Specifying Soil Volumes to Meet the Water Needs of Mature Urban Street Trees and Trees in Containers

Current recommendations

Current recommendations detailing appropriate soil volumes for trees have been culled from a variety of sources in the literature and are presented for comparison in Table 1. Many of these estimates are quite high, up to 7000 ft³ and would be next to impossible to achieve in most street tree plantings. Some of these recommendations are either simple rules of thumb, or are based on plant factors other than empirically determined water use rates. Further questions and considerations come readily to mind. Are changing regional climatic conditions accounted for in these estimates and is the amount and timing of rainfall integrated in some meaningful way? Are the changing water holding capacities of different soil types accomodated? Over what period of time will this soil volume support the tree and where will the water come from? Are these methods based on whole tree water use rates and do they account for species and canopy size differences? It would also be very useful if whole tree water loss estimations were standardized on one common plant parameter. Soil estimates could then be linked directly to this measurement. No one of these soil volume estimations really addresses all of these concerns together.

Where does the water for trees in urban areas come from?

Water is added to the soil mainly through precipitation. For the global hydrological cycle, precipitation equals evaporation. However, for discrete areas this is not always true, as an examination of modified climatic diagrams created for a range of United States cities shows. For these cities, atmospheric evaporative demand almost always exceeds precipitation, especially during the period of greatest tree growth, May through October (Figure 1). Atmospheric evaporative demand rises steadily over the growing season, peaking mainly in July, less frequently in June. This only represents the potential evaporation both from the soil and transpiration from plants (evapotranspiration) that could occur given prevailing atmospheric conditions. Actual transpiration from plants can be much less.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Climatic graphs of mean monthly precipitation and evaporation rates for eight major U.S. cities. The Y axis is in inches of precipitation and evaporation, the X axis is in months.

Moreover, not all precipitation is particularly effective. While most of the moisture in the soil available to trees is obviously derived from precipitation, not all precipitation increases soil moisture. Significant amounts may be evaporated before reaching the ground, may be intercepted by the canopy foliage, lost by surface runoff, or percolated beyond the root zone (5, 34).

Therefore, the proportion of summer precipitation that actually becomes available for plant use is the result of complicated interplay between atmospheric evaporative demand, the duration and intensity of rainfall, tree canopy size and structure, and the waterholding and drainage capacities of the soil. As an alternative, summer soil water storage values could be calculated if a soil profile description and textural classification were known for the area of interest. This information is extremely difficult to obtain for disturbed, heterogeneous urban soils. We can therefore use precipitation rates only as a general estimate of the water available for tree uptake for any defined period of time.

Estimating whole tree water use with pan evaporation data

There are few studies that have quantified the water demands of trees. Kramer (21) estimated that a 35’ height tree with an actual leaf surface area of 2000 ft² tree might lose up to 35 gallons of water a day. Vrecenak and Herrington (46) estimated 250 gallons a day for a 64’ canopy diameter tree of average density. For comparison, a typical 4’×4’×3’ (depth) tree pit with a loam textured soil having an available waterholding capacity of 12% and total volume of 48 ft³ could hold approximately 45 gallons of water, which the larger tree would use in a little over two hours. Trees growing in these pits will fare poorly as they get larger and die, unless the roots are able to move out of this constraining volume of soil and “break out” into amenable soils nearby.

It is not always possible to directly measure water loss. More common are indirect methods using climatic data and these methods have largely been developed for agronomic crops. Currently, over thirty different mathematically derived weather-based formulas have been developed for the sole purpose of predicting evapotranspiration and calculating the subsequent irrigation needs of these crops (10, 31, 39). These formulas vary both in complexity and in the type and quantity of data required. Application of some of these formulas to modeling the water use of single trees has been attempted by a few studies but is still highly problematic (24, 42, 43, 45).

Alternatively, one simple and reasonably accurate approach to estimating crop water use has been through the use of an evaporation pan (6, 48, 41). Nine types of pans are in current usage, the most common and considered the standard however, is the U.S. Weather Bureau Class A pan (Figure 2). This metal pan is round with a diameter of 47½” (120.65 cm), 10” deep (25.4 cm), and is placed slightly above ground level (13, 31). It is filled with water and a micrometer gauge measures daily water level changes that are a result of free surface water evaporation from the pan. Typically, evaporation from a pan integrates the major environmental influences, sunlight, temperature, wind and humidity. Atmospheric evaporative demand can then be calculated.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Evaporation from the pan compared with transpiration from the four tree species for a sampling of dates.

