Change in agriculturally used land and related habitat loss: a case study in eastern Estonia over 50 years/pollumajanduslikult kasutatud maa muutused 50 aasta valtel ja sellega seotud elupaikade kadumine Ida-Eesti naitel.

INTRODUCTION

Land use changes are affected by very different processes: intensification and extensification of agricultural practices, marginalization and abandonment of agricultural land, development of infrastructure, pressure from tourism (Calvo-Iglesias et al., 2006). Various complex social, economic, and political factors implicate land use changes (Nikodemus et al., 2005). Many studies concern the results of the abandonment of arable land (i.e. the agricultural land under different crops) and the reforestation process (Burgi & Turner, 2002; Calvo-Iglesias et al., 2006; Gellrich & Zimmermann, 2007; Tasser et al., 2007).

The natural overgrowing and forestation of open habitat areas are among the primary processes that caused landscape change in Europe during the 20th century. The rationalization of agriculture in the middle of the 20th century led to the intensification of land use, but on the other hand also to the abandonment of traditional agricultural land and to natural forestation (Strijker, 2005; Henle et al., 2008). The area of forests has increased in Europe by 10% in the last 40 years, the proportion of agricultural land has fallen by 11 %, and the area of permanent pastures has decreased by 11 % (European Environmental Agency, 1995). For example, in the last 50 years semi-natural communities have largely disappeared in Sweden due to the development of economically oriented agricultural landscape. More recent changes include a general reduction in agricultural land and in grassland and an increase in conifer plantations (Ihse, 1996). Changes in landscape structure have been shown to have a clear impact on the biodiversity of insects (Sepp et al., 2001; Schneider & Fry, 2005).

Several changes occurred in Estonian landscape in the 20th century, but the general trend is reduction in agricultural land and increase in the area of forests. At the beginning of the 20th century an intense change of forest land to agricultural land took place (Jogiste et al., 2005). The proportion of forest land had by then decreased to less than 30%. However, between 1918 and 2000, agricultural land area shrank from 65% to 30% (Peterson & Aunap, 1998; Mander & Reintam, 2001). In the same period, the area of forest land increased from 21% to 52%. All this happened as a result of the combined effects of land reform, population deportations, collectivization of agriculture, and urbanization. Changes in the political system, of which there were four in Estonia during the 20th century, also had an effect (Palang et al., 1998). One of the major changes in land use occurred after World War II, when the process of forestation of abandoned agricultural lands began. In addition to the natural successional process, forest was also planted on former fields (Jogiste et al., 2005). However, the relative importance of planting and natural forestation process has not been recorded.

The most recent processes in the Estonian landscape concerning the agricultural land use are forestation, abandonment and subsequent overgrowing of grasslands with shrubs and young trees, expansion of old-fields at the expense of abandoned arable land, and reduction in the proportion of arable land (Aaviksoo & Meiner, 2001). Peterson & Pussa (2001) estimated that 15-30% of the former arable land was lying abandoned between 1985 and 2000. Between 1995 and 2005 the area of arable land in Estonia fell by 283 000 ha, but the area of managed grasslands increased by 126 000 ha. Until 1999 the area of arable land increased, but since 2000 it has begun to fall again (Statistics Estonia, 2006).

When grasslands are abandoned and become forested, their vascular plant species richness generally declines. The reduction in species richness is associated with the length of time after abandonment. The speed of decrease in the total number of species might depend also on tree cover (Pykala et al., 2005). However, the abandonment and following succession of grasslands are still more preferable than intensifying their management with the help of fertilizers, pesticides, etc. (Jongman, 1996; Strijker, 2005).

Vegetation maps are good sources of information for investigating land use changes. Geo-botanical mapping of Estonian vegetation began in 1934. It was concluded after World War II, from 1948 to 1955. Vegetation units were defined and mapping took place across the whole of Estonia (Laasimer, 1965). Information was gathered on the species composition and dominants of the vegetation units as well as on environmental conditions (relief, soils, hydrological regime). The information is sufficient for classifying the different vegetation units as species rich or species poor.

