About effect of weeds on spectral reflectance properties of winter wheat canopy

Among the poorly studied factors affecting the spectral reflectivity of crops and, consequently, the success of detection of their condition based on remote sensing data is crop weedness. On the basis of field survey data, the effect of weed infestation on winter wheat spectral reflectance at different stages of vegetation was analyzed using the example of individual fields in the Tula region with chernozems, grey forest, and alluvial arable soils. Under field conditions, crop weedness, spectral reflectance of crops, weeds, winter wheat leaves and soil determined using FieldSpec® HandHeld 2TM field spectro-radiometer (ASD, Inc., USA) was assessed several times during the growing season, and the crop canopy surface was photographed. The decoding of the photos showed that the projective weed coverage on the crop canopy surface is low enough at the beginning and middle of the wheat growing season, but increases significantly since the beginning of leaves yellowing. At the same time, the projective coverage of weeds in the field with chernozems was minimal at the beginning and middle of the growing season, and maximal - by the end of the growing season. Projective coverage of weeds on fields with grey forest and alluvial arable soils did not differ statistically, but on alluvial soils it increased significantly by the end of wheat vegetation. Using the spectral mixing model, the contribution of weeds infestation to the integral reflection of light by crops in the visible and near infrared bands of the electromagnetic waves was estimated. It has been found that despite the rather high weedness of winter wheat canopy in the spring-summer growing season, its projective coverage on the surface of the crop canopy is small. The magnitude of the projective cover of weeds on the surface of crop canopy weakly depends on soil conditions, and is more determined by other factors (history of fields use, crop rotation, etc.). The effect of weed vegetation on the spectral reflectivity of crop canopy changes over time. It is minimal at the peak of the growing season, accounting for several percent for all wavelengths of the visible and near IR range. At the beginning of the post-winter vegetation period, the contribution of weed vegetation to the spectral reflectance of crop canopy can reach 10-20%, and at the end of the vegetation season, weed vegetation can predetermine the spectral appearance of crops at most wavelengths of the considered range. The greatest contribution is observed in all cases in the near IR (710-730 nm) and in the green (520-560 nm) spectral region, but at certain times there are local maxima of the contribution and in the blue spectral region of electromagnetic waves (400-420 nm). The data obtained open up the possibility for the development of new vegetation indices for remote monitoring of crops, which will be less affected by weedness than those traditionally used (for example, NDVI). Conversely, on the basis of the data obtained, special vegetation indices can be proposed for the remote detection of the weedness of winter wheat canopy. © 2020 Russian Academy of Agricultural Sciences. All rights reserved.