Assessing contemporary global land cover products to study land cover change in Poland

The availability of global land cover products has been increasing recently, however, its regional quality differs substantially. In this work, the contemporary global datasets including Google’s Dynamic World (GDW), ESA’s World Cover (ESA WC), Esri Land Cover (ELC) map were compared in Poland to assess their usefulness for the forest cover change studies in Central-European conditions. We grouped land cover categories of the three global datasets into 6 land cover classes (forest, semi-natural vegetation, cropland, built-up area, water and other). To assess the accuracy of the global products we used the EU Land Use/Cover Area Frame Survey (LUCAS) points. The results showed that precision for forest class was higher for ESA WC (0.92) than for the other two products (ELC: 0.85; GDW 0.41). Forest class accuracy was the highest for ELC (0.92) and lower for ESA WC (0.90) and GDW (0.81). Overall, ELC had the highest F1 score for the forest class (0.76), slightly higher than ESA WC (0.71), with GDW showing a significantly lower value of 0.57. Our analysis indicated that for forested areas ELC performed better than two other global products, suggesting its usefulness for forest cover change studies in Central Europe.

Acknowledgements:

This research was funded in whole or in part by the National Science Centre, Poland (UMO-2024/53/N/ST10/02518). For the purpose of Open Access, the author has applied a CC-BY public copyright license to any Author Accepted Manuscript (AAM) version arising from this submission.