Return to METRIA 12

Water Use of Acer rubrum cultivars: Quantifying Canopy-Level Water Flux Using Measurement and Modeling Techniques

William L. Bauerlea and Christopher J. Postb

aDepartment of Horticulture, Clemson University, Clemson, SC 29634-0375, USA
bDepartment of Forest Resources, Clemson University, Clemson, SC 29634, USA

aCorresponding author:
William L. Bauerle
Fax: (864) 656-4960
Phone: (864) 656-7433

A paper from the Proceedings of the 12th Metropolitan Tree Improvement Alliance Conference (METRIA 12), Landscape Plant Symposium: Plant Development And Utilization, held in Asheville, NC, May 23-25, 2002, co-sponsored by the North Carolina State University, North Carolina Division of Forest Resources, USDA Forest Service Southern Region, North Carolina Landscape Association, North Carolina Association of Nurserymen, The Landscape Plant Development Center, The North American Branch of The Maple Society, and The International Ornamental Crabapple Society.


Increasing concern over water conservation in urban landscapes provided the impetus to assess whole-tree water use of red maple, one of the most popular trees in the urban landscape. The objective of this study was to quantify canopy-level water flux of individual red maple trees using both stem-flow gauges and modeling techniques. Surprisingly, little information is available on woody ornamental water use. The results of this study suggest that estimates of daily transpiration simulated by the model were in agreement with sap flow measurements when soil water was not limiting. Moreover, preliminary evidence suggests that individual tree water use can be estimated in landscape conditions.


Simulation models are important tools for analysis and interpretation across scales of organization. Moreover, mathematical models afford us the ability to synthesize and integrate the complex and non-linear forcing of environmental variables on biological systems (Baldocchi and Harley, 1995). To date, the Penman-Monteith evapotranspiration model, a common combination equation for estimating crop water use, is the primary basis for both single-layer and multilayer models of canopy evaporation. Single layer models, commonly called "big-leaf", treat the canopy as a plane in which a physiological and aerodynamic resistance is ascribed to water vapor transfer. An alternative approach, "multilayer modeling", considers the canopy microclimate explicitly. Multilayer models divide the canopy into a finite number of horizontal layers in an attempt to describe the evaporation from individual parts of the canopy. Regardless of the type of model, accurate quantification of canopy level water use is of utmost importance, especially in canopies as complex as trees. Overall, the chosen model should be based on sound physical principals that integrate physiological and micrometeorological processes.

Water requirements of urban landscapes are becoming an increasing cause for concern as demand for water in metropolitan areas rises. Big-leaf models treat the canopy as a partly wet single plane. Unfortunately, the detailed and complex spatial structure of the actual canopy microclimate in metropolitan areas violates the assumption that the canopy was no more than a partly wet plane at the lower boundary of the atmosphere, thus eliminating the possibility of precisely estimating water use with a big-leaf approach. Application of complex sub-models that integrate environment and physiological processes, although difficult to apply, may be the only way to accurately estimate plant water use in urban settings.

The key objective of this paper is to illustrate the ability to quantify water use of individual trees by scaling from individual canopy layers, the sum of which equate to whole canopy water loss, and to test estimates against field data. A less common modeling approach, multilayer or vertically structured, is used to divide the canopy into individual layers. Estimates of water loss at each layer are scaled to the crown in an attempt to address the level of detail that should be incorporated into an individual crown water use model. The model provides predictions of water use that are significantly better than big-leaf estimates, opening up the possibility of quantifying urban landscape water use on a plant-by-plant basis.

Materials and Methods

South Carolina grown Acer rubrum (red maple) cultivars were shipped to Clemson University, transferred to an outdoor gravel pad, and fit with pressure-compensating micro emitters (ML Irrigation Inc., Laurens, SC). Prior to arrival, cultivars were transplanted into 15 gal spin-out treated plastic pots containing a mixture of pine bark and sand (20:1, v/v), fertilized with 8.3 kg m-3 of NutricoatTM 20N-3.0P-8.3K type 360 (Chiso-Asahi Inc., Japan), and grown on an outdoor gravel pad. Upon arrival, all pots were watered to saturation and permitted to drain for 18 hrs. Containerized red maple saplings were spaced 1.5 m center-to-center and irrigated daily to container capacity. Randomly selected plants of two cultivars, October Glory® and Summer Red®, were chosen for continuous sampling of sap flow. To eliminate evaporation from the soil surface and/or water penetration in case of rain, white plastic bags were cut and sealed to the stem with Parafilm (American National CanTM, Greenwich, CT). The bottom ends of the bags were left open and secured to the pots with an elastic fit. Wrapping the exterior of each container with aluminum foil reduced radiation load on containers. Commercially available sap flow gauges (Dynamax, Inc., Houston TX) were used for all measurements. The gauges have been described in detail elsewhere (Steinberg et al., 1990). Whole tree sap flow rates and microclimatic conditions were measured simultaneously. Meteorological data were collected at a height of 3 m using a Campbell Scientific Weather Station located on the north side immediately adjacent to the experimental plot and within 0.25 m of canopy level.

