Advertisement
Research Report

Grow Tube Material Selection Influences Thermal Microenvironments and Cold Hardiness Risk in Dormant Grapevines

Worasit Sangjan, Elizabeth C. Gillispie, M. Jacob Schrader, Michelle M. Moyer, Madison Shaw, Devin A. Rippner
Am J Enol Vitic.  2026  77: 0770018  ; DOI: 10.5344/ajev.2026.25051

Abstract

Background and goals Grow tube material selection plays a crucial role in shaping thermal microenvironments around young grapevines; however, quantitative guidance for selecting and removing grow tubes during winter months in cold climates remains limited. This study investigated how grow tube material and installation influence internal microclimates during dormancy, and the resulting implications for grapevine cold hardiness.

Methods and key findings Temperatures were monitored across five grow tube treatments (paper and plastic materials installed at two configurations [buried and raised], plus a no-tube control) over two dormancy seasons. These in-field temperature data were used as inputs to a temperature-based cold hardiness model to simulate responses for four Vitis vinifera cultivars. Solar radiation drove material-dependent thermal differences: plastic tubes, acting as thermal storage systems, produced warming effects of up to ~3.40°C during cold, sunny afternoons, with buried plastic installations generating the highest internal temperatures. At night, temperatures in all tubes returned to near-ambient levels, resulting in larger diurnal temperature swings in plastic tubes than in paper or no-tube controls. Modeling showed that plastic materials had warmer midwinter hardiness thresholds (~0.10 to 0.30°C) and more rapid spring deacclimation than paper-based materials. In contrast, paper tubes enhanced chilling accumulation and maintained thermal profiles aligned with natural dormancy conditions.

Conclusions and significance Overwintering vines in paper tubes without removal minimally affects dormancy temperatures and preserves chilling conditions similar to those in natural conditions. In contrast, overwintering in plastic tubes, particularly when partially buried, creates warmer daytime conditions that reduce chilling accumulation and may raise the risk of premature deacclimation in spring. This study provides insight into selecting grow tube material under continuous overwinter installation, balancing protective benefits with physiological risks during winter dormancy.

Introduction

Successful establishment of young grapevines (Vitis vinifera) is essential for long-term vineyard productivity and resilience. During the establishment phase, vines are particularly susceptible to abiotic and biotic stressors, including herbicide drift, animal browsing, and extreme temperatures, as well as physical factors such as mechanical injury (Moyer et al. 2024). Grow tubes are widely used in vineyards to mitigate these risks by providing a physical barrier around the developing trunk (Dami et al. 2005, Fiola 2021a).

Grow tubes, an elongated shelter often made from plastic or biodegradable materials such as waxed paper, modify the vine’s immediate microclimate by increasing temperature and humidity around the vine. This greenhouse-like environment can stimulate early shoot elongation, while the presence of grow tubes can improve weed management strategies and simplify overall vineyard floor maintenance (Olmstead and Tarara 2001, Fiola 2021b). While grow tubes are typically removed before winter to help acclimatize young grapevines to colder temperatures, labor constraints may hinder timely removal. Further, as grow tubes are often reinstalled yearly for 2 to 4 yr in young vineyards to support extended trunk development and provide protection during environmental stress, growers may choose to forgo removal depending on labor availability and perceived risk (Tarara et al. 2014).

Overwintering dormant vines in grow tubes, however, may compromise bud cold hardiness in colder regions, as elevated internal temperatures can delay cold acclimation in the fall or promote premature deacclimation during the spring dormancy transition, thereby increasing the risk of cold injury during fluctuating winter temperatures (as reported at https://smallfruits.org/2020/10/some-vineyard-tasks-to-prepare-for-winter/). High humidity and reduced airflow may also increase disease pressure. Although grow tubes are often assumed to buffer vines against thermal extremes, their effects on bud cold-hardiness dynamics particularly under cold-prone field conditions remain unclear (Olmstead and Tarara 2001, Gale and Moyer 2017).

Cold hardiness in grapevines develops through two dormancy phases: endodormancy and ecodormancy. Endodormancy, triggered by shortening day lengths and declining temperatures, involves internal physiological changes that enable tissue to survive freezing temperatures. This is followed by ecodormancy, where buds remain dormant due to cold exposure but become responsive to warming (Lavee and May 1997, Londo and Kovaleski 2025). These transitions are highly temperature-sensitive and vary by cultivar. Exposure to artificially warm microclimates such as those created by grow tubes may interfere with bud acclimation or promote premature spring deacclimation, thereby elevating the risk of winter injury (Ferguson et al. 2014, Cragin et al. 2017, Campos-Arguedas et al. 2026).

While grow tubes are designed to modify the microenvironment around young grapevines, their influence on the microenvironment during dormancy (late fall to early spring) has not been quantified under real-world conditions. Much of the existing research focuses on vine morphology or long-term growth outcomes, particularly during the growing season, with limited attention to short-term temperature fluctuations in the dormancy period that may significantly affect cold-hardiness transitions (Cragin et al. 2017, Thomas et al. 2017, Fiola 2021a, 2021b, North et al. 2024). Physiologically important events such as early morning cooling or midday warming are rarely captured due to the coarse temporal resolution of previous studies (Bordelon and Blume 2000, O’Brien et al. 2025). Furthermore, few investigations have compared different tube materials or installation approaches under standardized environmental settings, limiting growers’ ability to balance protective benefits against potential cold-related risks during dormancy (Olmstead and Tarara 2001, Tarara et al. 2013). This lack of detailed empirical evidence constrains efforts to evaluate the trade-offs between thermal protection and physiological risk during the dormant season.

In this study, the overall goal was to investigate how grow tube material and installation influence the internal microclimate during grapevine dormancy, and the potential implications for cold hardiness. Specifically, we explored the extent to which different grow tube materials and installation configurations modified internal temperatures relative to ambient conditions and a no-tube control; whether these microclimate shifts influenced chilling unit and growing degree day (GDD) accumulation; and how the resulting temperature dynamics could affect spring deacclimation risk, as assessed using a temperature-based cold-hardiness model. By linking fine-scale temperature patterns to physiological risk factors, this study offers a foundation for evidence-based recommendations on grow tube deployment in colder regions of cool-climate viticulture.

Materials and Methods

Study area

The experiment was conducted in the newly established Washington Soil Health Initiative Research Vineyard at the Washington State University-Irrigated Agriculture Research and Extension Center (WSU-IAREC) in Prosser, WA (46°15′N; 119°43′W). The vineyard was planted in June 2023 to Cabernet Sauvignon (V. vinifera) on 1103P rootstock (Vitis berlandieri × Vitis rupestris), with 2.74 m row spacing, 1.52 m vine spacing within rows, and drip irrigation. The soils were a tilled Warden silt loam (coarse-silty, mixed, superactive, mesic Xeric Haplocambids) and hilled to ~5 cm above the graft unions.

On-site climate conditions

The site has a semiarid continental climate (Köppen BSk), characterized by cold winters, warm autumns, and pronounced diurnal temperature variation (Beck et al. 2018). Weather data for the vine dormancy period were sourced from the WSU AgWeatherNet-Prosser station for 2023 to 2024 and from an on-site ATMOS 41 station (METER Group) for 2024 to 2025. A summary of seasonal conditions is presented (Table 1).

View this table:
Table 1

Summary of seasonal climate conditions during the vine dormancy period in Prosser, WA for 2023 to 2024 and 2024 to 2025. Weather data for 2023 to 2024 were obtained from the Washington State University AgWeatherNet-Prosser station and data for 2024 to 2025 were collected from an on-site ATMOS 41 weather station (METER Group).

Experimental design and treatments

A randomized complete block design with four replicates (blocks) was employed to assess the thermal effects of various grow tube materials and installation configurations during the late fall to early spring period (25 Oct to 30 March) of two consecutive seasons. In 2023 to 2024, grow tubes were installed around 1-yr-old vines and secured with bamboo stakes (Figure 1A). Each treatment replicate contained three vines, but only one temperature sensor was installed per treatment replicate. In 2024 to 2025, the same treatments were installed (one temperature sensor per treatment replicate) and secured with bamboo stakes but without vines, to avoid potential winter damage and loss of vines that were scheduled for field establishment in the subsequent growing season (Figure 1B).

Two photographs show vineyard rows with no grow tube, paper and plastic grow tubes, a temperature probe, and a HOBO data logger. The two photographs labeled A and B show an experimental setup among rows of leafless vines supported by stakes and wires. Panel A shows a person standing beside the vines and holding a grow tube. The labels No grow tube, Paper grow tube, and Plastic grow tube identify three treatments along the row. Panel B shows multiple vine rows fitted with grow tubes and irrigation lines. Two circular annotations identify a HOBO data logger attached near a vine and a Temperature probe positioned near the ground.
Figure 1

Experimental setup for monitoring grow tube microclimate in the vine dormancy period. (A) 2023 to 2024 installation, showing three treatments; (B) 2024 to 2025 installation, showing the sensor installation.

