Vegetation Stress Indicators Derived from Multispectral and Multitemporal Data

Paper number

IAC-06-B1.P.1.02

Author

Dr. R. Kancheva, Solar-Terrestrial Influences Laboratory, Bulgarian Academy of Sciences, Bulgaria

Coauthor

Mrs. D. Borisova, Solar-Terrestrial Influences Laboratory, Bulgarian Academy of Sciences, Bulgaria

Year

2006

Abstract

During the last years destructive processes caused by natural disasters or human activity are in the focus of the scientific research and occupy the attention of social communities and government authorities. A great variety of projects has been developed aimed at environmental monitoring and control. Recent developments in environmental studies are greatly connected with worldwide ecological problems related to anthropogenic impacts on the biosphere and first of all on vegetation. Advanced monitoring and alerting techniques, on-time information extraction, modeling and forecasting technologies are a preposition for successful data application and decision support in environmental studies. The interrelated nature of many environmental problems has imposed the need of multipurpose programs, data integration and information sharing between different databases.
Remote sensing technologies are widely used for natural resources management, crop assessment, land covers change detection, ecosystems preservation and many other world significant problems. Two issues are of essential importance for the application of airborne and satellite data: the development of efficient algorithms for data analysis and the explicit information about land covers spectral behaviour under different conditions, both associated with the higher reliability of the derived information. In agriculture remote sensing is applied for assessing plant development processes and growth conditions. Along with other databases provision, remote sensing is a tool that is used for retrieving plant agronomical variables in order to evaluate crop current state and make predictions. Especially valuable are the spatio-temporal aspects of the remotely sensed data in change detection and identifying of stress situations.
With all this in view our paper focuses on different techniques for handling data from multispectral and multitemporal measurements analyzing plant spectral signatures in terms of plant state and response to stress factors. Results are presented from a study devoted to crop state assessment using spectral-biophysical modeling approach. The investigations were carried out on various agricultural species (winter wheat, spring barley, peas, alfalfa) grown under different conditions (soil properties, fertilization, heavy metal pollution). Soil acidity, nutrient deficiency and toxic contamination were stress factors that affected crop development. Their impact was evaluated through different plant spectral signatures (reflectance, absorption, transmittance, color characteristics, fluorescence) which were examined for their ability to serve as sustainable stress indicators at different stages of plant ontogenesis. Statistical relationships were established between the stress factors, plant spectral and biophysical response thus attaching a quantitative measure to crop stress spectral indicators. A comparison was made between the stress bioindicators (reduced biomass and leaf area index, chlorophyll inhibition, senescence effects) and a variety of spectral stress indicators demonstrating a very good agreement. The obtained results indicated that growth conditions caused significant variations of plant spectral signatures which were used not only to discriminate between stressed and healthy vegetation but also to quantitatively assess the stress impact. Multispectral data were successfully used for crop agrodiagnostics in terms of plant growth parameters and yield prediction.  The analysis of multitemporal data provided for crop monitoring throughout the growing season and for evaluation of the stress impacts at different stages of plant development.
The value of the study is in its dedication to such an important issue as vegetation monitoring in respect to stress detection. The state assessment of natural vegetation resources or agricultural crops is of equal importance for biodiversity preservation, farmland management and precise agriculture running. Such studies serve as a solid foundation for incorporating geo-spatial-temporal observations into an integrated informational system for various applications ranging from databases development to operational monitoring and decision making.

Abstract document

IAC-06-B1.P.1.02.pdf

Manuscript document

IAC-06-B1.P.1.02.pdf (🔒 authorized access only).

To get the manuscript, please contact IAF Secretariat.