21846_3_Kivinen.indd Regional distribution and biodiversity perspectives of Finnish grasslands SONJA KIVINEN Kivinen, Sonja (2005). Regional distribution and biodiversity perspectives of Finnish grasslands. Fennia 183: 1, pp. 37–56. Helsinki. ISSN 0015-0010. The preservation of viable agricultural landscapes is essential in Finland and throughout Europe, because a considerable proportion of threatened species are associated with rural habitats, particularly various grasslands. In this study, the regional distribution of Finnish grasslands was studied using an existing nation- wide digital land-use database. Further, the occurrence of traditional rural bi- otopes and selected butterfly and bird species living in different grassland habi- tats was investigated in relation to agricultural landscape mosaic. The results showed that grasslands were typically abundant in versatile agricultural regions, land uplift shores as well as in agriculturally marginal landscapes experiencing decrease of agricultural activities. Grasslands were generally scarce in inten- sively cultivated agricultural landscapes. The occurrence and siting of traditional rural biotopes in grassland networks vary greatly regionally, abundant and well- connected traditional rural biotopes being found e.g. in river valleys and few isolated biotopes in intensively cultivated or agriculturally marginal, forested regions. Furthermore, the nationwide occurrence of selected butterfly and bird species was positively related to the abundance of grasslands. Overall, grass- lands enrich the Finnish agricultural landscape both at the habitat and species level. In addition to species-rich semi-natural grasslands, other low-productive grasslands also hold significant potential in the maintenance of diverse agricul- tural environments. Sonja Kivinen, Finnish Environment Institute, Research Programme for Biodiver- sity, PO Box 140, FI-00251 Helsinki, Finland. E-mail: sonja.kivinen@ymparisto. fi. MS submitted 05 November 2004. Introduction Agricultural practices and land use have changed rapidly during recent decades throughout Europe. Intensification and marginalization introduce im- balance to the agriculture-environment relation- ship and diminishing diversity in agricultural envi- ronments has been widely reported (Matson et al. 1997; Krebs et al. 1999; Robinson & Sutherland 2002). Species diversity is threatened mainly by habitat loss, increasing isolation of remaining hab- itat fragments and decline of habitat quality (e.g. Waldhart 2003). Species-rich semi-natural grasslands were wide- spread in Europe at the time of low-intensity farm- ing systems. These non-cultivated communities result from forest clearances and their persistence requires human management by grazing, mowing or burning (Briggs & Courtney 1989). Natural dis- turbances, such as fires and floods also creating grassland habitats in the forested landscape, are nowadays largely suppressed by humans (Pykälä 2000). Today, only remnants of semi-natural grass- lands are found in Europe owing to fertilization, cultivation, abandonment and afforestation of these habitats. Artificially planted and fertilized cultivated grasslands have largely replaced semi- natural grasslands in agricultural landscapes (Alanen 1997; Pärtel et al. 1999; Ihse & Lindahl 2000; Poschlod & WallisDeVries 2002). Because a significant part of European biodi- versity is associated with semi-natural grasslands, their rapid decline is recognized as a serious threat to high numbers of rare or decreasing plant and animal species (Norderhaug et al. 2000; Luoto et al. 2001; Söderström et al. 2001; Kiviniemi & Eriks- 38 FENNIA 183: 1 (2005)Sonja Kivinen son 2002; van Swaay 2002; Cousins et al. 2003; Duelli & Obrist 2003). In Finland, one quarter of all known species live in agricultural environments (Pykälä & Lappalainen 1998). A considerable pro- portion, 28% of endangered species live prima- rily in traditional rural biotopes or other cultural environments and 39% of disappeared species used primarily these habitats (Rassi et al. 2001). Traditional rural biotopes are habitats created by traditional practices of animal husbandry, includ- ing the most valuable species-rich grasslands in Finnish agricultural landscapes. The area of these semi-natural grasslands has decreased to less than 1% of that registered one century ago (Vainio et al. 2001). Nowadays it is also essential to identify other ru- ral habitats that may increase the capacity of agri- cultural landscapes to sustain biodiversity, in addi- tion to semi-natural grasslands (Krebs et al. 1999; Fahrig 2001; Robinson & Sutherland 2002; Benton et al. 2003). Different kinds of non-cropped habi- tats, such as old cultivated fields and pastures, set- aside land as well as field margins and grassy edge zones between fields and forests have been found to have significant value for plants (Corbet 1995; Freemark et al. 2002; Waldhart et al. 2003), butter- flies and other invertebrates (di Giulio et al. 2001; Lee et al. 2001; Steffan-Dewenter & Tscharntke 2001; Jeanneret et al. 2003b; Haysom et al. 2004; Pöyry et al. 2004; Pywell et al. 2004), birds (Pain et al. 1997; Vickery et al. 2001; Peach et al. 2004; Virkkala et al. 2004), and small mammals (de la Peña et al. 2003). These habitats can be consid- ered as an intermediate, biodiversity-supporting level between intensively managed, species-poor arable land and species-rich, but scarce semi-nat- ural grasslands. The quality of an individual grassland patch is mainly determined by its size, management his- tory and physical conditions (e.g. climate, bedrock and soil properties) as well as its relation to the sur- rounding landscape mosaic (Dauber et al. 2003; Opdam et al. 2003). Small and isolated patches are expected to have an increased risk of popula- tion extinction (MacArthur & Wilson 1967; Levins 