A three dimensional model was parameterized for an experimental containerized Acer rubrum site near Clemson, SC USA (lat. 34° 40' 8"; long. 82° 50' 40"). The canopy, treated as a paraboloidal crown, was divided into discrete layers in order to describe the partitioning of evapotranspiration between various parts of the crown. Using the physiological and physical parameters, the transpiration (E) values were computed and compared with the observed values. Predicted transpiration values were also compared to potential evapotranspiration of the Penman-Monteith combination equation. These values were compared with modeled estimates and observed values of transpiration from sap flow measurements on a fifteen-minute basis


Estimated light absorption and transpiration are illustrated by Figure 1 at seven different layers throughout the canopy of two replicates each of both October Glory® (OG) and Summer Red® (SR) cultivars of red maple. Cumulative water use of all layers was in agreement with sap flow estimates of whole tree transpiration at fifteen-minute intervals. Light absorption diminished at increasing depth within the canopy due to the light extinction coefficient of the model. Table 1 shows the difference in estimates between the single and multilayer approach as well as field collected sap flow measurements for the four trees under study. The single layer model, the Penman-Monteith combination equation, tended to drastically overestimate water use. The multilayer model on the other hand, slightly overestimated water use in three of the four cases. However, values of canopy scale water use collected by sap flow measurements were close to those estimated by the multilayer approach. Figure 1 also illustrates the stratification of leaf area through the seven layers of canopy. A linear increase in leaf area with increasing canopy was a result of the parabolic crown shape.

Table 1. Photosynthetic photon flux density at the top of the canopy (PPFDt, µmol m-2 s-1), average estimated photosynthetic photon flux density absorption (PPFD a, µmol m-2 s-1), multilayer modeled transpiration (Ee, g m-2), measured transpiration (Em, g m-2), and single-layer modeled transpiration (Es, g m-2) of irrigated cultivars of red maple. Results are given as a fifteen minute average at solar noon.

Cultivar PPFDt PPFDa Ee Em Es
SR1 1535 1022 42.19 38.76 150.58
OG1 1535 1036 31.41 29.8 150.58
SR2 1535 1143 39.28 43.32 150.58
OG2 1535 1135 36.32 31.59 150.58


Canopy Layer

Figure 1. The model estimates of photosynthetic photon flux density (PPFD, µmol m-2 s-1), estimated transpirational water loss (mmol m-2 s-1), and leaf area distribution (m-2) at multiple layers through out the canopy of irrigated cultivars of red maple.


While information on the latent heat dissipated by agronomic crops is available in the literature, there is much less information on the transpiration of ornamental trees. The complex nature of tree canopies does not lend itself to an easy modeling approach, such as those present in uniform agronomic crops. Most of the models for predicting transpiration are based on the thermal balance equation of the canopy, where the main difference occurs in the calculation of the solar absorption coefficient and stomatal resistance. The models generally use the Penman-Monteith equation (Monteith, 1965), which considers the vegetation as a single big-leaf. The one-layer model uses meteorological data to calculate the latent heat flux from knowledge of the aerodynamic and surface resistance. Wullschelger et al. (2000) found a clear relationship between canopy transpiration and both radiation and atmospheric humidity deficit in large red maple trees. The relationship suggests that aerodynamically well-coupled canopies, present in red maple and most tree species, neccesitates a multilayer approach in their attempt to estimate transpiration.
Modeling landscape transpiration in metropolitan areas is a monumental task. Inhomogeneous vegetation, heat islands, and various other urban uncertainties complicate the issue. Having said this, we feel progress in this area is imminent given the recent water deficits in the Eastern USA. It appears that current models can be improved upon when fluxes and profiles within a canopy are incorporated into calculations. In the case of individual plant estimates, single-layer models do not appear to function as well as multiplayer models, especially when it is necessary to resolve detail within the canopy of a single tree. The multilayer approach, conversely, is applicable when the detail and complex spatial structure of the canopy microclimate is relevant.

Literature Cited

Return to METRIA 12

Web Crafter: Anne S. Napier ~ Email:

Format updated October 23, 2009