Abstract
Transpiration of woody shrubs appears to increase with decreases in plant density within production beds as plants are randomly removed for sale. To assess potential impact on irrigation management, this observation was tested with market-sized plants in suspension lysimeters at specific levels of canopy closure. Canopy closure was defined as the percentage of cumulative projected two-dimensional canopy area of individual plants per unit ground area on which they were placed. In 1997, evapotranspiration (ETA) of plants in 26.6-L containers was comparable from isolated plants up to 67% canopy closure. At full canopy closure (100%), ETA was 40% less than 67% closure or lower. When repeated in 2003, results were similar for similar-sized plants and for two sizes smaller (11.4- and 3.8-L containers). ETA response to canopy closure was independent of height from 0.5 to 1.5 m tall. At full canopy closure, whole plant transpiration was equivalent to that measured from only the upper 40% (by height) of the canopy under full sun. This was independent of plant size. Implications for water conservation during production and plants’ irrigation needs in landscapes are discussed.
During selection of woody ornamentals for market, it is common in smaller nurseries to selectively choose the largest and fullest plants. Selection of the largest plants is considered beneficial in several ways. It promotes good customer relations. Larger plants in containers tend to require more irrigation, so their removal is thought to reduce irrigation frequency, or at least spot hand-watering. Furthermore, reduced plant density is thought to be associated with reduced levels of shoot diseases. However, at some undetermined reductions in plant density, irrigation requirements for sections that have been selectively harvested have been noted to be higher than in similar, adjacent unharvested blocks (R.C. Beeson Sr., Co-owner, Beeson's Rhododendron Nursery, personal communication).
Rose (1984) proposed a theoretical model for transpiration from isolated forest trees. In brief, canopy transpiration of isolated trees was partitioned into two fractions. One was termed vertical transpiration. This was defined as movement of water vapor to the air above the canopy from upper leaves responding to solar radiation, wind, and vapor pressure deficit (VPD). In biophysical terms, this would be the vertical flux density of water vapor moving by eddy diffusion, similar to uniform swaths of grass or dense-canopy agronomic crops (Nobel, 1999). The other fraction was termed horizontal transpiration, driven principally by VPDs caused by eddy movement horizontally through the shaded lower canopy (advection; Nobel, 1999). Rose's model predicted that as forest canopy cover approached 100%, i.e., complete canopy closure, air movement through the lower canopy would decline. At 100% canopy closure, where there is complete overlap among tree canopies, most if not all tree transpiration was predicted to be restricted to a thin, upper layer of a tree canopy; thus, total tree transpiration would be less than that of an isolated tree (Rose, 1984). No actual measurements accompanied this theoretical model.
In 1996, a study was completed that quantified water use during production of three representative woody shrub species of marketable size for an entire year (Beeson, unpublished data). Decades-long below-normal rainfall was prompting legislation limiting irrigation in both nurseries and landscapes. The only representative data were the research nearing completion where plants were grown in closed canopies blocks. Knowing differences in water use existed between closed canopies and isolated plants, experimental validation and defining the theoretical limits of the model proposed by Rose appeared to be an avenue to extend the results of the concluding project to landscape situations and nurseries searching for ways to conserve water. It also appeared to be a way to explain observations noted decades earlier in rhododendron nursery production. The objectives of this research were to verify theoretical reductions in plant evapotranspiration with increases in canopy closure and to begin defining the depth of the upper layer that constituted vertical transpiration. In 2003, objectives were to verify results recorded in 1997 and to test whether results were dependent on plant heights common in woody shrub nurseries.
Material and Methods
Phase I.
