Variable rate irrigation scheduling based on high-resolution short-wave infrared sensing and internet-of-things.

Summary:  Vineyards face climate change, increasing temperatures, and drought affecting vine water status. Water deficit affects plant physiology and can ultimately decrease yield and grape quality when it is not well managed. Monitoring vine water status and irrigation can help growers better manage their vineyards. However, when field measurements, such as stem water potentials (SWP), can be precise, they are time-consuming. In addition, they do not allow for easy assessment of spatial variability, which is a critical factor for water status management. Remote sensing tools can help map plant water status in space and time and streamline data acquisition over whole vineyards several times during the season. In this project, we monitored a variably irrigated vineyard several times during the season with a hyperspectral NIR/SWIR camera mounted on a UAV.

A study was put in place in a Vitis vinifera L. cv. Cabernet-Sauvignon vineyard in the San Joaquin Valley, CA, USA, where a variable rate automated irrigation system was installed to irrigate vines with twelve different water regimes in four randomized replicates, totaling 48 experimental zones. The purpose of this experimental design was to create variability in grapevine water status, in order to produce a robust dataset for modeling purposes. Throughout the growing season, spectral data within these zones was gathered using a Near InfraRed (NIR) – Short Wavelength Infrared (SWIR) hyperspectral camera (900 to 1700nm) mounted on an Unmanned Aircraft Vehicle (UAV). Given the high water-absorption in this spectral domain, this sensor was deployed to assess grapevine stem water potential, a standard reference for water status assessment in plants, from pure grapevine pixels in hyperspectral images. The stem was acquired simultaneously in the field from bunch closure to harvest and modeled via machine-learning methods using the remotely sensed NIR-SWIR data as predictors in regression and classification modes (classes consisted of physiologically different water stress levels). Hyperspectral images were converted to bottom of atmosphere reflectance using standard panels on the ground and through the Quick Atmospheric Correction Method (QUAC) and the results were compared. The best models used data obtained with standard panels on the ground and allowed predicting stem values with an R2 of 0.54 and an RMSE of 0.11 MPa as estimated in cross-validation, and the best classification reached an accuracy of 74%. This project aims to develop new methods for precisely monitoring and managing irrigation in vineyards while providing useful information about plant physiology response to deficit irrigation.

As the drought conditions become more familiar to California grapevine growers, the need for effective water management solutions has heightened. Understanding the physiological responses of the plant at various intensities and timings can still provide great context for more advanced developments.

LEVERAGING BERRY PHYSIOLOGY TO MITIGATE LATE-SEASON DEHYDRATION

Summary: Warm, dry conditions can exacerbate late-season berry shrivel, reducing yield and altering berry sensory properties. Late-season shrivel occurs when the berries undergo programmed cell death, and the water released from the ruptured cells is drawn from the fruit to the canopy by a water potential gradient. The goal of our project was to test whether we could reduce shrivel by interrupting these processes. First, we tested whether short (2 week) pulses of increased irrigation before or at the expected onset of cell death would reduce water stress- induced oxidative damage and delay or reduce the rate of programmed cell death, reducing shrivel at harvest. Second, we tested whether blocking the petiole water transport tissue (xylem) would reduce shrinkage by impeding water transport to the canopy. We implemented conventional (following standard commercial practices), early-pulse (before expected onset), and late-pulse (at expected onset) irrigation treatments on mature Cabernet Sauvignon vines in an experimental vineyard at UC Davis in 2022 and 2023. We monitored vine water stress, berry cell death and shrivel, reactive oxygen species concentrations, markers for cell oxidative damage, and berry gene expression. We also used glue to partially block the pedicle for berries in the conventional irrigation treatment and compared diameter shrinkage to undamaged berries. The late irrigation treatment significantly reduced the rate of cell death and the magnitude of berry shrivel at harvest compared to the conventional treatment. However, the early irrigation treatment did not significantly impact cell death or shrivel. Concentrations of the reactive oxygen species H2O2 and indicators of cell oxidative damage increased at the same time as cell death, consistent with a role in programmed cell death, but were not significantly different among treatments. We are still analyzing berry chemistry, cell death, and shrivel from 2023. Blocking the pedicle xylem did not significantly reduce shrinkage. Overall, these findings show that the time of onset of cell death is not impacted by water stress but using a short pulse of irrigation near the onset of cell death can slow the rate of cell death and reduce berry shrinkage at harvest.