Treatments consisted of white waxed-paper grow tubes and beige double-walled plastic grow tubes (Valley Wide Cooperative). Paper tubes measured 77 cm in height × 10 cm in diameter, while plastic tubes measured 75 cm in height × 8 cm in diameter. Each tube type was installed either with the bottom buried ~5 cm below the soil surface (Buried condition) or raised ~10 cm above the soil surface (Raised condition), along with a no-tube control (Figure 1).

Temperature measurements-HOBO sensors

Air temperature inside each grow tube treatment and in the no-tube controls was recorded using external temperature sensors connected to HOBO Pro v2 data loggers (models U23-003 and U23-004; Onset Computer Corporation). Each block included four grow tube treatments plus one no-tube control (five total per block), with each treatment replicate monitored using a single temperature probe, resulting in 20 probes across the experiment. Probes were mounted 10 to 15 cm above the soil surface and ~2 cm from bamboo stakes to avoid contact with vines, stakes, or grow tube surfaces (Figure 1B). Data were recorded at 5-min intervals throughout both dormancy periods and were periodically downloaded using HOBOware Pro software (Onset Computer Corporation).

Data preparation

Temperature data from HOBO Pro v2 loggers (collected at 5-min intervals) were exported as CSV files and processed using Python 3 (ver. 3.9.6; Python Software Foundation 2025). Timestamps were standardized and aligned across sensors and each record was assigned a unique identifier for the treatment replicate. Missing values were imputed using replicate means to maintain consistent temporal resolution. On-site weather station data (collected at 15-min intervals from the WSU AgWeatherNet-Prosser station [2023 to 2024] and the on-site ATMOS 41 station [2024 to 2025]), including ambient air temperature, solar radiation, and wind speed, were merged with HOBO records by aligning timestamps to 15-min bins (tolerance of ±7.5 min on HOBO data) for direct integration (Figure 2, Supplemental Figures 1 to 2).

A workflow diagram outlines data collection, preparation, integration, and three approaches for analyzing grow tube temperatures. A workflow diagram has the columns Process and Description and progresses through Data collection, Data preparation, Data integration, and Data analysis, 3 approaches. Data collection uses two data sets: weather station data comprising ambient air temperature, solar radiation, and wind speed from on-site stations at 15-minute intervals; and HOBO grow tube data comprising air temperature inside each grow tube treatment and the no-tube control treatment at 5-minute intervals. For the grow tube and no-tube data, Data preparation standardizes timestamps and aligns them across sensors, assigns unique treatment-replicate identifiers, and imputes missing values using replicate means. Data integration aligns grow tube, no-tube, and ambient weather data into 15-minute bins and merges the data sets based on synchronized timestamps. Data analysis includes three approaches. Approach 1, Temperature differential analysis, first groups environmental conditions. Thermal phase classifies days into four phases using daily mean ambient air temperature and solar radiation thresholds. Time block divides each day into four diurnal blocks based on solar radiation patterns. The thermal metrics are delta T subscript a, the difference between the temperature inside a grow tube and the ambient air temperature at the same bin, and delta T subscript n, the difference between the temperature inside a grow tube and the average of the no-tube control treatment at the same bin. Approach 2, Cold hardiness modeling, estimates cultivar-specific cold tolerance using daily mean temperatures as inputs to the Ferguson et al. 2014 model. Approach 3, Chilling unit and growing degree day accumulation, estimates seasonal chilling units and heat accumulation using traditional growing degree days with a base of 10 degrees Celsius and cultivar-specific heating degree-days derived from the cold hardiness model.
Figure 2

Workflow summarizing data collection, preparation, integration, and the three analytical approaches (temperature differentials, cold hardiness modeling, and thermal accumulation) examined in this study.

Data and statistical analysis

Three analyses were conducted to evaluate the effects of the grow tube’s material and installation on the vine during dormancy and dormancy transitions (Figure 2).

Temperature differential

Temperature differential analysis assessed the thermal effects of grow tubes relative to both ambient air and no-tube controls. Two grouping strategies were used to classify environmental conditions: thermal phases and time blocks. Thermal phases were defined using daily mean ambient air temperature and solar radiation thresholds, and each day was classified into four categories based on the combined seasonal distributions. Time blocks were derived from diurnal solar radiation patterns to segment each day into distinct thermal periods (Figure 2). Both grouping strategies were based on data collected from the on-site weather stations.

Two thermal metrics were computed for each grow tube measurement. The first (Equation 1), ΔTn, represents the difference between the air temperature inside the grow tube (Ttube) and the average temperature of the no-tube control (Tno-tube), both measured using HOBO sensors, at the same timestamp: ΔTn=Ttube−Tno−tube Eq. 1

The second (Equation 2), ΔTa, represents the difference between the air temperature inside the grow tube (Ttube) and the ambient air temperature (Tambient) obtained from the onsite weather stations: ΔTa=Ttube−Tambient Eq. 2

ΔTa quantified how grow tube temperatures differed from overall atmospheric conditions. Because ambient air was measured at the weather station height rather than at the grow tube (HOBO sensor) height, ΔTa was used only as a general atmospheric reference. All treatment comparisons relied on ΔTn, which uses a no-tube control measured at the same height.

Cold hardiness modeling

Daily mean temperature data for each grow tube treatment were analyzed using the cold hardiness model described previously (Ferguson et al. 2014). A custom Python script was developed to reproduce the model’s structure and parameterization, allowing modeling of cultivar-specific acclimation and deacclimation responses. Unique models were developed for four grapevine cultivars with differing cold sensitivities: Chardonnay (V. vinifera), Concord (Vitis labrusca), Cabernet Sauvignon (V. vinifera), and Mourvèdre (V. vinifera). Continuous daily temperature inputs were generated by merging on-site ambient air temperatures from weather stations (1 Sept to 24 Oct) with treatment-specific temperatures from HOBO sensors inside grow tubes (25 Oct onward), then aggregating to daily mean temperatures to meet model input requirements.

Chilling unit and GDD accumulation

The effects of grow tubes on altering experienced chilling degree days and GDD accumulations were also assessed. Chilling degree day accumulation was estimated using a temperature-based chilling metric generated within the Ferguson cold hardiness model (Ferguson et al. 2014). Although the original Ferguson model does not explicitly output chilling degree days, the Python-based implementation retains the internal daily acclimation calculations, enabling the generation of a cumulative chilling-related thermal value that reflects temperature exposure during dormancy progression. GDD accumulation was quantified using two approaches: cultivar-specific heating degree days from this same cold hardiness modeling, representing thermal inputs that contribute to deacclimation based on cultivar-specific thresholds; and traditional GDDs calculated from grow tube temperature data using the daily average of maximum and minimum temperatures minus a 10°C base temperature, accumulated only when daily averages exceeded 10°C (McMaster and Wilhelm 1997).

Model diagnostics and post-hoc analysis

Data transformations and model selection

All statistical analyses were conducted in R (ver. 4.5.2; R Core Team 2025) using a framework tailored to each analysis. Multiple data transformations (original, logarithmic, square root, and cube root) were tested for each response variable and the optimal transformation was selected based on model fit using Akaike Information Criterion (lowest AIC value; Akaike 1974) and residual normality (Shapiro-Wilks test, α = 0.05). Final models were fitted using the selected transformations.

Statistical modeling approaches

Temperature differential metrics were analyzed using linear mixed-effects models (LMMs) via the ‘lme4’ package (Bates et al. 2015) with the bobyqa optimization method in R. Data from all treatments were analyzed together using unified transformation and modeling approaches, allowing direct comparisons across environmental conditions. Three modeling frameworks were applied: treatment × thermal phase; treatment × time block; and treatment × thermal phase × time block. Random effects included treatment replicates and measurement dates to account for repeated measures and temporal correlation.

Predicted cold hardiness values (°C) for each cultivar described above were examined separately for three dormancy phases: fall acclimation (Julian days 250 to 330), winter maintenance (Julian days 330 to 90 of the following year), and spring deacclimation (Julian days 35 to 90 for the 2023 to 2024 season; Julian days 50 to 90 for the 2024 to 2025 season). These phases were operationally defined based on the seasonal patterns of modeled cold hardiness (fall decline, midwinter plateau, and spring increase) and were applied consistently across treatments and seasons for statistical comparison. For dynamic phases (fall and spring), linear regression was applied within each treatment–replicate combination to estimate the rate of change in predicted cold hardiness (°C/day). Winter maintenance was characterized by computing the mean predicted cold hardiness for each treatment–replicate combination. Statistical analysis used two structures: treatment comparisons within cultivars; and cultivar comparisons within treatments, with separate models fitted for each structure. Dynamic phases were analyzed using LMMs, whereas winter values were analyzed using one-way analysis of variance (ANOVA) (Quinn and Keough 2002). Treatment-specific chilling unit and GDD accumulations were compared using cultivar-specific one-way ANOVAs, with analyses performed independently for each cultivar.