In this study we analyse the changes in the agricultural land. We aim at giving the spatial persistence and extent of the agriculturally utilized land where management change has taken place. We describe the extent of the agricultural land created since the 1950s. We asked about the contribution of natural and semi- natural habitats in the creation of agricultural land. We will also ascertain the extent of the land use changes regarding species-rich and species-poor communities (forests, grasslands, fens) by comparing the map of vegetation from the middle of the 20th century with the map depicting land use in 2002. In addition, our goal was to assess the spatial extent of forestation processes embracing both natural forestation and forest planting.

MATERIAL AND METHODS

Study area

The study area is located in eastern Estonia, in Jogeva County, mostly in Saare Municipality but partly also in Kasepaa and Pala municipalities. The soils are less fertile than in neighbouring areas. The main crops of the arable lands are cereals (wheat, barley, rye) and in recent years, also oilseed rape. The area was defined on the basis of one grid square (128.2 km2) of the vegetation map drawn up at the beginning of the 1950s. The size of the study area was 11.1 km x 11.6 km. The area includes two major settlements (Voore and Saare), the former in the south (341 residents) and the latter in the south-west (206), and ten dispersed villages, each with 50-70 residents (Valla luhitutvustus. http://www.saarew.ee/index.php?lang=est&main id =93,95 (visited 10 Jan. 2007)).

Data

Vegetation map

A topographic map of 1:42 000 scale dating from 1896-1917, but later referred to by the single year of 1900 (Palang et al., 1998), was used as the mapping base of the vegetation map of the 1950s. The vegetation units within a grid square (128.2 km2) were always mapped and described by the same recorder; thus no bias in the data is expected. The classification of vegetation used for mapping relied on the classification by Lippmaa (1937).

For mapping the vegetation, all the different vegetation units were traversed and identified in the field and their plant communities were described. The vegetation units were demarcated on the map during the fieldwork. The number of replications of vegetation analyses per vegetation unit was related to its area: one vegetation analysis plot per each 0.25 [km.sup.2] of a vegetation unit was described. The plot size varied between open habitats and forest-like biotopes. A vegetation plot encompassed a 100 [m.sup.2] area of the tree and shrub layer in forests and wooded meadows. The plot size used for the grass layer in forest and in meadows varied from 4 [m.sup.2] to 10 [m.sup.2]. Species composition was determined by recording all species and their abundance on a five point scale.

Because of the scale of the map used for vegetation mapping, vegetation units do not have very clearly defined borders. However, this applies to the whole study and thus the error is the same all over the study.

Classification of arable lands and grasslands on the vegetation map of the 1950s

The agricultural lands consist of two differently managed types: arable lands and grasslands. The arable lands are used for cultivating various crops, while the grasslands produce forage for livestock. However, in the vegetation mapping the (natural or) semi-natural grasslands and cultivated grasslands were not separated. The latter are usually fertilized and cultivated with heycrops in Estonia. Compared to semi-natural grasslands, the cultivated ones have much lower species diversity (Sammul et al., 2003).

Arable lands are classified on the vegetation map according to their soil types. For this purpose soil samples were taken from each arable land and the soil profile and acidity level were recorded. This made it possible to determine what kind of forest or fen type could have grown in the place of the arable land earlier. In the study area arable lands of four types of origin were distinguished: those created from (1) mixed spruce, spruce, and fresh boreo-nemoral deciduous forests; (2) pine forests on sandy soil; (3) calcium-poor mires; and (4) pine forests on poor paludified soil.

Meadows and wooded meadows were grouped into six vegetation unit types: (1) species-rich meadows, (2) species-poor meadows, (3) dry species-rich wooded meadows, (4) dry species-poor wooded meadows, (5) paludified species-rich wooded meadows, and (6) paludified species-poor wooded meadows.

