Improved estimates of tree-shade effects on residential energy use

Abstract

Tree-shade alters building cooling and heating loads by reducing incident solar radiation. Estimates of the magnitude of this effect, and how it is influenced by urban forest structure (e.g. tree size and location), are difficult due to the complexity inherent in tree–sun–building interactions. The objective of this paper is to present a simplified method for making these estimates appropriate for neighborhood and larger scales. The method uses tabulated energy use changes for a range of tree types (e.g. size, shape) and locations around buildings (lookup tables), combined with frequency of occurrence of trees at those locations. The results are average change in energy use for each tree type that are not explicitly dependent on tree location. The method was tested by comparison to detailed simulations of 178 residences and their associated trees in Sacramento, California. Energy use changes calculated using lookup tables matched those from detailed simulations within ±10%. The method lends itself to practical evaluation of these shading effects at neighborhood or larger scales, which is important for regional assessments of tree effects on energy use, and for development of tree selection and siting recommendations for proposed energy conserving planting programs.

Introduction

The rapidly increasing pace of world-wide urbanization hastens the need for improved understanding of environmentally beneficial urban forestry practices. Urbanization, especially if haphazard, can result in increased urban temperature, energy use, carbon dioxide emissions from fossil fuel power plants, municipal water demand, ozone levels, and human discomfort and disease [5]. These problems are accentuated by global climate change, which may double the rate of urban warming [2]. The extent to which urban forestry can mitigate these effects depends in large part on development of better tools to quantify the cost-effectiveness of alternate strategies and demonstrate their potential benefits [6], [16], [22].

Urban forests modify climate and building energy use through (1) shading, which reduces the amount of radiant energy absorbed and stored by built surfaces; (2) evapotranspiration, which converts liquid water in plants to vapor, thereby cooling the air; and (3) wind speed reduction, which reduces infiltration of outside air, effectiveness of ventilation, and convective cooling of building surfaces [25]. Alterations in long-wave radiation between a building and its surroundings from addition of trees have small effects compared to those from shade [8]. The focus of this paper is an improved method for estimating amount and timing of building shading from trees and its effects on cooling and heating energy use. Simple characterization of shading effects is complicated by the many possible permutations of building type, building surface orientation, tree location with respect to each building surface, tree size, canopy density, solar angle (time of day), season and microclimate. Given the difficulties associated with a measurement program that includes all of these factors, simulation models have been a necessary and practical alternative for evaluating these effects.

Simulation models that account for tree configuration (species, age and location), building characteristics (e.g. window area, building orientation, level of insulation), and weather conditions can be used to estimate effects of tree-shade on heating and cooling energy use [7], [8], [15], [27]. Their use on a large scale has been limited due to their complexity and data requirements. Buildings representative of a range of construction practices have been used to account for differences in building energy use characteristics; housing stock data for all regions of the US are available (e.g. [21]). Weather data, important model inputs, are available for most regions of the country [10], [21].

Irradiance reductions are a function of extent and transmissivity of tree crowns. Tree-shade has been modeled as a uniform irradiance reduction constant over time [15], as a plane horizontal building shade uniformly distributed around a building [8], as a horizontal cylinder simulating a continuous row of trees [30], as rectangular solids placed strategically next to a building [8], and as three-dimensional solids of revolution: spheroids, paraboloids, cylinders, etc. [17], [31]. Canopy transmissivity is accounted for using shade coefficients which typically range from 0.5 to 0.9 for leaf-on and from 0.1 to 0.3 for leaf-off periods [11], independent of the method used to describe extent of shade.

Changing tree size (determined by species and age) or location (defined by tree–building distance and tree azimuth with respect to a building) results in dramatic variation in amount and timing of building shade [26]. Tree azimuth is true compass bearing of a tree relative to a building. Effects of mature, medium-sized, deciduous trees have most often been modeled [8], [14], [26], [30]. McPherson [12] treats trees of 7.3, 11.0, and 15.2 m height. These sizes can be interpreted as corresponding roughly to tree ages of 20, 30 and 45 years, or mature trees of small, medium and large size. Species differences are accounted for primarily by assigning trees to mature size classes, and distinguishing between evergreen and deciduous leaf patterns. More precise information on species-dependent size and structural characteristics of urban tree crowns is becoming available [20].

Tree azimuth is most often accounted for by placing tree(s) in various combinations adjacent to frequently sunlit building walls, e.g. those facing east, south or west. Generally only a few azimuths are considered [8], [12], [14], [30]; a full range of azimuths is less commonly treated [26]. Tree–building distances of 2–5 m are most common [8], [14], [30], where trees approach but do not overhang the roof. McPherson [12] treated distances of 3.6, 6.7 and 10.4 m for one scenario in Chicago. Jones and Stokes Associates Inc. [9] proposed savings inversely proportional to the square of the distance to the building, and proportional to the square of canopy diameter. In terms of real-world applications, there is little available information about tree location with respect to buildings of different types as a function of species and age. Ground surveys of selected properties [27] and interpretation of aerial photographs [13], [18] have been used to tabulate frequency of occurrence of trees at various locations around residences in Sacramento, California.

The actual incremental shade on a building surface from addition of a tree as part of a planting program is in part determined by the amount of shade on that surface from existing trees, from other program trees, or from adjacent structures such as fences or nearby buildings. Simpson and McPherson [27] calculated shading from each tree and structure within 18 m. It is difficult in practice to explicitly account for all solar obstructions, especially those near the horizon, so that a global reduction in solar radiation that accounts for existing trees and structures is often used. While resulting effects on energy use are relatively insensitive to the exact value of the reduction factor used [25], failure to account for existing shade inflates the impact of proposed tree planting.

Energy use changes resulting from addition of trees will be diminished to the extent of coincident shading from pre-existing trees. Reductions from multiple trees have been accounted for implicitly in tabulated values (Jones and Stokes Associates Inc. [9]). The reduction factor is approximately 5% per tree, based on Simpson and McPherson’s [27] observation that an added tree produced changes in energy use that were 20–30% less than that for the first tree (200 kWh per tree for cooling and 1.0 GJ for heating annually) when there were a total of three–six trees already present.

While direct comparison of measured and simulated effects of tree-shade on building energy use are few, measurements generally tend to confirm the magnitude of simulation results. Meier [19] reviewed several studies where air-conditioning savings from landscaping were directly measured, concluding that savings of 25–50% are likely through use of trees in the landscape. Akbari et al. [1] compared air-conditioning energy use before and after positioning 16 containerized trees (8 were ∼6 m high, and 8 ∼2.5 m high) so as to shade the southeast and southwest exposures of two residences in Sacramento. Measured savings, determined by comparing energy use before and after the addition of shade trees, averaged about 30% for both sites. Simulated savings were conservative, underestimating measured usage on days with higher cooling loads by up to 50%, with better agreement for lower cooling loads. Simpson [24] found good short term agreement between measured and simulated energy use for 1/4-scale model buildings surrounded by turf and rock ground covers.

Planting trees throughout a city will lead to changes in the energy balance that effect the climate of the entire city, primarily air temperature, wind speed, and vapor pressure. On balance, these large scale effects have been found to reduce building energy use [3], [8], [29]. Wind speed reductions were found to often lead to increases in cooling load, but these were much smaller than savings from either shade or reduced air temperature. Measurements made in and around 0.05–1.1 ha urban wooded sites [23], 25 ha park [4], and 10–40 ha residential areas [28] confirm air temperature reductions from trees that extend outside the treed area up to four times the width of the site. These measurements also showed that vapor pressure differences were small between treed and adjacent untreed locations.