Model diagnostics

Model assumptions were evaluated separately for each modeling framework. After model fitting, assumptions for LMMs were assessed, including residual normality and homoscedasticity; for cold hardiness slope models, the normality of random effects was also evaluated. For ANOVAs, assumptions were tested using the Shapiro-Wilks test for residual normality (‘stat’ package) and Levene’s test for homogeneity of variance (‘car’ package). Diagnostic procedures were matched to each modeling structure: pooled diagnostics were used for unified models (temperature differential analysis), while cultivar-specific or treatment-specific diagnostics were applied for structure-specific models (cold hardiness and chilling/GDD analyses).

Model assumptions were evaluated using residual diagnostics appropriate to each model structure. For LMMs, residual normality and homoscedasticity were assessed; for ANOVAs, Shapiro-Wilks and Levene’s tests were applied. Corresponding diagnostic plots are provided (Supplemental Figures 3 and 4).

Post-hoc analysis

For all analyses, when treatment effects were statistically significant (p < 0.05), post-hoc pairwise comparisons were performed on estimated marginal means (‘emmeans’ package; Lenth 2023) using Tukey’s honest significant difference adjustment for multiple comparisons. Compact letter displays were generated from these Tukey-adjusted means using the ‘multcomp’ package to facilitate interpretation of group differences.

Results

Temperature data from both dormancy seasons were processed and aligned to 15-min intervals for analysis. Seasonal weather conditions differed modestly between the two dormancy periods (Table 1). Mean fall (September to November) temperature was higher in 2024 to 2025 than in 2023 to 2024 (11.7°C versus 10.5°C). Winters (December to February) in both seasons were cold, with mean temperatures near or below 2°C and an extreme minimum of −19.2°C recorded in 2023 to 2024. Spring (March) represented the warmest portion of dormancy, averaging 7.4°C in 2023 to 2024 and 8.9°C in 2024 to 2025. Precipitation was greatest during winter (up to 135 mm), whereas solar radiation was lowest in winter (~52 W/m2) and highest in spring, reaching mean values of 170 W/m2 in 2023 to 2024 and 136 W/m2 in 2024 to 2025. All code and data sets used in this study are publicly available (GitHub and Ag Data Commons).

Temperature differences

Data grouping

No significant differences were observed between years in daily mean temperature, diurnal temperature range, or solar radiation (Figure 3A to 3C; t-test, p > 0.1). Therefore, a unified set of thermal-phase and time-block thresholds was applied across both seasons.

A seven-panel figure compares seasonal temperature and solar radiation distributions, thresholds, and hourly patterns. A seven-panel figure labeled A through G presents environmental temperature and solar radiation data. Panel A contains box plots of daily mean air temperature in degrees Celsius for 2023 to 2024 and 2024 to 2025. Both medians are near 3 degrees Celsius; 2023 to 2024 has several low outliers extending to about minus 16 degrees Celsius and several high outliers near 14 degrees Celsius, while 2024 to 2025 ranges from about minus 8 to 17 degrees Celsius without visible outliers. A horizontal line marks the overall mean of 3.5 degrees Celsius. Panel B contains box plots of daily diurnal range in degrees Celsius. Both seasons have medians near 8 degrees Celsius and similar interquartile ranges, with values extending from approximately 1 to 20 degrees Celsius in 2023 to 2024 and 0 to 17.5 degrees Celsius in 2024 to 2025. A horizontal line marks the overall mean of 8.1 degrees Celsius. Panel C contains box plots of daily mean solar radiation in watts per square meter. Both medians are near 65 watts per square meter. Values extend from approximately 7 to 215 watts per square meter in 2023 to 2024, with outliers near 230 and 248, and from 0 to 198 watts per square meter in 2024 to 2025, with outliers from approximately 205 to 215. A horizontal line marks the overall mean of 75.1 watts per square meter. Panel D is a histogram of daily mean air temperature in degrees Celsius, with Count on the vertical axis. Most observations occur between approximately 0 and 10 degrees Celsius, peaking near 1 degree Celsius, with sparse values extending from about minus 15 to 17 degrees Celsius. Vertical reference lines mark Mean, 3.5 degrees Celsius; Median, 3.3 degrees Celsius; Q1, 0.7 degrees Celsius; Q3, 6.4 degrees Celsius; Cold dormancy, 0.7 degrees Celsius; and Warm dormancy, 6.4 degrees Celsius. Shading represents plus or minus 1 standard deviation. Panel E is a histogram of daily diurnal range in degrees Celsius, with Count on the vertical axis. Values extend from approximately 0 to 20 degrees Celsius, with the highest count near 8 degrees Celsius. Reference lines mark Mean, 8.1 degrees Celsius; Median, 8.1 degrees Celsius; Q1, 4.7 degrees Celsius; and Q3, 11.2 degrees Celsius. Shading represents plus or minus 1 standard deviation. Panel F is a histogram of daily mean solar radiation in watts per square meter, with Count on the vertical axis. Counts are greatest below approximately 60 watts per square meter and generally decline toward 250 watts per square meter. Reference lines mark Mean, 75.1 watts per square meter; Median, 64.3 watts per square meter; Q1, 32.1 watts per square meter; Q3, 100.6 watts per square meter; Low solar, 32.1 watts per square meter; and High solar, 100.6 watts per square meter. Shading represents plus or minus 1 standard deviation. Panel G is a dual-axis line graph with Hour of day from 0000 to 2400 on the horizontal axis, Air temperature in degrees Celsius on the left vertical axis, and Solar radiation in watts per square meter on the right vertical axis. Mean, minimum, and maximum temperature remain lowest before sunrise, rise through midday, and decline during the evening. Mean temperature rises from approximately 1 degree Celsius at 0600 to 7 degrees Celsius near 1400. Minimum temperature rises from approximately minus 19 degrees Celsius before 0800 to minus 14 degrees Celsius near 1400. Maximum temperature rises from approximately 13 degrees Celsius near 0700 to 25 degrees Celsius between 1400 and 1700. Mean, minimum, and maximum radiation remain near zero overnight, increase after sunrise, peak around midday or early afternoon, and return near zero by 2000. Maximum radiation reaches slightly above 800 watts per square meter, mean radiation reaches approximately 300 watts per square meter, and minimum radiation remains near or below approximately 20 watts per square meter. Vertical boundaries divide Morning warming, 0700 to 1100; Peak heat, 1100 to 1500; Afternoon cooling, 1500 to 1900; and Night period, 1900 to 0700.
Figure 3

Grouping of environmental conditions during dormancy across two seasons (2023 to 2024 and 2024 to 2025) using on-site weather station data. (A to C) Seasonal distributions of (A) daily mean air temperature, (B) daily diurnal temperature range, and (C) daily mean solar radiation. (D to F) Histograms showing distributions used to define thresholds for classifying cold versus warm days and cloudy versus sunny conditions. (G) Diurnal patterns of mean air temperature and solar radiation for defining four-time blocks.

Thermal phases were defined using thresholds based on the combined distribution of daily mean temperature and solar radiation (Figure 3D and 3F). Days with average temperatures below 3.5°C were classified as “cold,” and days with solar radiation ≥100 W/m2 were classified as “sunny,” resulting in four thermal phases: Cold Cloudy, Cold Sunny, Warm Cloudy, and Warm Sunny.

Within-day variability was further divided into four time blocks based on diurnal patterns of air temperature and solar radiation (Figure 3G): Morning Warming (0700 to 1100 hr), Peak Heat (1100 to 1500 hr), Afternoon Cooling (1500 to 1900 hr), and Night Period (1900 to 0700 hr), defined by inflection points in the average temperature and radiation curves.

Treatment effects

Thermal responses of grow tube treatments were evaluated using two metrics: ΔTn and ΔTa. Temperature differential data were log-transformed prior to fitting LMMs (Supplemental Tables 1 and 2). Treatment effects were significant for both ΔTn and ΔTa across all modeling frameworks (p ≤ 0.001; Supplemental Tables 1 and 2). Full-factorial (treatment × thermal phase × time block) LMMs also revealed significant interaction effects, indicating that the grow tube’s thermal responses varied across different environmental conditions and diurnal cycles.

