Forest Area Dynamics on the Eastern Border of Latvia (1967–2015)

FOREST AREA DYNAMICS ON THE EASTERN BORDER OF LATVIA (1967–2015)

Zigmārs Rendenieks

Rural landscapes of Latvia have experienced multiple shocks during the 20^th century – two World wars, deportations of rural inhabitants, forced collectivization. These events impacted land use patterns, facilitating farmland abandonment, which in most cases resulted in natural reforestation of former agricultural lands. Unique data from the declassified Corona spy satellite provided high-resolution images (up to 2.5m) and land cover information from 1967 and Landsat 5 and Landsat 8 images for 1989 and 2015, respectively. We used photogrammetric orthorectification for Corona images and, after mosaicking of all images we performed image segmentation and assigned forest and non-forest classes to image objects. To determine changes in land cover, we produced two change maps depicting stable forest, stable non-forest, forest gain and forest loss areas for 1967–1989 and 1989–2015. Our classification accuracy reached 90% and 93%, respectively. We found that during the Soviet rule (1967–1989) adjusted forest area increased by 20.1%, but in the following period (1989–2015) by 7.2% of the studied area. The main conclusion from these results is that land use change during the Soviet era needs to be studied more extensively since most of land cover changes happened during that time.  Agricultural lands on the eastern border of Latvia are characterized as marginal for farming and have a higher risk of abandonment compared to the rest of the country. We also found that the overall increase in forested areas resulted from de-fragmentation of forests by increasing core areas of forest tracts and merging of smaller forest patches. However, the ecological consequences of such changes remain unclear. The use of Corona images enable the extension of high-resolution time series in land cover change research and provides the opportunity for more comprehensive analysis of land use change processes and of the forces driving it in past and the future.