Proceedings of the 2nd Workshop of the EARSeL Special Interest Group on Remote Sensing of Land-Use & Land-Cover, 28th-30th September 2006, Bonn, Germany [online publication] Tobias Kuemmerle, Patrick Hostert, Theresa Schiller, and Volker C. Radeloff
Land use is an important driver of global environmental change and has resulted in wide-spread degradation and loss of ecosystem structures and services (Foley et al. 2005). Monitoring land-cover change and assessing its drivers is therefore of great international concern (Gutman et al. 2004; Lambin and Geist 2006), but the understanding of how people influence land change is still far from complete (Rindfuss et al. 2004). Remote sensing is the most important tool to provide information on where land changes occur (Lambin and Geist 2006). Traditional methods to quantify land change from multitemporal remote sensing images often rely on classifying the image into discrete land cover classes, or alternatively into change classes (Lu et al. 2004). Although these methods are useful to assess land cover conversions such as deforestation or urbanization, they potentially overlook changes in within-class heterogeneity (Coppin et al. 2004). This is unfortunate, because modifications of land cover are widespread. Consequently, there is an urgent need for developing robust and repeatable change detection methods that rely on continuous rather than discrete data, and thus allow for monitoring land cover modifications (Southworth et al. 2004).
Following the Fall of the Iron Curtain in 1989, Eastern European countries transitioned from planning economies to market oriented systems. This transition has drastically affected land management and land use decisions, and resulted in widespread land cover changes, such as the abandonment of farmland (Peterson and Aunap 1998), and changes in forest cover (Augustyn 2004; Bicik et al. 2001). Moreover, the transition has triggered modifications of land cover and changes in landscape pattern, particularly concerning farmland. Before 1990, most of Eastern Europe’s farmland was managed by the state in large-scale agricultural co-operatives. Since 1990, all Eastern European countries have implemented land reforms to break up the large-scale farming structures and to privatize the agricultural sector (Lerman et al. 2004). These land reforms, in combination with inheritance practices and the underlying ownership pattern, resulted in a split-up of the large socialistic fields into smaller parcels, and led to the physical fragmentation (hereafter called parcelization) of farmland in many areas (Sabates-Wheeler 2002; van Dijk 2003). Land reform strategies differed strongly among Eastern European countries. However, not much is known about the extent and spatial pattern of post-socialist parcelization and it remains largely unclear how different land ownership structures and land reforms affected the parcelization of farmland in Eastern Europe.
We selected the border triangle of Poland, Slovakia, and Ukraine because all three countries had different land ownership patterns and land management policies in socialist times (Table A-1), which in turn led to different land reform strategies after the system change (Augustyn 2004; Lerman et al. 2004). Moreover, the region was part of the Austro-Hungarian Empire for a period of around 150 years before 1914 with relatively homogeneous land management (Turnock 2002). Differences in farmland parcelization among countries are therefore likely due to either socialist or post-socialist land management (Kuemmerle et al. 2006), making the area particularly well suited to study the effects of land reforms on parcelization.
Table A1: Land ownership of agricultural land and privatization strategies of the countries in the study area (Lerman et al. 2004).
Country |
Land Ownership |
Potential |
Privatization |
Land market |
Poland |
Private and state owned |
All |
Sell state land (plots) |
Buy/sell, lease |
Slovakia |
Collectivized (cooperatives) |
All |
Restitution (plots) |
Buy/sell, lease |
Ukraine |
State owned |
All |
Distribution (shares) |
Only lease until 2005 |
Monitoring and quantifying parcelization is challenging, because statistical or cadastral data often do not exist, or data are of limited or unknown liability (Filer and Hanousek 2002). Using remote sensing images is promising because consistent data from before and after 1990 exist, but studying parcelization requires the quantification of changes in the structural pattern within farmland. Image texture measures are interesting to address this challenge, because texture measures capture the spatial and structural arrangement of image objects by quantifying the spatial variability of grey levels within a local neighborhood (Haralick et al. 1973). As such texture measures can be used to characterize heterogeneity within land cover classes (St-Louis et al. 2006), and may be well suited to quantify the parcelization of farmland in Eastern Europe. In summary, we were interested in assessing the extent and spatial pattern of post-socialist farmland parcelization in the border triangle of Poland, Slovakia and Ukraine. Our specific objectives were to:
Our analysis was based on three Landsat images, representing spring, summer, and early autumn, for each of the two time periods 1985-88 (before the system change) and 2000 (10 years after the system change). We used 4 Landsat TM images (30th April 1985, 2nd October 1986, 27th July 1988, and 21st August 2000) and 2 Landsat ETM+ images (6th June 2000, and 30th September 2000) from path 186 and row 26. All images were co-registered and corrected for relief displacement using a semi-automatic method (Hill and Mehl 2003) and the SRTM digital elevation model as a base map (Kuemmerle et al. 2006a). The TM and ETM+ data were atmospherically corrected using calibration coefficients and a modified 5S radiative transfer model that incorporated a terrain dependent illumination correction (Hill and Mehl 2003).
