Mechanical Harvesting Has Little Effect on Water Status and Leaf Gas Exchange in Citrus Trees

Mechanical harvesting of citrus trees can cause physical injuries, such as shedding of leaves, exposing roots, and scuffi ng bark. Although mechanical harvesting usually has not reduced yield, physiological consequences to the tree from these visible injuries have not been investigated. We hypothesized that physical injuries to tree canopies and root systems from a properly operated trunk shaker would not cause short-term physiological effects. Tree water status and leaf gas exchange of mature ʻHamlin ̓and ʻValencia ̓sweet orange [Citrus sinensis (L.) Osb.] trees that were harvested by a trunk shaker were compared to hand-harvested trees. A trunk shaker was operated with adequate duration to remove >90% of mature fruit or with excessive shaking time under various environmental conditions and drought stress treatments throughout the harvest season. Mid-day stem (Ψstem) and leaf (Ψleaf) water potentials along with leaf gas exchange were measured before and after harvest. Trees harvested by the trunk shaker did not develop altered water status under most conditions. Trees harvested with excessive shaking time and/or with limited soil water supply developed low Ψstem resembling Ψstem of drought-stressed trees. However, water potential of all treatments recovered to values of the well-irrigated, hand-harvested trees after rainfall. In addition, mechanical harvesting did not reduce CO2 assimilation, transpiration, stomatal conductance, water use effi ciency, or photosystem II effi ciency as measured by chlorophyll fl uorescence. Thus, despite visible injuries, a properly operated trunk shaker did not result in any measurable physiological stress. Mechanical harvesting of fruit crops gained its fi rst major thrust in the1930s (OʼBrien et al., 1983), and the principles of machinery-operated harvesting as well as its economical value have been developed ever since. Numerous types of harvest machines have been designed and commercially applied in various fruit industries including citrus in Florida (Hedden et al., 1983; Whitney, 1995). The major benefi t of mechanical harvesting relies on its effi ciency and lower costs in comparison to harvesting by hand, thereby increasing grower profi t. However, the adoption of mechanical harvesting in Floridaʼs citrus industry has been unexpectedly slow. By 2003 <3% of its 240,000 ha of processed orange groves were mechanically harvested (Brown, 2005). Many factors contribute to the lack of widespread adoption of mechanical harvesting by the Florida citrus industry. Among these factors is the apparent violent shaking that trees experience during harvesting with trunk or canopy shakers. This is usually the growerʼs fi rst impression of mechanical harvesting. Depending on the type of the harvest machines, the skill of the operators, the weather, the grove conditions, and the physiological status of the trees, a harvest machine can cause visible physical injuries to the trees. These include shedding of leaves, fl owers, and young fruit, breaking branches, scuffi ng of bark, and exposing roots (Halderson, 1966). Nevertheless, long-term studies revealed that fruit yield of citrus trees was affected little by mechanical harvesting comReceived for publication 25 Jan. 2005. Accepted for publication 21 Mar. 2005. We thank John Crum from FMC Corp. for use of the trunk shaker and data acquisition systems. Fernando Alferez, Baylis Carnes, Jill Dunlop, Igor Kostenyuk, Luis Pozo, Ana Redondo, Shila Singh, Zhencai Wu, and Rongcai Yuan assisted in harvest. Thanks to Jackie Burns and Richard Buker for critical reviews. Mention of a trademark does not imply endorsement of the product named, or criticism of similar products not named. This work was supported by UF/IFAS and partially funded by the Florida Dept. of Citrus. Florida Agricultural Experiment Stations Journal Series No. R-10706. 