From global to regional scale: Remote sensing-based concepts and methods for mapping land-cover and land-cover change in tropical regions

A great number of studies have been dealing with land-cover mapping of tropical regions using earth remote sensing technology recently. This is partly due to a growing number of operational sensor systems for both scientific and commercial use and also because of an increasing demand for land-cover information relevant to global environmental issues and international policy instruments (e.g. the Kyoto protocol). Within this context, the present article discusses the state of the art of data processing and analysis for the assessment of broad scale land-cover and land-cover change in tropical regions. Current global scale land-cover maps are compared with regional satellite mapping products (Landsat/ETM+) for a test region in the humid tropics of Central Sulawesi, Indonesia. The paper suggests the land cover classification system (LCCS) to be used as the conceptual basis for future land-cover analysis in Sulawesi because it delivers a consistent and comparable scale-independent class structure for satellite image-based land-cover mapping and monitoring. The results of the comparative analysis of land-cover and land-cover change document the inhomogeneity, inconsistency and hence high uncertainty of existing estimates. The outcome of the harmonized and generalized land-cover products for two base years (1992 and 2000) indicates considerable disagreements in area estimates and spatial distributions of land-cover classes for a single date that in some cases exceed the detectable changes between years. Future work aiming at a long-scale operational land-cover mapping of tropical environments has to account for (a) a further harmonization of existing and planned land-cover definitions and products, (b) the regional validation of products and (c) the implementation of a multi-level standardized technical and conceptual classification workflow for ecosystem mapping and monitoring in tropical regions.