Computer Science – Performance
Scientific paper
Jun 2002
adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=2002esasp.474e..11m&link_type=abstract
Proceedings of the SPECTRA Workshop - The concept of a space-borne Earth Observation Mission addressing the terrestrial componen
Computer Science
Performance
Scientific paper
The presentation analyses the role of land surface parameters, in particular those, which can be derived from hyperspectral/directional remote sensing data, for land surface process models. Land surface process models are used to understand and predict the dominant cycles of energy, water, carbon, nutrients and humans on the plant on the local, regional and global scale. They address environmental issues of great importance like the carbon budget and the availability and quality of water as a basis for life. Land surface process models use land surface parameters to characterize the properties of the land surface and to solve the underlying physically based models. Among these parameters are vegetation type, leaf area index (LAI), fraction of absorbed photosynthetically active solar radiation, biomass, soilmoisture or chlorophyll-content. The main characteristics of the cycles on the land surface are complexity as well as large temporal dynamics and spatial heterogeneity on all considered scales. Conventional, state of the art modelling of land surface processes usually derive the temporal and spatial distribution of the parameters involved from interpolation of point measurements, which either leads to large errors or creates prohibitive sampling efforts. Recently land surface process models have learned to treat spatial processes in a spatial way and are now prepared to digest spatially explicit input information e.g. from remote sensing soruces. Remote Sensing data and especially hyperspectral/directional data can be used to derive land surface parameters. Their main advantage for land surface process modelling is, that they can implicitly measure the temporal dynamics and spatial heterogeneity of the reflection of the land surface. Parameter models convert the directional reflectance spectra into spatial fields of land surface parameter values. They in turn can be used as spatially distributed inputs to the process models. In the classical approach (and usually with multispectral (TM or SPOT data)) these parameter models use regressions between measured reflectances in spectral bands (or band ratios) and land surface parameters. Examples of LAI determined from LANDSAT-TM data and successfully assimilated into a crop growth model are presented. Hyperspectral data offers the possibility to analyse specific absorbtion features of substances like water or chlorophyll to determine their abundance on the land surface. Directional measurements allow the derivation of structural information from remote sensing measurements. To fully exploit the information content of these complex measurements they have to be assimilated in radiative transfer models, which can adjust the values of the parameters they use in order to describe the radiative transfer within a canopy (e.g. LAI, biomass) until model outputs and measurements coincide. Best results can be achieved when the radiative transfer model is coupled with a plant growth model, which can suggest a plausible range of parameter values from plant growth modelling based on the weather and soil conditions. An example is shown, where the radiative transfer model GeoSAIL is coupled with the plant growth model PROMET-V to successfully determine the heterogeneous distribution of biomass, carbon flux and nitrogen leaching from a 1000 km2 test area. The biggest challenge still lies in the representation of the complexity of natural and manmade land surfaces. Hyperspectral/directional remote sensing data offers the chance to considerably improve the quality of the retrieved land surface parameters and to retrieve new and important parameters of the land surface, which were not accessible with remote sensing measurements before, like chlorophyll of water content of vegetation. Proper assimilation of the measured spectra into radiative transfer and process models can significantly improve their performance. SPECTRA therefore is a necessary tools to improve the understanding of land surface processes and in this way be able to predict future environmental development as well consequences of alternative environmental management actions.
Bach Heike
Mauser Wolfram
Schneider Kai
No associations
LandOfFree
The use of hyperspectral / directional data in land surface process models does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with The use of hyperspectral / directional data in land surface process models, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and The use of hyperspectral / directional data in land surface process models will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-1893805