Nighttime Evapotranspiration from Alfalfa and Cotton in a Semiarid Climate

Nighttime evapotranspiration (ETN) has typically been neglected in estimating water loss from land surfaces. Our objective was to quantify the contribution of ETN to daily (24-h) ET (ET24) of irrigated and dryland cotton (GossypiumhirsutumL.) and irrigated alfalfa (Medicago sativa L.) grown in a semiarid climate. The results were then examined using a Penman–Monteith ETmodel which separates control of ET into its radiation (equilibrium) and atmospheric demand (imposed) components. Nighttime ET was measured at Bushland, TX using weighing lysimeters containing monolithic soil cores of Pullman clay loam (fine, mixed, superactive thermic Torrertic Paleustoll) for alfalfa in 1998 and cotton in 2001. Measured ratios of ETN to ET24 ranged from an average of 3% for a dryland cotton crop to 7.2% for irrigated alfalfa over a season. In the largest events, ETN was as much as 12% of ET24 with single nighttime losses approaching 2 mm. Model calculations showed that virtually all ETN was the result of imposed atmospheric conditions, primarily vapor pressure deficit. However, ETN was also related to sensible heat transfer to the canopy. Nighttime ET can be an important part of total ET of irrigated crops in a semiarid environment. THE contribution of ETN to ET24 is largely ignored in methods that estimate water loss from land surfaces, especially those that use only solar radiation as the energy source to vaporize water. But, the limited research that has been performed has shown that ETN accounted for 7% of ET24 from tall grass prairie vegetation (Sugita and Brutsaert, 1991), 4.1% of the mean ET24 from a willow (Salix viminalis L.) stand (Iritz and Lindroth, 1994), and from 1.7% to about 14% of ET24 for an irrigated alfalfa (Medicago sativa L.) field in a semiarid mountain valley as wind speed increased (Malek, 1992). The ETN proportion of ET24 measured by weighing lysimeters for alfalfa was 8% in North Carolina as reported by England (1963), while Rosenberg (1969) found that it varied from 7 to 21% in spring and 0 to 15% in summer for the central Great Plains. Nighttime ET has been shown to include water losses from plants. Reported nighttime transpiration losses were 19% of daily totals for kiwifruit [Actinidia deliciosa (A.Chev.) C.F. Liang et A.R. Ferguson] vines and 6% for apple trees (Malus sylvestris Mill. ‘Red Delicious’) (Green et al., 1989), and 8% of daily losses in unstressed wheat (Triticum aestivum L.) grown in a dry environment (Rawson and Clarke, 1988). Partially open stomata during the nightwere found in cotton (Sharpe, 1973), and in kenaf (Hibiscus cannabinus L.) (Muchow et al., 1980). Other ETN sources identified were soil water (Iritz and Lindroth, 1994), dew (Malek, 1992), and canopyintercepted rainfall (Pearce et al., 1980). Using 24-h totals from a weighing lysimeter, Jackson et al. (1983) concluded that the ETof wheat predicted from one time of day meteorological measurements needed to be multiplied by 1.1 to account for ETN, but that the accuracy of the multiplier depended on prevailing environmental conditions. Penman (1948) presented a general formula for the rate of open water evaporation and later applied to bare soil and grass that was a function of meteorological elements such as temperature, vapor pressure, wind, and radiation. Monteith (1965, 1981) later added resistances (e.g., aerodynamic and surface) to fluxes of momentum, heat, and water vapor through the system. It can reasonably be assumed that different processes determine nighttime and daytime ET in what later became known as the Penman–Monteith ET model. McNaughton and Jarvis (1983) presented a form of the Penman–Monteith ETmodel which helps examine these processes. It consisted of an equilibrium ET (ETeq) component, in which ETwas a function of available energy at the surface, and an imposed ET (ETimp) component, in which ETwas a function of bulk atmospheric conditions. The contributions of the two components were then weighted by a plant–atmosphere decoupling factor V. The equation was written as lE 5 V[D(Rn 1 G)/(D 1 g)] 1 (12V)[(rcpVPD)/(grs)]