In May 1996, plants of Viburnum odoratissimum Ker Gawl (viburnum) grown in 11.4-L black polyethylene containers (1600; Lerio Corp., Kississimee, FL) were transplanted into 26.6-L containers (B7; Lerio Corp.) at the University of Florida's Central Florida Agricultural Research and Education Center in Sanford, FL. The substrate was purchased from a commercial company (Sunrise Landscape, Longwood, FL) and was the same composition as in the 11.4-L containers, consisting of 64% composted pine bark fines (less than 25 mm), 27% Florida sedge peat, and 9% coarse sand amended with 2.9 kg·m−3 dolomitic limestone and 0.86 kg·m−3 micronutrients (Peter's Fritted Trace Elements; Scotts Company Inc., Marysville, OH). Plants were top-dressed with 120 g of a 18N–2.6P–9.9K controlled-release fertilizer (Osmocote 18-6-12; Scotts Company Inc.) and placed in full sun on black polyethylene woven ground cloth (VJ Growers Supply, Apopka, FL). Plants were then irrigated twice daily until November by microirrigation using spray stakes (0.42 L·min−1; Roberts Irrigation, Plover, WI). Irrigation from November until early January was once daily. Plants were spaced 0.8 m on center within rows and 1.2 m between rows and pruned every 4 to 5 weeks using hand pruners (Felco 2; Geneveys sur Coffrane, Switzerland) during growth to promote a full, dense canopy and plant height of over 1 meter.
In early Jan. 1997, the canopy's widest width and that perpendicular to the widest width were measured for each plant. Widths were multiplied to calculate a two-dimensional rectangular projected canopy area (PCA) for each plant that was attached to it.
Three tripod suspension lysimeters were placed near the center of a 6.15 × 6.15-m square production pad that was independently irrigated using two impact sprinkler heads placed in diagonal corners and raised to 1.5-m height (Beeson, 2006). Lysimeters were staggered to maximize air movement and sun exposure while remaining well within the confines of a 3.7 × 4.9-m rectangular area in the center of each pad, where the Christensen coefficient of irrigation uniformity was greater than 0.85 (Haman et al., 1996). Lysimeters consisted of a load cell (SSM-100; Interface Force, Inc., Scottsdale, AZ) attached to the top of a 2.5-m tall tripod. A basket, suspended from the load cell, held a plant ≈3 cm above the ground cloth. Load cells were connected to multiplexers (AM 416 and AM32; Campbell Scientific, Inc., Logan, UT) that were connected to and controlled by a data logger (CR10; Campbell Scientific, Inc.). Mass of each basket and plant was recorded every 30 min. The lysimeter/pad setup was replicated three times.
Before 2 Jan. 1997, nine plants were selected for full, uniform canopies and placed in the nine lysimeters. Remaining plants were used to fill a pad area to one of three levels of canopy closure: 33%, 67%, or 100%. Percent canopy closure (%CC) was achieved by placing border plants in each defined pad area (22.9 m2) until the specified amount of %CC was in place. This included the PCA of lysimeter plants. Border plants were placed to achieve a visual uniform coverage of the %CC within a pad area. In addition, a low (10%, termed 0%) %CC was included where only the three lysimeter plants remained on a pad area. Actual evapotranspiration (ETA) of each lysimeter plant was calculated daily.
From 2 Jan. until 3 Feb. each of four %CC levels was established for 3 or 4 d before randomly changing %CC to a different level. After all four levels were achieved, the process was repeated again in random order. Level of %CC was consistent across all three pads each time. Addition or removal of plants to change %CC occurred either during twilight the last day of a cycle or before dawn the day of a new cycle. All plants were irrigated to container capacity using overhead sprinkler irrigation beginning at midnight each day.
On 4 Feb., border plants were removed and leaves from the lower 60% of the height of each lysimeter plant were removed by hand. ETA was then calculated for these plants for 7 d. On the eighth day, lower leaves of an additional 20% of the plant's height were removed and ETA was calculated for an additional 7 d. During this time, any new leaves from latent buds were removed every other day.
Over the 49-day period, reference evapotranspiration (ETo) was calculated daily from an onsite weather station using the Campbell Scientific, Inc. Application Note 4 subprogram. This calculates ETo using the full ASCE Penman-Montieth equation with resistance (Allen et al., 1989). Days with rain were corrected using the procedure outlined by Beeson (2006). ETA data were normalized by dividing each day by the respective ETo.
Phase II.