1To whom reprint requests should be addressed; e-mail ktli@ufl .edu. pared to trees that were hand harvested (Hedden et al., 1988). In cherry (Prunus cerasus L.) trees, however, mechanical harvesting hastened tree decline and cut their productive life in half (Burton et al., 1986). Although there is no evidence that visible injuries could seriously weaken citrus trees, tree health after mechanical harvesting remains a major concern to orange growers. Previous studies have concentrated on fi nding the potential cause of mechanical injuries (Abdel-Fattah, et al., 2003; Gurusinghe and Shackel, 1995) as well as improving machinery design and harvest effi ciency (Affeldt et al., 1989; Fridley, 1983; Peterson, 1998). It is apparent that currently available harvest machines have been greatly improved and tree injures caused by modern shaking machines are less severe than they once were. On the other hand, it also appears that unless a more sophisticated means of mechanical harvesting is employed, some visible tree injuries are unavoidable. This underscores the need to document the treeʼs physiological responses to injuries by mechanical harvesters. Despite the fact that secondary infections after immediate injuries have been blamed for tree decline in mechanical harvested cherry, almond [Prunus dulcis (Mill.) Webb.] (De Vay et al., 1968), and peach [Prunus persica (L.) Batsch.] (Glenn, et al., 1995) trees, physiological reactions to physical injuries by harvest machines to these fruit trees or to citrus, have not been documented. In 2003, we initiated the fi rst program dedicated to investigating the effects of mechanical harvesting on tree physiological activity in citrus trees. The objectives of this program were to 1) quantify the relationship between harvest machines and the potential of short and long-term physiologically important injuries, 2) examine the physiological effects of mechanical harvesting under different weather conditions and tree water status over the 6-month-long harvest season, and 3) develop safe operating conditions for mechanical harvesting to avoid tree injury in an effort to accelerate the adoption of mechanical harvesting in the Florida citrus industry. 508.indd 661 8/17/05 4:09:07 PM 662 J. AMER. SOC. HORT. SCI. 130(5):661–666. 2005. We measured tree water status and leaf gas exchange characteristics after harvesting by a trunk shaker, one of the most popular harvest machine types in Florida. We hypothesized that the severity of injuries caused by a normally operating trunk shaker would not induce signifi cant physiological stresses in well-managed citrus trees. In addition, if a tree that was under environmental stresses or was actively growing, it likely would be more vulnerable to any physiological stress induced from mechanical harvesting with an improperly operated shaker. Materials and Methods Three grove sites of ʻHamlin ̓or ʻValencia ̓sweet orange (C. sinensis) trees on Carrizo citrange [C. sinensis x Poncirus trifoliate (L.) Raf.] or Swingle citrumelo (C. paradise Macf. x P. trifoliate) rootstocks at the Univ. of Floridaʼs Citrus Research and Education Center in Lake Alfred were used in this study. The soil type of these sites is Candler fi ne sand with low water holding capacity and low organic matter. Each site was subdivided by rows or sections to form a random block design. Trees received normal grove management, irrigation, pest control, and fertilization. During the dry season between late October and May, irrigation by microsprinklers was applied at the rate of 170 to 227 L/tree twice per week as needed. Harvest dates were the normal harvest period for either variety in Florida. ̒ Hamlin ̓is harvested between November and February, whereas ̒ Valencia ̓is harvested between February and June. Bloom time for both varieties is normally in mid-March (Davies and Albrigo, 1994) TRUNK SHAKER. A prototype (Dotan Ltd., Migdal Haeʼmek, Israel) FMC linear-type trunk-shaking system (FMC Corp., Lakeland, Fla.) was used in this study. The padded clamp shaker head was equipped with 70.8 kg of unbalanced weight and connected by a power take off to a tractor engine operating at 2100 rpm. This weight and power combination was selected to generate a shaking frequency of 4 