Forests were masked out using unsupervised clustering (see Kuemmerle et al. 2007). The three images for each time period were stacked and transformed into principal components to emphasize phenological differences between the images, to enhance the signal to noise ratio, and to reduce storage space and computation time. We retained principal component 1-8 and carried out image segmentation on each image stack separately using a region-growing algorithm (Baatz and Schäpe 2000). Texture measures were calculated for each segment and we gathered a set of representative samples for the two classes “high parcelization”, and “low parcelization” based on field visits and very-high resolution data (3 IKONOS images and 14 Quickbird images were available for these purposes). The segmented images were then classified using texture measures and the maximum likelihood classifier to derive parcelization maps for the two time periods. We used post-classification comparison of these parcelization maps to delineate a change map and summarized parcelization changes for our study area.
Agricultural parcelization differed markedly among the countries Poland, Slovakia, and Ukraine in socialist times, likely due to different ownership patterns that in turn led to different land reforms. In Poland, much of the farmland was privately owned even before 1990, resulting in smaller farm sizes and parcels, and thus in a high share of highly parcelized land (Figure A-1). Slovakia was dominated by large parcels, because all agricultural land was managed in large-scale co-operatives. In Ukraine, a heterogeneous pattern of large-scale and fine-scale agriculture was observed in the lowland areas. The mountain valleys showed a highly parcelized farmland pattern already before 1990 (Figure A-1), because population density is high in these areas and many people depend on subsistence farming (Augustyn 2004; Turnock 2002).
Figure A1: Changes in the parcelization of farmland in the border region of Poland, Slovakia, and Ukraine (UTM reference system with WGS84 datum and ellipsoid). | ||
Concerning changes in parcelization in the post-socialist period, Poland did not experience much change, because ownership did not change substantially (Lerman et al. 2004), except for some areas close to the border to Slovakia where state farms managed all land. On the other hand, the land use pattern changed considerably in Ukraine and Slovakia. In Slovakia, the share of large parcels was still very high in 2000, suggesting that the large-scale agricultural enterprises still managed most of the land, although cooperatives were transformed into private enterprises (Lerman et al. 2004). Most of the farmland was restituted to former owners; yet, these owners often chose to lease their land back to the cooperatives (Lerman et al. 2004). Some parcelization occurred, mainly close to settlements in the south of the study area. In Ukraine, where land was distributed among the workers of the cooperatives much of the land became parcelized in post-socialist times, in particular farmland close to settlements, and especially close to the cities of Uzhgorod, Mukacheve and Sambir (Figure A-1). Decreases in parcelization occurred mainly in mountain valleys and are possibly connected to the abandonment of farmland due to outmigration from these regions.
This research demonstrated that image texture can be a useful tool to map farmland parcelization and to quantify modifications within land cover classes. Our results showed distinct differences in parcelization among the three countries in our study area. In Poland, not much has changed, because private land ownership existed before 1990. Where land reforms were implemented, they led to marked changes in parcelization. Changes were strongest in Ukraine, where land was distributed among the people, whereas restitution in Slovakia partly preserved the large-scale farming sector because many owners leased their land to agricultural enterprises. Further research is required to quantitatively link parcel size and texture measures, and to validate parcelization changes based on high-resolution remote sensing data, aerial images, or cadastral maps.
Augustyn, M. (2004). Anthropogenic changes in the environmental parameters of the Bieszczady Mountains. Biosphere Conservation, 6, 43-53.
Baatz, M., & Schäpe, A. (2000). Multiresolution Segmentation – an optimization approach for high quality multi-scale image segmentation. In Strobl, J., Blaschke, T. & Griesebner, G. (Eds.), Angewandte Geographische Informationsverarbeitung XII (pp. 12-23). Heidelberg: Wichmann Verlag
Bicik, I., Jelecek, L., & Stepanek, V. (2001). Land-use changes and their social driving forces in Czechia in the 19th and 20th centuries. Land Use Policy, 18, 65-73.
