Page Banner

United States Department of Agriculture

Agricultural Research Service


item Detar, William
item Penner, John
item Funk, Howard

Submitted to: Transactions of the ASAE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/5/2006
Publication Date: 6/1/2006
Citation: Detar, W.R., Penner, J.V., Funk, H.A. 2006. Airborne remote sensing to detect plant water stress in full canopy cotton. Transactions of the ASAE. Vol. 49(3):655-665.

Interpretive Summary: Timeliness of irrigation is important in cotton production. Irrigating either too early or too late can reduce the yields. Because of the natural variability of soils in a large field, timing the irrigation based on what is visible from the edge of a large field can be very misleading. In this two-year study remote sensing provided high-resolution maps of the water stress levels in Acala cotton, using airplane-mounted multispectral and hyperspectral cameras. Half of the plots in two 2.6 ha. experimental fields on sandy soil were water stressed and those in the other half were properly irrigated, using a subsurface drip irrigation system. Flights were made after full canopy cover was established. All the data from 9 flights were combined into one large file for use with multiple regression to find which wavelengths best predicted the temperature rise of the canopy, as determined by a thermal infrared camera, also on board the airplane. The best 2-band model had wavelengths centered at 686 nm and 850 nm. A modified NDVI was also presented.

Technical Abstract: Airborne remote sensing data, using hyperspectral (HSI), multispectral (MSI), and thermal infrared cameras (TIR), were collected for two seasons, with a different variety of Acala cotton and a different experimental field for each season, for a total of 9 flights. The TIR camera was used to detect the elevated canopy temperature that occurs when the plant is water stressed. The degree of stress, as measured by the rise above the lower baseline of the crop water stress index (CWSI), was well correlated to several new vegetation indices. Both linear and nonlinear multiple regression was used to find the best-fitting wavelengths in the range of 429 to 1010 nm for one-, two-, three- and four-parameter HSI models. The best-fitting two-parameter wavelengths were centered at 686 nm and 850 nm, and the best three were 686, 811, and 860 nm. The normalized difference vegetation index (NDVI) was modified by putting extra weight on the red term, and provided a better fit than an un-modified NDVI. The main finding was that plant water stress in Acala cotton at full canopy can indeed be detected with airborne remote sensing.

Last Modified: 06/23/2017
Footer Content Back to Top of Page