Estonian Basic Map

The source of information on the current land use in Estonia is the Estonian Basic Map, which in digital form is at a scale of 1:10 000. The Basic Map was constructed from aerial photographs. We used the Basic Map to compare the current land use with that depicted on the vegetation map from the 1950s. In our study area the aerial photographs were taken in 2002. The Basic Map provides information about the location and size of the following land-use types: grasslands (including cultivated and natural grasslands, which cannot be distinguished), fens, forests, young forests, arable lands, scrubland, and other open areas. Young forests are defined as areas covered with young trees (<4 m) whose cover is more than 30%. The same rule applies to clearings (clearings are defined as young forest). For assessing the spatial extent of forestation we used both of the land use types of the Estonian Basic Map: forest and young forest. Scrubland is defined as an area with at least 50% coverage of shrubs. The coverage of trees in the scrubland must be less than 30%, otherwise it classifies as forest (Estonian Land Board, 2003).

Map analysis

To compare the current and historical land use, the vegetation map of the study area was digitalized. All analysis of the map was performed using Mapinfo (7.5, Mapinfo Corporation), mainly by map layer querying for different vegetation units. The queries were used to delimit land use types within the study area and fmd their aggregate areas. The subsequent work included the analysis of qualitative and quantitative changes. The overlaps of areas of every historical vegetation type with the current land use type layers on the Basic Map were determined. The areas of the overlaps were identified and the percentage change for every vegetation unit was calculated. It is important to point out that the mapping scales of the two maps are different. This affects the results but the impact is the same all over the area.

Management changes in arable land and meadows

The arable land and meadows were divided into four categories according to data of the 1950s: arable land on fertile soil, arable land on infertile soil, species-rich meadows, and species-poor meadows. The change of land use in these four categories by 2002 was described and the proportions of land use types in 2002 were compared using [chi square]-test.

RESULTS

General changes in land use

It appears that in the 1950s arable land accounted for almost half of the land (Fig. 1). Various grasslands made up over a tenth of the studied area. By 2002 the share of arable land had fallen by 20.7% and that of grasslands by 3.5% (Fig. 2). Over the 50 years the forest coverage had grown by almost a half. Of the land that was forested in the 1950s 90.4% was forest land at the same place in 2002. Of the arable land 46.4% had remained arable land and 11% of the grassland was grass-land in 2002.

Changes in agriculturally used land

The percentage coincidence of the historical and current land use types distinguished on the vegetation map and the Basic Map is presented in Table 1. In all, 46.4% of the total area of the historical arable land had retained its former agricultural function, i.e. had remained arable land or had been converted to grassland. About 200%, depending on the type of arable land, had overgrown with forest, with young forest making up 3-8%.

The changes in species-rich and species-poor grasslands and fertile and infertile arable land are shown in Table 1. Comparison of the mean values of transition of the four groups revealed that the changes were not random ([chi square]-test, p < 0.001). Of fertile and infertile arable land quite equal parts had remained arable land or turned into forest. Major differences were observed in changes into grassland. The proportion of arable land that had changed to grassland was larger on infertile soil than on fertile soil. More of species-rich grassland had turned into forest compared to species-poor grassland. The changes in agricultural land use are depicted in Fig. 3, which shows the proportion of grassland and arable land.

Creation of new agricultural land

Arable land

New arable land, totalling 526.5 ha, had been created almost everywhere in the area. These areas are mostly near villages. Some historical fen areas were arable land in 2002. In all cases, fens had been taken into use partially (primarily fen margins) and as a result of amelioration. Of species-rich vegetation types grasslands and of species-poor types fens gave the largest percentage of the new arable land (Table 2). The changing of different habitats to arable land was not random (p < 0.01).

Grassland

Another type of agricultural land the change of which we were interested in was grassland. During the about 50 years under study, 359.6 ha of grassland was formed. It includes both natural and cultivated grasslands. The creation of grasslands was affected in part by human settlement and the growth of villages. Larger grasslands were most often created near settlements and close to major roads. Table 3 demonstrates that grasslands were created almost equally at the expense of species-rich and species-poor communities (forests and fens). If we consider the total area of the respective vegetation types, then species-rich communities form a greater part, especially in the case of fens. A significant difference (p < 0.001) is visible in the changing of different types of forests and fens to grassland. Table 3 shows that a large proportion of the grassland was formed from fens. We can find new agricultural land all over the study area (Fig. 4).

