|Kustas, William - Bill|
Submitted to: Dahlia Greidinger International Symposium
Publication Type: Proceedings
Publication Acceptance Date: 5/25/2009
Publication Date: 2/15/2010
Citation: Agam, N., Ben-Gal, A., Kustas, W.P., Cohen, Y., Alchanatis, V. 2010. Thermal remote sensing of crop water status: pros and cons of two different approaches. Dahlia Greidinger International Symposium, March 2-5, 2009, Haifa, Isreal. p. 367-375. Interpretive Summary:
Technical Abstract: Recent climate change has lead, in many places around the world, to a decrease in the availability of fresh water resources. This limited water availability decreases the cost-effectiveness of irrigated agricultural crops, and increases the desirability of practices that reduce applied water without decreasing quantity and/or quality of yields. Routine monitoring of crop water status may provide useful information allowing growers to irrigate only when and where needed and thereby conserve water. Continuously growing availability of airborne and spaceborne data has led to development of various methods utilizing thermal remote sensing to detect and monitor water status in agricultural crops. In large, these methods can be divided into two main approaches. The first approach, used since the 1960s, is based on the understanding that canopy temperature is indicative of crop water status. Generally, canopy temperature is normalized to upper and lower bounds, representing non-transpiring and fully transpiring leaves, respectively, to calculate a crop water stress index (CWSI). The normalization allows comparison between the CWSI under different environmental conditions. This approach is simple to apply and requires relatively few inputs, but necessitates thermal images at high spatial resolution since the remotely sensed temperature must represent the canopy only, isolated from the surrounding soil. Such high resolution thermal images are not currently routinely available. The second approach is based on more complex physical models, of which the prime input is thermal images. Numerous models have been developed, which can be further divided into two main categories. The first is based on the "big leaf" theory, according to which the land surface is assumed homogeneous and is treated as a whole. In the second category are the “two-source” models, which relate to the vegetation and the soil separately. The physical models in general, and the two-source models particularly, require more inputs than the CWSI approach, but can utilize thermal images at a coarser spatial resolution. Such imagery is regularly available from several satellite systems. A discussion of the pros and cons of each of the two approaches follows a brief description of their principles and utility.