Abstract Withholding alfalfa (Medicago sativa L.) irri-

Withholding alfalfa (Medicago sativa L.) irrigations during the summer, a practice referred to as summer irrigation termination (SIT), can conserve substantial amounts of water in long-season desert environments; however, plant mortality associated with SIT may be substantial. Proper timing of re-irrigation is critical for minimizing mortality and yield reductions following SIT. Procedures that would permit probable mortality prediction during drought stress would improve management efficiency with SIT. This study was conducted to determine (1) whether plant mortality occurs once the moisture content of the plant woody stem portions (crown) falls below some critical threshold and (2) if such a threshold could be used to predict the likelihood of plant mortality during SIT. Crown samples were taken from single, spaced, fieldgrown plants in Tucson, Arizona, at the end of a 84-day SIT period in 1994. A crown moisture content of about 42% was identified as a likely threshold critical for crown tissue SIT survival. This value was then used to predict whole-plant mortality of alfalfa grown in solid-seeded plots comparable to commercial fields. Crown samples were taken at five locations within the field along a solid gradient that was related to plant mortality. At each sampling location, the proportion of samples with less than 42% crown moisture was used to predict plant mortality. Predicted mortality slightly overestimated actual mortality but differences between predicted and observed mortality were significant for only one of five sampling locations. Alfalfa growers may be able to use this simple method of crown moisture determination to prevent permanent yield reductions by initiating irrigation before substantial portions of crowns fall below the threshold moisture content of 42%. Introduction The annual water consumption of the perennial forage crop alfalfa (Medicago sativa L.) can be extremely high in longseason desert environments where nearly continuous yearround growth is possible. For example, irrigation application can exceed 200 cm per year in the southwestern United States (Robinson and Teuber 1992). Rising irrigation costs and water demand from municipalities have forced alfalfa growers in this region to use water more efficiently (Frate et al. 1991). During the period from July to September, peak water demand from municipalities and other crops coincides with low yields and low water use efficiency in alfalfa (Kipnis et al. 1989). In addition, hay quality and prices in the southwestern United States are typically lowest during this period (Hood 1992). This combination of factors drastically reduces the profitability of alfalfa production from July to September, and in extreme cases can lead to net losses for growers (Husman 1992). Discontinuation of irrigation during this period, forcing plants into semidormancy, could increase overall water use efficiency in alfalfa and make up to 800 mm of irrigation water available for other, more economical uses with minimal effects on farm income (Ottman et al. 1996; Robinson and Teuber 1992). In addition to water conservation efforts, summer irrigation termination (SIT) has been considered in areas where heavy insect infestations severely affect the production of fall vegetable crops (Wrona 1992). Withholding alfalfa irrigation from mid-summer to fall could interrupt the nearly continuous year-round cropping pattern in long-season desert environments. Alfalfa can be the main host for insects such as the sweetpotato whitefly (Bemisia tabaci) between summer crop harvests and fall crop plantings (Natwick et al. 1992). Depriving whiteflies of a preferred host in early fall may reduce whitefly populations to a level where biological control agents or insecticides will be an effective control in fall crops. In order to use SIT successfully, plant mortality must be limited to levels that do not lead to a permanent reducIrrig Sci (1997) 17: 87–91 © Springer-Verlag 1997 Received: 10 May 1996 Matthias Wissuwa · S. E. Smith · Michael J. Ottman Crown moisture and prediction of plant mortality in drought-stressed alfalfa ORIGINAL PAPER M. Wissuwa · S. E. Smith (1⁄2) · M. J. Ottman Department of Plant Sciences, University of Arizona, Tucson, AZ 85721, USA e-mail: [email protected] tion in forage yield. Studies conducted to evaluate the effect of SIT on the productivity of alfalfa stands have produced mixed results. In Cyprus (Metochis and Orphanos 1981), central California (Frate et al. 1991), and central Arizona (Ottman et al. 1996), 56–63 days of SIT were associated with reduced forage yields only in the first postSIT harvest. However, in the Imperial Valley of California (Wrona 1992) and in western Arizona (Ottman et al. 1996), stand density and post-SIT yield were significantly reduced on soils with a low water-holding capacity or if SIT was prolonged to October (>110 days) (Ottman et al. 1996; Wrona 1992). New herbaceous shoots originate from the perennial basal (woody) portion of the stem following SIT, an area referred to as the crown (Steward 1926). Even after all herbaceous shoots have died, plants can recover from drought stress if crown tissue is capable of producing new shoot growth from axillary buds within the woody crown tissue. The degree of mortality in crown tissue (100% crown mortality = whole-plant mortality) will be a major factor affecting post-SIT productivity of individual plants and whole alfalfa stands. Crown mortality during SIT is likely to depend on the amount of water lost from this tissue. Crown moisture content may therefore indicate whether or not a plant will survive SIT. Where soil conditions may lead to high plant mortality or under conditions that lead growers to extend the SIT period beyond 110 days, irrigation timing will be of utmost importance to minimize mortality. Reports of high mortality rates in commercial fields that were subjected to SIT in the Imperial Valley (Wrona 1992) indicate the need for an objective criterion