Site description and plant materials

The experimental site was comprised of fourteen 10-m x 10-m plots. During May 1999, each plot was planted with trees, shrubs, and ground covers based on the number and life forms of vegetation in xeric style residential landscape plots surveyed in the Phoenix metropolitan area. A 2.5-m wide strip of unvegetated soil was maintained between each plot. Each 100-m2 plot was divided into equal quadrants and planted with one each of 19-L container grown Eucalyptus microtheca F. Muell. Blue Ghost (coolibah) and Quercus virginiana var. fusiformis (Small) Sarg. (Texas live oak) in opposite corner quadrants. Six each of 3.8-L containergrown clones of Nerium oleander L. Sister Agnes (oleander) and Leucophyllum frutescens var. green cloud Berl. (Texas sage) spaced approximately 1.5 m on center and grouped in opposite diagonal quadrants of the plot were installed in the remaining two quadrants. Rosmarinus officinalis L Prostratus (prostrate rosemary) was used as a ground cover and one each of 3.8L clones were planted in each plot near the center in the quadrant occupied by L. frutescens . To facilitate establishment, plants in all plots were drip irrigated at a similar rate (416 L m-2 yr-1) for the first six months after transplanting. To reduce effects of possible soil nutrient heterogeneity, all plants were fertilized initially and subsequently each April with controlled release fertilizer 20N-4.4P-4K plus micronutrients (Best Paks, J.R. Simplot Co, Boise, Idaho).

Experimental design

Beginning January 2000, plots were subjected to a two by four factorial treatment combination of drip irrigation rate and shrub pruning regime for 3 years. Irrigation treatments were defined as low rate (mean 814 L m-2 yr-1) and high rate (mean 1954 L m-2 yr-1). Pruning treatments included hedging Nerium and Leucophyllum every 6 weeks or 6 months, yearly renewal pruning those taxa to the ground, or leaving them unpruned. Irrigation treatments were blocked in seven rows of two plots, one of each irrigation rate treatment, running in a north to south direction, with the four shrub pruning treatments randomly assigned to entire plots within that scheme. Water was delivered by an automatic irrigation system to each plant through drip emitters situated under the canopy near the base of each shrub or ground cover and under the canopy near the trunk of each tree. During the first year, irrigation to tree species was supplied by two 3.8 L hr-1 emitters each and to shrubs and ground covers by one 3.8 L hr-1 emitter each. During the second year, the number of emitters supplying Eucalyptus was increased to four, and during the third year to six. Also during the third year, the number of emitters to Quercus was increased to three and to both shrub species was increased to two. The volume of water applied to each plot was monitored using totalizing water meters (Precision Manufacturing Company, Orlando, Fla.). The volume of irrigation to each plant was calculated as the total volume of water delivered to a plot over time divided by the fraction of total emitters on the plot for each plant. The irrigation rate was calculated as volume per m2 of plot area per year. Irrigations were scheduled to give plants in the high irrigation rate plots twice as much water as those in the low rate irrigation rate plots and both were ramped seasonally. The irrigation rate treatments were based on a previous assessment of landscape irrigation application rates by local residents. Precipitation was measured via tipping bucket (Model TE525WS, Campbell Scientific, Ogden, Utah) and soil moisture content was monitored at a depth of approximately 30 cm in unirrigated soil and for three replicates of each irrigation rate treatment using soil moisture probes (ThetaProbe, Dynamax, Inc, Houston, Texas). Moisture probes were situated below the canopy near the base of each of the two pruned shrubs species and approximately 1 m from the based of E. microtheca . Precipitation and soil moisture data were recorded with a 23X micrologger (Campbell Scientific). Mean annual soil moisture content (v/v) was 6.1, 12.9, and 16.7percent for dry soil, low irrigation rate, and high irrigation rates, respectively. For three years pruning treatments were applied to N. oleander and L. frutescens in two plots for each irrigation treatment (4 plots x 3 pruning treatments). Shrubs were either hedged every six weeks, hedged every six months during February and August or renewal pruned to the ground yearly during February (n equals 2 plots and 12 pseudo replicates of each shrub for each pruning treatment). In addition, shrubs in one plot within each irrigation treatment were left unpruned as controls (n equals 1 plot and 6 pseudo replicates of each shrub species for each irrigation x pruning treatment combination). Tree canopies were thinned and the base elevated yearly and similarly and Rosmarinus was left unpruned on all plots.