In Mar. 2001, rooted cuttings of V. odoratissimum were transplanted into 11.4-L black polyethylene containers (C1200; Nursery Supply, Chambersburg, PA) using the substrate composition as stated previously (Florida Potting Soil, Inc., Orlando, FL) and placed in full sun on a black woven polyethylene ground cloth at the University of Florida's Mid-Florida Research and Education Center in Apopka, FL. Plants were fertilized with 50 g of a controlled-release fertilizer (18N–2.6P–9.9K Osmocote; Scotts Company Inc.) at transplanting and again in July and then 40 g of 14N–6P–11.6K Osmocote (14–14–14; Scotts Company Inc.) in November. Irrigation was through overhead impact sprinklers. Plants were pruned every 4 to 6 weeks during the growing season. They were placed container tight initially and then spaced on 0.4-m centers in Nov. 2001 to promote high commercial quality. In Oct. 2001, another set of rooted cuttings of V. odoratissimum were transplanted into 11.4-L containers (C1200; Nursery Supply) and grown as described previously with containers respaced in July 2002.
Plants initially transplanted in Mar. 2001 were transplanted in Apr. 2002 into 26.6-L black polyethylene containers (C2800; Nursery Supply) using the same substrate composition as before. These plants were placed in a different production area and microirrigated twice daily using two 0.42-L·min−1 spray stakes (Roberts Irrigation) until December. Plants were fertilized with 120 g 18N–2.6P–9.9K Osmocote (Scotts Company Inc.) at transplanting and again in August. In late May 2002, rooted cuttings of V. odoratissimum were transplanted into 3.8-L black polyethylene containers (C300; Nursery Supply) using the same substrate composition. These were fertilized with 12 g of 18N–2.6P–9.9K Osmocote (Scotts Company Inc.) and placed container-tight under overhead irrigation. They were fertilized again in October with 8 g 14N–6P–11.6K Osmocote (Scotts Company Inc.) and respaced on 0.25-m centers. Like the larger plants, these were pruned every 4 to 6 weeks during the growing season.
In Jan. 2003, commercially marketable-sized plants of three distinct heights were achieved based on the Florida Grades and Standards (DACS, 1994). Plants in the 3.8-L containers averaged ≈0.55 m tall, whereas those in 11.4- and 26.6-L containers averaged 0.80 m and 1.23 m tall, respectively. Root ball height added an extra 0.12, 0.20, and 0.30 m to the overall height of the plants in 3.8-, 11.4-, and 26.6-L containers, respectively. The same methodology used in 1997 was followed in 2003 with the PCA of each plant determined. Mean PCAs of lysimeter plants were 0.237, 0.436, and 0.970 m2 for plants in 3.8-, 11.4-, and 26.6-L containers, respectively. Nine plants of each container size were selected and placed in the same suspension lysimeters, although basket size was scaled to container size. Three pad areas were occupied for each container size. An identical production area to one used at the Sanford site in 1997 was constructed at the Apopka site in 1999 with the data logger control system relocated to the Apopka site. The CR10 data logger was replaced with a CR10X (Campbell Scientific Inc.).
Beginning 15 Jan. 2003, all pads were at 100%CC. Through 23 Feb. 2003, ETA at each of the four levels of %CC was measured for 4 to 5 d before %CC level was randomly switched and the process repeated. Like in 1997, each pad area was at the same %CC level. From 25 Feb. through 5 Mar. 2003, ETA was measured from plants at 0%CC with the lower 60% (of plant height) of leaves removed. From 7 to 13 Mar. 2003, ETA was measured from plants with only the upper 20% of leaves attached. Reference evapotranspiration (ETo) was calculated daily from an onsite weather station using the Campbell Scientific Inc. App. Note 4 subprogram as described in 1997.
Leaf areas of leaves removed from each lysimeter plant in 2003 were measured with a LI-1500 leaf area meter (LI-COR Biosciences, Lincoln, NE). Leaf area index (LAI) was calculated by dividing each plant's leaf area by its respective PCA.
For data analysis, %CC and the two levels of reduced leaf area were considered treatments. Normalized ETA was analyzed by treatment separately by year and container size. For both years, mean normalized ETA of the 0%CC was divided into normalized ETA of the other three %CC levels with the resulting percentage of isolated plants analyzed as a two-level factorial for combined years. Factors were container size and %CC. Normalized ETA from 2003 was further divided by leaf area to calculate a normalized ETA/m2 with units of mL ETA/mm ETo/m2 leaf area. These values were analyzed as a factorial of container size and treatments. Regression analysis of the relationship between the independent variable LAI and dependent variable normalized ETA/m2 was calculated using the LAIs of each individual plant and the corresponding mean normalized ETA/m2 across the different days for each treatment level. Slopes of regression lines were compared using a single degree of contrast and mean separation, where appropriate, was by Fisher's protected least significant difference (Snedecor and Cochran, 1980). All statistical analysis was by SAS Version 9.2 (SAS Institute, Cary, NC).
