Is dynamic world a contender in global land-cover making race? A swift field assessment from Kastamonu, Turkiye

Abstract

Alterations in land-cover significantly influence global climate fluctuations. To utilize land resources rationally and sustainably, it is essential to identify the open-source remote sensing capabilities, the resulting products, and assess their geographical accuracies. This study conceptualized over Kastamonu province in northwestern T & uuml;rkiye, focused on comparing three of the high-resolution (10 m) land-cover products; Environmental Systems Research Institute (ESRI) 2022, European Space Agency (ESA) World-Cover 2021 and Google-The World Resources Institute, Dynamic Word (DW) 2022, and 2022 Google Earth imagery were utilized for spatial comparisons. The overall accuracy (OA) and Kappa coefficient were computed, along with additional accuracy assessment metrics. OAs of land-cover maps (local accuracy), from highest to lowest, were ESRI2022; 76 %, ESA2021; 75.8 % and DW2022; 73.4 %. The Kappa coefficients for the three land-cover maps were calculated as 0.703 (very good) for ESA2021 and 0.69 and 0.68 (very good) for ESRI2022 and DW2022, respectively. The maximum user accuracy value was recorded at 92.23 % for the crops and 92.21 % for the built area classes in ESA2021. A comparison was also conducted among the corresponding class definitions. The most exemplary portrayal was observed in the categories of water, trees, and crops. Consequently, ESRI, ESA, and DW datasets were found to be fairly comparable to one another and can serve as auxiliary data in research pertaining to water, forestry and cultivated land resources.