Introduction

Water-resource planners and engineers, and farmers often need to estimate the amount of water that is consumptively used on croplands, and estimates of cropland water use are important for computing a water budget for WRIA 1. Although the term "consumptive use" has been defined to include both evaporative losses and the relatively small amount of water consumed by incorporation in the tissues of growing plants, evaporative losses represent by far the largest consumptive use of water on these lands (Jensen and others, 1990). Thus, the problem of estimating consumptive use is practically one of determining evaporation. Evaporation from terrestrial sites is typically referred to as "evapotranspiration" (Et), which is the sum of evaporative losses by transpiration by plants, evaporation of water from wetted foliage, and evaporation from the soil.

A field with crops presents several types of surfaces that are exposed to the atmosphere and from which evaporation can take place, including surfaces of leaves and stems of the crop and of weeds, and the soil surface. However, for many purposes these multiple surfaces can be treated as a single effective "surface" that exchanges heat, water, and momentum (from wind) with the atmosphere. Thus, a surface of interest could be a field of corn or wheat, or bare soil for which the Et rate is sought by an investigator.

Measuring Et

Many Et measurement techniques are available with which to reliably determine cropland Et. Some of the more scientifically rigorous techniques employ theory of atmospheric turbulence to infer rates of Et based on micrometeorological measurements that are made in the turbulent air that passes over a surface of interest. One such technique, eddy correlation (Campbell, 1977), relies on rapid measurements of air moisture content and vertical wind speed to compute the flux of water vapor from a surface, and can be used to determine Et rates with an accuracy of ± 10 percent under ideal conditions.

Evapotranspiration can also be accurately determined using lysimeters. A lysimeter is a container of fixed volume that contains soil and that can be planted with a crop of interest. Et is determined from the water balance of the lysimeter, and by monitoring both the amount of water in the soil and the amount of water added by precipitation or irrigation during a known period of time. A properly installed and maintained lysimeter can reliably yield Et rates for many types of field crops.

Modeling Et

For most instances where Et rates are sought, available instrumentation or human resources are not sufficient to apply any of the Et measurement techniques such as those described above. Instead, investigators use one or more of the many empirically based Et models that have been developed. The type of model most commonly used, and the type described here, is of the form:

where

Et is the rate of evapotranspiration from the surface of interest, such as a field of raspberries, expressed in inches of water per unit of time (for example, inches per day or per month);

k is an empirical, surface-dependent coefficient, dimensionless; and

Et_r is a reference Et rate that is determined or estimated for the site of interest.

Jensen and others (1990, p. 57) define reference Et as "the rate at which water, if available, would be removed from the soil and plant surface of a specific crop, arbitrarily called a reference crop." The purpose of Et_r is to provide an Et standard that controls for effects of spatial and temporal climate variations on the Et rate from the surface of interest. The two most commonly used reference crops are perennial grass and alfalfa.

Summaries of long-term average Et_r for grass and alfalfa at the Washington State University (WSU) Research and Extension Unit near Mount Vernon, Washington, and grass consumptive use at Clearbrook, Washington, are presented in Reference Evapotranspiration. Grass Et_r was computed by the Washington Public Agricultural Weather System (WPAWS) using the so-called "FAO-24 Penman" technique (M.J. Hattendorf, Director, WPAWS, oral commun., January, 2000). The FAO-24 Penman technique (Doorenbos and Pruitt, 1977) was developed after Penman's (1948) combination energy-balance and aerodynamic equation. Alfalfa Et_r was computed by WPAWS using the Kimberly Penman technique (M.J. Hattendorf, Director, WPAWS, oral commun., January, 2000), which is another variant of Penman's original equation (Jensen and others, 1990). Both of these Penman-type equations employ the environmental variables of net radiation, soil heat flux, air temperature and water vapor pressure, and wind speed that are typically measured at a height of about 6 feet above the reference crop. Grass consumptive use, which for the purpose of equation (1) is functionally equivalent to Et_r, was computed by the temperature-based FAO-24 Blaney-Criddle equation (Doorenbos and Pruitt, 1977) from daily average temperature at Clearbrook, Washington (Hydrosphere Data Products, Inc., 1999).

The coefficient k describes the relation between the Et rates from the surface of interest and from the reference crop. When the surface of interest is a crop during its growing season, k is referred to as a crop coefficient (k_c). The crop coefficient embodies characteristics of the crop, soil, climate, and cultural practices, such as irrigation scheduling, that influence evaporative loss. Because crop density, height, leafiness, and physiological activity usually vary during the growing season, and these factors affect the Et rate, crop coefficients commonly vary during the growing season. Thus, when an investigator wishes to estimate the Et rate at a particular time during the growing season from a crop of wheat, for example, that person can do so by applying equation (1) with locally available Et_r and a value of k_c that is appropriate for wheat during the particular time during the growing season. Published crop coefficients for some important commercially grown crops in WRIA 1 are given or described in Crop Coefficients.