[FIGURE 3 OMITTED]

Forestation processes

Analysis of the new forested areas reveals that many of them are areas located next to existing forest (Fig. 5). At the same time, there are areas where forest appeared in the middle of a field. The less fertile areas were mostly situated near roads and were relatively large pieces of land. Forestation had occurred primarily at some distance from houses. Out of species-poor meadows and wooded meadows, those located in riversides had partly become forested by 2002, especially in the northern part of the study area, away from larger human settlements. Small patches of young forests had developed evenly over the whole study area. In many cases logged areas had become forested.

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

Forestation had taken place on 669.9 ha of species-rich grasslands (incl. wooded meadows) and on 373.8 ha of species-poor ones. In addition, 1504.5 ha of fertile arable land and 442.6 ha of less fertile arable land had become forested. Analysis of different arable land and grassland types changed into forest indicated that the proportions of different types were not random (p < 0.001).

DISCUSSION

Forestation is the main process that caused changes in the agriculturally used land in the study area. Even a brief look at the maps of land use patterns reveals that the forest coverage significantly increased over 50 years. The proportion of forest almost doubled in the area. Tullus (2000) and Mander & Reintam (2001) observed the same trend for the whole country.

In the study area there are numerous arable fields where natural overgrowing can be observed. As the number of overgrowing arable fields is quite large and these did not show up yet as young forest on the 2002 Basic Map, the proportion of young forest will increase. Moreover, overgrowing has been hastened by the dissolution of a farming cooperative in 2000. The fields abandoned at the end of the 1990s and at the beginning of the new millennium need to be further investigated. The incidence and degree of overgrowing should be ascertained in order to study the processes in recent history. Many areas have ceased to be managed because there is no longer an economic need for that.

The forestation process is not so intensive near the larger settlements as on the land between the smaller settlements where the arable lands and grasslands are no longer managed and have begun to overgrow. In addition, the afforestation of arable land received financial support in 2005 and 2006 (from the Estonian Agricultural Registers and Information Board), as a result of which less fertile arable lands were planted with forest trees. Some effect of this support mechanism is already discernible, but its spatial extent and distribution need further recording.

Positive effects of the abandonment and natural forestation of arable land are considered to be the stabilization of soils, carbon sequestration, and a temporary increase of biodiversity (Gellrich & Zimmermann, 2007). However, as a result of the expansion of forest area, a great loss of species-rich grasslands was observed on the area. No managed wooded meadows had preserved in the study area: by 2002 all of them had become overgrown. Abandoned grasslands that begin to overgrow with shrubs become poorer in species than traditionally managed grasslands (Jongman, 1996; Ockinger et al., 2006; Gellrich & Zimmermann, 2007). Hansson & Fogelfors (2000) reported the change of wooded meadows to woodland in 15 years. An Estonian study recorded the onset of shrub overgrowth 4-6 years after the cessation of agricultural production (Lauringson et al., 2000).

The area of grasslands decreased only by 3.5%. However, only 11% of the grassland was located at the same place as 50 years ago. This shows that the area of valuable permanent semi-natural grasslands has decreased considerably, and probably so has the species richness. In this study we described only two time slots, and due to scarcity of data we could not analyse changes in detail in between these two. The selection of land for creating new agricultural land had a rather clear trend. The border areas of forests and fens were taken into use, and sites close to villages and major roads were often preferred. Fens were the most suffered habitat type, their area considerably decreased due to creating new arable land. Therefore further analysis is needed to see if the abundance of different species inhabiting fens has declined. As biodiversity is a very important topic in the EU, in the Action Plan (Commission of the European Communities, 2006) "Halting the loss of biodiversity by 2010 - and beyond" different actions are recommended.

