i.e., highest marketable yields) of both butterhead and romaine cultivars from our 2015 spring lettuce trial were selected for inclusion. Butterhead lettuce cultivars included: ‘Adriana’, ‘Sylvesta’ and ‘Skyphos’. Romaine lettuce cultivars included: ‘Green Forest’, ‘Salvius’ and ‘Super Jericho’. The butterhead cultivar ‘Red Cross’ and the romaine cultivar ‘Red Rosie’ were included due to their unique red leaf tissue color.

All the samples were harvested from the 2016 high tunnels PD2 lettuce plots 2 to 4 days before the survey date. All samples were stored, cleaned and prepared in a similar manner. Volunteer survey participants were solicited by project staff in person with a verbal request based on an approved recruitment script. Only individuals greater than 18 years old were allowed to participate.

First, all participants were asked ten general questions about their consumer and buying preferences regarding lettuce. Second, participants were randomly assigned to evaluate either butterhead or romaine lettuce. Each participant was given four, lettuce samples to evaluate. They were asked to rank each sample on a Likert scale of 1 to 5 for sweetness, bitterness, overall taste, and crunchy texture. Next, participants were asked to rate on a Likert scale of 1 to 5, four full-sized lettuce heads (butterhead or romaine) for color and visual appearance. And lastly, participants were asked to provide an overall rating of each lettuce sample they evaluated based on taste, texture, and color. Selected demographic information was also asked including: age, household size and annual household income. A $5 token, which could be used as cash at the ‘Athens Farmers Market’ was given to each participant.

Data about demographics and general lettuce buying and consumer were combined for the butterhead and romaine lettuce surveys. Six questions that evaluated taste and visual appearance were analyzed separately for butterhead and romaine lettuce. The mean and standard deviation were calculated for questions with quantitative categories. The frequency and percentage of each category was calculated for questions with qualitative choices. And the mean was calculated for each category associated with question #10 that asked participants to rank which factors are most important when they consider buying lettuce on a scale of 1 to 6 (with 1 = most important and 6 = least important). Therefore a mean close to 1 was considered a stronger factor; a mean close to the 6 was considered a weaker factor when buying lettuce. The frequency and percentage of total responses was generated for each qualitative taste or visual assessment including indicators of sweetness, bitterness, overall taste, crunchy texture, color and overall rating for each assigned sample. Paired T-tests were used to identify differences among consumer preferences within the butterhead group and within the romaine group. Frequencies were also generated for responses to selected questions by demographic group.