Results and Discussion
Viburnum odoratissimum is a fast-growing woody evergreen shrub with leaves generally 12 cm long and 5 cm wide rated to winter-hardiness zone 8 (Halfacre and Shawcroft, 1989). When shoots are pruned to two to four nodes every 4 to 6 weeks during production, a mature canopy shape that is broadly oval with a relative flat top (Halfacre and Shawcroft, 1989) is achieved. Leaf density is greater in the upper portion of the canopy. This species was chosen for its canopy form and because it is dormant from the first freeze (normally early December) to early March each year in central Florida. Thus, leaf area was constant during the experiments. Subfreezing temperatures are infrequent as is rainfall during the winter months. ETo is normally 2.3 to 3.6 mm daily during January to mid-March.
In 1997, there were no differences in normalized ETA (mL·mm−1 ETo) among isolated plants (0%CC) or those at 33%CC or 67%CC (Fig. 1). However, when the canopy was 100% closed, normalized ETA declined by ≈40%. Removing all but the upper 40% (by height) of leaves resulted in equivalent normalized ETA to that of whole plants at 100%CC. Retaining only the upper 20% (by height) of canopy leaves resulted in lower normalized ETA than that of whole plants at 100%CC.
Normalized actual evapotranspiration (ETA) of Viburnum odoratissimum of marketable size in 26.6-L containers at four levels of canopy closure, 1997 data. The 40% and 20% upper canopy levels are normalized ETA of the uppermost 40% and 20% of leaves by height, respectively. Error bars represent se from nine plant replicates.
Citation: HortScience horts 45, 3; 10.21273/HORTSCI.45.3.359
Normalized actual evapotranspiration (ETA) of Viburnum odoratissimum of marketable size in 26.6-L containers at four levels of canopy closure, 1997 data. The 40% and 20% upper canopy levels are normalized ETA of the uppermost 40% and 20% of leaves by height, respectively. Error bars represent se from nine plant replicates.
Citation: HortScience horts 45, 3; 10.21273/HORTSCI.45.3.359
Normalized actual evapotranspiration (ETA) of Viburnum odoratissimum of marketable size in 26.6-L containers at four levels of canopy closure, 1997 data. The 40% and 20% upper canopy levels are normalized ETA of the uppermost 40% and 20% of leaves by height, respectively. Error bars represent se from nine plant replicates.
Citation: HortScience horts 45, 3; 10.21273/HORTSCI.45.3.359
Results in 2003 were identical to those obtained in 1997 for plants in the 26.6-L containers. There was a substantial decrease in normalized ETA between 67%CC or less and 100%CC (Fig. 2) for the 26.6-L containers. Normalized ETA was higher in 2003 than in 1997. Plants were also larger than in 1997. Like in 1997, normalized ETA of the 40% (by height) upper leaves was equivalent to normalized ETA of entire plants at 100%CC, with both larger than that of the upper 20% of leaves. Responses of plants to changes in %CC in the 11.4- and 3.8-L containers were nearly identical to those of the 26.6-L containers in 2003 (Fig. 2). Differences among sizes are those of normalized volumes.
Normalized actual evapotranspiration (ETA) of Viburnum odoratissimum of marketable size in three container sizes at four levels of canopy closure, 2003 data. The 40% and 20% upper canopy levels were the normalized ETA of the uppermost 40% and 20% of leaves by height, respectively. Each bar is the mean of nine plants. Bars with different letters within a container size are significantly different (α = 0.05) based on Fisher's protected least significant difference.
Citation: HortScience horts 45, 3; 10.21273/HORTSCI.45.3.359
Normalized actual evapotranspiration (ETA) of Viburnum odoratissimum of marketable size in three container sizes at four levels of canopy closure, 2003 data. The 40% and 20% upper canopy levels were the normalized ETA of the uppermost 40% and 20% of leaves by height, respectively. Each bar is the mean of nine plants. Bars with different letters within a container size are significantly different (α = 0.05) based on Fisher's protected least significant difference.