Agronomic crop canopies are however, qualitatively different from an isolated tree canopy. When soil water is not limiting, and atmospheric conditions are primarily determining the rate and amount of whole tree water loss, can a proportional relationship be established between water evaporation from the surface of a pan and transpiration from the surface of a leaf? Conveniently, evaporation from the pan is measured as inches of water lost per square inch of pan surface, which can be converted to milliliters of water per square centimeter of pan (ml/cm²). Likewise, transpiration in plants can also be characterized as ml of water lost per cm² of leaf surface area.

These graphs were derived from data in Farmsworth and Thompson (12); Farmsworth et al. (13) and NOAA (32). The data represents calculations of evapotranspiration and pan evaporation.

It must be emphasized again that what the pan predicts is the potential transpiration that can occur from a plant under the prevailing atmospheric conditions, the actual amount will generally be lower. This is because of differing physical and aerodynamic properties between a pan and a leaf. Factors that increase evaporation from the pan compared to the plant are: water absorbs more heat than a leaf, heat may be transferred from the metal sides of the pan, heat may be stored and released it at night from the pan, and microclimatic conditions existing directly above the pan may be different than those above the plant (31). And too, as noted previously, soil and plant resistances can also significantly lower transpiration relative to pan evaporation.

In a previous study, the relationship between pan evaporation and gravimetrically determined water loss from tree canopies was derived for a variety of tree species over two growing seasons in Ithaca, N.Y. (27) These species, representing a range of leaf sizes were Amelanchier ‘Robin Hill Pink’, serviceberry; Sophora japonica ‘Regent’, Japanese pagoda tree; Tllia americana ‘Redmond’, basswood; and Fraxinus americana ‘Autumn Purple, white ash. The results of this experiment yielded a significant regression equation, whereby 85% of the variability in whole tree water loss could be accounted for simply with knowledge of total tree canopy area (or leaf area) and pan evaporation. Pan evaporation, therefore, was a significant predictor of whole tree water loss on a daily basis for a range of atmospheric conditions. Knox (19) also found a strong correlation between pan evaporation and water use among five woody species growing in one gallon containers. However, instead of actual leaf area, a growth index was included with pan evaporation.

Also in our previous study we found that whole tree water loss relative to pan evaporation was not statistically different for the four species. On any given day, over comparable surface areas, water transpired from the trees generally averaged 30% of the water evaporated from the pan (Figure 2). In addition, though many studies discuss the possible effect of smaller leaf sizes on reduced water losses (26, 33, 40), in this study, leaf or leaflet size was not a good predictor of water loss. Transpiration increased only as overall canopy area increased, even though these four trees represent a gradient of leaf sizes from 5 to 46 cm². It would appear then that individual correlations between each species and pan evaporation may not have to be established to accurately describe whole tree water loss.

This 30% seems like a low value compared with the ones already derived for other trees, such as 25-50% for pecan (28), 40-135% for various fruit and nut trees (48), and 60-70% for apples in a semi-arid region (25). It must be remembered though that these other values included evaporation from the ground surface as well, which was eliminated in this study. Using small field grown liners, Ponder (37) found that replacing only 25% of net evaporation from a Class A pan produced plants that were not significantly smaller than plants grown with higher replacement rates. Our study also showed that the ratio of transpiration to pan evaporation decreased rapidly with increasing canopy size, dropping to about 20% in the larger trees (Figure 3). This is probably due to the effects of greater mutual leaf shading in these trees, which resulted in reduced water losses per cm² of leaf area. Therefore, while larger trees lose more water on a whole tree basis, they lose less per cm² of leaf area. This would indicate that as a tree canopy continues to mature, this ratio could in fact be much lower.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

The change in the ratio of transpiration to pan evaporation as canopy size increases. All species aregrouped. Standard errors as noted.

A Methodology to Determine Adequate Soil Volumes

Knowing now that there is strong relationship between pan evaporation and whole tree water loss and that a tree is expected to typically lose only 20% of what the pan loses, a methodology can be formulated. All of the calculations will be based on a hypothetical tree with a crown diameter (width) of 20’, and an approximate height of 35’. This tree will be growing in Ithaca, N.Y. The intent of three steps that follow is to present the mathematical calculations of whole tree water loss, soil volume, and pit configuration in a logical order with informed discussion offered on the various decisions that must be made as one precedes through this methodology.