that will allow growers to schedule irrigation based on plant mortality estimates. The objective of this study was to test the hypothesis that crown mortality occurs when crown moisture content falls below some critical threshold level. Parts of crowns from field-grown plants were sampled to determine their moisture content and the remaining unsampled crown area was used to assess mortality during SIT. Data from these experiments were used to determine a crown moisture value at which death will occur. After this threshold value had been established, it was used to predict alfalfa plant mortality in a second field trial exposed to SIT. Materials and methods The experimental site was located at the Campus Agricultural Center, Tucson, Arizona. Plant material used to established a crown moisture content threshold were single, spaced plants of a composite population of very nondormant Arabian and North African ecotypes (AZ91-AC). Seed was produced in a bee cage on plants that represented bulked open-pollinated Syn-1 seed from the six highest-yielding entries (described in Smith et al. 1991) in a 2-year field trial of extremely nondormant Arabian and North African ecotypes at Tucson, Arizona. Individual plants of this composite were sown in two contiguous field plots, which were treated equally and considered replicates, on a soil classified as Pima clay loam (Fine-silty, mixed (calcareous), thermic Typic Torrifluvent) on 30 September 1993. Each plot contained 240 plants (ten rows, each with 24 plants), which were spaced 25 cm within and 50 cm between rows. The two outside rows and three plants at both ends of rows were treated as borders. Plots were sprinkle irrigated until seedlings were established and flood irrigated at approximately 120% of cumulative evapotranspiration thereafter (60 mm of irrigation every 5–10 days), except for the interval from 18 July to 10 October 1994, when irrigations were withheld for a 84-day SIT period. Cumulative evapotranspiration during the SIT period was 558 mm and precipitation 85 mm (data obtained from the Arizona Meteorological Network at the Campus Agricultural Center, Tucson, Arizona). The volumetric soil moisture content was measured with a neutron probe at depths of 15 and 105 cm on days 2, 42, and 84 of the SIT period. Forage was harvested when most plants were at approximately 10% bloom five times prior to the SIT period (10 cm stubble). We considered the day of the last pre-SIT harvest to be day 1 of SIT because the next irrigation would have been scheduled for that day. On day 4 of SIT, the percentage of the crown area that produced new shoots was estimated visually. This procedure was repeated after the first post-SIT harvest in December 1994. From both estimates, crown mortality was calculated as a percentage of the productive pre-SIT crown area. On days 5, 42, and 84 of the 1994 SIT period, one branch (approximately 3 cm in length and 2 g in weight) of the woody, lower part of the crown was sampled on 60 plants in both plots (three alternate rows per plot with 10 plants sampled per row). Crown samples were wrapped in aluminum foil, stored in plastic bags, and placed on ice for transportation to the laboratory. Prior to fresh weight determination, all nonwoody plant material was removed from crown samples. This included dry stubble and any herbaceous shoot tissue. Samples were shock-frozen in liquid nitrogen and stored at –80 °C. Dry weight was determined after lyophilization for 24 h. Crown moisture content was calculated as percent moisture in the fresh sample. Crown mortality in neighboring rows was used to determine whether sampling influenced survival and crown mortality. A onesided t-test was conducted to compare mean crown mortality of sampled and unsampled rows. The following year (1995), crown moisture content was used to predict mortality in field plots that closely resembled conditions in commercial alfalfa fields in this region. The experiment was located at the Campus Agricultural Center, Tucson, Arizona, on a soil classified as a Pima-Agua Complex: Pima clay loam [Fine-silty, mixed (calcareous), thermic Typic Torrifluvent]; Agua sandy clay loam [Fine-loamy over sandy, mixed (calcareous), thermic Typic Torrifluvent]. Six experimental populations were sown in a latin square design on 30 March 1995 at a seeding rate of 6.7 kg ha. Plots contained five 1-m-rows with 15-cm spacing between rows and 30 cm between plots. Initially, 60 seeds m were sown; these were later thinned to 22 plants m. A 7-row border seeded with the cultivar Lew (20 kg ha) surrounded the experimental area. The field was flood irrigated at approximately 130% of cumulative evapotranspiration (70 mm per irrigation every 5–10 days) until 1 August 1995, when a 71-day SIT period was imposed. Cumulative evapotranspiration during the SIT period was 471 mm and precipitation 81 mm (data obtained from the Arizona Meteorological Network at the Campus Agricultural Center, Tucson, Arizona). Plots were harvested twice prior to the SIT period when most plants were at approximately 10% bloom (10 cm stubble). Live plants in each plot were counted on day 3 of SIT and following the first post-SIT forage harvest on 17 November 1995. An ANOVA was conducted and demonstrated no differences in plant mortality between experimental populations in the latin square. Soil variability strongly affected the plant response during the SIT period along a gradient from southwest (all leaves desiccated) to northeast (the majority of leaves remained turgid) within the experimental field. This gradient allowed us to take crown samples from areas with the experimental field that had experienced different degrees of drought stress. Crown samples were taken from five locations along this gradient prior to initiating irrigations. These five locations represented the whole range of stress intensities observed in our field and were considered nonrandomly arranged treatments. We designated these treatments (locations) A– E (A = plants most stressed, E = least stressed). 88