PP, LA and WUEI

The size of all plants was measured every three months. For trees, height, diameter, canopy area and canopy base height were measured with a clinometer and steel tapes. Tree canopy area was calculated as A equals pi(0.5 * ((NS + EW)/2))2 where A is canopy area, NS is the measured north/south dimension of canopy, and EW is the measured east/west dimension of canopy, assuming a tree canopy to be circular with a radius of half the average diameter. Shrubs were also measured before and after each pruning event. For shrubs and ground covers, canopy volume (height x width EW x width NS) was measured. All pruned biomass and accumulated litter (green waste, GW) from each plant was collected at pruning events, oven dried for 7 days at 70 degree C and weighed. Leaf surface area to dry weight (DW) relationships were established using regression analysis by measuring surface area of 100 leaves of each species using a digitizing camera system interfaced to a computer and drying and recording mass after leaves were dried for 3 days at 60 degrees C (Appendix A). At the end of three years, all shrubs and ground covers were pruned to the ground and fresh weight (FW) of all pruned material measured. All plants with final FW less than 2 kg were dried and weighed, as were all Rosmarinus and 3 replicates of Nerium and Leucophyllum from each plot. The remaining Nerium and Leucophyllum were sub-sampled and total DW was calculated based on FW/DW ratios of each plant. Leaf material was separated from woody material of harvested plants and leaf DW was used to estimate final canopy leaf area (LA). For Eucalyptus, standing biomass was estimated by the equation: loge DW equals -3.8604 + 0.9644 loge (DBH2 x height) and LA was estimated via canopy area, assuming a leaf area index (LAI) of 2.19. Standing biomass of Quercus was estimated using the equation: DW equals -0.086 + 0.05 (d) where d is the trunk diameter at 15 cm and LA was calculated via canopy area assuming an LAI of 0.66. Plot PP was calculated as the sum of estimated tree standing biomass and total biomass removed via pruning and destructive harvest. Irrigation water use efficiency (WUEI) was calculated for each plant as PP per total volume of irrigation water applied for three years. Precipitation was assumed to be equal to all plants and had negligible affect on available soil moisture relative to irrigation so was not considered in WUEI calculations. For trees, LA over time was estimated via measured canopy area and LAI. Regressions of final LA on standing shrub canopy volume were used to estimate shrub LA over time. Average LA for each plant was calculated from all measurements made during the 3 year study, and total plot average LA was the sum of the average LA for all plants on a plot. Average LAI was calculated for the shrubs as estimated LA/canopy area over three years.

Plant gas exchange

On two consecutive days every 3 months from October 2000 until October 2002, maximum plant gas exchange fluxes were measured for the most recently physiologically mature plant tissues and consisted of partial leaves (Nerium and Eucalyptus), whole leaves (Quercus), or small shoot segments (Leucophyllum and Rosmarinus) of all plants. Measurements were made on cloudless days during a 3 hour time frame previously determined to be that of optimal conditions for maximum gas exchange fluxes for that season. Measurements were made the day after a watering event for each irrigation treatment (two sampling days) to reflect fluxes under conditions of maximum water availability. Measurements were made with a quarter-liter chamber connected to a portable infrared gas analyzer (Model LI 6250, LI-COR Inc., Lincoln, NE, USA) operated in closed system mode. The chamber thermocouple was positioned in direct contact with the abaxial surface of measured leaves and stems to measure shoot temperature. To reduce the potential errors associated with measuring plant gas exchange in an enclosed chamber, particularly with respect to leaf conductance (gs) and transpiration (E) as a measure of actual leaf gs and E under field conditions, a sampling protocol for gas exchange measurements that consisted of short measurement intervals (10-30 seconds) to minimize changes in the chamber microenvironment during the sampling interval, followed by exhaustion of the chamber volume to return to ambient conditions, was employed. Typical chamber microenvironmental conditions were as follows: initial CO2 concentration ranged from 355 to 409 ppm and drawn down was less than 1 ppm per second; flow rate was adjusted to maintain chamber relative humidity within 10% of ambient and for a given measurement leaf to air temperature differential was less than 0.5C. Instrument derived values for E compared with values calculated from measured gs and vapor pressure deficit (VPD) via the formula: E equals gs * VPD/p, where p is local barometric pressure were similar via paired t-test, so instrument derived values for E were used for subsequent scaling activities. For each of the 9 seasonal 2-day sampling periods, 10 leaf or shoot replicate measures for each irrigation x pruning interaction (n equals 80 each species) of Nerium and Leucophyllum and 20 replicate measures for each irrigation treatment (n equals 40 each species) for other species were made. Total number of gas exchange measurements made for each 2-day seasonal sampling period was 280 except the first two measurement periods, during which control plants and those pruned yearly were indistinguishable. For large leaves, a known surface area was enclosed in the chamber and smaller leaves or shoot segments were harvested and total one sided surface area was measured with a digitizing camera system interfaced to a computer to calculate fluxes per unit leaf area.

Scaling and integration of A and E

Total plant carbon acquisition potential (CAP) and transpiration potential (TP) were estimated as the maximum integral of A or E using trapezoidal numerical integration. Seasonal A and E data were used in conjunction with the model developed by Martin and Stabler to scale the amplitude of diurnal patterns of gas exchange fluxes to calculate daily CAP and TP.