During the dormant season, James and others (1988) indicate Et can be computed from equation (1) by replacing k_c with an evaporation coefficient, k_e, which James and others indicate approximately equals 0.9 for western Washington. This coefficient is intended to describe the relation between Et_r and dormant-season Et, such as might include a small amount of plant transpiration from perennials, evaporation of intercepted precipitation, and evaporation from the soil. An investigator must be able to distinguish between the growing and dormant seasons to select the appropriate coefficient, k_c or k_e. James and others (1988) provide techniques for estimating the times of the beginning and ending of the growing season, and these are summarized in Crop Coefficients.

Advantages of Et Modeling

Modeling of crop Et has some important advantages over measuring it, and the widespread use of Et modeling for water budgeting and related work partly reflects these advantages:

1. Simple Et models, such as equation (1), are much less data-, labor- and equipment-intensive to apply than are Et measurement techniques. Data that are used to compute Et_r, which can include air temperature and water vapor pressure, wind speed, solar or net radiation, and soil heat flux, are relatively simple to collect using automated weather stations. Crop coefficients can sometimes be obtained from the literature. In contrast, procedures and equipment for measuring Et by eddy correlation or by using lysimeters are demanding and time-consuming.

2. Et modeling can be conducted retrospectively to construct histories of Et. Past crop Et can be estimated if one has a knowledge of cropping history and crop coefficients appropriate for past crops.

Some Limitations and Sources of Error for Modeled Crop Et

Most crop Et models are developed empirically by measuring Et for a given crop, soil type, climate, and irrigation regime, and by comparing measured Et with Et_r. Application of a published model to any of the myriad of possible combinations of crop, soil type, climate, and irrigation regime is almost always accompanied by a risk of using the model for circumstances that are substantively different from those under which it was developed. Because of the strongly empirical basis of Et models, their misapplication can produce inaccurate and misleading Et estimates. Given below are some practices that will minimize the risk of misapplication:

1. Using the same technique or equation for computing Et_r when applying a model as was used for developing the model. Different techniques for computing Et_r each yield different values and it is important that crop coefficients that have been computed using one Et_r technique not be applied to estimate Et from Et_r computed using another technique (Jensen and others, 1990, p. 56).

2. Computing Et_r using environmental variables measured over an extensive area that is occupied by the reference crop or vegetation that is similar to the reference crop. Air temperature and vapor pressure, and wind speed measured near a surface strongly reflect interactions of the surface with the atmosphere (Brutsaert, 1982, p. 214). As a result, the measured values of air temperature and vapor pressure, and wind speed will depend on the nature of the surface over which they are measured. Because Et_r is computed using some or all of these environmental variables, the value of Et_r also will depend on the nature of the surface. For example, Et_r computed from environmental variables measured over an urban area with many arid surfaces, such as dry roofs and dry pavement, will likely not be the same as Et_r computed from measurements made over the reference crop. If the arid-surface Et_r is then used with equation (1), the resulting Et estimates can be substantially biased.

3. Using appropriate crop coefficients. Obtaining crop coefficients from the literature for a particular cropland setting can be problematic because studies to determine k_c's for all possible crops, soil, climates, and irrigation regimes have not been conducted and published. Moreover, the transferability of crop coefficients among different crop varieties is generally not known. One means to assure the crop coefficients are appropriate is to develop them for local crops, climate, soils, and irrigation practices.

References Cited

Brutsaert, W., 1982, Evaporation into the atmosphere: Boston, Mass., D. Reidel, 299 p.

Campbell, G.S., 1977, An introduction to environmental biophysics: New York, Springer-Verlag, 159 p.

Doorenbos, J., and Pruitt, W.O., 1977, Guidelines for predicting crop water requirements: Rome, Italy, Food and Agricultural Organization of the United Nations, FAO Irrigation and Drainage Paper 24, 2nd ed., 156 p.

Hydrosphere Data Products, Inc., 1999, Climatedata^TM vol. 10.3, NCDC Summary of the Day, West 2: Boulder, Colorado, Hydrosphere Data Products, Inc., CD-ROM.

James, L.G., Erpenbeck, J.M., Bassett, D.L., and Middleton, J.E., 1988, Irrigation requirements for Washington- estimates and methodology: Washington State University Extension Bulletin 1513, 37 p.

Jensen, M.E., Burman, R.D., and Allen, R.G., eds., 1990, Evapotranspiration and irrigation water requirements: New York, American Society of Civil Engineers, 322 p.

Penman, H.L., 1948, Natural evaporation from open water, bare soil, and grass: Proc. R. Soc. London, Ser. A193, p. 120-146.