For ΔTn, plastic treatments exhibited higher internal warming than paper treatments across most conditions (Figure 4). During Cold Sunny and Warm Sunny phases (Figure 4A), the Plastic Buried treatments showed the strongest warming, with increases of +1.97 to +2.26°C relative to the no-tube control. Across time blocks (Figure 4B), the greatest warming was observed during the Peak Heat time block, with Plastic Buried reaching +1.63°C relative to the no-tube control. In contrast, all other treatments across both thermal phases and time blocks showed much lower ΔTn values, with maximum increases of only ~+1.0°C relative to the no-tube control. The combined thermal phase × time block analysis (Figure 4C) revealed the largest ΔTn values during Cold Sunny × Afternoon Cooling, where Plastic Buried reached +3.40°C relative to the no-tube control. Buried treatments produced greater warming than Raised treatments, with differences of ~+0.50 to +1.50°C during sunny conditions.

A three-panel illustration uses bar graphs and heatmaps to compare grow-tube temperature differences by thermal phase, time block, and treatment. The three-panel illustration compares delta T subscript n on a transformed scale among Paper Buried, Paper Raised, Plastic Buried, and Plastic Raised treatments. Lowercase letters indicate statistical groupings, and error bars accompany the bars. Panel A contains four bar graphs for thermal phases, with transformed delta T subscript n on the vertical axis from 0 to about 3 and a temperature-difference legend from minus 2 to positive 2 degrees Celsius. In treatment order Paper Buried, Paper Raised, Plastic Buried, and Plastic Raised, Cold Cloudy has temperature differences of minus 0.185, minus 0.232, 0.048, and minus 0.131 degrees Celsius, with groups a, b, b, and b; Cold Sunny has minus 0.038, minus 0.210, 2.255, and 0.840, with groups a, b, c, and c; Warm Cloudy has minus 0.229, minus 0.151, 0.034, and 0.120, with groups a, a, b, and b; and Warm Sunny has minus 0.213, minus 0.525, 1.971, and 1.064, with groups a, b, c, and d. Panel B contains four bar graphs for time blocks and a temperature-difference legend from minus 1 to positive 1 degree Celsius. In the same treatment order, Morning Warming has minus 0.270, minus 0.318, 0.857, and 0.394 degrees Celsius, with groups a, b, c, and c; Peak Heat has minus 0.157, minus 0.413, 1.635, and 0.783, with groups a, b, c, and d; Afternoon Cooling has minus 0.152, minus 0.214, 0.642, and 0.215, with groups a, b, c, and c; and Night Period has minus 0.192, minus 0.191, minus 0.067, and minus 0.070, all in group a. Panel C contains heatmaps for Cold Cloudy, Cold Sunny, Warm Cloudy, and Warm Sunny, with Treatment on the vertical axis, Time block on the horizontal axis, and a temperature-difference legend from minus 2 to positive 2 degrees Celsius. Each cell presents the temperature difference in degrees Celsius, transformed value, and grouping letter. For Cold Cloudy, the Morning Warming, Peak Heat, Afternoon Cooling, and Night Period cells are: Plastic Raised, minus 0.026 and 2.704 b, 0.004 and 2.706 b, minus 0.125 and 2.697 b, and minus 0.146 and 2.696 ab; Plastic Buried, 0.262 and 2.723 b, 0.531 and 2.740 a, 0.137 and 2.715 a, and minus 0.030 and 2.703 a; Paper Raised, minus 0.213 and 2.691 c, minus 0.143 and 2.696 b, minus 0.255 and 2.688 b, and minus 0.220 and 2.691 b; and Paper Buried, minus 0.174 and 2.694 bc, minus 0.023 and 2.704 b, minus 0.199 and 2.692 ab, and minus 0.176 and 2.694 ab. For Cold Sunny, Morning Warming, Peak Heat, and Afternoon Cooling are: Plastic Raised, 0.846 and 2.760 b, 0.820 and 2.759 b, and 1.197 and 2.782 b; Plastic Buried, 2.166 and 2.841 a, 2.253 and 2.846 a, and 3.404 and 2.910 a; Paper Raised, minus 0.286 and 2.686 c, minus 0.150 and 2.695 d, and 0.171 and 2.717 c; and Paper Buried, minus 0.224 and 2.690 c, 0.055 and 2.709 c, and 1.078 and 2.775 b. For Warm Cloudy, Morning Warming, Peak Heat, Afternoon Cooling, and Night Period are: Plastic Raised, 0.168 and 2.717 a, 0.147 and 2.715 b, 0.162 and 2.716 a, and 0.069 and 2.710 a; Plastic Buried, 0.152 and 2.716 a, 0.429 and 2.734 a, 0.247 and 2.722 a, and minus 0.137 and 2.696 b; Paper Raised, minus 0.156 and 2.695 b, minus 0.110 and 2.698 c, minus 0.245 and 2.688 b, and minus 0.136 and 2.696 b; and Paper Buried, minus 0.291 and 2.686 b, minus 0.158 and 2.695 c, minus 0.294 and 2.686 b, and minus 0.222 and 2.691 b. For Warm Sunny, Morning Warming, Peak Heat, and Afternoon Cooling are: Plastic Raised, 1.076 and 2.775 b, 1.080 and 2.775 b, and 0.917 and 2.765 b; Plastic Buried, 1.636 and 2.809 a, 1.927 and 2.827 a, and 2.419 and 2.855 a; Paper Raised, minus 0.690 and 2.658 d, minus 0.620 and 2.663 d, and minus 0.105 and 2.698 d; and Paper Buried, minus 0.476 and 2.673 c, minus 0.258 and 2.688 c, and 0.165 and 2.716 c.
Figure 4

Thermal differences of grow tube treatments (white waxed-paper grow tubes [Paper] and beige double-walled plastic grow tubes [Plastic]) relative to no-tube controls (ΔTn). Each tube type was installed either with the bottom buried ~5 cm below the soil surface (Buried) or raised ~10 cm above the soil surface (Raised). Shown on the transformed (Logarithm) and corresponding °C values. Data are from across two dormancy seasons (2023 to 2024 and 2024 to 2025). (A) Mean ΔTn across thermal phases and (B) across time blocks. (C) Summary heatmap presenting transformed ΔTn, statistical groupings, and temperature differences across all thermal phase-time block combinations. Lowercase letters indicate Tukey’s honest significant difference groupings within time block; treatments sharing the same letter are not significantly different at α = 0.05.

ΔTn displayed a consistent diurnal pattern across treatments (Figure 4B). Values increased from Morning Warming to Peak Heat, declined during Afternoon Cooling, and reached their lowest levels at Night. Plastic Buried showed the largest daytime change, rising from morning to peak heat (+0.86 to +1.63°C relative to the no-tube control), followed by Plastic Raised (+0.39 to +0.78°C). Paper treatments exhibited minimal daytime change (≤+0.1°C). Across all treatments, nighttime ΔTn values were near zero, and no treatment differences were detected during the Night Period.

ΔTa was interpreted as a reference relative to ambient atmospheric conditions. Plastic Buried remained consistently warmer than ambient air across all conditions (Supplemental Figure 5). The largest ΔTa values occurred during the Warm Sunny phase (+4.64°C) and the Peak Heat time block (+3.42°C), both relative to ambient air (Supplemental Figure 5A and 5B). Cold Sunny and Warm Sunny phases combined with Morning Warming, Peak Heat, and Afternoon Cooling showed the greatest warming for Plastic Buried (+3.64 to +4.75°C relative to ambient air). Plastic Buried remained warmer than Plastic Raised across these combinations by ~+1.00 to +2.00°C (Supplemental Figure 5C). Paper-based treatments remained close to ambient air across all conditions.

Cold hardiness modeling

Cold hardiness responses to grow tube treatments were evaluated across fall acclimation, winter maintenance, and spring deacclimation during two dormancy periods (2023 to 2024 and 2024 to 2025). Significant treatment and cultivar effects were detected across all phases (p ≤ 0.001; Supplemental Tables 3 to 6). Treatment-specific LMMs showed significant treatment × phase interactions for all cultivars (p ≤ 0.001), indicating that treatment effects differed across dormancy phases. Cultivar × phase interactions were also significant (p ≤ 0.001), reflecting consistent genotypic differences across phases regardless of treatment. Winter maintenance showed the strongest statistical differentiation for both treatment and cultivar effects, as supported by one-way ANOVA (p ≤ 0.001). Alternative transformations were evaluated; because model fit was comparable across approaches, results are presented on the original scale for interpretability.

Across both seasons, grow tube treatments showed similar cold hardiness behavior during fall acclimation and winter maintenance but diverged during spring deacclimation (Figures 5 and 6, Supplemental Figure 6). During fall acclimation, all treatments exhibited comparable hardening rates and no statistically significant differences in acclimation slopes were detected. Plastic Buried exhibited a slightly lower hardening rate (0.002 to 0.007°C/day) relative to other treatments. During winter maintenance, Plastic Buried consistently maintained slightly higher modeled cold hardiness thresholds than the remaining treatments (~+0.1 to +0.3°C).