VRI Scheduling System

Summary: Vineyards face climate change, increasing temperatures, and drought affecting vine water status. Water deficit affects plant physiology and can ultimately decrease yield and grape quality when it is not well managed. Monitoring vine water status and irrigation can help growers better manage their vineyards. However, when field measurements, such as stem water potentials (SWP), can be precise, they are time-consuming. In addition, they do not allow for easy assessment of spatial variability, which is a critical factor for water status management. Remote sensing tools can help map plant water status in space and time and streamline data acquisition over whole vineyards several times during the season. In this project, we monitored a variably irrigated vineyard several times during the season with a hyperspectral NIR/SWIR camera mounted on a UAV. We worked in a Cabernet Sauvignon vineyard in the San Joaquin Valley of California equipped with an automated irrigation system. We created forty-eight independent watering zones and applied twelve different amounts of water replicated four times in a randomized block scheme. Water amounts were fractions of the grower allocation and applied as sustained and regulated deficit irrigation strategies. Hyperspectral images in 112 bands from 900 nm to 1700 nm were collected using a UAV every two weeks from June to harvest. Contemporarily, we measured vine water status through SWP, stomatal conductance (gs) and net assimilation (AN). For the analysis, the images were segmented to extract the canopy signal and converted to reflectance, then used to predict the field water status measurements using machine learning models. Models were evaluated using coefficients of determination (R2 ), and root mean square error (RMSE). Feature importance was also computed to determine the importance of each band in the model. Field measurements of stem water potential ranged from -2.0 to -1.14 MPa. The canopy signal was segmented from the soil background using a classifier with an accuracy of 99.7%. We tested random forest, gradient boosting machine, and support vector machine algorithms in a preliminary analysis to predict SWP values. The most performant model was the random forest, and it was able to predict SWP values with an R2 of 0.6 and an RMSE of 0.1 MPa as assessed in a 5-fold crossvalidation procedure. The most important bands for model prediction were 1146 nm, 1153 nm, 1321 nm, 1363 nm, and 1434 nm, all situated in water absorption domains. These promising results demonstrate that SWIR images can monitor the field’s vine water status and inform irrigation management with high resolution. Besides the work on remote sensing, we took advantage of the irrigation trial to determine the primary effects of sustained and regulated deficit irrigation on plant water status, berry composition, gas exchange, and yield components. There was little to no effect of irrigation treatment on plant performance and overall quality. The results of this study did follow the general trends for grapevines in semi-arid environments, although continued data acquisition is needed to assess the carry-over effect of water reduction on plants. As the drought conditions become more familiar to California grapevine growers, the need for effective water management solutions has heightened. Understanding the physiological responses of the plant at various intensities and timings can still provide great context for more advanced developments.

