Evaluation of a Pan Evaporation Model for Estimating Post-Planting Street Tree Irrigation Requirements

Introduction

Arborists are acutely aware of the need for maintaining proper soil water relations during tree establishment, particularly in urban areas where site conditions can dramatically increase daily transpiration and water use. However, until recently, little has been done to scientifically quantify water requirements of recently planted street trees. As a result, municipal foresters and landscapers typically provide only general recommendations, such as “irrigate frequently during dry periods” or “irrigate weekly during the first year” to crews or residents seeking post-planting watering tips. Current budget restraints require that many cities recruit nearby homeowners to irrigate newly planted street trees. Without specific and easily followed instructions, homeowners often irrigate only when leaves wilt or scorch, and trees are already under considerable stress at this time. Irrigation is often excessive after this point, as frantic homeowners make too-frequent water applications to visibly affected, yet seemingly unresponsive trees. Stress is exacerbated as soil oxygen needed for normal respiratory function is temporarily displaced from the saturated soil. The net effect of this type of reactive watering approach can be a prolonged establishment period, reduced growth and vigor, and abnormally high mortality rates, often attributed to “transplant shock” rather than poor irrigation practices.

A newly planted street tree, moved balled and burlapped (B&B), may have only 5% of its original root system length to supply water to the entire tree (4). Water stress at transplanting is a major factor responsible for poor establishment (5) and disease predisposition (15). Root growth during the first growing season after transplanting, especially early in the season, may be minimal on some species (6). Favorable root ball water relations are therefore of paramount importance during the first year of establishment.

In urban sites, evaporative demand is high (16), but water movement into root zones is usually disrupted (3). As a result, street trees may be highly dependent on irrigation, whereas their rural counterparts may more easily exist on rainfall. In Seattle, where only 18% of the annual rainfall is received during the growing season, failure to irrigate newly transplanted street trees can result in survival rates as low as 20%. Survival rates approach 100%, however, when trees are irrigated for the first two years (2).

Although some estimates have been made of whole tree water loss by individual trees (reviewed by Lindsey and Bassuk (10)), little or no published research has quantified the post-planting water requirements of street trees during establishment. The difficulty in directly measuring whole-plant transpiration has historically favored the use of more indirect methods which utilize meteorological data such as temperature and relative humidity to estimate evapotranspiration in agricultural crops. However, attempts to apply these formulas to determine transpiration rates of individual trees have been largely unsuccessful. One useful method for determination of evapotranspiration (ET) is the evaporation pan (14). The evaporation pan measures daily water level changes in a large pan of water as a result of evaporation. Atmospheric factors which influence evapotranspiration also influence water loss in the evaporation pan, and actual crop ET can be indexed against pan evaporation for future calculations.

Total leaf area and pan evaporation may account for as much as 85% of the variability of whole tree water loss (9). Knox (8) and Lindsey (9) found a strong correlation between transpiration rates and pan evaporation for several woody species. Lindsey and Bassuk (10) found that transpiration rate of larger trees was 20% that of pan evaporation. Ponder et al. (12) found no significant difference in growth between trees irrigated at rates equaling 25% of pan evaporation over those receiving higher amounts.

Assuming minimal evaporative loss of soil water through mulch and minimal water movement into the root ball across the root ball:backfill interface, the daily whole-tree water use of recently planted street trees can be determined through quantitative measurements of root ball water loss. Recommendations for an irrigation regime can therefore be developed from knowledge of daily root ball water loss and the amount of water held in the root ball. The recommendation could then be adjusted on-site according to local rainfall. The objective of this study was to develop and test a model that would predict an easily followed irrigation regime for recently planted street trees in Norfolk, VA.

Materials and Methods

An irrigation model was developed and field tested on two street tree species in Norfolk, VA during August and September of 1994. The model was comprised of two primary components. The first was a pan evaporation model for estimation of daily whole tree water loss (10). The second was an estimation of irrigation requirements based on replacement of water predicted to be lost from root balls.

Results and Discussion

The 2.8 in (7 cm) water needed every three to four days to maintain root ball moisture within the available range represents a very significant amount of rain. In southeastern Virginia, rainfall events of this magnitude are relatively rare, occurring only during major storms, whereas rainfall events totaling 2 in (5 cm) which would be sufficient to wet the root ball to a 14 in (35 cm) depth (effective rooting area) occur on average every 28 days throughout the growing season (local state climate office). According to our model, almost 22 in (55 cm) rain (77 gal or 29 I) per month would be needed to support each tree during the growing season. The volume calculated through the model would appear to be grossly excessive, particularly since the mean monthly rainfall of 4.3 in (11 cm) received during the growing season seems sufficient to support undisturbed 3-in caliper trees growing in wooded areas. However, unlike their urban, newly-transplanted counterparts, trees in wooded areas typically have unrestricted roots which can exploit a much larger soil water reservoir.

Irrigation regime did not affect caliper during the study period (data not shown). Pear and birch LAI were 2.8 and 1.3, respectively. PE during the experiment (obtained from the nearby state weather station) was similar to historical levels (0.18 in (0.45 cm) vs 0.19 in (0.47 cm)). Recalculation of the model using the slightly lower PE resulted in a change of only 0.2 gal water transpired/day in August.