Citation: HortScience horts 45, 3; 10.21273/HORTSCI.45.3.359
Normalized actual evapotranspiration (ETA) of Viburnum odoratissimum of marketable size in three container sizes at four levels of canopy closure, 2003 data. The 40% and 20% upper canopy levels were the normalized ETA of the uppermost 40% and 20% of leaves by height, respectively. Each bar is the mean of nine plants. Bars with different letters within a container size are significantly different (α = 0.05) based on Fisher's protected least significant difference.
Citation: HortScience horts 45, 3; 10.21273/HORTSCI.45.3.359
To test for an effect of plant height on the response to %CC, normalized ETA for each container size was divided by the mean normalized ETA of plants at 0%CC to calculate a percentage of normalized ETA of isolated plants for each level of canopy closure. An interaction between container sizes and difference levels of %CC would indicate a three-dimensional response of ETA to %CC. It was hypothesized that taller plants would exhibit greater reductions in ETA with canopy closure because a greater depth of canopy would be enclosed. Such differences among container sizes did not occur (Fig. 3), although canopy height varied 2.5-fold between the smallest and largest sized plants. Thus, the effect of canopy closure was independent of plant size and canopy height for these plants. In all cases, the percentage of normalized ETA of isolated plants, for those at 100%CC, was lower than those at the other %CC levels. There were few other differences within a container size. Percentages of normalized ETA of isolated plants with increasing canopy closure for plants in 1997 were the same as those in 2003 with the exception that in 2003, the value at 0%CC was higher than those of the 33% and 67%CC. Thus, effects on normalized ETA by changes in %CC are independent of plant height up to at least 1.5 m.
Percent normalized actual evapotranspiration (ETA) of isolated plants (0%CC) at increasing levels of canopy closure for three sizes of plants designated by container size (L). Each bar is the mean of nine plant replicates. The 26.6 -‘97 data were collected in 1997. All other data were collected in 2003. Bars with different letters within a container size are significantly different (α = 0.05) based on Fisher's protected least significant difference. CC = canopy closure.
Citation: HortScience horts 45, 3; 10.21273/HORTSCI.45.3.359
Percent normalized actual evapotranspiration (ETA) of isolated plants (0%CC) at increasing levels of canopy closure for three sizes of plants designated by container size (L). Each bar is the mean of nine plant replicates. The 26.6 -‘97 data were collected in 1997. All other data were collected in 2003. Bars with different letters within a container size are significantly different (α = 0.05) based on Fisher's protected least significant difference. CC = canopy closure.
Citation: HortScience horts 45, 3; 10.21273/HORTSCI.45.3.359
Percent normalized actual evapotranspiration (ETA) of isolated plants (0%CC) at increasing levels of canopy closure for three sizes of plants designated by container size (L). Each bar is the mean of nine plant replicates. The 26.6 -‘97 data were collected in 1997. All other data were collected in 2003. Bars with different letters within a container size are significantly different (α = 0.05) based on Fisher's protected least significant difference. CC = canopy closure.
Citation: HortScience horts 45, 3; 10.21273/HORTSCI.45.3.359
When compared with isolated plants, plants at 100%CC transpired ≈40% less water than those up to 67%CC independent of plant size (Fig. 3). With decreases in %CC, there would have been increased air movement and radiation to greater depths in the canopies, especially for the largest plants. These would have driven greater flux densities of water vapor from lower depths in the canopy and account for the higher normalized ETA at less than 100%CC. However, it is intriguing that normalized ETA achieved near maximum values when %CC declined only by 33% and that this was consistent independent of the 2.5-fold differences in canopy depth. The path length for eddy diffusion of water vapor would have been much shorter for plants in the 3.8-L containers than those in the 26.6-L ones, yet relative ETA was the same. Being closer to the ground, wind speeds would have been less (Nobel, 1999), perhaps reducing the speed of eddy diffusion. In lieu of other explanations, the results suggest that 67%CC is a sufficiently porous canopy to invoke comparable transpiration of subtending leaves to that of isolated plants of at least 1 m in depth. Evaporation from container substrates would have contributed some to these higher water vapor flux densities. However, evaporation contributions would have been proportionally very small based on reported values from plantless 11.4-L containers under slightly warmer spring conditions (Beeson, 2004; Table 1).
