Reprinted from Crop Science, Volume 34, Issue 1, January 1, 1994, pages 192-198.
Publisher website: http://www.crops.org
Journal website: http://crop.scijournals.org
NOTE: At the time of publication, the author David J. Wehner was affiliated with the University of Illinois at Urbana-Champaign. Currently, March 2008, he is Dean of the College of Agriculture, Food and Environmental Sciences at California Polytechnic State University - San Luis Obispo.
Estimation of the lower limit of the canopy-air temperature differential, (Tc–Ta)LL, is required for calculation of an empirically-based crop water stress, index. This research determined the complexity of model needed for accurate estimation of (Tc–Ta)LL for several field grown cultivars of Kentucky bluegrass (Poa pratensis L.) and for creeping bentgrass (Agrostis stolonifera L. var. palustris (Huds.) Farw.). Regression models using vapor pressure deficit of the air (VPD), net radiation load (Rn), and wind speed (WS) were developed for predicting (Tc–Ta)LL. The best one to three-variable regression models for predicting (Tc–Ta)LL used variable groups of VPD (r2 = 0.47); VPD and Rn (R2 = 0.66); and VPD, Rn, and WS (R2 = 0.82). Models developed for predicting (Tc–Ta)LL on individual Kentucky bluegrass cultivars, across Kentucky bluegrass, and across both species were tested on a validation data set. Models using only VPD accounted for <2% of variation in actual (Tc–Ta)LL of nonwater-stressed turf. Models using VPD and Rn developed from pooled Kentucky bluegrass data, individual Kentucky bluegrass cultivars, or Kentucky bluegrass and creeping bentgrass data accounted for an average of 15, 13, and 14% of variation in actual (Tc–Ta)LL, while models using VPD, Rn, and WS accounted for an average of 62, 62, and 64%, respectively. On creeping bentgrass, the Kentucky bluegrass model and dual species model introduced a large amount of bias to predicted (Tc–Ta)LL. At sites where environmental conditions are highly variable, the effects of VPD, Rn, and WS must be taken into account to accurately predict (Tc–Ta)LL of turfgrass. A single model appears appropriate for prediction of (Tc–Ta)LL across Kentucky bluegrass cultivars; a separate model for creeping bentgrass is required.
Agronomy and Crop Sciences