Reducing Late Season Berry Dehydration

Summary: Warm, dry conditions can exacerbate late-season berry shrivel, reducing yield and altering berry sensory properties. Late-season shrivel occurs when the berries undergo programmed cell death, and the water released from the ruptured cells is drawn from the fruit to the canopy by a water potential gradient. The goal of our project was to test whether we could reduce shrivel by interrupting these processes. First, we tested whether short (2 week) pulses of increased irrigation before or at the expected onset of cell death would reduce water stress induced oxidative damage and delay or reduce the rate of programmed cell death, reducing shrivel at harvest. Second, we tested whether blocking the petiole water transport tissue (xylem) would reduce shrinkage by impeding water transport to the canopy. We implemented conventional (following standard commercial practices), early-pulse (before expected onset), and late-pulse (at expected onset) irrigation treatments on mature Cabernet Sauvignon vines in an experimental vineyard at UC Davis. We monitored vine water stress, berry cell death and shrivel, reactive oxygen species concentrations, markers for cell oxidative damage, and berry gene expression. We also used glue to partially block the pedicle for berries in the conventional irrigation treatment and compared diameter shrinkage to undamaged berries. The late irrigation treatment significantly reduced the rate of cell death and the magnitude of berry shrivel at harvest compared to the conventional treatment. However, the early irrigation treatment did not significantly impact cell death or shrivel. Concentrations of the reactive oxygen species H2O2 increased at the same time as cell death, consistent with a role in programmed cell death, but were not significantly different among treatments. We are still analyzing oxidative cell damage and berry transcriptomics, which we expect to provide more insight into the drivers of the treatment differences. Blocking the pedicle xylem did not significantly reduce shrinkage. Overall, these findings show that the time of onset of cell death is not impacted by water stress but using a short pulse of irrigation near the onset of cell death can slow the rate of cell death and reduce berry shrinkage at harvest.

Characterizing Willamette Valley Soil Moisture and Grapevine Response under Drying Seasonal Conditions

Soil moisture, weather data, and vine growth response were measured in 2020 in one vineyard location that had Pinot noir of the same vine age, clone and rootstock growing in three soil types, including volcanic soils (Saum), sedimentary soils (Dupee), and marine sediment soils (Willamette Woodburn). Soil sensors measured soil moisture, soil temperature, and electrical conductivity for each soil type. Soil probes were installed to a depth of 18 and 36 inches under[1]vine and in the middle of the alley between rows. Soil moisture remained relatively consistent through much of spring, with the start of soil moisture decline beginning in mid-June. This occurred shortly after bloom, and continued throughout the rest of summer, when there was little to no precipitation. Vine growth measures of leaf area and lateral count did not vary in-season, but dormant season pruning weights show that the most vigorous vines grew in the Willamette[1]Woodburn soil. Vine vegetative vigor was similar for Dupee and Saum. Soil moisture decline was greatest at the 18” depth and varied less at the 36” depth, and the greatest decline occurred with Willamette-Woodburn, suggesting that the higher vigor vines required more water from the soil profile than vines in the other two soil types. Leaf water potential did not show clear differences in vine water stress of the three soil types. Berry weight lagged slightly for Willamette-Woodburn, but there were no differences in the overall growth curve through development. By harvest, yields were similar from each soil type. However, the Willamette Woodburn had lower Brix and sugar per berry compared to the other two soil types. Data analysis from the 2020 season continues as of this reporting. This research will continue through two additional growing seasons (2021 and 2022).

Evaluation of interactive effects of mechanical leafing and deficit irrigation on berry composition and wine chemistry of Vitis vinifera cv. Cabernet Sauvignon Grown in the San Joaquin Valley of California

Mechanical leafing at bloom and berry set either on one side or both sides of the canopy does not affect the final yield. Leafing, either on bloom or berry set, improves the anthocyanins accumulation during the ripening and increases the harvest berry anthocyanins. Light exposure resulted from bloom leafing only lasts for approximately 2 weeks, and such a short period of light exposure during/after bloom is enough to increase the anthocyanins accumulation and final berry anthocyanins. Overexposure from berry set leafing might promote the anthocyanins degradation at the end of berry ripening.

Water deficit during cell growth stage I, from berry set to veraison, reduces berry size and ultimate yield, although the decline of berry size and yield depends on the severity of water deficit during the stage I. Water deficit increases the harvest berry anthocyanins, although its increase mainly results from the high skin/pulp ratio associated with small berry size.

From research wine micro-fermentation, resulted wine color follows the similar pattern of harvest berry color.

Developing solar-induced chlorophyll fluorescence as a ground-based and remotely-sensed physiological indicator of grapevine stress

Upon having access to the funds, we initiated work in June 2019 and plan to build off of these findings to continue efforts and analysis in 2020-2021. We conducted a dry down experiment on potted grapevines to evaluate the relationship between active and passive (i.e. ground based equivalent of SIF) fluorescence and used this experiment system to test modification to a gas exchange system for comparing active and passive fluorescence while simultaneously measuring gas exchange. We also utilized existing field installations, specifically at the Ripperdan GRAPEX-Grape Remote sensing and Atmospheric Profiling and Evapotranspiration eXperiment site, which contains a variable rate irrigation system where water can be delivered down to 30m by 30m pixels and contains 4 treatments (two stressed and two well-watered) for comparison of active and passive fluorescence responses of grapevines under stress conditions.

