Assessing the suitability of GlobeLand30 for land cover mapping and sustainable development in Malaysia using error matrix and unbiased area Estimation.

Global Land Cover (GLC) maps have been extensively used as background data for numerous applications such as land-use planning, climate change monitoring, and urban planning. While countries in Europe and North America have greatly developed land cover maps, several developing countries in Asia and Africa lack such vital data sources. GlobeLand30's 30 m resolution dataset has the potential to meet the mapping needs of these developing regions. However, differences in accuracy of the dataset between countries can be significant making it unsuitable to adopt regional accuracy directly for assessing individual country's accuracy level. Therefore, it is critical to develop reliable localized assessments. This research assesses the accuracy of GlobeLand30 for Malaysia using multiple evaluation approaches. Seven classes of land cover data were evaluated using 197 samples in the country. An overall accuracy of 63.45% and 65.70% was obtained from the Error Matrix using sample counts and the new unbiased area estimator respectively. We conclude that the overall accuracy of GlobeLand30 in tropical developing countries like Malaysia is generally lower than the accuracy in developed countries and arid developing countries thereby requiring significant improvement for optimal use. However, despite the errors due largely to misclassification, the GlobeLand 30 data has the potential to complement other LULC data to estimate and monitor vital land-use changes in Malaysia and the ASEAN region in general for sustainable development. [ABSTRACT FROM AUTHOR]