ACKNOWLEDGEMENTS

This work was funded by grant No. P6062PKPK06 from Estonian University of Life Sciences and target funding SF0170052s08. We would like to thank Marek Sammul, Kalevi Kull, and Kaire Lanno for their helpful comments, and Ilmar Part for language revision. We would also like to thank Petr Petrik and one anonymous reviewer.

Received 17 December 2007, in revised form 14 Apri12008

REFERENCES

Aaviksoo, K. & Meiner, A. 2001. Satellite monitoring of Estonian landscape. In Development of European Landscapes. TALE European Conference Proceedings. Yol. I. Publ. Inst. Geogr. Univ. Tartu., 92 (hander, U., Pritsmann, A. & Palang, H., eds), pp. 233-238. Institute of Geography, University of Tartu, Tartu.

Burgi, M. & Turner, M. G. 2002. Factors and processes shaping land cover and land cover changes along the Wisconsin River. Ecosystems, 5, 184-201.

Calvo-Iglesias, M. S., Fra-Paleo, U., Crecente-Maseda, R. & Diaz-Varela, R. A. 2006. Directions of change in land cover and landscape patterns from 1957 to 2000 in agricultural landscapes in NW Spain. Environ. Manage., 38, 921-933.

Commission of the European Communities. 2006. Halting the loss of biodiversity by 2010--and beyond. Sustaining ecosystem services for human well-being. COM (2006) 216. http://eurlex.europa.eu/LexUriServ/site/en/com/2006/com2006_0216enOl.pdf (visited 10 Apr. 2008).

Estonian Land Board. 2003. The Guide for Digital Mapping of Estonian Basic Map. Eesti pohikaardi 1:10 000 digitaalkaardistuse juhend. http://www.maaametee/index.php?fang id=1&page id=15&menu_id=36 (visited 30 May 2007).

European Environmental Agency. 1995. Dobris assessment. Copenhagen. http://reports.eea.europa.eu/92-826-5409-5/en/ (visited 7 Apr. 2008).

Gellrich, M. & Zimmermann, N. E. 2007. Agricultural land abandonment and natural forest re-growth in Swiss mountains: A spatially explicity economic analysis. Agric. Ecosyst. Environ., 79, 65-76.

Hansson, M. & Fogelfors, H. 2000. Management of a semi-natural grassland; results from a 15-year-old experiment in southern Sweden. J. Yeg. Sci.,11, 31-38.

Henle, K., Alard, D., Clitherow, J., Cobb, P., Firbank, L., Kull, T., McCracken, D., Moritz, R. F. A., Niemela, J., Rebane, M., Wascher, D., Watt, A. & Young, J. 2008. Identifying and managing the conflicts between agriculture and biodiversity conservation in Europe--a review. Agric. Ecosyst. Environ., 124, 60-71.

Ihse, M. 1996. Monitoring cultural landscape in Sweden - methods and data for changes in landuse and biotopes. In Ecological and Landscape Consequences of Land Use Change in Europe. Proceedings of the First ECNC Seminar on Land Use Change and Its Ecological Consequences, 1(x18 Febr. 1995 Tilburg, The Netherlands (Jongman, R. H. G., ed.), pp. 103-129. Ramona, Brussels.

Jogiste, K., Vares, A., Uri, V. & Tullas, H. 2005. Baltic afforestation. In Restoration of Boreal and Temperate Forests (Stanturf, J. A. & Madsen, P., eds), pp. 225-234. CRC Press/Lewis Publishers, Boca Raton.

Jongman, R. H. G. 1996. Introduction. In Ecological and Landscape Consequences of Land Use Change in Europe. Proceedings of the First ECNC Seminar on Land Use Change and Its Ecological Consequences, 1(x18 Febr. 1995 Tilburg, The Netherlands (Jongman, R. H. G., ed.), pp. 9-15. Ramona, Brussels.