Mean normalized actual evapotranspiration (ETA)/m2 leaf area of Viburnum odoratissimum and associated leaf area index (LAI) for commercially marketable plants in different container sizes at different levels of canopy closure and reduced leaf areas.z
Normalized ETA from 2003 was divided by the respective leaf areas of each plant to control for differences in plant size (normalized ETA/m2). Within a container size, normalized ETA/m2 did not vary between isolated plants up to 67%CC (Table 1). In all sizes, normalized ETA/m2 at 100%CC was lower and 64% to 66% of the average normalized ETA/m2 at the lower %CCs. Accounting for differences in plant leaf areas, there was a linear decline in normalized ETA/m2 with increasing plant LAI at all %CC as plant size increased (Table 1; Fig. 4). Rate of decline with increasing LAI was similar at 67%CC or less (mean = −60.3) as were the intercepts. Rate of decline with increasing LAI slowed at 100%CC and was less than that at lower %CC. This decline is most likely the result of increased shading of lower and/or interior leaves with increases in plant size and %CC and reduced eddy diffusion and reductions in VPD at canopy closure.
Relationship between leaf area index and normalized actual evapotranspiration (ETA)/m2 for three plant sizes and four levels of canopy closure. Plant sizes were based on container sizes of 3.8, 11.4, and 26.6 L, from left to right, respectively. Each point is the mean of nine plant replicates.
Citation: HortScience horts 45, 3; 10.21273/HORTSCI.45.3.359
Relationship between leaf area index and normalized actual evapotranspiration (ETA)/m2 for three plant sizes and four levels of canopy closure. Plant sizes were based on container sizes of 3.8, 11.4, and 26.6 L, from left to right, respectively. Each point is the mean of nine plant replicates.
Citation: HortScience horts 45, 3; 10.21273/HORTSCI.45.3.359
Relationship between leaf area index and normalized actual evapotranspiration (ETA)/m2 for three plant sizes and four levels of canopy closure. Plant sizes were based on container sizes of 3.8, 11.4, and 26.6 L, from left to right, respectively. Each point is the mean of nine plant replicates.
Citation: HortScience horts 45, 3; 10.21273/HORTSCI.45.3.359
The second objective of this study was to begin defining the depth of the upper layer that constituted vertical transpiration. Vertical transpiration is greater than the horizontal component. The shallower the depth of vertical transpiration, the greater the response to canopy closure. The initial removal of the lower 60% of canopy height was targeted for a 50% removal of total leaf area, accounting for reduced leaf density at lower levels. The lower 60% of canopy height contained 52% to 57% of total leaf area in 2003. LAIs for the upper 40% were correspondingly approximately half that of whole plants (Table 1). The upper 20% by height contained 14% to 21% of total leaf area for plants in 2003. For the partitioned upper canopy, normalized ETA/m2 was higher than that of any of the whole plants’ %CC levels (Table 1). Previous research found that defoliation or bagging of branchlets can increase transpiration or stomatal conductance (gs) of remaining exposed leaves. However, the response appears related to the percentage of total leaf area prevented from transpiring. For individual structural branches of large Pseudotsuga menziesii (Brooks et al., 2003) and branches of Pinot noir grapevines (Candolfi-Vasconcelos et al., 1994), reducing transpiration to near zero for large portions (greater than 50%) of existing leaves had little to no effect on gs or transpiration rate of remaining fully exposed leaves on the select branches. When 30% of total leaf area was enclosed on second-year grafted peach, there was no increase in gs or transpiration, yet enclosing 60% of leaves increased both by 30% for remaining exposed leaves (Solari et al., 2006). Similarly, for Populus tremuloides Michx. (greater than 3 m tall), defoliating 50% of an entire tree's canopy induced a 25% increase in gs (Hart et al., 2000). No further increase in gs occurred even if defoliation was extended to 98% (Hart et al., 2000). With the 52% to 57% defoliation here, whole plant ETA increased 31% (11.4 and 26.6 L) to 46% (3.8 L), in agreement with previous research. A complicating factor was that removal of lower canopy leaves increased exposure of substrate upper surfaces to conditions