Eight line graphs show modeled cold hardiness for four grapevine cultivars under five grow-tube treatments across two seasons. The eight line graphs show modeled cold hardiness for four grapevine cultivars under five treatments during 2023 to 2024 in panels A through D and 2024 to 2025 in panels E through H. The horizontal axis is Julian day, marked 250, 275, 300, 325, 350, 365, 25, 50, 75, and 90. The vertical axis is Temperature in degrees Celsius, ranging from minus 30 to minus 5. Panels A and E represent Cabernet Sauvignon, B and F represent Chardonnay, C and G represent Concord, and D and H represent Mourvedre. In panels A through D, the five lines begin between approximately minus 10 and minus 13 degrees Celsius near Julian day 250, decline after day 295, and reach their lowest levels between approximately minus 22 and minus 29 degrees Celsius from late December through early February. The lines then rise gradually, followed by a steep rise near days 70 to 85. Concord reaches the lowest temperature, near minus 29 degrees Celsius, followed by Chardonnay near minus 26, Cabernet Sauvignon near minus 25, and Mourvèdre near minus 22. Treatment lines separate during spring warming, with Plastic Raised reaching the highest final temperature and Paper Raised generally remaining lowest. In panels E through H, the lines similarly decline from approximately minus 10 to minus 13 degrees Celsius near day 250 to minima between approximately minus 22 and minus 29 degrees Celsius, remain nearly level until about day 50, and then rise sharply. Concord again reaches the lowest minimum, followed by Chardonnay, Cabernet Sauvignon, and Mourvedre. Treatment separation increases near days 70 to 85, with Plastic Buried or Plastic Raised generally reaching the highest final temperature and Paper Raised generally reaching the lowest. The legend lists No-Tube, Paper Buried, Paper Raised, Plastic Buried, and Plastic Raised.
Figure 5

Modeled cold hardiness for four grapevine cultivars under five grow tube treatments (white waxed-paper grow tubes [Paper] and beige double-walled plastic grow tubes [Plastic]; each tube type was installed either with the bottom buried ~5 cm below the soil surface [Buried] or raised ~10 cm above the soil surface [Raised], along with a no-tube control [No-Tube]) during two dormancy seasons: 2023 to 2024 (A to D) and 2024 to 2025 (E to H). Julian day 250 = 7 Sept; Julian day 365 = 31 Dec.

Six graphs compare fall and spring slopes and winter cold hardiness across four grapevine cultivars, five treatments, and two seasons. Six graphs in rows A and B compare Fall acclimation, Winter maintenance, and Spring deacclimation for Cabernet Sauvignon, Chardonnay, Concord, and Mourvèdre under five treatments. Row A represents 2023 to 2024, and row B represents 2024 to 2025. Fall acclimation uses Slope in degrees Celsius per day on the vertical axis. All slopes are negative. In A, approximate cultivar ranges are minus 0.155 to minus 0.149 for Cabernet Sauvignon, minus 0.143 to minus 0.136 for Chardonnay, minus 0.165 to minus 0.157 for Concord, and minus 0.141 to minus 0.134 for Mourvèdre. In B, the ranges are minus 0.139 to minus 0.133, minus 0.131 to minus 0.121, minus 0.148 to minus 0.135, and minus 0.130 to minus 0.117, respectively. Plastic Buried generally has the least negative fall slope, while Paper Buried or Paper Raised generally has the most negative slope. Winter maintenance uses Mean hardiness in degrees Celsius on the vertical axis from minus 28 to minus 22. In both rows, mean hardiness is approximately minus 24.3 for Cabernet Sauvignon, minus 24.9 for Chardonnay, minus 28.3 for Concord, and minus 21.7 for Mourvèdre, with little separation among treatments. Spring deacclimation uses Slope in degrees Celsius per day on the vertical axis. In A, approximate ranges are 0.19 to 0.30 for Cabernet Sauvignon, 0.22 to 0.37 for Chardonnay, 0.26 to 0.42 for Concord, and 0.19 to 0.31 for Mourvèdre. In B, the ranges are 0.28 to 0.37, 0.38 to 0.46, 0.43 to 0.54, and 0.26 to 0.38, respectively. Plastic Buried generally has the greatest spring slope, while Paper Buried or Paper Raised generally has the lowest. Lowercase letters a, b, c, and d mark significant differences among treatments within cultivars. The legend lists Paper Buried, Paper Raised, No Tube, Plastic Buried, and Plastic Raised.
Figure 6

Statistical summary of cold hardiness metrics for three dormancy phases of each grapevine cultivar under five grow tube treatments (white waxed-paper grow tubes [Paper] and beige double-walled plastic grow tubes [Plastic]; each tube type was installed either with the bottom buried ~5 cm below the soil surface [Buried] or raised ~10 cm above the soil surface [Raised], along with a no-tube control [No-Tube]) during two seasons: 2023 to 2024 (A) and 2024 to 2025 (B). The boxplot shows the median and quartiles slope for fall and spring, as well as the mean predicted hardiness for winter. Different letters indicate significant differences among treatments within cultivars (Tukey’s adjustment, α = 0.05).

Spring deacclimation showed the clearest separation between treatments. Plastic treatments exhibited significantly faster deacclimation than paper treatments, with Plastic Buried showing the strongest effects, particularly during the 2024 to 2025 period. Deacclimation rates for Plastic Buried were higher by +0.02 to +0.05°C/day across most cultivars. In contrast, paper-based treatments hardened slightly faster in fall, reached lower maximum cold hardiness thresholds during winter, and showed slower deacclimation rates in spring.

Chilling unit and GDD accumulation

Thermal accumulation responses to grow tube treatments were evaluated using three metrics: chilling units, cultivar-specific GDD-like heating units from the Ferguson model, and traditional GDD (base 10°C). One-way ANOVAs showed strong treatment effects across all metrics (p ≤ 0.001; statistical values provided in Supplemental Table 7). Alternative transformations were evaluated for this analysis and results are presented on the original scale for interpretability.

Chilling unit accumulation differed significantly among treatments for every cultivar and season (p ≤ 0.001; Figure 7). Paper treatments consistently accumulated more chilling units than plastic treatments across all cultivars, with Plastic Buried showing the lowest accumulation. Differences between paper and plastic treatments typically ranged from 20 to 60 units. Paper Buried and Paper Raised did not differ statistically, whereas plastic installations exhibited cultivar- and season-specific differences between the Buried and Raised positions. Mourvèdre showed the clearest treatment (paper versus plastic) separation in both seasons.

Six box plot groups compare chilling units and growing degree days among four cultivars and five grow tube treatments over two seasons. Six groups of box plots in rows A and B compare five treatments: Paper Buried, Paper Raised, No-Tube, Plastic Buried, and Plastic Raised. Row A represents 2023 to 2024, and row B represents 2024 to 2025. The chilling units graphs have Cabernet Sauvignon, Chardonnay, Concord, and Mourvèdre on the horizontal axis and chilling units in degree Celsius-days on the vertical axis. In A, values are approximately 1040 to 1090 for Cabernet Sauvignon, 1040 to 1090 for Chardonnay, 920 to 990 for Concord, and 990 to 1100 for Mourvèdre. In B, the corresponding values are approximately 1130 to 1180, 1100 to 1150, 990 to 1040, and 1070 to 1140. Across cultivars, Paper Buried and Paper Raised generally have higher values, followed by No-Tube and the plastic treatments. The growing degree day-like heating graphs have growing degree days in degree Celsius-days on the vertical axis. In A, values are approximately 90 to 145 for Cabernet Sauvignon, 155 to 240 for Chardonnay, 155 to 240 for Concord, and 65 to 125 for Mourvèdre. In B, the corresponding values are approximately 90 to 130, 155 to 190, 155 to 195, and 60 to 95. Plastic Buried and Plastic Raised generally have higher values than the paper treatments and No-Tube. The traditional growing degree day graphs combine all cultivars under the horizontal axis category All. In A, Paper Buried and Paper Raised are near 290 degree Celsius-days, No-Tube is near 302, Plastic Buried has a median near 320, and Plastic Raised has a median near 325 with the widest range, approximately 290 to 375. In B, Paper Buried, Paper Raised, and No-Tube are near 315 to 320, Plastic Buried has a median near 353 and extends to approximately 390, and Plastic Raised has a median near 330. Lowercase letters above the plots indicate significant differences among treatments within each cultivar.
Figure 7

Cumulative chilling and growing degree day (GDD) metrics for four grapevine cultivars under five grow tube treatments (white waxed-paper grow tubes [Paper] and beige double-walled plastic grow tubes [Plastic]; each tube type was installed either with the bottom buried ~5 cm below the soil surface [Buried] or raised ~10 cm above the soil surface [Raised], along with a no-tube control [No-Tube]) during two dormancy seasons (September to March): 2023 to 2024 (A) and 2024 to 2025 (B). Different letters indicate significant differences among treatments within each cultivar (Tukey-adjusted, α = 0.05).