Discussion

A summary of the steps involved and the data needed to compute these steps are given in Table 2. Several points need to be emphasized. The highest mean monthly pan evaporation value was used to calculate daily water use, and this represents the extreme condition. Generally, water use may be much lower on a dialy basis over the whole growing season. The highest water use typically occurs in July. This might be the month to target for supplemental irrigation, if at all. Also it will be at least 10-15 years before the tree used in the example reaches a size requiring the full use of all available water in this soil volume. The implication is that this volume is selfsupporting for this number of years. When the maximum tree size used to make the calculations has been reached, the tree water supply needs should be assessed if one anticipates significantly more growth. At this point it should be determined if summer soil water storage appears to be occurring in sufficient amounts, or whether supplemental irrigation needs to be applied.

Importantly, this methodology also allows one to work from the other direction. If given an existing volume of soil in a tree pit, vault or container, one can decide what size tree this volume will reasonably support. Up to this point it has been assumed that the entire volume of soil provides usable rooting space. Obviously when planting directly into existing soils in urban areas, good soil structure may be lacking, i.e. roots may not be able to penetrate compacted soil. Appropriate soil remediative action must take place then before planting.

It should be emphasized strongly that for these volumes to work, tree pits, extended shared space, and containers all must be mulched. A coarse textured mulch, 3-4 inches deep, with a particle size roughly that of pea gravel, will conserve over 80% of the precipitation that accumulates in the soil (16). Groundcovers used under the tree canopy, especially turf, quite effectively compete for water with tree roots. Currently, it is hard to predict the amount of this additional water loss, and so these plantings should be avoided unless planted areas are irrigated.

The relationship of soil volume needed per unit area of crown projection has been computed for the six representative cities (Table 3). Omitting Phoenix, AZ (a city experiencing exceptionally high atmospheric evaporative demand coupled with low precipitation), rounding these values up yields a general estimate of 2 ft³ of soil per 1 ft² of tree crown projection. This figure is in agreement with other related work. Re-interpretation of the estimation given by Vrecenak and Herrington (46), in an energy budget analysis of whole tree water loss, yields 1.6 ft³ of soil per 1 ft² of crown projection. Bakker (3) deriving transpiration rates from annual forestry values using a multiplier, calculated 2½ ft³ of soil per 1 ft² of crown projection.

If the total volumes derived from this methodology are hard to obtain at the desired planting site, then perhaps supplemental irrigation should be installed. Likewise, trees that are smaller at maturity and need less total soil could be planted. The best alternative is to modify adjoining soils under paved areas and then cover them with pervious paving. This paving will help ensure vital oxygen diffusion and water infiltration through the soil (11). Currently, aggregate-based tree pit soil mixes that can be compacted for use under these paved areas and yet still allow adequate root growth are being developed and tested on site (1,44).

A final caveat concerns the reliability of using pan evaporation values that are not specifically tied to one urban site, where microclimatic conditions result in evaporation values that can be very different from weather station data (47, 18,14). Most evaporation values are not obtained from airports or research stations, areas typically outside of the city proper. Predicting the size of any given site specific “urban effect” is highly problematic. The built environment is complicated and atmospheric demand conditions are still largely unquantified. This methodology though, is meant to offer just a general approximation of supportive soil volumes. More localized pan evaporation readings would be ideal but they are hard to obtain. Just as likely, informed and intuitive adjustments could be made in the field by the professional. If one suspects that a given planting site is subject to greater atmospheric demand than the pan evaporation values indicate, either larger evaporation values could be substituted, or a shorter rain/irrigation period could be specified.

Summary

Street trees live on average 7-10 years, with trees in containers living only 2-5 years (29). In Seattle, 80% of unirrigated newly planted street trees died within two years (9). Soil, overly wet or too dry, or even more simply, the lack of soil, can account for many tree survival problems. The challenge is to engineer a larger and more suitable soil environment, especially for the inner city street tree. Unfortunately, outdated installation details, planting specifications and procedures are often still being used. Successful urban planting must be properly informed by a new landscape technology based on the broadening body of scholarly urban tree research.

This soil volume methodology, and the subsequent recommendation of 2 ft³ of soil for every ft²of crown projection, is an attempt to transfer a vital part of this burgeoning technology into the hands of interested professionals. Hopefully, the resulting applications of this soil volume methodology can enhance current attempts to “green” our cities, making them more aesthetically pleasing, livable, and ecologically sound environments.