Interaction of red blotch virus (GRBV) and deficit irrigation on grapevine water relations, disease development, and vine productivity

The second year of a field experiment with two irrigation treatments – wet (W) and dry (D) – and two vine disease statuses – healthy (GRBV-) and infected (GRBV+) – was continued in a commercial vineyard to understand the interaction between GRBV infection and deficit irrigation on disease development, vine productivity, and fruit quality. W vines were irrigated at 100% of crop evapotranspiration (ETc), while D vines received water at 50% ETc. Within each irrigation treatment, GRBV- and GRBV+ vines (split-plot) that were previously identified in early 2017 based on symptomology data from 2016. The identified vines were confirmed as GRBV+ and GRBV- by PCR-based assays. GRBV- vines from 2017 were re-tested in early 2018 to confirm disease status.

In both years, measurements of vine water status (midday stem water potential; Ψstem) were made at regular intervals throughout the growing season beginning just after berry set until just before harvest. Similarly, disease severity was recorded every week after the first symptom appearance was observed on GRBV+ vines. At harvest, berry samples were collected for berry size and compositional analyses; and vine yield and yield components were determined.

With respect to vine water relations and gas exchange, there were no significant interactions among experimental treatments. Irrigation treatment and disease status both impacted these aspects of vine physiology, but they acted independently, with water deficits consistently reducing vine water status, and GRBV infection consistently increasing it. In other words, GRBV infection had the same effect on vine water status (stem water potential) under both well-watered and deficit conditions. However, the significant impact of GRBV infection on vine water status only arose post-veraison – at the same time that foliar symptoms became visible. The increase in post-veraison water status under GRBV+ conditions was likely a consequence of reduced stomatal conductance, which in turn reduced net photosynthesis.

Berry development was similarly impacted by the treatments independently, with consistently larger berries in W and GRBV+ vines. This was observed at nearly every sample date in each year, but differences between vines of different disease status only became significant post-veraison. TSS were also only significantly different between GRBV- and GRBV+ vines post-veraison, and there was a limited impact of irrigation treatment. In contrast, pH and TA were more variable among treatments and years, suggesting that GRBV has a limited effect on organic acid metabolism.

Irrigation treatment and disease status impacted skin and seed phenolic concentrations to varying degrees over two seasons. Whereas irrigation treatment and disease status impacted skin phenolic concentration together, disease status alone impacted seed phenolic concentration. In both years, skin anthocyanin concentration was increased with deficit irrigation – in both GRBV- and GRBV+ vines – but only increases in GRBV- vines were statistically significant. Conversely, skin tannin and iron-reactive phenolic concentrations were variably affected by treatments, and results were not consistent between years. In seeds, there were no effects of the irrigation treatments, but disease status significantly reduced both tannin and iron-reactive phenolic concentrations. This effect was consistent between years. All together, these results suggest that the genetic control of phenolic metabolism by GRBV infection is stronger than the environmental control due to vine water deficits. Furthermore, experimental results suggest that keeping vines well-watered may mitigate some of the negative effects of GRBV infection, but ultimate changes in secondary metabolism due to GRBV infection may necessitate using infected fruit for different wine programs (e.g. rosé and/or sparkling) or blending with lots from healthy vineyards.

In contrast to 2017 data on disease severity, significant differences on rate of disease progression as well as disease severity were observed in 2018 between the wet and dry treatments. The vines in wet irrigation treatment showed significantly low disease severity at harvest and two weeks prior to harvest. The differences were observed as significant increase in vine canopy in irrigation treatment compared to vines in dry treatment; as a result percent of symptomatic leaves in wet treatment vines were less compared to dry treatment vines. Even though, the virus expression remained same (symptomatic) within wet and dry treatments, it would be informative to assess the status of virus (quantity) within each treatment. Furthermore, the carry over effect of less severe vines on vine health as well as fruit qualities would be an additional information on long-term management of GRBV infected vines.