GDD accumulation differed significantly between paper and plastic treatments (p ≤ 0.001; Figure 7). For traditional GDD analysis (base 10°C), plastic treatments generally accumulated more degree days than paper treatments; however, significant differences across both seasons occurred only between Plastic Buried and Paper Raised. For cultivar-specific GDD-like heating units, plastic treatments accumulated substantially more thermal units (~15 to 50 units) than paper treatments, with larger differences observed during the 2023 to 2024 season. Within material types, Buried installations tended to accumulate more thermal units than Raised installations, although this pattern was statistically significant only for plastic treatments in several cultivars.

Discussion

Material performance and environmental context dependencies

The dual-metric temperature ΔTn and ΔTa analysis revealed that solar radiation rather than air temperature alone drives material-based differences in grow tube thermal performance (Figure 4A and 4C, Supplemental Figure 5A and 5C). The 100 W/m2 solar radiation threshold, which distinguishes between sunny and cloudy conditions, was identified when plastic materials exhibited maximum warming effects, even overriding cold ambient conditions (<3.5°C) during Cold Sunny phases. Others have previously observed this “greenhouse” effect from grow tubes; in fact, it is primarily considered a positive effect of using grow tubes during the growing season, as it promotes vine growth (Kjelgren 1994, Hall and Mahaffee 2001, Tarara et al. 2014). Time-block analysis revealed that plastic materials function as thermal storage systems, absorbing solar energy during Peak Heat periods and retaining heat during Afternoon Cooling, when ambient temperatures naturally decline (Figure 4B and 4C, Supplemental Figure 5B and 5C). This creates cumulative warming effects that exceed instantaneous solar heating, with maximum temperatures occurring during sunny Afternoon Cooling combinations rather than during peak radiation hours. Such heat accumulation, which coincides with a decrease in solar radiation intensity, was previously observed (Kjelgren 1994) using plastic grow tubes to promote plant establishment during the summer.

Installation condition effects were material-dependent. Plastic treatments showed clear microclimate differences between Buried and Raised configurations under high solar radiation, with Buried installations likely enhancing heat retention through reduced ventilation. Reduced ventilation in Buried configurations has previously been documented to influence microclimate, including decreased powdery mildew incidence (Hall and Mahaffee 2001). Paper materials demonstrated minimal temperature sensitivity to installation height and exhibited more conservative thermal behavior, which was largely independent of solar conditions. Limited literature exists on material-driven differences in microclimate within grow tubes. Earlier research (Tarara et al. 2013) summarized findings from Munnell (2003), who reported that opaque cardboard grow tubes were cooler than ambient air, a pattern consistent with conservative warming behavior observed in the paper-based tubes. A different study hypothesized that opaque grow tubes, like those made of paper or cardboard, would have lower temperatures under solar radiation due to reduced shortwave radiation transmittance (Olmstead and Tarara 2001). Additionally, paper and cardboard grow tube materials have the potential to absorb and release water, which could modulate temperature through evaporative cooling or phase-change warming upon freezing (Snyder 2000).

These distinct thermal behaviors underscore the importance of selecting the appropriate materials for grow tubes. Growers seeking generally warmer microclimates may opt for plastic materials, which can benefit in-season vine establishment by facilitating vine growth, but at the risk of potentially decreased winter cold hardiness and increased spring frost risk due to enhanced spring deacclimation. Paper materials maintain environments closer to natural dormancy conditions while providing physical protection, but do not enhance in-season microclimate conditions. Material choice thus represents an important management decision, with installation configuration becoming critical only for materials that promote solar radiation-induced heating.

Timing of grow tube removal is also an important management consideration, as many growers remove tubes in late summer or early autumn to facilitate acclimation prior to winter. However, labor shortages may limit timely grow tube removal. While this study evaluated continuous overwinter installation to isolate material and installation effects, future research examining staged removal timing would further refine region-specific grower recommendations.

Day-night thermal dynamics and implications for cold hardiness

Diurnal temperature patterns showed clear material-dependent behavior relative to the no-tube control (Figure 4B). Plastic grow tubes warmed substantially during the daytime under high solar radiation, consistent with a greenhouse-like shortwave trapping effect described in an earlier study (Olmstead and Tarara 2001) and summarized by Tarara et al. (2013). At night, however, plastic tubes returned to near-ambient temperatures due to radiative heat loss and their limited insulating capacity (Snyder 2000). This combination (daytime heating followed by unbuffered nighttime cooling) is relevant for cold hardiness, as elevated daytime temperatures can promote more rapid deacclimation. Nighttime conditions, however, remain fully governed by ambient temperatures (North et al. 2024).

In contrast, paper grow tubes remained thermally conservative, closely matching the no-tube control throughout the diurnal cycle. Their minimal daytime warming and lack of nighttime buffering reflect low shortwave transmissivity and lower heat-retention capacity, as described earlier for opaque or paper-based materials (Olmstead and Tarara 2001). As a result, paper tubes are less likely to induce temperature-driven deacclimation, thereby maintaining cold-hardiness patterns that are more similar to those under natural ambient conditions.

Cold hardiness modeling and treatment effects

Cold hardiness modeling using grow tube temperatures as inputs to the Ferguson model revealed treatment-driven differences among cultivars (Figure 5 and Supplemental Figure 6). Modeled acclimation and deacclimation patterns generally aligned with the expected cultivar classifications described by Ferguson et al. (2014), and no cultivar × grow tube material interactions were detected. This indicates that differences among treatments arose from the microclimate created by the grow tubes rather than from cultivar-specific physiological responses.

Plastic grow tubes notably altered multiple phases of the dormancy cycles, but through distinct mechanisms. During midwinter, plastic treatments consistently maintained higher cold hardiness thresholds (+0.1 to +0.3°C) than the no-tube control, indicating slightly reduced midwinter hardiness relative to paper and no-tube conditions. Although statistically significant, the magnitude of this shift was modest and would likely be biologically meaningful only under conditions in which minimum temperatures approach cultivar-specific injury thresholds. This effect reflects the warmer internal temperatures documented during the winter maintenance period. In contrast, acclimation and deacclimation rates were influenced primarily during transitional periods. Plastic Buried showed slightly slower fall acclimation and substantially faster spring deacclimation, consistent with the well-documented sensitivity of hardiness transitions to warmer microclimates (North et al. 2024). Paper treatments showed minimal effects on either acclimation or deacclimation rates, consistent with their more conservative thermal behavior.

The installation of the grow tubes further shaped modeled outcomes. Plastic Buried installations produced the warmest internal temperatures and when modeled resulted in less cold-hardy midwinter thresholds and faster spring deacclimation than Plastic Raised installations. The paper grow tube again showed minimal sensitivity to installation height. These patterns indicate that burying plastic tubes amplifies thermal modifications beyond material effect alone, and raising plastic tubes during winter may help reduce treatment-induced warming effects and better preserve modeled cold hardiness in colder regions. Adjusting installation height during winter may represent a practical alternative to full tube removal and reinstallation in some vineyard systems; however, its feasibility will depend on vineyard labor capacity and management logistics.

The modeling suggests that both material type and installation configuration influence distinct components of dormancy development and that treatment choice should be matched to cultivar sensitivity and regional winter risk. Field validation, including direct bud LT50 measurements, is still needed to confirm how these modeled responses translate into whole-vine cold hardiness under natural conditions.

Thermal accumulation and dormancy phase transitions

Thermal accumulation patterns showed that the effects of grow tubes on dormancy were clearly dependent on both material and cultivar (Figure 7). Paper treatments consistently accumulated more chilling units and fewer heating units than plastic treatments across all cultivars and both seasons, with installation height further modifying these responses: Plastic Buried produced the lowest chilling and highest heating, Plastic Raised showed intermediate values, and paper tubes remained largely insensitive to installation position. Mourvèdre exhibited the largest and most consistent material-driven separation, with a 40 to 70°C·day reduction in chilling under Plastic Buried relative to paper treatments across both seasons. Heating accumulation increased under plastic tubes by a similar magnitude across all cultivars, indicating that the warming response was material-driven rather than cultivar-specific.

These patterns reveal a cultivar × treatment interaction primarily in chilling accumulation: Mourvèdre and to a lesser extent, Concord, were more affected by chilling depletion, whereas Cabernet Sauvignon and Chardonnay showed smaller and less consistent reductions (Ferguson et al. 2014, North et al. 2024). These dynamics suggest that paper grow tubes are better suited to maximizing chilling accumulation than plastic tubes, particularly for cultivars requiring sustained chilling.

