02156naa a2200217 a 450000100080000000500110000800800410001910000170006024501460007726000090022352014400023265000240167265000290169665000280172565000210175370000190177470000170179370000210181070000190183177300880185011359952025-10-16       2025    bl uuuu     u00u1 u    #d1 aSILVA, D. A.  aFilling the gap between pasture and forestsbclassifying initial-stage forests in the subtropics using satellite data.h[electronic resource]  c2025  aForests cover an important portion of terrestrial surface and much of its cover is formed by secondary forests in different stages of succession. Classifying and mapping the succession stages of these forests is fundamental for environmental planning, estimating carbon stocks and biodiversity, and monitoring forest dynamics; however, it is still a challenge for both ground and remote sensing surveys. We aim to find the best sensor and band combination that can appropriately differentiate intermediate and advanced stage forests from initial stage forests and pastures in Santa Catarina, Brazil. For this end, we used Operational Land Imager (OLI) and Multi-Spectral Instrument (MSI) data in combination with ground-truth data from the FlorestaSC Program to test the difference of spectral response for each land-cover class in each band and spectral index, the separability of the classes using Bhattacharyya Distance and to classify the land-cover over ∼4 km2 of land. We found that most bands and spectral indexes have different responses to the land-cover classes, and separability varied greatly between sensors and land covers. However, the classification was successful with overall accuracies above 88%. MSI data and spectral indexes using red-edge bands had a better performance differentiating the three classes in more fragmented and anthropized landscapes (overall accuracy 93.17 % versus 88.52 % using OLI data).  aland use/land cover  amultispectral instrument  aoperational land imager  asecondary forest1 aVIBRANS, A. C.1 aBOLLMANN, G.1 aNICOLETTI, A. L.1 aLINGNER, D. V.  tJournal of Applied Remote Sensing, Bellingham, WAgv. 19, n. 3, p. 034502-1-, 2025.