The removal of the experimental vineyard site between 2018 and 2019 precluded some of the confirmatory and deeper data collection that was originally planned. Unfortunately, this is an all too common occurance with respect to recent research efforts on GRBV. The new site was on a heavier soil, thus irrigation differences were not manifested until just prior to harvest. However, several of the vine physiological responses to GRBV infection that were observed at the first site in 2017 and 2018 were also observed at the new site in 2019. These included: (1) higher water status, (2) lower photosynthetic rate and stomatal conductance, and (3) lower sugar and color in fruit. Additionally, there were no effects of GRBV on yield or yield components, as in 2017/2018. These consistencies across the two sites underscore the main effects of GRBV on grapevines and future research efforts should be targeted at exploring the underlying mechanisms behind them.

Assessing Rootstock Biology and Water Uptake through Proximal Sensing under Different Wetting/Drying Conditions

Considering the importance of the water status as a driver of the whole plant physiology at the vineyard scale, developing new technologies for the space-time measurement of roots distribution and water uptake is crucial for the development of efficient precision viticulture practices. This project (2017-2022) is devoted to the development of a non-invasive technique (electrical resistivity tomography, ERT), as a tool to compare dynamic changes in root growth and water uptake patterns, and to apply this technique to the study of several commercially available rootstocks under varied irrigation delivery methods (drippers, micro-sprinklers) and water regimes (sustained deficit irrigation, rain fed, fully watered). The project was conducted in a vineyard at UC Davis specifically planted for the study of the response of rootstocks to watering systems. According to the timetable of the funded proposal the first year consisted of two parallel objectives: i) the calibration of ERT to soil water and roots distribution iii) the physiological monitoring of different rootstocks under different water amounts and delivery methods.

In a vineyard  planted to Chardonnay on different rootstocks representing the most common parentage classes in a randomized complete block design, six soil pits were dug for the description of the soil profile, the sampling of the soil for chemical-physical analysis and the installation of TDR probes for the monitoring of soil moisture. The pits were then refilled, and the soil allowed to resettle for several months before the measurements commenced. The section of the field used for calibration purposes was not irrigated throughout the season in order to cover a full range of moisture levels for that soil (from 24% to 8% vol.). In close correspondence of these pits, at the soil surface 300 stainless steel electrodes were installed at a distance of 0.62 m (2 feet) for the ERT monitoring. A single ERT measurement was able to cover 2 plants on 2 different rootstock, which represented an experimental unit. The water status of the plants within the experimental units was monitored using infrared thermography and pressure chamber measurements. Photosynthesis and stomatal conductance were monitored through a pressure bomb. The electrical resistivity was calibrated to the soil water content obtained by TDR measurements, and two different pedoelectrical models were fitted to the data: Archie, in a canonical and linearized form, and Waxman and Smits models. Their performances were tested on 20% of the data from the whole dataset that were excluded from the model fitting and then used as a test set of unobserved data. The three modeling approaches gave similar results on the test set, with a slight decrease in performance from the log-log linearization (RMSE = 1.22% vol., R2 = 0.73) and the Archie law (1.22 %vol., R2 = 0.73), to the Waxman-Smits model (RMSE = 1.23% vol., R2 = 0.73). The Archie law was then chosen, and confirmed to be well adapted to predominantly sandy soils such as the experiment site. This calibration was used to transform the ERT images from soil resistivity to soil moisture maps under different rootstocks in 2D and 3D. This is the first report in the world that 3D images of soil moisture were developed in a vineyard, and the first time that this technique was used to compare rootstock physiology in general. Great differences were found between contiguous rootstocks in their spatial use of water. The presented data show how lateral heterogeneity in soil moisture could reduce the efficiency of spot measurements, as those obtained with soil probes, and soil moisture sensors in determining irrigation needs. In the experimental region, despite not irrigating for the whole summer, grapevines did not suffer from water stress  until it reached the harvest date where the midday stem water potential were at -0.7 MPa across all four rootstocks and no significant differences were found. Chardonnay/110R had lowest assimilation rate (P<0.0017) and stomatal conductance (P<0.05) but had greater berry weight, cluster weight and yield per vine, whereas Chardonnay grafted on 140Ru had the lowest yield among four rootstocks. At harvest Chardonnay /110R also had the greatest total soluble solids and lowest titratable acidity. The results of Ravaz Index (kg yield/kg of pruning weight) indicated that Chardonnay /110R and 140Ru achieved vine balance where the Ravaz Index were around 5. However, Chardonnay /101-14Mgt and 420A had  Ravaz Index lower than 3, indicating that excessive vegetative growth due to no water stress resulting in the unbalance of the grapevine. Our results suggested that although Chardonnay /110R had lower vigor, it had higher productivity and enhanced ripening of grapes, related to better vine balance under non-limiting water supply. Further information is needed to test the interaction between rootstocks and different irrigation regimes and their drought resistance. At the end of the season, the soil was sampled with an auger at different locations in the middle between contiguous ERT electrodes. The soil was sampled every 0.1 m, brought to the laboratory and oven-dried. Roots were physically separated from the soil, and their presence was assessed in a gravimetrically (mg of roots per g of soil). The presence of roots was negatively correlated to soil moisture obtained through ERT (r = 0.45), i.e. greater the amount of roots, lower the amount of water at the end of a dry period. This relations was then used to map the roots using ERT as ancillary variable for the spatialization, therefore enabling us to use the electrical images of the roots assess plant water use.