that would have increased evaporation and therefore overall ETA. This evaporation was estimated for each container size based on data from11.4-L containers without plants also measured in the spring (mean of 560 mL·mm−1 ETo/m2 substrate area; Beeson, 2004). Subtracting these estimated evaporations from ETA of these upper canopy-only plants did not lower the normalized ETA/m2 to that of entire plants at 67%CC or less (Table 1). Increases in ETA associated with defoliation were reduced to 32%, 11%, and 21% increases for plants in 3.8-, 11.4-, and 26.6-L containers, respectively. It is unclear whether these increases were defoliation-induced or simply from increased exposure of remaining leaves. Unlike previous research, defoliation was not random. Partitioned canopies likely had decreased boundary layers and higher VPD around leaves as a result of increased air movement both above and below the remaining canopies. Plant daily averaged transpiration increased slightly with the 60% height defoliation, increasing from 1.37, 1.15 and 0.82 mmol·m−2·s−1 for isolated plants to 1.91, 1.30, and 1.09 mmol·m−2·s−1 for defoliated plants in 3.8-, 11.4-, and 26.6-L containers, respectively. Effect of greater air movement would have been at the extreme at the upper 20% level, where normalized ETA/m2 was 92% to 201% higher than isolated plants. Based on the P tremuloides results, such increases in transpiration are likely not attributable to defoliation-induced increases alone. Although normalized ETA/m2 values for the 20% upper canopy appear high (Table 1), mean daily averaged transpiration for the 20% upper canopy was 5.1, 3.0, and 2.7 mmol·m−2·s−1 for the plants in 3.8-, 11.4-, and 26.6-L containers, respectively. The 5.1 mmol·m−2·s−1 was similar to midday values measured from exposed lower canopy leaves of Ulmus parvifolia in late September (Beeson, unpublished data). Rates reported here are most likely higher than what occurred in the upper 40% of the whole canopies at 100%CC. LAIs of whole plants in the 3.8-L containers were similar to the LAIs of the upper 40% canopy of plants in the 26.6-L containers. Yet when normalized ETA/m2 was divided by the respective LAI, values were higher (P < 0.014) for the smaller 3.8-L container-grown isolated whole plants (220 mL·mm−1 ETo/m2 leaf area/unit LAI) than larger and more elevated upper 40% canopies of the partitioned 26.6-L container-grown plants (184 mL·mm−1 ETo/m2 leaf area/unit LAI). Lower transpiration suggests greater resistance to water vapor diffusion in the larger canopies of the plants in 26.6-L containers. For whole plants at 100%CC, upper leaves would still receive most of the radiation energy. However, boundary layers at the upper canopy would be much thicker as a result of the depth of canopy below and the larger, more uniform and smoother surface provided by surrounding canopies, which would reduce wind speed at the leaf surface. Thus, the decline in normalized ETA/m2 with increasing plant size is consistent with the factors driving water vapor flux densities. It is unlikely that transpiration is mainly from the upper 40% by canopy height in closed canopy situations for larger plants or trees. However, it was uniform across plant sizes in this study that are representative of shrub production.
Changes in tree transpiration with changes in canopy closure have been reported in the forestry literature, usually associated with rapid changes occurring after forests are thinned (Medhurst et al., 2002). Bréda et al. (1995) reported increased transpiration of remaining trees after thinning substantially reduced canopy closure. Increased transpiration occurred without changes in LAI, similar to the results reported here. Increased transpiration was proposed as a result of greater coupling of a tree's entire canopy to atmosphere rather then just the upper portion. The same response was shown here at %CC less than 100%. Earlier Morikawa et al. (1986) had correlated transpiration of forest trees with solar radiation and VPD. In a thinned forest, there is greater light penetration to the lower canopy and VPD would fluctuate more. Both of these factors would decrease the boundary layer resistance to water diffusing from leaves (Nobel, 1999). Teklehaimanot et al. (1991) reported increased boundary layer conductance with increasing tree spacing attributable principally to increased wind speeds in lower-density plantings.