Laasimer, L. 1965. Eesti NSYtaimkate. Valgus, Tallinn. Lauringson, E., Kull, T., Vipper, H. & Talgre, L. 2000. Mahajaetud pollud biosusteemis. Agraarteadus, 11, 232 248.

Lippmaa, T. 1937. Eesti vegetatsioonikaardi koostamise alused ning senise too tulemusi. Eesti Loodus, 2, 592.

Mander, U. & Reintam, L. 2001. Development of Estonian landscapes. In Development of European Landscapes. TALE European Conference Proceedings. Yol. I. Publ. Inst. Geogr. Univ. Tartu., 92 (Mander, U., Pritsmann, A. & Palang, H., eds), pp. 25-31. Institute of Geography, University of Tartu, Tartu.

Nikodemus, O., Bell, S., Grine, I. & Liepins, I. 2005. The impact of economic, social and political factors on the landscape structure of the Vidzeme Uplands in Latvia. Landscape Urban Plan., 70, 577.

Ockinger, E., Eriksson, A. K. & Smith, H. G. 2006. Effects of grasslands abandonment, restoration and management on butterflies and vascular plants. Biol. Conserv., 133, 291-300.

Palang, H., Mander, U. & Luud, A. 1998. Landscape diversity changes in Estonia. Landscape Urban Plan., 41, 163-169.

Peterson, U. & Aunap, R. 1998. Changes in agricultural land use in Estonia in the 1990s detected with multitemporal Landsat MSS imagery. Landscape Urban Plan., 41, 193-201.

Peterson, U. & Pussa, K. 2001. Forest area and agricultural land use change in the Eastern Baltic region. In Development of European Landscapes. TALE European Conference Proceedings. Yol. I. Publ. Inst. Geogr. Univ. Tartu., 92 (Mander, U., Pritsmann, A. & Palang, H., eds), pp. 301-303. Institute of Geography, University of Tartu, Tartu.

Pykala, J., Luoto, M., Heikkinen, R. K. & Kontula, T. 2005. Plant species richness and persistence of rare plants in abandoned semi-natural grasslands in northern Europe. Basic Appl. Ecol., 6, 25-33.

Sammul, M., Kull, K. & Tamm, A. 2003. Clonal growth in a species-rich grassland: results of a 20-year fertilization experiment. Folia Geobot., 38, 1-20.

Schneider, C. & Fry, G. 2005. Estimating the consequences of land-use changes on butterfly diversity in a marginal agricultural landscape in Sweden. J. Nature Conserv., 13(4), 247-256.

Sepp, K., Mand, M. & Truu, J. 2001. The influence of landscape pattern on bumble bee communities. In Development of European Landscapes. TALE European Conference Proceedings. Yol. I. Publ. Inst. Geogr. Univ. Tartu., 92 (Mander, U., Pritsmann, A. & Palang, H., eds), pp. 89-92. Institute of Geography, University of Tartu, Tartu.

Statistics Estonia. 2006. Environment 2005 Yearbook. Statistikaamet, Tallinn.

Strijker, D. 2005. Marginal lands in Europe - causes of decline. Basic Appl. Ecol., 6, 99-106.

Tasser, E., Walde, J., Tappeiner, U., Teutsch, A. & Noggler, W. 2007. Land-use changes and natural reforestation in the Eastern Central Alps. Agricult. Ecosyst. Environ., 118, 115-129.

Tulles, H. 2000. Pollumajandusmaade metsastamine ja metsastumine. Agraarteadus, 11, 22-27.

Silja Kana ([email]), Tuu Kull, and Merit Otsus

Institute of Agricultural and Environmental Sciences, Estonian University of Life Sciences, Rua mnt. 181, 51014 Tartu, Estonia

([email]) Corresponding author, [e-mail]

Fig. 1. Land use in the 1950s.

Arable land   46.4%
Grassland     13.3%
Forest        28.3%
Other           12%

Note: Table made from pie chart.

Fig. 2 Land use in 2002.

Forest        55.4%
Scrubland      1.1%
Other          8.1%
Arable land   25.7%
Grassland      9.8%

Note: Table made from pie chart.