Conclusion

This study demonstrates that grow tube material and installation configuration strongly influence thermal microclimates during grapevine dormancy, with clear implications for modeled cold hardiness. Plastic tubes acted as solar-driven thermal buffers, producing the greatest daytime warming when partially buried. Raising the tube base off the soil reduced, but did not eliminate, this effect. At night, temperatures inside all tubes returned to near-ambient levels, resulting in larger diurnal temperature swings in vines enclosed in plastic tubes than in those in paper tubes or no-tube controls.

Cold hardiness modeling showed that plastic tubes produced modestly warmer midwinter hardiness thresholds (~0.1 to 0.3°C) and accelerated spring deacclimation across cultivars. Paper tubes maintained thermal conditions more closely aligned with natural dormancy environments. Thermal accumulation responses further confirmed that paper tubes enhanced chilling and limited heat exposure, while plastic tubes, particularly buried installations, reduced chilling accumulation and increased heating units. Additional field validation across cultivars and climates is needed to refine the practical winter grow tube recommendation.

CRediT Authorship Contributions

WS: Formal Analysis, Software, Validation, Visualization, Writing – Original Draft; WS, EG, MJS, MM, MS, DR: Writing – Review & Editing; EG, MS, DR: Data Curation; EG, MJS, MM, MS, DR: Investigation; EG, DR: Project Administration; MM, DR: Conceptualization, Funding Acquisition, Methodology, Supervision; DR: Resources

Supplemental Data

The following supplemental materials are available for this article in the Supplemental tab above:

Supplemental Table 1 Linear mixed-effects model (LMM) diagnostics, transformation selection, and Type III tests for temperature differential analysis (ΔTn) on combined 2023 to 2024 and 2024 to 2025 data. Results reported on log-transformed data. AIC, Akaike Information Criterion; Original, original data; Log, logarithmic transformation data; SQRT, square root transformation data; CBRT, cube root transformation data; ICC, intraclass correlation coefficient, representing the proportion of total variance explained by random effects.

Supplemental Table 2 Linear mixed-effects model (LMM) diagnostics, transformation selection, and Type III tests for temperature differential analysis (ΔTa) on combined 2023 to 2024 and 2024 to 2025 data. Results reported on log-transformed data. AIC, Akaike Information Criterion; Original, original data; Log, logarithmic transformation data; SQRT, square root transformation data; CBRT, cube root transformation data; ICC, intraclass correlation coefficient, representing the proportion of total variance explained by random effects.

Supplemental Table 3 Linear mixed-effects model (LMM) diagnostics, transformation selection, and Type III tests for fall and spring cold hardiness analysis by cultivar and season. Results reported on original (untransformed) data. Fall and spring are dynamic cold hardiness phases; therefore, predicted cold hardiness slopes (°C/day) were compared among grow tube treatments within each cultivar using LMMs. AIC, Akaike Information Criterion; Original, original data; Log, logarithmic transformation data; SQRT, square root transformation data; CBRT, cube root transformation data; ICC, intraclass correlation coefficient, representing the proportion of total variance explained by random effects.

Supplemental Table 4 One-way analysis of variance (ANOVA) diagnostics and assumption testing for winter cold hardiness analysis by cultivar and season. Winter is a static cold hardiness phase; therefore, mean predicted cold hardiness was compared among grow tube treatments within each cultivar using one-way ANOVA. Results reported on original (untransformed) data.

Supplemental Table 5 Linear mixed-effects model (LMM) diagnostics, transformation selection, and Type III tests for fall and spring cold hardiness analysis by grow tube treatment and season. Fall and spring are dynamic cold hardiness phases; therefore, predicted cold hardiness slopes (°C/day) were compared among cultivars within each grow tube treatment using LMMs. Results reported on original (untransformed) data. AIC, Akaike Information Criterion; Original, original data; Log, logarithmic transformation data; SQRT, square root transformation data; CBRT, cube root transformation data; ICC, intraclass correlation coefficient, representing the proportion of total variance explained by random effects.

Supplemental Table 6 One-way analysis of variance (ANOVA) diagnostics and assumption testing for winter cold hardiness analysis by grow tube treatment and season. Winter is a static cold hardiness phase; therefore, mean predicted cold hardiness was compared among cultivars within each grow tube treatment using one-way ANOVA. Results reported on original (untransformed) data.

Supplemental Table 7 One-way analysis of variance (ANOVA) diagnostics and assumption testing for cumulative chilling and growing degree day (GDD) metrics analysis by cultivar and season. Chilling units, GDD-like heating units, and GDD (base 10°C) were compared among grow tube treatments within each cultivar and season using one-way ANOVA. Results reported on original (untransformed) data.

Supplemental Figure 1 Time series of 15-min temperature data from ambient air (on-site weather stations) and air inside each grow tube treatment (white waxed-paper grow tubes [Paper] and beige double-walled plastic grow tubes [Plastic]; each tube type was installed either with the bottom buried ~5 cm below the soil surface [Buried] or raised ~10 cm above the soil surface [Raised]) and the no-tube control (No-Tube) for each replicate (HOBO Pro v2 sensors) during the 2023 to 2024 dormancy period.

Supplemental Figure 2 Time series of 15-min temperature data from ambient air (on-site weather stations) and air inside each grow tube treatment (white waxed-paper grow tubes [Paper] and beige double-walled plastic grow tubes [Plastic]; each tube type was installed either with the bottom buried ~5 cm below the soil surface [Buried] or raised ~10 cm above the soil surface [Raised]) and the no-tube control for each replicate (HOBO Pro v2 sensors) during the 2024 to 2025 dormancy period.

Supplemental Figure 3 Residual diagnostics for linear mixed-effects model analysis of thermal differences of grow tube treatments relative to no-tube controls (ΔTn). (A) Diagnostic plots for the original data; (B) diagnostic plots after applying the logarithmic transformation.

Supplemental Figure 4 Residual diagnostics for linear mixed-effects model analysis of thermal differences of grow tube treatments relative to ambient air (ΔTa). (A) Diagnostic plots for the original data; (B) diagnostic plots after applying the logarithmic transformation.

Supplemental Figure 5 Thermal differences of grow tube treatments (white waxed-paper grow tubes [Paper] and beige double-walled plastic grow tubes [Plastic]) relative to ambient air (ΔTa). Each tube type was installed either with the bottom buried ~5 cm below the soil surface (Buried) or raised ~10 cm above the soil surface (Raised). Shown on the transformed (Logarithm) and corresponding °C values. Data are from across two dormancy seasons (2023 to 2024 and 2024 to 2025). (A) Mean ΔTa across thermal phases and (B) across time blocks. (C) Summary heatmap presenting transformed ΔTa, statistical groupings, and temperature differences across all thermal phase-time block combinations. Lowercase letters indicate Tukey’s honest significant difference groupings within time block; treatments sharing the same letter are not significantly different at α = 0.05.

Supplemental Figure 6 Modeled cold hardiness for four grapevine cultivars under five grow tube treatments (white waxed-paper grow tubes [Paper] and beige double-walled plastic grow tubes [Plastic]; each tube type was installed either with the bottom buried ~5 cm below the soil surface [Buried] or raised ~10 cm above the soil surface [Raised], along with a no-tube control [No-Tube]) during two dormancy seasons: 2023 to 2024 (A to E) and 2024 to 2025 (F to J). Julian day 250 = 7 Sept; Julian day 365 = 31 Dec.

Data Availability

The data underlying this study were deposited into Ag Data Commons at https://doi.org/10.15482/USDA.ADC/30680408. The source code for data preparation, cold-hardiness modeling, and statistical analysis, as well as documentation, is available on GitHub (https://github.com/WorasitSangjan/Grow-Tube-Microclimate-Analysis).

Footnotes

  • This research was supported in part by an appointment to the Agricultural Research Service (ARS) Research Participation Program administered by the Oak Ridge Institute for Science and Education (ORISE) through an interagency agreement between the U.S. Department of Energy (DOE) and the U.S. Department of Agriculture (USDA). ORISE is managed by ORAU under DOE contract number DE-SC0014664. This study was funded by USDA-ARS projects 2072-30500-001-000D, 2072-21000-057-000-D (Rippner, Schrader); USDA National Institute of Food and Agriculture, Hatch project 7005262, and Washington State University (Moyer). We thank Collins Wahkoli, Alexis Mace, and Cassandra Orozco for their assistance with data collection and organization. The mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the USDA. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and should not be construed to represent any official USDA or U.S. Government determination or policy. USDA is an equal opportunity provider and employer.

  • Sang jan W, Gillispie EC, Schrader MJ, Moyer MM, Shaw M and Rippner DA. 2026. Grow tube material selection influences thermal microenvironments and cold hardiness risk in dormant grapevines. Am J Enol Vitic 77:0770018. DOI: 10.5344/ajev.2026.25051

  • By downloading and/or receiving this article, you agree to the Disclaimer of Warranties and Liability. If you do not agree to the Disclaimers, do not download and/or accept this article.

