↓99 |
The first overarching goal of this thesis was to advance the understanding of broad-scale land use determinants by studying the natural experiment that occurred in Eastern Europe after the system change in 1990. The political, economic, and societal transition resulted in fundamental changes in the framework of underlying driving factors of land use decisions. Studying land cover dynamics in the post-socialist period therefore allowed for assessing of how such changes became manifested in landscapes. Cross-border comparisons were a useful approach to uncouple the effect of driving factors that changed similarly in all countries (e.g., economic conditions) from country-specific factors (e.g., different land reform strategies). Studying the border region of Poland, Slovakia, and Ukraine in the Carpathians, was particularly interesting, because it allowed for comparing the effects of the three different ownership patterns and land reform strategies that occurred in Eastern Europe after 1990.
↓100 |
Research question I: Did the Polish, Slovak, and Ukrainian regions of the study area differ from each other concerning land use and landscape patterns?
Chapter II clearly showed that the three countries took diverse pathways after the demise of the Austro-Hungarian Empire, resulting in marked differences in land cover and landscape patterns. Concerning forests, this became apparent in differences in forest cover (i.e., highest in Poland), the altitudinal distribution of forests (e.g., considerably lower forest cover in Ukraine at higher altitudes), forest composition (e.g., relatively high share of conifer plantations in Ukraine), and forest fragmentation (i.e., lowest in Poland). Agriculture was most abundant and less fragmented in Slovakia compared to Poland and Ukraine. In 2000, grassland dominated Polish and Slovak mountain valleys, whereas Ukrainian mountain valleys had a considerable share of cropland. Given the countries common history and the environmentally relatively homogeneous background of the region, these differences are most likely attributable to either socialist or post-socialist land use changes.
Research question II: What were the changes in land use in the post-socialist period and did land use change differ among the three countries in the study area?
↓101 |
Chapters III, IV, and V investigated the question whether post-socialist land use changes resulted in converging or diverging trends in terms of land cover and landscape pattern when comparing among the three countries in the study area. Land cover change was widespread between 1988 and 2000, and the forest change analysis, the farmland abandonment map, and the quantification of land use pattern clearly showed that trends differed considerably among the three countries. Overall, forest cover changes were relatively moderate in Poland and Slovakia, especially when including afforested areas that were most widespread in Slovak mountain valleys. In Ukraine, however, forest cover decreased, forests today are considerably more fragmented than in the late 1980s, and protected areas were less effective compared to Poland and Slovakia.
Concerning agriculture, the main result was widespread abandonment of farmland in all countries. Yet, the rates and spatial patterns of abandonment differed markedly, both among countries and when comparing different regions within countries (e.g., different altitudinal zones or land tenure regimes). Farmland abandonment was most widespread in Slovakia and on former state land in Poland, whereas relatively low abandonment rates were found in areas where subsistence farming dominated (e.g., Ukrainian mountain valleys, or areas where private farms dominated in Poland). These areas were also characterized by a highly heterogeneous landscape pattern with small fields, whereas the plains in Slovakia and Ukraine as well as some Polish mountain valleys were characterized by large fields. Generally, small fields indicating parcelization tended to occur in the vicinity of settlements, especially around larger cities (e.g., Uzhgorod, Mukacheve; see also Appendix A) and this zone of small-scale farming was often followed by a ring of abandoned farmland. Areas further away from larger settlements tended to become abandoned where population density was relatively low (e.g. Poland and Slovakia) and parcelized where many people depend on subsistence farming (e.g., Ukraine).
Overall, Poland and Slovakia showed a converging trend in the post-socialist period, characterized by already high or increasing forest cover in the foothill and mountainous zone, farmland abandonment in mountain valleys, and relatively low abandonment rates in the plains. In contrast, Ukraine clearly diverged from these two countries in terms of land cover and landscape pattern after the system change. Forest cover decreased, abandonment was relatively widespread at all altitudes, and human pressure in mountain valleys was considerable. This trend will likely amplify further in the future, because Poland and Slovakia are now members of the European Union with a uniform framework of land management policies and environmental standards.
