Argronomic tools to support a whole system approach to reducing water risk in rain dependent and supplementary irrigated viticulture
Abstract
The question of vineyard water management in Australia is often associated with irrigation, but in many cooler and intermediate climate regions, rainfall will make the greater contribution to the seasonal water requirement. This project investigated the extent of management control over this rainfall derived water, with a focus on factors that influence grapevine root distribution, ground cover competition and the possibilities to improve drought resilience without increasing reliance on irrigation. Through a combination of field based studies with contrasting established vineyards in the Orange Wine Region, and water balance modelling to simulate potential drought adaptation strategies, it was shown that there is still significant scope within existing water budgets to adapt to future increases in water demand. A requirement, and topic for future research, may be the development of flexible floor management strategies to follow rainfall extremes, and approaches to encourage root exploration for a greater share of available soil water capacity when new vineyards are established.
Summary
This project examined the extent of management control over water use in rain dependent viticulture with the aim of identifying strategies to improve drought resilience of current and future vineyards. It was undertaken against a backdrop of increasing use of permanent ground cover, and a need to better understand how the benefits of ground cover can be balanced against increased competition in water limited environments. A field-based component of the project, using sites in commercially operated vineyards, addressed the question of cover crop water use traits and the impact of irrigation, rainfall and soil water competition on grapevine root distribution. This was complemented by additional field and controlled environment studies of individual cover crop species. A water balance modelling component, undertaken in collaboration with the Hochschule Geisenheim University (HGU) in Germany, then used this information to simulate soil water dynamic responses to varied ground cover management practices. By extending these simulations using longer term historical climate data, possible adaptation strategies were then compared based on the risk of exceeding soil water stress thresholds at different stages through the season.
The project built on a two year lead-in study funded through the former National Wine and Grape Industry Centre, where soil water dynamics and grapevine water status were monitored at three Shiraz vineyards in the Orange wine region through the 2019/20 and 2020/21 seasons. Soil water content was measured with a manual capacitance probe using 1.5-1.6 m deep access tubes, and plant water status with a pressure chamber to record stem water potential. No treatments were applied at the vineyards, but there was a contrast of water availability with one rain-fed site and two irrigated sites. There was also a contrast in floor management, with one vineyard having full permanent plant ground cover and two having permanent ground cover in the mid-row only with herbicide used in the undervine area. Support provided by Wine Australia through CSU 1702-9 allowed the continuation of these measurements through the 2021/22 and 2022/23 seasons. The new project also supported the upgrade of site instrumentation, and a detailed study of grapevine root distribution. This component of the work used soil coring in combination with next generation DNA sequencing to produce transect maps of root length density from the vine row to mid-row. It allowed the root depth of each plant species found in the vineyards to be determined and linked to water use characteristics of the leaves. Microbial profiling of the same root samples provided further insight into the impact of floor management, irrigation and soil depth on root function.
The first soil moisture measurements in winter 2019 showed that the profiles had been significantly depleted during the preceding seasons to at least a depth of 1.6 m, meaning that the vineyards entered the third and final season of the drought with soils at well below field capacity. The water that was available was used by late spring at two of the three vineyards, and plant water potential measurements showed that the vines were in severe water stress by the end of December. Irrigation volumes at the third vineyard site were just sufficient to avoid the severe water stress threshold, but at all three vineyards the impacts of water stress were clear with reduced shoot growth and berry weights, low yeast assimilable nitrogen and rapid increases in juice sugar concentrations. The water deficits were only relieved by late summer storms and then a return to more average autumn and winter rainfall. Three subsequent La Niña events, following the declaration of the first in late September 2020, then shifted water availability to the other extreme.
Field capacity was estimated at each site from periods of stable water content during winter of the subsequent seasons with the aim of establishing the total amount of water that could be stored within the rootzone at each vineyard. For the rain-fed vineyard, where the soil was most suited to the use of capacitance probes, the total profile sum was estimated at 250 mm. Soil shrinkage with clay layers around the access tubes at the other two vineyards prevented an accurate lower limit being established, and therefore a full profile sum could not be determined with confidence. However, values of at least 300 mm would be expected given the soil types and root depth at these vineyards. On a volumetric basis this is equivalent to 2.5 to 3 ML/ha of plant available stored water in the profile when full, which is well beyond the capacity of most vineyards in the region to compensate for with irrigation in the event of low winter rainfall and drought.