Table 1. Territorial transition of the land use of the 1950s
(four types and their subtypes) into the land use types in 2002 (%)

                                      Vegetation types in 2002

                                      Arable            Young
Vegetation types in 1950s              land    Forest   forest

1. Arable land (fertile soil)          47.9     27.6      3.0
2. Arable land (infertile soil)        41.0     35.0      7.1
    Arable land created from           40.9     38.5      4.9
      species-poor fen
    Arable land on sandy soil          40.8     33.3      7.8
    Arable land on peaty soil          44.6     17.4     26.2
3. Species-rich grasslands             10.0     68.5      8.9
    Species-rich open grasslands        2.6     47.1     18.6
    Dry species-rich wooded meadows    17.3     61.9      5.6
    Paludified species-rich wooded      5.4     79.2      9.5
      meadows
4. Species-poor grasslands              8.5     50.3      2.5
    Species-poor open grasslands       10.0     50.1      2.8
    Dry species-poor wooded meadows     7.5     48.6      2.3
    Paludified species-poor wooded      4.1     59.8      1.4
      meadows

                                         Vegetation types in 2002

                                                  Minero-
                                                  trophic
Vegetation types in 1950s             Grassland    swamp    Scrubland

1. Arable land (fertile soil)            0.9       12.7        1.2
2. Arable land (infertile soil)          9.6        1.2        0.6
    Arable land created from             7.9        2.5        0.6
      species-poor fen
    Arable land on sandy soil           10.6        0.4        0.6
    Arable land on peaty soil           11.4        0.0        0.0
3. Species-rich grasslands               7.6        0.4        0.9
    Species-rich open grasslands        21.9        0.9        0.8
    Dry species-rich wooded meadows      8.5        0.2        1.7
    Paludified species-rich wooded       3.5        0.4        0.2
      meadows
4. Species-poor grasslands              16.0        0.2        6.3
    Species-poor open grasslands        18.2        0.5        2.7
    Dry species-poor wooded meadows     13.6        0.0       12.5
    Paludified species-poor wooded      13.8        0.0        0.0
      meadows

                                        Vegetation types in 2002

                                      Other      Roads
                                       open       and
Vegetation types in 1950s              area    buildings   p-Value

1. Arable land (fertile soil)          3.6        3.2      <0.001
2. Arable land (infertile soil)        4.5         1       <0.001
    Arable land created from           4.0        0.7
      species-poor fen
    Arable land on sandy soil          4.9        1.5
    Arable land on peaty soil          0.3       0.05
3. Species-rich grasslands             2.9        0.8      <0.001
    Species-rich open grasslands       7.3        0.9
    Dry species-rich wooded meadows    3.3        1.6
    Paludified species-rich wooded     1.5        0.3
      meadows
4. Species-poor grasslands             13.7       2.5      <0.001
    Species-poor open grasslands       14.1       1.6
    Dry species-poor wooded meadows    13.8       1.8
    Paludified species-poor wooded     11.3       9.5
      meadows

Table 2. Conversion of vegetation types into arable land: change in
relation to the total area of new  arable land and in relation to the
total area of the vegetation type (%)

                           Percentage of total   Percentage of total
                           area of new arable    area of vegetation
Vegetation type in 1950s          land                  type

Species-rich types:
  Fen                              5.8                  22
  Grassland                       16.5                  10
  Forest                           2.7                   6
Species-poor types:
  Fen                             17.4                  14.4
  Grassland                       11.4                   8.5
  Forest                          15.2                   3.9

Table 3. Conversion of vegetation types into grassland: change in
relation to the total area of new  grassland and in relation to the
total area of the vegetation type (%)

                              Percentage of      Percentage of total
                            total area of new    area of vegetation
Vegetation type in 1950s        grassland               type

Species-rich:
  Forest                          12.1                   2.8
  Fen                             18.2                  29.7
Species-poor:
  Forest                          14.2                   2.5
  Fen                             20.9                  11.8