This is data from a multi-site field trial of overwinter hairy vetch cover cropping in high tunnels. Data was collected for three years at each site. Data in the set includes hairy vetch productivity, yield, and marketability of the subsequent tomato crop, and soil health outcomes including soil nitrate, microbial biomass C and N, total C and N, enzyme activity, potentially mineralizable N, and permanganate oxidizable C. Data has been cleaned to exclude a site-year in which hairy vetch failed to overwinter, and includes metadata.

HT plots were arranged in a randomized complete block design, dividing replicate blocks among tunnels to provide four replicates at each site. Treatments were applied in rotation with summer tomato production and consisted of: 1) a fall-planted, overwintering hairy vetch cover crop, terminated in spring and incorporated to provide partial N credit toward summer cash crop fertility (+vetch), supplemented with application of commercially available pelleted organic fertilizer, and 2) a standard-practice control with no cover crop over winter, receiving a full-rate spring application of pelleted organic fertilizer at standard rates (-vetch). The standard-practice winter control consisted of cool-season spinach (Spinacia oleracea L.) production in KS and KY, and fallow in MN. All site management practices followed USDA National Organic Program standards. Vetch and tomato planting dates were selected to reflect typical commercial grower practices in each region (full table of seeding dates is available in the associated publication). Vetch establishment failed in MN in fall 2017, possibly due to insufficient irrigation. Consequently, no biomass data was collected in 2018, and an additional trial year (2020) was conducted at this site.

Winter management

A hairy vetch (variety not stated, Albert Lea Seed, Albert Lea, MN) cover crop was established in +vetch plots each fall by hand broadcasting at a rate of 33.63 kg ha^-1 and incorporating with a hard rake to improve seed to soil contact. Seed was inoculated with N-Dure inoculant (Verdesian Life Sciences, NC) by moistening seeds with a 1:4 sucrose solution, mixing with the recommended rate of inoculant, and allowing seeds to air dry overnight. Vetch was watered as needed to maintain a moist soil surface until full germination, then periodically to provide approximately 2.54 cm of water weekly until water supplies had to be turned off for winter in MN and KS; irrigation continued to be supplied to all plots in KY. In MN, floating row cover (AgroFabric Pro50, Tifton, GA) was installed prior to predicted nighttime temperatures falling below -1 °C, layered above thin metal hoops to keep fabric approximately 0.5 m above the soil surface and limit contact with vetch leaves, minimizing risk of frost injury at contact freeze points. In KS and KY, no row cover was used.

At MN, -vetch plots remained fallow following removal of harvested tomato plants. In -vetch plots at KY and KS, winter spinach (‘Corvair’, Johnny’s Selected Seeds, Winslow, ME) was direct seeded at 32,280 seeds m^-2 using a push-type seeder (Jang JP-1, Jang Automation Co, South Korea) at the time of HT vetch planting; no additional fertilizer was applied to spinach. Spinach was harvested repeatedly as leaves reached marketable size until bolting occurred, and was terminated by mowing and tilling at the time of vetch termination. Total harvested fresh weight of spinach at KS was 8.5 kg m^-2 in Y1 and 2.4 kg m^-2 in Y2, and was not measured in Y3; total harvested fresh weight at KY was 6.9 kg m^-2 in Y1, 8.3 kg m^-2 in Y2, and 5.5 kg m^-2 in Y3.

To characterize the soil thermal environment created by the HT during this period, vetch plots inside the HT and in an adjacent outdoor field at the MN site were instrumented with soil temperature data loggers (Decagon EM50, Decagon Devices Inc., Pullman, WA). A logger was inserted to 10 cm depth in each plot at vetch planting and removed at vetch termination; data were not collected during the tomato growing period. Although HT soil did freeze briefly overwinter (daily maximum temperatures below 0 ℃), soil temperatures inside the HT were both higher and more variable than soil temperatures in the outdoor field, and spring soil warmup occurred much earlier inside the HT.

Cover crop biomass

Vetch biomass was sampled immediately prior to cover crop termination at all sites, and total biomass and C:N ratios were quantified on a per hectare basis. Two quadrats of a known size (MN and KY: 0.25 m², KS: 0.1 m²) were randomly tossed into each plot, avoiding a 30 cm edge border. All biomass in quadrat was clipped at the soil surface, pooled for each plot, and sorted between weeds and vetch. Biomass was oven dried at 65 °C, ground to 1 mm, and analyzed for percent C (%C) and percent N (%N) using a Vario PYRO cube combustion analyzer (Elementar, Langenselbold, Germany; Tabatai & Bremner, 1990). Vetch biomass-N content was not analyzed for Y2 samples from KS due to shipping issues; for that site-year, we estimated biomass-N based on biomass productivity and the average percent N content of biomass from the same site analyzed in Y1 and Y3.

Summer management

Prior to tomato planting, vetch was terminated by mowing with a walk-behind lawnmower or walk-behind tractor (BCS America, OR) fitted with a flail mower attachment, allowed to dry on the surface for approximately one week, and incorporated using a tiller attachment on a walk-behind tractor. Beds were then fertilized with pelleted organic fertilizer composed of meat, bone, feather, and blood meal (NatureSafe 8-5-5, Darling Ingredients, Irving, TX). To meet the recommended rate of 112 kg N ha^-1, pelleted fertilizer was applied in -vetch plots at 1400 kg ha^-1. In +vetch plots, an N credit was calculated from biomass samples described in Section 2.4 and deducted from the pelleted fertilizer application rate. This credit was calculated as vetch biomass in kg ha^-1 × (percent biomass-N / 100) × 0.5, based on the assumption that approximately 50% of biomass-N becomes available in the season following incorporation (full table of fertilizer application rates is available in the associated publication) (Andrews et al., 2012). No N credit was attributed to belowground vetch biomass.