Coppin, P., Jonckheere, I., Nackaerts, K., Muys, B., & Lambin, E. (2004). Digital change detection methods in ecosystem monitoring: a review. International Journal of Remote Sensing, 25, 1565-1596.
Filer, R.K., & Hanousek, J. (2002). Data watch - Research data from transition economies. Journal of Economic Perspectives, 16, 225-240.
Foley, J.A., DeFries, R., Asner, G.P., Barford, C., Bonan, G., Carpenter, S.R., Chapin, F.S., Coe, M.T., Daily, G.C., Gibbs, H.K., Helkowski, J.H., Holloway, T., Howard, E.A., Kucharik, C.J., Monfreda, C., Patz, J.A., Prentice, I.C., Ramankutty, N., & Snyder, P.K. (2005). Global consequences of land use. Science, 309, 570-574.
Gutman, G., Janetos, A.C., Justice, C.O., Moran, E.F., Mustard, J.F., Rindfuss, R.R., Skole, D., Turner II, B.L., & Cochrane, M.A. (Eds.) (2004). Land Change Science. Dordrecht: Kluwer Academic Publishers.
Haralick, R.M., Shanmuga., K., & Dinstein, I. (1973). Textural features for image classification. IEEE Transactions on Systems, Man, and Cybernetics, SMC3, 610-621.
Hill, J., & Mehl, W. (2003). Geo- and radiometric pre-processing of multi- and hyperspectral data for the production of calibrated multi-annual time series. Photogrammetrie-Fernerkundung-Geoinformation (PFG), 2003, 7-14. In German.
Kuemmerle, T., Hostert, P., Perzanowski, K., & Radeloff, V.C. (2006). Cross-border comparison of land cover and landscape pattern in Eastern Europe using a hybrid classification technique. Remote Sensing of Environment, 103, 449-464.
Kuemmerle, T., Hostert, P., Radeloff, V.C., Perzanowski, K., & Kruhlov, I. (2007). Post-socialist forest disturbance in the Carpathian border region of Poland, Slovakia, and Ukraine. Ecological Applications, 17, 1279-1295.
Lambin, E.F., & Geist, H.J. (Eds.) (2006). Land Use and Land Cover Change. Local Processes and Global Impacts. Berlin, Heidelberg, New York: Springer Verlag.
Lerman, Z., Csaki, C., & Feder, G. (2004). Evolving farm structures and land-use patterns in former socialist countries. Quarterly Journal of International Agriculture, 43, 309-335.
Lu, D., Mausel, P., Brondizio, E., & Moran, E. (2004). Change detection techniques. International Journal of Remote Sensing, 25, 2365-2407.
Peterson, U., & Aunap, R. (1998). Changes in agricultural land use in Estonia in the 1990s detected with multitemporal landsat MSS imagery. Landscape and Urban Planning, 41, 193-201.
Rindfuss, R.R., Walsh, S.J., Turner, B.L., Fox, J., & Mishra, V. (2004). Developing a science of land change: challenges and methodological issues. Proceedings of the National Academy of Sciences of the United States of America, 101, 13976-13981.
Sabates-Wheeler, R. (2002). Consolidation initiatives after land reform: responses to multiple dimensions of land fragmentation in Eastern European agriculture. Journal of International Development, 14, 1005-1018.
Southworth, J., Munroe, D., & Nagendra, H. (2004). Land cover change and landscape fragmentation - comparing the utility of continuous and discrete analyses for a western Honduras region. Agriculture Ecosystems & Environment, 101, 185-205.
St-Louis, V., Pidgeon, A.M., Radeloff, V.C., Hawbaker, T.J., & Clayton, M.K. (2006). High-resolution image texture as a predictor of bird species richness. Remote Sensing of Environment, 105, 299-312.
Turnock, D. (2002). Ecoregion-based conservation in the Carpathians and the land-use implications. Land Use Policy, 19, 47-63.
van Dijk, T. (2003). Scenarios of Central European land fragmentation. Land Use Policy, 20, 149-158.
© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme. | ||
DiML DTD Version 4.0 | Zertifizierter Dokumentenserver der Humboldt-Universität zu Berlin | HTML-Version erstellt am: 17.04.2008 |