The root transect mapping showed that grapevine roots were present to the depth of the soil moisture monitoring, but the roots of a number of ground cover species where able to match this. Perennial grasses were the dominate group at all of the vineyards, with the roots of two introduced pasture species, Festuca arundinacea and Paspalum dilatatum, and the native Bothriochloa macra at the warmer site, consistently found between 1.3 and 1.6 m. Roots of the perennial herb Plantago lanceolata could also reach this depth, and even short lived annual roots of Avena barbata and Trifolium subterranean were found to at least 1 m deep, before the grapevines reached bud-break. Most of the other species present also appeared to occupy at least the top 1 m of the profile. These findings suggest that in the case of established ground cover, the full sum of water potentially available to grapevine roots is also accessible by the other plants. As a number of the annual and perennial species identified are active months before the grapevines develop a full canopy, they can also utilise this water before the grapevine roots have sufficient transpiration-driven water demand to compete.
For lateral root exploration, it was found that grapevine roots did not occupy the full volume of soil under the mid-row. As the proportion of irrigation relative to rainfall derived water increased, grapevine roots were mostly aligned with the vine row leaving the soil under the mid-row dominated by the roots of other plants. Where water was only supplied by rainfall there were more roots in the mid-row compared to the vine row, but these were deeper in the profile. It appeared that the most extensive exploration of grapevine roots occurred with the combination of some irrigation, but then higher rainfall, to encourage exploration of the mid-row. Even then, the top soil was dominated by the roots of the ground cover species. The root length density on the 0-10 cm depth increment was over 20 cm/cm3, providing an extremely competitive environment for vine roots, and well above the values thought to be needed to extract all the plant available water from the surrounding soil.
The possibility of planting cover crop species that use less water is often discussed, but the findings of this project suggest that this would be difficult to achieve in practice in higher rainfall regions. In general, all species that had green leaves had equal or higher rates of transpiration than grapevine leaves, and as a mixed population, could access the same soil water and for more of the season. For example, in addition to deep roots, many of the C3 annual and perennial grasses commence growth through winter and early spring and use a significant amount of the stored soil water early in the season. The C4 grasses and perennial C3 grasses then have the collective capacity to respond more rapidly to summer rainfall, or maintain green cover on the vineyard floor after the grapevine canopy has senesced in autumn. A cover crop plot trial on the Charles Sturt University Orange campus also demonstrated the challenges of trying to establish a specific species, and have it persist in the presence of a seed bank or neighbouring population of plants that are more competitive.
Some of the benefits of retaining permanent ground cover, in addition to soil physical and chemical attributes include their lower maintenance, improved trafficability in the mid-row, suppression of less desirable weeds, and reducing vigour in wetter seasons. The preliminary results of the microbial profiling from the current study also suggest the presence of ground cover causes a relative shift towards more favourable bacterial and fungal taxa. For example, the families Bradyrhizobiaceae and Rhizobiaceae, which include many symbiotic nitrogen fixing species, the family Rhodospirillaceae and genera Burkholderia and Azospirillum which include free-living nitrogen fixing species that can associate with grasses. Both of the families were more abundant in roots under permanent ground cover than where herbicides were used at the trial sites, potentially providing a source of nitrogen for the vine roots below. Conversely, the fungal family Nectriaceae, which contains a number of root pathogens was associated with roots at all sites and all depths, but was more abundant where there were more grapevine roots. Where there was a side-by-side comparison of long term permanent mid-row ground cover with bare soil under the vine row, soil carbon was also slightly higher in the mid-row. All of these factors have the potential to enhance the long term sustainability of a vineyard, so the key question appears to be the development of agronomic practices that retain as many of the benefits of ground cover as possible while having the flexibility to move soil water balance in favour of the grapevines during dry seasons or drought.