Hz with a maximum trunk displacement of 6.5 cm. The shaking frequency and force output were evaluated on adjacent trees with a set of accelerometers (PCB 35B33; PCB Piezotronics, Depew, N.Y.) and high-speed portable data acquisition systems (WavePort/PE8; Iotech, Cleveland). Trunk displacement was measured directly using a measuring tape. EXPERIMENT 1. Grove site one contained forty-fi ve 14-yearold ʻHamlinʼ/ʻSwingle ̓trees in an east-west oriented row. Trees were 3.5 m tall and spaced at 6.1 m between rows and 4.6 m within rows. The trunk diameter averaged 14 cm at 30 cm above ground. Thirty uniform trees were grouped into fi ve blocks of six trees. Irrigation was withheld for 3 weeks before harvest and no rainfall was recorded during this period. On 7 Jan. 2004, two trees in each block were harvested by hand (Hd) and three trees were harvested by the trunk shaker. Hand harvesting removed 100% of the fruit whereas mechanically harvesting removed >90% of the fruit. Two trees harvested by the trunk shaker were subjected to 10 s shaking time (10S) and the other tree to 20 s (20S). The remaining tree in each block was not harvested and served as an un-harvested control (C). Irrigation was resumed in this grove the day after harvest except on one of the hand-harvested trees (HdD) and one of the 10 s mechanical harvested trees (10SD) in each block. To test effects of drought stress, irrigation was withheld from these trees for one additional month. To characterize physiological responses of trees, mid-day stem water potential (Ψstem), leaf gas exchange, and leaf chlorophyll fl uorescence were measured 1 d before harvest and every 1 to 3 d thereafter until 6 Feb. 2004. Details of physiological measurements are described below. EXPERIMENT 2. Grove site two contained thirty-eight 12-yearold ʻHamlinʼ/ʻCarrizo ̓trees in fi ve north-south oriented rows. Trees were 3.7 m tall and spaced at 4.5 to 9 m between rows and 3 m within rows. The trunk diameter averaged 12.6 cm at 30 cm above ground. Twenty-fi ve uniform trees were grouped by rows into fi ve blocks of fi ve trees. Irrigation was withheld from this grove site for 2 weeks before harvest on 28 Jan., but 33.78 mm rain fell on 18 and 19 Jan. and another 13.46 mm on 27 Jan. Consequently, soil was well watered prior to harvest. Two trees in each block were harvested by hand (Hd) whereas three others were harvested using the trunk shaker as above. One mechanically harvested tree in each block was shaken for 10 s (10S) and the others for 30 s (30S) shaking time. After harvest, irrigation was resumed in this grove except on one of the hand harvested (HdD) and the 30S trees (30SD) in each block, in which irrigation was withheld for one additional week. Mid-day Ψstem and leaf gas exchange was measured on 23 and 29 Jan. and 5 Feb. as described below. EXPERIMENT 3. Grove site three contained forty 15-year-old ʻValenciaʼ/ʻSwingle ̓trees in a north-south oriented row. Trees were 3.5 m tall and spaced at 4.56 m between rows and 2.13 m within rows. The trunk diameter averaged 36.8 cm at 30 cm above ground. Twenty-fi ve uniform trees in full bloom were grouped into fi ve blocks of fi ve trees. Soil was well watered since irrigation at the rate of 227 L/tree was applied daily to this grove site from 11 to 15 Mar. plus 36.07 mm rain fell on 15 Mar. and 55.12 mm on 16 Mar. Two trees in each block were harvested by the trunk shaker for 10 s on 17 Mar. One tree was hand harvested on 19 Mar. and another by the trunk shaker for 20 s. The remaining tree in each block was not harvested (C). After harvest, a portion of the bark was removed from the main trunk of one of the 10S trees in each block using a sharp knife to simulate severe bark injury (10SB) by the trunk clamp. An average of 7.5 × 21.2-cm bark area ≈20 to 40 cm above the soil surface on both the north and south side of the main trunk was removed. This corresponded to the area that was in contact with the shaker clamp pads. The total width of the removed bark equaled 42% of the trunk circumference. After 19 Mar., irrigation was withheld from trees that had been