Ziel
- Application of Spectral Mixture Analysis to Remotely Sensed Data with Different Spectral Dimensionality
- Identification of Areas Affected by Land Degradation Processes and their Characterization in Terms of Ecological Relevant Components
- Monitoring Land Degradation Trends
Progress to end 1991
This section not applicable - new project.
Detailed description of work foreseen in 1992 (expected results)
A major problem in land degradation assessment and monitoring is the difficulty to remotely identify all key components (green vegetation, dry vegetation and woody components, soil) of natural surfaces. While considerable research has focussed on quantifying the amount of green vegetation only few attempts have been made to assess the woody and senescent components though these can represent a significant portion of the biomass within a vegetation community.
In spectral mixture analysis fundamental spectral components are defined in terms of laboratory or field spectra (endmembers) of well characterized materials, and image pixels are modeled as mixtures of these endmembers. In many aspects, spectral mixture analysis appears superior to conventional methods of image enhancement or spectral classification. It can be equally applied to spectral data of varying dimensionality (i.e. operational broadband sensors or experimental imaging spectrometers) and may therefore be considered a potentially valuable tool to
a) remotely identify and characterize effects of land degradation processes in mediterranean regions (soil erosion, accidental or deliberate fires, water shortage)
b) monitor affected areas to detect further degradation or any positive response to conservation measures
c) provide the link between remotely sensed data with different spectral dimensionality, such as operational satellite systems (TM) and present and future imaging spectrometers (AVIRIS, GERIS, MERIS, MODIS, HIRIS) The work in 1992 will primarily concentrate on the application of spectral mixture analysis to imaging spectrometer data sets (GERIS, AVIRIS) acquired over EMAP's test site on land degradation monitoring (Ardeche/Gard, France) in 1989 and 1991. It will include actions on
1) the identification of degraded land units
2) the characterization of degraded areas in terms of their ecologically relevant components (green and dry vegetation, woody materials and rock/soil)
3) establishing correlations between the fractions of these components and ecologically relevant parameters such as percent vegetation cover etc.
4) the comparison of these results with operational satellite data from broad-band optical systems such as the Landsat Thematic Mappers.
Additional emphasis will be placed on evaluating the potential of spectral mixture analysis for the calibration of data from different optical sensors and/or long times series of optical imagery from land surfaces. Strong efforts will apply to the selection of suitable new test site(s) and the identification of national institutes that are interested to cooperate with JRC.
The results may provide environmental information on a spatial scale and level of detail which can hardly be obtained by conventional methods (i.e. ground mapping or aerial photographs). They will apply especially to mediterranean type landscapes, and as such strongly complement research activities of DGXII in the field of "Technological and Natural Risks" (IV.3 Desertification in the Mediterranean area). As these topics can also be expected to form a main point of emphasis in the activities of the future "Environment Agency", it is hoped that the proposed pilot study may lead to future "Support Commission" activities.
The funding for this project is primarily requested for the acquisition of some additional equipment needed for a more dedicated analysis of hyperspectral data (i.e. SUN workstation). Some additional funds will be required for the acquisition of satellite imagery
Short description of evolution of work in 1993
Based upon the results from this first phase, it is intended to concentrate on the application of spectral mixture modelling for an analysis of long time series of Landsat Thematic Mapper data which will then be available for 9 years. One or several test sites will be selected within the mediterranean basin according to the presence of important processes that cause degradation of valuable land (overgrazing, soil erosion, woodland burning, excessive use of water resources by agriculture, tourism and urban agglomerations)
For this phase, it is suggested to activate the links with the second exploratory research project on the analysis of Landsat MSS time series.
Programm/Programme
Thema/Themen
Data not availableAufforderung zur Vorschlagseinreichung
Data not availableFinanzierungsplan
Data not availableKoordinator
21020 Ispra
Italien