↓102 |
What can be learned from the research in this thesis about the role of broad-scale political and socio-economic driving forces of land use change? Assessing post-socialist land use change in the border region of Poland, Slovakia, and Ukraine did not only emphasize the pivotal role of such driving forces for land use decisions, but also provided compelling evidence that widespread land use change is triggered where these boundary conditions change. Furthermore, this study showed that abrupt changes in driving forces immediately translate in rapid land use change. For example, economic depression, weakened institutions, and lower level of control explain the uniform pattern of increased forest harvesting in the first years of the transition period. Also, the rapidly decreasing profitability of agriculture under free-market conditions along with general population trends in Eastern Europe (i.e., migration of young people away from rural areas) resulted in the widespread abandonment of farmland, particularly were farming conditions are marginal.
Even more importantly, the cross-border comparisons carried out in this research allowed for separating out the effect of specific driving forces of land use change, particularly the role of changes in land ownership and land reforms. Land use change was more widespread where ownership patterns changed drastically. For instance, all other factors being equal, farmland abandonment rates were twice as high on former state land in Poland compared to areas that had always been in private ownership. Moreover, this study also underpins the importance of tenure stability (Geist et al., 2006). Farmland abandonment was lowest where land tenure was stable (i.e., private land in Poland) and highest where land tenure was insecure, for example where former owners were difficult to locate (e.g., Slovakia). Concerning forests, the findings in this study strongly support the assumption that neither state forestry (as in Poland and Ukraine) nor private forestry (as partly in Slovakia) are clearly better in lowering harvest rates and in guarding forest ecosystems. Rather the strength of institutions and the pace at which they are reformed is important, and good institutions tend to result in stable or increasing forest cover (Dietz et al., 2003;Tucker and Ostrom, 2005).
Different land reforms resulted in markedly different outcomes concerning land use change. Restitution led to widespread abandonment in marginal areas (e.g., Slovak mountain valleys), because land disinterest among former owners was substantial. On the other hand, restitution practically preserved the large-scale socialist farming structure in areas where favorable farming conditions prevailed (seeCsaki et al., 2003). Auctioning of farmland (i.e., former state-owned land in Poland) resulted in extensive set-aside areas, because most farmland was bought for speculative purposes. In contrast, abandonment rates were lower and parcelization was high where land was distributed, particularly in areas where people’s livelihoods depend on subsistence agriculture (e.g., in Ukrainian mountain valleys).
↓103 |
The uniform environmental setting and the common history of the study area were important boundary conditions for comparing land use change across borders. Generalizing post-socialist land use trends observed in the border triangle of Poland, Slovakia, and Ukraine to areas outside the Carpathians or to the country level should be carried out with care. While much of the land use changes (and its driving forces) observed in the Polish, Slovak, or Ukrainian region of the study area may still be typical at the country level, a Carpathian study area can not account for the heterogeneity of environmental, societal, and economic conditions within a country.
The second goal of this thesis was to assess the fate of Carpathian ecosystems in the post-socialist period. Overall, human pressure has considerably decreased after the system change and many areas in the study area are essentially undergoing a process of rewilding. Farmland abandonment and land use extensification in rural areas provide opportunities for afforestation and increased carbon sequestration, and forest species may benefit from recent land use changes, particularly area demanding top herbivores and carnivores. However, farmland abandonment will likely decrease Carpathian biodiversity in the long run as landscapes characterized by low-intensity land use in mountain valleys are diminishing (Baur et al., 2006). Moreover, increased fragmentation of mature forests is of growing concern, particularly in the Ukrainian Carpathians where illegal logging is coupled to corruption, similar to other regions of the world (Buksha et al., 2003;Geist et al., 2006;WWF, 2007).
The findings of this study underpin the essential role of broad-scale underlying driving factors on local land use decisions. Because changes in the framework of broad-scale land use determinants affect many land managers, resulting land use changes are widespread and may have a strong effect on earth system functioning when aggregated to regional or global scales. For example, farmland abandonment was widespread in all three countries and probably occurred at similar rates in other areas of Eastern Europe and the former Soviet Union, too (Peterson and Aunap, 1998;Ioffe and Nefedova, 2004;Baur et al., 2006;Müller and Munroe, 2007). This study showed that much of these lands will revert back to forests, and this may have a profound effect on regional carbon balances. Broad-scale boundary conditions for land use decisions are likely equally important in other parts of the world where institutional and socio-economic change occurs more slowly. Interpreting the institutional and socioeconomic transition that occurred in Eastern Europe as a natural experiment provided useful insights into the relative importance of some of these boundary conditions. Such insights are urgently needed to guide decision makers in designing a policy framework that balances trade-offs between immediate human needs and the long-term capacity of the earth system to provide humanity with multiple ecosystem services (Foley et al., 2005;Bennett and Balvanera, 2007;Kareiva et al., 2007).