Beds were covered with black polyethylene or woven poly mulch, with irrigation provided via a single line of drip irrigation placed underneath the mulch. Determinate tomatoes adapted for HT conditions (‘BHN-589’, Johnny’s Selected Seeds, Winslow, ME) were transplanted at a spacing of 0.46 m apart within-row, and 1.5 m, 0.9 m, and 1.2 m between rows in KS, KY, and MN, respectively. Plants were pruned to remove suckers below the first fruit cluster and trellised using the stake-and-weave system. Fields were scouted regularly to manage pests and disease. Minimal supplemental in-season minerals and fertility were provided via fertigation based on results of foliar nitrate testing conducted on a bi-weekly basis after first flowering using a hand-held nitrate meter (Laqua Twin nitrate meter, Horiba LTD, Kyoto). An organic-approved fungicide (EF400, AnjonAG Inc., Weatherford, TX) was applied on a 7–10 d interval if fungal disease was detected.

Tomatoes were harvested once per week when fruits reached the breaker stage of ripeness or greater. Yield was recorded as number and fresh weight of marketable and unmarketable fruit based on USDA grading standards, with USDA grade #1 and #2 fruit classified as marketable (USDA Tomato Grading Standards, 1997).

Soil sampling

Soil samples were collected at two temporal scales: annually, prior to vetch planting each fall to capture changes across years resulting from the cover crop system; and at three additional timepoints to identify within-season fluctuations: at legume termination (Termination), and approximately two and four weeks post-termination (Termination +2w and Termination +4w). At each timepoint, five soil cores were collected from each plot to a depth of 15 cm and aggregated into a single sample per plot. A portion of each field-moist sample was sieved to 2 mm and subdivided into subsamples stored at 4 °C (for fresh-soil analysis) and -80 °C (for subsequent microbial analysis); the remaining sample was dried at 35 °C to constant weight and ground to 2 mm for dry-soil analyses.

Soil C and N analyses

Soil total C and N concentrations were measured on samples from the fall timepoints via combustion, as described above for cover crop biomass (Section 2.4). Soil mineral N and POX-C concentrations were determined for soils from all sampling timepoints via colorimetric analysis as described in Perkus et al. (2022). Briefly, for mineral N, fresh soil samples were shaken with 40 mL 1 M KCl for 1 h at 240 rpm, then filtered. For POX-C, dry soil was shaken and incubated in a 0.2 M KMnO₄ solution, after which a diluted aliquot of the supernatant was analyzed for absorbance at 550 nm. PMN was determined using a 7 d anaerobic incubation as described by Drinkwater et al. (1997), with the modification that extractions were conducted using 1 M rather than 2 M KCl. Briefly, 10 g fresh soil saturated with 10 mL DI H₂O was held in the dark at 37 °C. After 7 d, samples were shaken with 30 mL 1.33 M KCl for 1  h at 240 rpm, then filtered with Whatman No. 1 filter paper. To determine extractable NO₃–N and NH₄–N concentrations, microplate colorimetric analyses on soil extracts were performed using a SpectraMax 190 microplate reader (Molecular Devices, LLC, San Jose, CA, USA; Doane & Horwáth 2003; Sinsabaugh et al., 2000). Soil PMN was calculated as the difference between NH₄–N concentration before and after incubation. Samples from the Y1 termination+2w timepoint and from Y4 in MN were excluded from PMN analysis due to flawed incubation conditions.

Microbial biomass and enzyme activity

Microbial biomass C and N concentrations were quantified using the chloroform fumigation-extraction method (Vance et al., 1987; Gregorich et al., 1990; McDaniel et al., 2014). Briefly, 10 g soil per plot was shaken in duplicate with 40 mL 0.5 M K₂SO₄ for 4 h. One duplicate received 0.5 mL of chloroform (CHCl₃) to lyse microbial cells, while the other served as the non-fumigated control; both were then filtered. Extracts (3 mL) were mixed with 17 mL 0.01 M HCl and analyzed for total organic C and total N concentrations on a Shimadzu analyzer (Kyoto, Japan). Potential enzyme activity was determined by the fluorometric method described by German et al. (2011) and Johnsen & Jacobsen (2008). Frozen soil samples were used to evaluate four extracellular hydrolytic enzymes: β-glucosidase, cellobiohydrolase, N-acetylglucosaminidase (NAGase), and acid phosphatase. Briefly, 0.5 g soil was homogenized in 50 mL Tris buffer using a bath sonicator for 5 min, then soil slurries were dispensed into 96-well plates containing 50 µL of a 1500 µM 4-methylumbelliferone (MUB)-linked substrate per well. Plates were incubated for 1 h at room temperature, and fluorescence of released MUB was measured on a microplate reader (Biotek, USA) set at an excitation wavelength of 365 nm, and emission wavelength of 450 nm. Enzyme activities were calculated from standard curves generated with known MUB concentrations