harvested until late April, while a regular irrigation schedule was continued on the un-harvested control trees. Mid-day Ψstem and leaf water potential (Ψleaf) were measured as described below every one to three days between 18 Mar. and 4 Apr., and again on 20 and 28 Apr. Soil water content was measured gravimetrically immediately after Ψ measurements. Five soil samples from the root zone were collected at each measurement date at depths of 0 to 30 cm and 30 to 60 cm. Soil water content was calculated by the difference between fresh and oven-dried soil weight. LEAF AND STEM WATER POTENTIAL MEASUREMENTS. Mid-day Ψstem and Ψleaf were measured with a pressure chamber (Scholander et al., 1965) between 1330 and 1500 HR; Ψstem was measured on leaves that had been enclosed in plastic bags and covered by aluminum foil at least 3 h before measurement. Two exposed mature leaves from the last mature summer or spring fl ushes were used for each mid-day Ψstem measurement. Ψleaf was measured on uncovered, transpiring leaves immediately following Ψstem measurements on two similar adjacent leaves. LEAF GAS EXCHANGE MEASUREMENTS. Leaf gas exchange was measured with a portable photosynthesis system (LI-6200; LI508.indd 662 8/17/05 4:09:12 PM 663 J. AMER. SOC. HORT. SCI. 130(5):661–666. 2005. COR, Lincoln, Nebr.) on selected clear days between 0930 and 1130 HR. Photosynthetic photon fl ux density (PPFD) was greater than 900 μmol·m–2·s–1 during all measurements. Leaf temperature in the measurement cuvette was usually 2 to 3 °C above ambient air temperature. Relative humidity (RH) inside the cuvette was close to ambient RH. Measurements were made across all treatments in two 1-h blocks to allow testing for effects of changing environmental conditions over time. On each date measurements were taken on two to three healthy light-exposed mature leaves on the last summer or spring fl ushes on each tree. Net assimilation of CO2 (ACO2), stomata conductance (gs), transpiration (E), and leaf water use effi ciency [WUE (micromoles CO2 per millimole H2O)] was calculated. LEAF CHLOROPHYLL FLUORESCENCE MEASUREMENTS. Chlorophyll fl uorescence was measured with a pulse-modulated fl uorometer (OS-1-FL; Opti-Science, Tyngsboro, Mass.) on selected days between 1300 and 1330 HR. Two leaves from each tree were dark-adapted with leaf clips for at least 2 h prior to measurement to ensure the photosystem II reaction centers were in an active open status. The minimal fl uorescence emission from the darkadapted leaf area (Fo, nomenclature from van Kooten and Snel, 1990) was excited by low intensity (<1 μmol·m–2·s–1) modulated 655-nm light and was detected in the 700to 750-nm ranges. After Fo was recorded the photosystem was saturated by a high intensity (≈15,000 μmol·m–2·s–1), 350to 690-nm light pulse for 1 s to induce the maximal fl uorescence from the dark-adapted leaf area (Fm). Variable fl uorescence from the dark-adapted leaf area (Fv) was calculated from (Fm – Fo) and maximum quantum effi ciency of PSII photochemistry was derived from Fv/Fm (van Kooten and Snel, 1990). Measurements were taken on every tree in Expts. 1 and 2. ENVIRONMENTAL DATA AND STATISTICAL ANALYSIS. Daily temperature, precipitation, and RH for the Lake Alfred site were obtained from Florida Automated Weather Network (Univ. of Florida/Institute of Food and Agricultural Sciences, 2005). Maximum daily vapor pressure defi cit (VPD) was derived from maximum day temperature and minimum RH. Data were subjected to analysis of variance (ANOVA). Each tree was an experiment unit. Measurements from each tree were averaged before analysis using general linear model or balanced model procedures. Data collected after harvest were also pooled and subjected to repeated measures ANOVA where appropriate. Mean values of treatments with or without standard error are presented. Signifi cant differences were determined at P ≤ 0.05, 0.01, or 0.001 and separated by Dunnʼs or Duncanʼs multiple comparison tests where appropriate. Due to lack of interaction between treatment and blocks overtime, only treatment and block effects are reported.