  • Received November 2025.
  • Accepted April 2026.
  • Published online August 2026

This is an open access article distributed under the CC BY 4.0 license.

References

  1. ↵
    1. Akaike H.
    1974. A new look at the statistical model identification. IEEE Trans Automat Contr 19:716–723. DOI: 10.1109/TAC.1974.1100705
    OpenUrlCrossRefPubMed
  2. ↵
    1. Bates D,
    2. Mächler M,
    3. Bolker B and
    4. Walker S.
    2015. Fitting linear mixed-effects models using lme4. J Stat Softw 67:1–48. DOI: 10.18637/jss.v067.i01
    OpenUrlCrossRefPubMed
  3. ↵
    1. Beck HE,
    2. Zimmermann NE,
    3. McVicar TR,
    4. Vergopolan N,
    5. Berg A and
    6. Wood EF.
    2018. Present and future Köppen-Geiger climate classification maps at 1-km resolution. Sci Data 5:180214. DOI: 10.1038/sdata.2018.214
    OpenUrlCrossRefPubMed
  4. ↵
    1. Bordelon BP and
    2. Blume J.
    2000. Grow th of g rapev ines w ith and without grow tubes. HortScience 35:424. DOI: 10.21273/HORTSCI.35.3.424E
    OpenUrlAbstract/FREE Full Text
  5. ↵
    1. Campos-Arguedas F,
    2. Kirchhof E,
    3. North MG,
    4. Londo JP,
    5. Bates T,
    6. van Leeuwen C et al
    . 2026. Cold hardiness dynamics predict budbreak and associated low-temperature threats in grapevine. New Phytol 250:2203-2220. DOI: 10.1111/nph.71074
    OpenUrlCrossRef
  6. ↵
    1. Cragin J,
    2. Serpe M,
    3. Keller M and
    4. Shellie K.
    2017. Dormancy and cold hardiness transitions in winegrape cultivars Chardonnay and Cabernet Sauvignon. Am J Enol Vitic 68:195–202. DOI: 10.5344/ajev.2016.16078
    OpenUrlAbstract/FREE Full Text
  7. ↵
    1. Dami I,
    2. Bordelon B,
    3. Ferree DC,
    4. Brown M,
    5. Ellis MA,
    6. Williams RN et al
    . 2005. Midwest Grape Production Guide. Bulletin 919. Ohio State University Extension, Columbus, OH. https://ohiograpeweb.cfaes.ohio-state.edu/grape-growing/midwest-grape-production-guide
  8. ↵
    1. Ferguson JC,
    2. Moyer MM,
    3. Mills LJ,
    4. Hoogenboom G and
    5. Keller M.
    2014. Modeling dormant bud cold hardiness and budbreak in twenty-three Vitis genotypes reveals variation by region of origin. Am J Enol Vitic 65:59–71. DOI: 10.5344/ajev.2013.13098
    OpenUrlAbstract/FREE Full Text
  9. ↵
    1. Fiola J.
    2021a. Grow Tubes for Grapevines: The Benefits. University of Maryland Extension. https://extension.umd.edu/resource/grow-tubes-grapevines-benefits/
  10. ↵
    1. Fiola J.
    2021b. Grow Tubes for Grapevines: The Drawbacks. University of Maryland Extension. https://extension.umd.edu/resource/grow-tubes-grapevines-drawbacks/
  11. ↵
    1. Gale EJ and
    2. Moyer MM.
    2017. Cold hardiness of Vitis vinifera roots. Am J Enol Vitic 68:468–477. DOI: 10.5344/ajev.2017.16109
    OpenUrlAbstract/FREE Full Text
  12. ↵
    1. Hall TW and
    2. Mahaffee WF.
    2001. Impact of vine shelter use on development of grape powdery mildew. Am J Enol Vitic 52:204–209. DOI: 10.5344/ajev.2001.52.3.204
    OpenUrlAbstract/FREE Full Text
  13. ↵
    1. Kjelgren R.
    1994. Growth and water relations of Kentucky coffee tree in protective shelters during establishment. HortScience 29:777–780. DOI: 10.21273/HORTSCI.29.7.777
    OpenUrlAbstract/FREE Full Text
  14. ↵
    1. Lavee S and
    2. May P.
    1997. Dormancy of grapevine buds - facts and speculation. Aust J Grape Wine Res 3:31–46. DOI: 10.1111/j.1755-0238.1997.tb00114.x
    OpenUrlCrossRef
  15. ↵
    1. Lenth R.
    2023. emmeans: Estimated Marginal Means, aka Least-Squares Means. R package version 1.11.2-8. https://cran.r-project.org/package=emmeans
  16. ↵
    1. Londo JP and
    2. Kovaleski AP.
    2025. Integrating cold hardiness and deacclimation resistance demonstrates a conserved response to chilling accumulation in grapevines. J Exp Bot 76:3158-3171. DOI: 10.1093/jxb/eraf045
    OpenUrlCrossRef
  17. ↵
    1. McMaster GS and
    2. Wilhelm WW.
    1997. Growing degree-days: One equation, two interpretations. Agr Forest Meteorol 87:291–300. DOI: 10.1016/S0168-1923(97)00027-0
    OpenUrlCrossRef
  18. ↵
    1. Moyer M,
    2. Davenport J,
    3. Skinkis P and
    4. Copp C.
    2024. Considerations and Resources for Vineyard Establishment in the Pacific Northwest. PNW634. Washington State University. https://pubs.extension.wsu.edu/product/considerations-and-resources-for-vineyard-establishment-in-the-inland-pacific-northwest/
  19. ↵
    1. Munnell JA.
    2003. Determining the microenvironmental factors of grow tubes that affect grapevine (Vitis vinifera) growth and development. Thesis, Washington State University, Pullman, WA.
  20. ↵
    1. North MG,
    2. Workmaster BA,
    3. Atucha A and
    4. Kovaleski AP.
    2024. Cold hardiness-informed budbreak reveals role of freezing temperatures and daily fluctuation in a chill accumulation model. J Exp Bot 75:6182–6193. DOI: 10.1093/jxb/erae287
    OpenUrlCrossRefPubMed
  21. ↵
    1. O’Brien P,
    2. Montague T and
    3. Cook M.
    2025. Insulative effects of vine shelters may impact growth potential and cold hardiness of young vines. In IVES Conference Series, GiESCO 2025.
  22. ↵
    1. Olmstead MA and
    2. Tarara JM.
    2001. Physical principles of row covers and grow tubes with application to small fruit crops. Small Fruits Rev 1:29–46. DOI: 10.1300/J301v01n03_04
    OpenUrlCrossRef
  23. ↵
    1. Python Software Foundation
    . 2025. Python Language Reference (version 3.9.6). https://www.python.org
  24. ↵
    1. Quinn GP and
    2. Keough MJ.
    2002. Experimental Design and Data Analysis for Biologists. Cambridge University Press, UK. DOI: 10.1017/CBO9780511806384
    OpenUrlCrossRef
  25. ↵
    1. R Core Team
    . 2025. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.r-project.org/
  26. ↵
    1. Snyder RL.
    2000. Principles of Frost Protection (Short Version – Quick Answer FP005). University of California. https://ucanr.edu/sites/default/files/2025-03/principles%20of%20frost%20protection.pdf
  27. ↵
    1. Tarara JM,
    2. Chaves B and
    3. Strik BC.
    2013. Above- and below-ground microclimate of grow tubes in an organic mulch-incorporated, raised bed system for blueberry. HortScience 48:1363–1369. DOI: 10.21273/HORTSCI.48.11.1363
    OpenUrlAbstract/FREE Full Text
  28. ↵
    1. Tarara JM,
    2. Chaves B and
    3. Strik BC.
    2014. Grow tubes change microclimate and bush architecture but have little effect on bush biomass allocation at the end of the establishment year in blueberry. HortScience 49:596–602. DOI: 10.21273/HORTSCI.49.5.596
    OpenUrlAbstract/FREE Full Text
  29. ↵
    1. Thomas AL,
    2. Harris JL,
    3. Bergmeier EA and
    4. Striegler RK.
    2017. Establishment technique and rootstock impact ‘Chambourcin’ grapevine morphology and production in Missouri. HortTechnology 27:248–256. DOI: 10.21273/HORTTECH03610-16
    OpenUrlCrossRef
Back to top
Print
View full PDF
Email Article
Citation Tools

Open Access
Grow Tube Material Selection Influences Thermal Microenvironments and Cold Hardiness Risk in Dormant Grapevines
Worasit Sangjan, Elizabeth C. Gillispie, M. Jacob Schrader, Michelle M. Moyer, Madison Shaw, Devin A. Rippner
Am J Enol Vitic.  2026  77: 0770018  ; DOI: 10.5344/ajev.2026.25051
del.icio.us logo Twitter logo Facebook logo Mendeley logo
Save to my folders

Jump to section