Continuous fields: A new approach for representing gradients of biochemical and structural properties of natural ecosystems

Landscape heterogeneity is often described as being made up of two forms, as a series of gradients with gradual differences in concentration or as a mosaic with patches and or corridors with abrupt/hard or soft discontinuities. Natural landscapes usually show both continuous and discrete behaviour. Land cover attributes have for a long time been described using discrete classification systems. Traditional discrete classification introduces classes with hard boundaries and this is in contrast with the heterogeneity of the earth’s surface.

Continuous fields (CF) of land cover properties rather than discrete ones with abrupt boundaries and unrealistic homogeneity are today considered a more realistic approach in land cover mapping. CF’s are a sub-pixel estimate of the percent ground cover and offer land cover data that is richer in information content. In recent years there has been an increasing drive in the need for the use of continuous data in land cover classification, but only a few studies have attempted to develop and apply continuous landscape metrics which are usually based on remotely sensed images or image based maps of continuous variables.

Charity Mundava from RSL worked on the development and evaluation of a concept to generate continuous field maps of columnar water content. The concept incorporates abundance images and thematic water content information derived from Airborne Prism EXperiment (APEX) imaging spectroscopy data and was tested for a high alpine test site in Switzerland, the Swiss National Park. Abundance maps of predominant land cover types (snow, rock, grass, and forest) are combined with individual thematic water content information of each land cover type. The retrieval of water content for snow, rock and grass is based on empirical-statistical methods; for the grass, water content is derived from the inversion of a physically based radiative transfer model.

Results show that continuous fields well depict gradients of biochemical and structural properties of natural landscapes and offer a useful alternative compared to traditional classification schemes. As CF’s enhance the visibility of change vectors, this information also proves useful in studies pertaining to ecology. The evaluation also highlights an important aspect, being that multiple continuous field maps can be generated for various products (i.e. molecules, structural properties, or even ecosystem processes).

Contact
Charity Mundava – charity.mundava@no.spam.wur.no.spam.nl
Alexander Damm – adamm@no.spam.geo.uzh.no.spam.ch
Mathias Kneubühler – kneub@no.spam.geo.uzh.no.spam.ch

PDF Copy
PosterPDF (PDF, 5.7 Mb)