K.N. Abbott, B. Leblon, J.P. Prive, T. Cori,M. Cunningham and P. Amirault
SERG Project # 2000/08
Executive Summary
Herbicide application efficiencies over conifer plantations for suppressing wild raspberry may be related to drought conditions. A promising tool to detect drought conditions over large areas is through optical and thermal infrared remote sensing. In our study, we investigated the potential of hyperspectral reflectance and surface radiative temperature to detect drought stress among raspberry plants under laboratory conditions. Spectral measurements acquired in the 400-2500 nm at a fine spectral resolution showed that control plants have a significantly lower reflectance than stressed plants, mainly in the near-infrared and shortwave infrared regions. We were able to define two hyperspectral indices that were correlated to leaf water potentials, p548/p720 and p1335/p588. The first index was related to indirect effects of drought stress, i.e. changes in chlorophyll absorption and content due to water shortage. The second index was related more directly to the direct spectral effect of water in the short-wave infrared bands. The first index worked the best for plants growing on promix, whereas the second works the best for plants growing over sand soils. Although these results are encouraging, they were obtained at the leaf level, and there is a need to calibrate more complex models for using these hyperspectral indices from images acquired by airborne imaging spectrometers, like CASI.
Surface radiative temperatures worked better at the plant level than at the leaf level, because the attenuation of Ts temporal fluctuations resulting from spatial integration over large surfaces makes Ts less sensitive to wind speed and thus less ephemeral when averaged for a large surface. Use of thermal infrared NOAA-AVHRR or LANDSAT-TM should then provide a valuable tool for mapping drought conditions over conifer plantations. For both kinds of sensors, other applications of such a research include the further design of smart herbicide sprayers, which use remote sensing tools for detecting areas where herbicide spraying will be successful.