Agricultural Water Management 98 (2011) 1523–1535 Contents lists available at ScienceDirect Agricultural Water Management j ourna l ho me page: www.elsevier.com/locate/agwat Using radiation thermography and thermometry to evaluate crop water stress in soybean and cotton S.A. O’Shaughnessy ∗ , S.R. Evett, P.D. Colaizzi, T.A. Howell USDA-ARS, Conservation and Production Research Laboratory, P.O. Drawer 10, Bushland, TX 79012, United States a r t i c l e i n f o Article history: Received 17 September 2010 Accepted 9 May 2011 Available online 15 June 2011 Keywords: Infrared thermometry Infrared thermography Empirical crop water stress index Water use Cotton yields Soybean yields a b s t r a c t The use of digital infrared thermography and thermometry to investigate early crop water stress offers a producer improved management tools to avoid yield declines or to deal with variability in crop water status. This study used canopy temperature data to investigate whether an empirical crop water stress index could be used to monitor spatial and temporal crop water stress. Different irrigation treatment amounts (100%, 67%, 33%, and 0% of full replenishment of soil water to field capacity to a depth of 1.5 m) were applied by a center pivot system to soybean (Glycine max L.) in 2004 and 2005, and to cotton (Gossyp- ium hirsutum L.) in 2007 and 2008. Canopy temperature data from infrared thermography were used to benchmark the relationship between an empirical crop water stress index (CWSI e ) and leaf water poten- tial ( L ) across a block of eight treatment plots (of two replications). There was a significant negative linear correlation between midday  L measurements and the CWSI e after soil water differences due to irrigation treatments were well established and during the absence of heavy rainfall. Average seasonal CWSI e values calculated for each plot from temperature measurements made by infrared thermome- ter thermocouples mounted on a center pivot lateral were inversely related to crop water use with r 2 values >0.89 and 0.55 for soybean and cotton, respectively. There was also a significant inverse relation- ship between the CWSI e and soybean yields in 2004 (r 2 = 0.88) and 2005 (r 2 = 0.83), and cotton in 2007 (r 2 = 0.78). The correlations were not significant in 2008 for cotton. Contour plots of the CWSI e may be used as maps to indicate the spatial variability of within-field crop water stress. These maps may be useful for irrigation scheduling or identifying areas within a field where water stress may impact crop water use and yield. Published by Elsevier B.V. 1. Introduction Crop sensing technologies have potential as tools for monitor- ing crop water status, predicting yields (Idso et al., 1978, 1980; Pinter et al., 1983), improving water use efficiency (Evett et al., 1996, 2001, 2006) and harvesting methods, and precisely managing irrigation (Wanjura et al., 1995). Useful information on crop canopy temperature and water relations can be derived from infrared ther- mography and thermometry. Infrared thermography has been used in agriculture as a non-invasive, versatile imaging tool to inves- tigate biotic stresses (disease or insect infestation), and abiotic stresses (e.g., nutrient and water deficit). Chaerle et al. (2006) combined thermal and chlorophyll fluorescence imaging to study spatial and temporal heterogeneity of leaf transpiration and pho- tosynthesis. These techniques helped identify pre-symptomatic responses (higher chlorophyll intensity co-located with thermal symptoms) and provided diagnosis of diseases (fungal and bacte- ∗ Corresponding author. Tel.: +1 806 356 5770; fax: +1 806 356 5750. E-mail address: susan.oshaughnessy@ars.usda.gov (S.A. O’Shaughnessy). rial infections) and abiotic stresses not yet perceptible in visible spectrum images. Stoll et al. (2008) used an infrared camera to observe thermal responses in grapevine infected with a fungus well in advance of visible symptoms. Bulanon et al. (2009) fused digital and thermal images taken of the same area to improve the identification of fruit for robotic harvesters. Studies involving the analysis of abiotic stresses with thermal imagery include those by Jones (1999) and Jones et al. (2002) in which field studies were designed to assess the consistency and repeatability of using thermal imagery to measure stomatal conductance in grapevine canopies. They concluded that thermog- raphy allows for semi-automated analysis of large areas of canopy with much more effective replication than can be achieved with porometry. Leinonen and Jones (2004) classified thermal images to identify leaf area, sunlit, and shaded parts of the canopy. Their methods provided improved estimates of temperature distribution across a canopy by separating out mixed pixels and reducing the effects of thermal contribution from background, and angle of view (Luquet et al., 2003). Leaf water potential measurements became routine in the 1960s with the commercialization of pressure chambers (Turner, 1988) 0378-3774/$ – see front matter. Published by Elsevier B.V. doi:10.1016/j.agwat.2011.05.005