In conclusion, for commercial quality evergreen shrubs, ETA declines rapidly, 40% for V. odoratissimum, when canopies approach complete closure somewhere above 67%CC. At this point, previous irrigation rates would be excessive and contribute to leaching of nutrients. The relationship between ETA and %CC was independent of plant height up to at least 1.5 m above ground level. These results suggest that for small isolated trees or shrubs, the upper 40% of a canopy transpires 60% of daily ETA. From a different production perspective, once sufficient numbers of plants have been randomly removed from a block of closed canopy plants or plants have been spaced to produce greater canopy width or fullness such that %CC approaches 67% or less, daily ETA can increase up to 60%. When transplanted into most landscapes, shrubs essentially become isolated plants. Thus, transpiration rates may increase up to 60% above that experienced in production beds. Previously, Spoomer (1974) showed that textural differences between container substrates and mineral soil led to excessive water loss from root balls, contributing to rapid development of water stress of recently planted container-grown shrubs. This research expands the explanation of why container-grown woody ornamentals transplanted into landscapes require more irrigation than during production.
Literature Cited
Allen, R.G. , Jensen, M.E. , Wright, J.L. & Burman, R.D. 1989 Operational estimates of reference evapotranspiration Agron. J. 81 650 662
Beeson R.C. Jr 2004 Modeling actual evapotranspiration of Ligustrum japonicum from rooted cuttings to commercially marketable plants in 12 liter black polyethylene containers Acta Hort. 664 71 77
Beeson R.C. Jr 2006 Relationship of plant growth and actual evapotranspiration to irrigation frequency based on management allowed deficits for container nursery stock J. Amer. Soc. Hort. Sci. 131 140 148
Bréda, N. , Granier, A. & Aussenac, G. 1995 Effects of thinning on soil and tree water relations, transpiration and growth in an oak forest [Quercus petraea (Matt.) Liebl.] Tree Physiol. 15 295 306
Brooks, J.R. , Schulte, P.J. , Bond, B.J. , Coulombe, R. , Domec, J.C. , Hinckley, T.M. , McDowell, N. & Phillips, N. 2003 Does foliage on the same branch compete for the same water? Experiments on Douglas fir trees Trees (Berl.) 17 101 108
Candolfi-Vasconcelos, M.C. , Koblet, W. , Howell, G.S. & Zweifel, W. 1994 Influence of defoliation, rootstock, training system and leaf position on gas exchange of Pinot noir grapevines Am. J. Vittic. 45 173 180
DACS (Dept. of Agriculture and Consumer Services) 1994 Grades and standards for nursery plants Fla. Dept of Agr. and Consumer Services Tallahassee, FL
Halfacre, R.G. & Shawcroft, A.R. 1989 Landscape plants of the southeast Sparks Press, Inc Raleigh, NC
Haman, D.Z. , Smaljstra, A. & Pitt, D. 1996 Uniformity of sprinkler and microirrigation systems for nurseries Fla. Coop. Extension Service Bulletin IFAS, Univ. of Florida
Hart, M. , Hogg, E.H. & Lieffers, V.J. 2000 Enhanced water relations of residual foliage following defoliation in Populus tremuloides Can. J. Bot. 78 583 590
Medhurst, J.L. , Battaglia, M. & Beadle, C.L. 2002 Measured and predicted changes in tree and stand water use following high-intensity thinning of an 8-year-old Eucalyptus nitens plantation Tree Physiol. 22 775 784
Morikawa, Y. , Hattori, S. & Kiyono, Y. 1986 Transpiration of a 31-year-old Chamaecyparis obtusa Endl. stand before and after thinning Tree Physiol. 2 105 114
Nobel, P.S. 1999 Physicochemical and environmental plant physiology 2nd Ed Academic Press San Diego, CA
Rose, C.W. 1984 Modeling evaporation: An approach to heterogeneous communities Agr. Water Manage. 8 203 211
Snedecor, G.W. & Cochran, W.G. 1980 Statistical methods 7th Ed The Iowa State Univ. Press Ames, IA
Solari, L.I. , Johnson, S. & Degong, T.M. 2006 Relationship of water status to vegetative growth and leaf gas exchange of peach (Prunus persica) trees on different rootstocks Tree Physiol. 26 1333 1341
Spoomer, L.A. 1974 Two classroom exercises demonstrating the pattern of container soil water distribution HortScience 9 152 153
Teklehaimanot, Z. , Jarvis, P.G. & Ledger, D.C. 1991 Rainfall interception and boundary layer conductance in relation to tree spacing J. Hydrol. (Amst.) 123 261 278