 

A New Embedded Sensor for Continuous Monitoring of Stem Water Potential in Grapevines

This project is to extend the application of a recently developed microtensiometer (MT) device, which measures water tension, to the measurement of stem water potential (SWP) in grapevines under field conditions. The objectives of this project were to test the accuracy of this device using standard laboratory methods, test the performance of sensors compared to pressure chamber measured SWP under field conditions, and help optimize sensor packaging and installation for accuracy and robustness in the field. Laboratory and greenhouse testing indicated that the sensors responded correctly to water potential, although in some cases, sensor response to changes in water potential was relatively slow and temperature sensitivity high, both of which might limit sensor performance in the field. Improvements in the sensor and in handling/installation have been made and are ongoing, but near-commercial prototypes were installed in field vines starting in August, 2018. A total of 14 sensors were installed on 6 mature Cabernet Sauvignon vines in the RMI vineyards at Davis, CA. These vines were minimally drip-irrigated, had not been irrigated for about 3 weeks, and did not receive any irrigation once the sensors were installed. Essentially all sensors exhibited similar daily patterns in SWP that were comparable to patterns shown by the pressure chamber, but there was large sensor-to-sensor variation in the range of water potential values reported. For these vines, the relative values of the sensors indicated that the highest SWP did not occur at predawn (prior to 6:00am in August), as is commonly assumed, but rather about an hour after dawn. The midday minimum water potential (maximum stress) occurred around 3:00 pm, about 2h after peak reference evapotranspiration (ETo). These findings are consistent with previous in-situ measurements in other woody perennial species. Each MT was calibrated against the corresponding pressure chamber measured SWP on an individual vine basis, giving consistent MT measured SWP values for the 5 (of 6) vines in the study that had enough calibration points. The average values from these vines showed a very clear pattern of overnight recovery to about -3 bars from August to November, when rains caused an increase to above -2 bars. Midday SWP values were around -7 bars, also recovering with normal seasonal leaf loss as well as rains. Midday MT measured SWP also fluctuated in parallel with the weather-related baseline (fully irrigated) SWP for about 3 months, indicating the possibility for continuous monitoring over relatively long periods. These data illustrate that automated and sensitive monitoring of SWP in grapevine is possible with these sensors, and should provide useful information for irrigation management as well as for physiological studies.