↓104 |
Several interesting research issues for follow-up research beyond the scope of this work evolved during the course of this thesis.
Assessing post-socialist land use change in the Romania part of the Carpathians would significantly broaden the picture of how Carpathian ecosystems changed since 1990. An initial case study suggested that farmland abandonment may be equally widespread in this part of the Carpathians, whereas forest disturbance was overall relatively low (Kuemmerle et al., 2007;Müller et al., 2007). Further research is certainly needed to gain a better understanding of land use trends in this part of the Carpathians, too. Similarly, studying post-socialist land use change for the Carpathians as a whole at sufficiently detailed scales would allow for improved assessments of habitat fragmentation and connectivity, and therefore for better assessing the consequences of land use changes for Carpathian wildlife and biodiversity. Moreover, an area-wide study of the Carpathians would facilitate comparisons of land change mapped from remote sensing images with official logging statistics that are often only available at highly aggregate levels. The discrepancy between these two data sources to assess forest change was one of the surprising results of this study and may be an important indicator of illegal logging in the region.
This study focused on quantifying land cover and landscape pattern change, and on qualitatively linking observed changes to its underlying drivers. A quantitative assessment of the underlying causes of land use change can give useful insights (see for exampleMüller and Sikor, 2006;Müller et al., 2007). Yet gathering the necessary datasets for statistical model building in a border region with three countries would have been beyond the scope of this work and should be subject to future research. Moreover, it would also be interesting to extend observed land use trends into the future and to develop alternative land use scenarios for the Carpathians under consideration of different policy and economic environments (Verburg et al., 2006;Westhoek et al., 2006).
↓105 |
On a technical level, a few issues show potential for further investigations. First, Riitters’ (2002) indices were useful to compare forest fragmentation among countries, but interpreting their ecological importance is challenging. Advancing Riitters concept based on morphological image processing is a promising research direction and should allow for a better linkage of landscape pattern and ecological processes (Vogt et al., 2007;Vogt et al., 2007). Second, the disturbance index concept (Healey et al., 2005) could be extended by developing a method to derive thresholds separating disturbed from undisturbed forests. Also, testing the sensitivity of the disturbance index to the initial normalization step will be important. Both issues would certainly break a path for the more widespread use of this method (Healey et al., 2005). Third, support vector machines have great potential for becoming a standard method for complex multitemporal classification problems. Their strongest advantage is the relatively small number of trainings samples required, while still being able to handle complex class distributions. Testing the SVM approach for different problems and evaluating the minimum number of training samples needed to derive robust classifications would greatly improve the applicability of this method. Last, this thesis showed the usefulness of image texture for mapping changes in land use pattern and field size. Initial tests indicated that texture may be image inherent, and that transferring statistical models among images and time periods may be difficult. Combining the quantitative modeling approach presented in Chapter V and the segmentation-based classification detailed in Appendix A may be a promising way of making use of the potential of both approaches.
Generally, this study showed the great potential of studying natural experiments to better understand the drivers of land use change. In Eastern Europe, the next natural experiment has already occurred with the accession of 10 former socialist countries to the European Union. These countries now comply to extensive environmental regulations, adopt new land management policies, and benefit from the agri-environment schemes of the European Union. Studying how Eastern Europe’s landscapes change under such fundamentally changing boundary conditions for land use decisions will be interesting. For example, much of the abandoned or set-aside land may be put back into production to receive subsidies. At the same time, increased urbanization and migrations from East to West may lead to increased abandonment rates. Assessing this natural experiment will give invaluable insights into the drivers of local land use decisions, particularly when comparing socialist centralized economies, the free-market period, and EU market conditions. This may ultimately contribute to both, an improved understanding of the coupled human-environment system, and a more detailed picture of ecosystem dynamics in this understudied region.
© 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 |