Comparing the validation results of the Markovian land cover model with the refined and roughened Modis data Winda Wijayasari (a), Faizal Immaddudin Wira Rohmat (b), Sparisoma Viridi (c*)
a) Department of Computational Science, Faculty of Mathematic and natural Science, Bandung Institute of Technology, Jl. Ganesa No. 10 Bandung 40132.
b) Water Resources Development Center, Bandung Institute of Technology, Jl. Ganesa No. 10 Bandung 40132
c) Department of Physics, Faculty of Mathematic and natural Science, Bandung Institute of Technology, Jl. Ganesa No. 10 Bandung 40132
Abstract
Land cover modeling is essential to design sustainable development, especially in developing countries like Indonesia. The land cover model using LANSAT satellite data required expertise in manual land cover classification processes. MODIS is a product of the Terra satellite launched by NASA and specially made using the land cover algorithm to minimize errors generated due to errors in the land classification process. Available MODIS resolutions are 250m, 500m, and 1 km. This study will discuss the effect of coarsening and refining the resolution of MODIS data on land cover modeling results. The method used to analyze land cover is the Markov chain. Smoothing of the grid size does not significantly affect the error of the land cover. However, roughing the grid increases the error of the land cover modeling results. Thus, the error value of the land cover model using MODIS data depends on the resolution limit.
Keywords: MODIS- Land cover- Markov Chain- Upper Citarum Watershed
