Monitoring farmland abandonment by multitemporal and multisensor remote sensing imagery
Final Report Abstract
The aim of the project was the development of adequate strategies to monitor post-Soviet land use changes, specifically farmland abandonment, using multitemporal SAR and multispectral remote sensing data. Eastern Europe is a prime example for drastic broad-scale land use changes resulting from the fundamental political and socio-economic changes after the collapse of the Soviet Union. Generally, our investigations showed different major land use trends after the end of the USSR with its widespread large-scale cropland. On the one hand there was the massive farmland abandonment in our study area in western Ukraine. On the other hand, subsistence agriculture emerged. The mapping framework is based on so-called "land management regimes", referring to (1) intensively managed large-scale cropland, (2) small-scale cropland as subsistence agriculture with a potentially low management intensity, and (3) farmland abandonment with no active management. To map the land management regimes, we acquired time series of Landsat and ERS-SAR satellite data and performed an extensive field campaign in western Ukraine to collect in-situ data. We developed a semi-automatic parameter selection to optimize image segmentation, which is a prerequisite for object-based image analysis. In order to select the optimal segmentation parameters, the user selects a range of parameters and the algorithm automatically performs the segmentations and subsequently assesses the best segmentation with regard to the classification accuracy. Second, we integrated optical and SAR data into an object-based image analysis, based on the approach that is mentioned above. Both the synergistic effects of the multisensoral data as well as the object/spatial information in addition to the spectral information of the images provided elementary information for improving the mapping of the land management regimes. Third, with the use of change trajectory analysis and the use of a long ERS SAR time series, we monitored the land management regimes between 1986 and 2010. Our results clearly showed substantial abandonment both in the 1990s and 2000s. More recently, with the integration of Ukraine in the world markets and the emergence of agri-business at the end of the 2000s, noticeable recultivation of abandoned farmland took place. Since the end of the planned-economy in 1991, small-scale cropland emerged. Mostly, it was directly converted from large-scale cropland. In order to compare the spatial patterns of the land management regimes with spatial indicators related to the marginality of farming, we overlayed our final results with additional data such as soil type, distance to markets, or elevation. All in all, we found a considerable potential for recultivation, although parts of the abandoned farmland were on soils not very well suited for agriculture.
Publications
- (2013). Optimization of object-based image analysis with Random Forests for land cover mapping. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, Vol. 6(6), 2492–2504
Stefanski, J., Mack, B., & Waske, B.
- (2014). Mapping and monitoring of land use changes in post-Soviet western Ukraine using remote sensing data. Applied Geography, Vol. 55, 155–164
Stefanski, J., Chaskovskyy, O., & Waske, B.
(See online at https://doi.org/10.1016/j.apgeog.2014.08.003) - (2014). Mapping land management regimes in western Ukraine using optical and SAR data. Remote Sensing, Vol. 6(6), 5279–5305
Stefanski, J., Kuemmerle, T., Chaskovskyy, O., Griffiths, P., Havryluk, V., Knorn, J., Korol, N., Sieber, A., & Waske, B.
(See online at https://doi.org/10.3390/rs6065279) - (2015): Monitoring land use dynamics with optical and radar remote sensing data in western Ukraine. Dissertation, Freie Universität Berlin
Stefanski, J.