Substitution of the actual LAI and CP measured for pear and birch controls, 2.8 LAI, 24.1 ft² CP and 1.3 LAI, 49.0 ft² CP, respectively, into the model equation yields a lower daily transpiration rate (1.8 and 1.7 gal for pear and birch, respectively) for the month of August than originally predicted (2.7 gal based on a LAI of 2.0 and CP of 50 ft²). For pear, rainfall in Norfolk, VA would therefore, on average, provide only 30 % of the tree’s daily water requirements. Actual average daily rainfall in August 1994 was only 0.12 gal (0.46 I), representing approximately 7% of the total daily water requirements for pear. Clearly, rainfall is insufficient to meet transpiration demand without irrigation. Rainfall throughout the experiment, however, was considerably lower than average. Only 2.3 in (6 cm) of rain was received at the project location during August and September in contrast to a 31-year regional average of 9.2 in (23 cm). Due to such variations, our model did not account for rainfall. Homeowners would therefore need to withhold irrigation in periods of heavy rainfall to avoid excessive irrigation.

A total of 130 gal (494 I) of water per tree was applied to each model tree during the experiment while the “as needed” pear and birch control trees required an average of 108 and 122 gal (410 and 4641) per tree, respectively. Under the conditions of this experiment where natural precipitation was 78% below normal for the month of August, these results indicate that the model closely approximates the actual water requirements of newly planted pear and birch. However, a simple comparison of the total volume of water applied to treatments and controls does not indicate if the 10 gal (38 I) applied at each irrigation was the amount actually transpired or lost from the root ball between water applications. The irrigation volume was based on the amount of water calculated to resaturate the root ball from a soil tension approximating the permanent wilting point (0.15 MPa), but we actually irrigated when soil tension dropped below 0.055 MPa.

To test the accuracy of the predicted water lost through transpiration, the volume of water lost from four soil core samples was determined from 0.03 to 0.1 MPa tension. From these data, volume of water lost per unit volume soil was calculated, and the volume of water loss in the soil core from saturation to 0.055 MPa was determined through linear interpolation. Upon this basis, the amount of water that would be lost in the root ball between saturation and 0.055 MPa was calculated to be 4.6 gal (17 I). The amount of water applied at each irrigation event could therefore theoretically be reduced from 10 to 5 gal (38 to 19 I) per irrigation if soil tension was 0.055 MPa instead of 0.15 MPa. However, 10 gal (38 I) water insures that the root ball will always be saturated at each irrigation, even if the root ball has dried close to the plant wilting point or if some runoff occurs. Although an irrigation threshold of 0.055 MPa soil tension was established for controls, average soil tension recorded in the root ball at the point of irrigation was slightly higher (i.e. the soil was drier), for both species (pear = 0.072 MPa; birch = 0.077 MPa) since root balls dried rapidly at that tension and irrigation at exactly 0.055 MPa was difficult. Irrigation intervals ranged from 2.6 to 3.3 days in pear and from 2.3 to 3.3 days in birch controls. Mean irrigation intervals (2.8 days for pear; 2.9 days for birch) were similar (data not shown). Utilizing the recalculated mean daily transpiration of 1.8 gal (6.8 I) for both species and the mean interval of 2.9 days for control trees, the model predicts whole tree water use between irrigation events of 5.0 gal (19 I) (1.8 gal. daily loss X 2.9 day interval), close to the estimate of 4.6 gal (17 I) from the laboratory analysis. Using the calculated root ball water lost from saturation to 0.055 MPa (4.6 gal or 17 I) and dividing by the newly calculated daily transpiration loss (1.8 gal or (6.81), trees should be irrigated every 2.6 days, one day less than the interval predicted with the original assumptions. We rounded the original interval down from 3.6 to 3 days, however.

Although the total water applied between species and treatments were similar, considerably more site visits were required to provide water on an “as-needed” basis. A total of 25 site visits were required to maintain proper soil moisture relations in the pear controls (24 for birch controls) versus 13 for both the pear and birch model treatments. Additionally, the average number of control trees irrigated per visit was higher in birch, indicating a closer match to the model for this species than for pear. The increased number of visits needed to irrigate either species on an “as-needed” basis indicate that irrigation policy based on a predetermined frequency and amount would be more cost effective for city employees to implement, and less confusing for homeowners to follow. In addition, the determination of when trees actually need water is difficult without proper instrumentation.

Conclusions

Because of extreme variability in soil composition and difficulty in conducting controlled research in urban areas, water relations in urban soils and its impact on street tree establishment is not well understood. In this study, the progressive decrease in soil moisture in root balls between irrigation events indicates that water movement from backfill into root balls did not prevent the development of water-limiting conditions. These observations further suggest that root balls, rather than backfill soil, serve as the principal water reservoir during the first growing season. Transpiration estimates calculated through the model also suggest that rainfall is insufficient to meet transpiration demand, indicating a need for irrigation during the first year of establishment to reduce tree water stress.

Our model predicted that newly planted 3-in (7.5 cm) caliper B&B pear and birch should be saturated every three days. The trees used in our study were planted approximately five months before the model was tested. Although casual excavation of a few trees revealed little root growth into the backfill at the beginning of the experiment, some new root growth likely occurred, especially by the end of the experiment. Different results may have been obtained had the model been tested immediately after planting. The saturation of root balls every three days, however, closely approximated actual need during the test period. The estimated volume of 10 gal was excessive, although it assured saturation. In areas where surrounding soil is poorly drained, over saturation may result in root anoxia. The irrigation model tested has direct application to the 24 % of large cities which operate a field nursery (7). Application of the model in other localities, which must purchase trees from commercial growers, would be more difficult because of the model’s dependency on root ball soil characteristics.

Water issues have serious implications for municipal reforestation programs. Many cities are periodically under water use restrictions which limit the liberty to irrigate. As a result, affected communities are forced to accept increased mortality rates. This model provides a reasonably simple methodology for quantifying supplemental irrigation requirements of individual trees based on site and biological factors. Equipped with this information, municipal foresters should be able to assess the need for supplemental irrigation and more effectively compete for increasingly scarce and costly water resources. Research is needed to determine the leaf area index and average transpiration rates for common street trees. The accuracy of this irrigation model under different site and soil conditions should be tested.