More and more evapotranspiration models, evapotranspiration crop coefficients and associated measurements of evapotranspiration (ET) are being reported in the literature and used to develop, calibrate and test important ET process models. ET data are derived from a range of measurement systems including lysimeters, eddy covariance, Bowen ratio, water balance (gravimetric, neutron meter, other soil water sensing), sap flow, scintillometry and even satellite-based remote sensing and direct modeling. All of these measurement techniques require substantial experimental care and are prone to substantial biases in reported results. Reporting of data containing measurement biases causes substantial confusion and impedance to the advancement of ET models and in the establishment of irrigation water requirements, and translates into substantial economic losses caused by misinformed water management. Basic principles of ET measuring systems are reviewed and causes of common error and biases endemic to systems are discussed. Recommendations are given for reducing error in ET retrievals. Upper limits on ET measurements and derived crop coefficients are proposed to serve as guidelines. The descriptions of errors common to measurement systems are intended to help practitioners collect better data as well as to assist reviewers of manuscripts and users of data and derived products in assessing quality, integrity, validity and representativeness of reported information. This paper is the first part of a two-part series, where the second part describes recommendations for documentation to be associated with published ET data.

Crop coefficient curves provide simple, reproducible means to estimate crop evapotranspiration (ET) from weather-based reference ET values. The dual crop coefficient ͑K c ͒ method of the Food and Agricultural Organization of the United States (FAO) Irrigation and Drainage Paper No. 56 (FAO-56) is intended to improve daily simulation of crop ET by considering separately the contribution of evaporation from soil. The dual method utilizes "basal" crop coefficients representing ET from crops having a dry soil surface and separately predicts evaporation from bare soil based on a water balance of the soil surface layer. Three extensions to the evaporation calculation procedure are described here that are intended to improve accuracy when applications warrant the extra complexity. The first extension uses parallel water balances representing the portion of the soil surface wetted by irrigation and precipitation together and the portion wetted by precipitation alone. The second extension uses three "stages" for surface drying and provides for application to deep cracking soils. The third extension predicts the extraction of the transpiration component from the soil surface layer. Sensitivity and analyses and illustrations indicate moderate sensitivity of daily calculated ET to application of the extensions. The dual K c procedure, although relatively simple computationally and structurally, estimates daily ET as measured by lysimeter relatively well for periods of bare soil and partial and full vegetation cover.

The FAO-56 procedure for estimating the crop coefficient K c as a function of fraction of ground cover and crop height has been formalized in this study using a density coefficient K d. The density coefficient is multiplied by a basal K c representing full cover conditions, K cb full , to produce a basal crop coefficient that represents actual conditions of ET and vegetation coverage when the soil surface is dry. K cb full is estimated primarily as a function of crop height. K cb full can be adjusted for tree crops by multiplying by a reduction factor (F r) estimated using a mean leaf stomatal resistance term. The estimate for basal crop coefficient, K cb , is further modified for tree crops if some type of ground-cover exists understory or between trees. The single (mean) crop coefficient is similarly estimated and is adjusted using a K soil coefficient that represents background evaporation from wet soil. The K c estimation procedure was applied to the development periods for seven vegetable crops grown in California. The average root mean square error between estimated and measured K c was 0.13. The K c estimation procedure was also used to estimate K c during midseason periods of horticultural crops (trees and vines) reported in the literature. Values for mean leaf stomatal resistance and the F r reduction factor were derived that explain the literature K c values and that provide a consistent means to estimate K c over a broad range of fraction of ground cover.