In the first step toward developing such strategies, the HGU water balance model was successfully converted to the southern hemisphere and validated against the measured soil water data from the rain-fed trial site. Modelled data for the other two vineyards were then used to calculate the relative shares of grapevine transpiration, cover crop transpiration and soil evaporation for the study period. Simulations for the rain-fed vineyard were also run using SILO interpolated weather data back to 1889, to compare the conditions of the recent drought with historical droughts in the long term climate record. The results of the modelling show that the relative share of grapevine water use reaches about one third at the maximum canopy extent, but ranged between 14 and 20% on a whole season basis due to the earlier growth of the ground cover plants and relatively wide row spacing that allows more radiation to reach the vineyard floor than is intercepted by the vine canopy. For the simulation run back to 1889, the 2017 to 2019 drought appeared to be the only period in the historical climate record where severe soil water deficits extended for three full consecutive years.
A set of four simulations was then run from 1970, comparing contrasting percentages of vineyard ground cover, shallower cover roots and narrower row spacing with the current vineyard parameters. The impacts of these practice changes were summarised as the number of days below mild and severe water stress thresholds, which were defined as when the percentage of available soil water fell below 40% and 10% respectively. None of the simulated adaption practices would have made a difference to the severity of the 2017 to 2019 drought, but having no ground cover or 50% ground cover, reduced the average risk of severe water deficit in the bud-break to veraison period from 19% to between 4 and 5%. This is where the most impact on yield would be expected. Reducing the row spacing from 3 m to 2 m did not reduce total water use as it exchanged light interception by ground cover plants with light interception by the grapevine canopy. However, yield relative to water use would have potentially been increased by a third with the increase in planting density. Simulating a shallow cover crop root system, although potentially difficult to achieve in practice, suggested an intermediate reduction in water risk where grapevine roots do not have to compete in the whole profile for water.
For existing vineyards, the work has highlighted that the water competition from cover crops is significant, and that there may be less scope to control this through species selection or existing management practices than is generally appreciated. In a higher rainfall region like Orange, this may not be a problem in all seasons, and irrigation would usually be sufficient to buffer shorter periods of drought. However, in the event of longer term dry periods that reduce irrigation storage, ground cover plants can rapidly deplete soil water in spring that could otherwise be used by the vines. For these situations, future work could focus on methods to more effectively reduce water competition on a temporary basis. If there is a permanent decline in water availability relative to water demand with climate change, permanent adaptations may be required. This could include the approach of using mid-row ground cover in every second row or alternating mid-rows of permanent perennial cover with a mid-row managed on annual basis according to water availability.
For future vineyards, there is more scope for adaptive changes to drought and rainfall variability, including the selection of new sites with higher water holding capacity that allows more buffering from high rainfall seasons to low, differing slope and aspect, and the ability to vary row spacing, orientation and canopy architecture to modify radiation interception. However, even for existing sites that are being replanted, there could be significant gains in storage capacity by managing the early seasons after planting to increase root depth and lateral spread. This could include the way irrigation is managed and cover crops are introduced, in the first few years following planting. The project results also suggest that if soil moisture monitoring is used, some measurements could be made away from the volume of soil wetted by irrigation, where they are not influenced by short term irrigation events that have minimal impact on total available soil water. This would track to longer term rainfall driven trends in soil water availability necessary for advanced drought planning, ideally with deep profile probes that can provide a volumetric water sum.
As a concluding point, there is still an important question as to what rainfall scenarios need to be prepared for in the future. While the 2017 to 2019 drought appeared to be the most severe on record, and tested irrigation supplies to the limit, the vines did recover in the next season. Is the priority for future research the possibility of another three-year drought or worse, a gradual long-term increase in water demand relative to water availability, or strategies that consider both? The study findings suggest that current practices could be adapted to a gradual rise in water demand without the need for more water, but possibly not for a drought that extended beyond three years without increased irrigation security. Similar to the climate risk studies undertaken with the HGU water balance model in Germany, further research could run management simulations forward in time with downscaled climate projection data. Understanding how future water demand will change compared to historic conditions could provide a clearer understanding as to how much existing practices would need to change in order to adapt to future water supply scenarios.