Environmental Modelling and Software 173 (2024) 105956 Available online 10 January 2024 1364-8152/© 2024 Elsevier Ltd. All rights reserved. Appraisal of EnMAP hyperspectral imagery use in LULC mapping when combined with machine learning pixel-based classifers Christina Lekka , George P. Petropoulos * , Spyridon E. Detsikas Department of Geography, Harokopio University of Athens, El. Venizelou 70, Kallithea, 17671, Athens, Greece A R T I C L E INFO Handling Editor: Daniel P Ames Keywords: Hyperspectral remote sensing Imaging spectroscopy EnMAP LULC Machine learning ABSTRACT The recent availability of satellite hyperspectral imaging combined with the developments in the classifcation techniques have paved the way towards improving our ability to obtain information on the spatiotemporal distribution of land use/land cover (LULC) at improved accuracy. In this context, the present study aims at evaluating the combined use of the recently launched Environmental Mapping and Analysis Program (EnMAP) hyperspectral satellite mission with two powerful machine learning (ML) classifers. In particular, the Support Vector Machines (SVM) and Random Forest (RF) are used synergistically with EnMAP imagery in performing LULC mapping at a typical Mediterranean setting located in northern Greece. Evaluation of the derived LULC maps is based on the computation of a series classifcation accuracy metrics. The McNemar’s chi-square statis- tical signifcance testing was also computed to confrm the statistical signifcance of the differences between the classifers. In overall, results showed that SVM slightly outperformed RF, exhibiting a higher overall accuracy, of 92.6% and 88.1%, respectively, whereas the statistical signifcance of the fndings was also attested by the McNemar’s statistical test results. To our knowledge, this study is one of the frst published so far focusing on exploring the capabilities of ENMAP imagery when combined with different ML pixel-based classifers in the context of LULC mapping. Our results, indeed provided useful insights on the potential of EnMAP datasets in deriving information on the spatiotemporal distribution of LULC at a highly fragmented Mediterranean land- scape, evidencing the EnMAP promising potential in this feld. 1. Introduction There is a growing awareness of the rapid alterations occurring in the world’s natural resources, resulting in signifcant implications for eco- systems, biodiversity, and human society on a global scale (FAO, 2022). The rapid rate of change particularly in land use and land cover (LULC) is driven by factors such as climate change, unregulated population growth and urban sprawl, land degradation and abandonment. In Mediterranean areas in particular, the signifcance of LULC changes and spatial distribution becomes even more pronounced (Guida et al., 2022). The Mediterranean region is ecologically diverse and vulnerable to environmental changes and experiences unique challenges in the face of climate-induced alterations. The abandonment-induced degradation can have severe repercussions on ecosystems and local communities (Nadal-Romero et al., 2023). Monitoring LULC is essential to better understand landscape dynamics and its transformation in LULC and addressing the complex challenges faced by Mediterranean regions (Elatawneh et al., 2012; Fragou et al., 2020). Such information plays a crucial role for guiding land use practices and developing sustainable adaptive strategies to mitigate the adverse effects of physical and human factor changes (Singh et al., 2020) and supports policy-makers at stra- tegic planning efforts in achieving environmental sustainability goals (Estoque, 2020; Giuliani et al., 2020; Andries et al., 2022). The rapidly growing advances in the feld of Earth Observation (EO), entailed a vast range of cutting-edge remote sensing platforms which enable for better understanding of the Earths’ surface features and their spatial distribution (Winkler et al., 2021; Zhao et al., 2022). In recent years, advancements particularly in hyperspectral EO technology have made it possible to perform LULC over large areas and achieve fne-grained land surface components discrimination with high accuracy (Pandey et al., 2020; Moharram and Sundaram, 2023). For numerous applications, researchers have benefted from the spectrally rich infor- mation found in hyperspectral data. For example, Vangi et al. (2020) assessed spectral discriminability of forest types in Italy using the new hyperspectral satellite PRISMA compared to Sentinel-2 MSI, allowing better discrimination across forest types. Later, Mzid et al. (2021) * Corresponding author. E-mail address: gpetropoulos@hua.gr (G.P. Petropoulos). Contents lists available at ScienceDirect Environmental Modelling and Software journal homepage: www.elsevier.com/locate/envsoft https://doi.org/10.1016/j.envsoft.2024.105956 Received 4 December 2023; Received in revised form 9 January 2024; Accepted 9 January 2024