1969; Hanski & Gilpin 1991). In other words, a comprehensive network of habitat patches of suf- ficient size enables species dispersal and reduces the local extinction risk (Saunders et al. 1991; Bender et al. 2003; Jordán et al. 2003). A prerequisite for monitoring and conservation of biodiversity is widespread and up-to-date infor- mation concerning the distribution of habitats and species. During recent decades new nationwide databases containing information of land use, land cover and species occurrences have been created both on the national and EU level (Mik- kola et al. 1999; FEI 2003; Weiers et al. 2004). These digital datasets provide new possibilities for biogeographical research and maintenance of di- versity. Here, the distribution of Finnish grasslands was studied and grassland patterns were further exam- ined from the biodiversity viewpoint by means of case studies. A database containing grassland habitats in 10 m resolution was created using dig- ital land use data and the nationwide abundance of grasslands was studied in 100 km2 grid squares. The occurrence of traditional rural biotopes in the regional grassland networks and their relation to general land cover was studied within six drainage basins. Using species atlas data recorded in the 100 km2 grid system, the relationships between the nationwide occurrences of two butterfly and four bird species and agricultural land use were examined. The main questions of the study were: (1) How are grassland habitats distributed in Finland and what factors are correlated with the observed pat- terns? (2) What is the occurrence and spatial loca- tion of traditional rural biotopes in regional grass- land networks in various agricultural landscapes? (3) Is there a spatial relationship between the oc- currences of the studied species and the abun- dance of grasslands and arable land? Background Characteristics of the Finnish landscape Although Finland is generally a low-lying region, the topography varies widely in a small scale. The main part of the relief results directly from the crystalline ancient bedrock (Seppälä 1986; Si- monen 1990). The most common deposit type is compact basal till that often forms a basis for other deposits. Hummocky moraines, eskers, deltas, and ice marginal and interlobate formations impart unique characteristics to the landscapes. Extensive fine-sediment deposits are found mainly in south- ern and western Finland (Fig. 1). Peat deposits are abundant particularly in western and northern Fin- land. Land uplift continually exposes new land in coastal areas, especially in the Bothnian Bay (Ku- jansuu & Niemelä 1990). FENNIA 183: 1 (2005) 39Regional distribution and biodiversity perspectives of Finnish… Fig. 1. a) The location of the study area in Finland (shaded in grey) and b) the distribu- tion of fine sediment soils (silt and clay). Finland belongs to the cold temperate climate zone with cool summers (Tuhkanen 1984). The mean annual temperature is highest in south- western Finland, decreasing towards the north- east. The length of the growing season (mean daily temperature > 5 °C) varies from 180 days in south-western Finland to 100 days in northern La- pland (Helminen 1987). Rainfall is rather evenly distributed throughout the seasons. The mean an- nual precipitation ranges from 400–450 mm along the western coast and northern Finland to 700– 750 mm in parts of southern and eastern Finland (Solantie 1987). Finland is located mainly in the boreal conifer- ous vegetation zone. Only the most south-western part of the country is situated in the hemi-boreal zone. The southern boreal zone covers southern and central Finland and parts of eastern Finland and the western coast. The middle boreal zone in- cludes most of western Finland and parts of east- ern Finland, whereas northern parts of Finland are mainly situated in the northern boreal zone (Ahti et al. 1968). Finnish landscape is largely dominated by forests and mires (Hämet-Ahti 1988; Ruuhijärvi 1988). Agricultural and horticultural land cover 7.3% of the total area of Finland (MAF 2004b). The history of grasslands in Finland The first signs of cultivation and livestock hus- bandry in Finland date back to the Neolithic Stone Age (5100−1500 B.C.) (Carpelan 1999). The signif- icance of agriculture increased during the Bronze Age (1500−500 B.C.) and slash-and-burn cultiva- tion was practiced over wide areas of southern and western Finland in the Iron Age (500 B.C.−1300 A.D.). Arable cultivation was started probably be- fore the end of Iron Age in south-western Finland (Huurre 1995; Vuorela 1999). The period of old traditional agriculture in Fin- land extended from the historical era until the 1870s (Soininen 1974). In the beginning of the 19th century, arable cultivation was practiced gen- erally in southern and western Finland, whereas most of eastern Finland was largely dependent on slash-and-burn cultivation. Agriculture was based on crop cultivation and livestock was kept prima- rily for manure to fertilize fields (Grotenfelt 1922; Soininen 1974). The availability of fodder during the long winter feeding period was a crucial fac- tor for cattle numbers and thus also for crop cul- tivation. Fodder was collected on natural mead- ows and the meadow area was increased through forest clearances and lake drainages. In summer- time, cattle grazed in forests and fallows and after mowing also in meadows (Linkola 1922; Soininen 1974). Cultivation and grazing opened up and diversi- fied landscape structure and created new habitat types, whereas before the establishment of agricul- ture the boreal landscape was altered mainly by natural disturbances, such as fires and storms. Dif- ferent physical conditions regulating agricultural intensity led to varying regional landscape mosa- 40 FENNIA 183: 1 (2005)Sonja Kivinen ics, ranging from more open agricultural regions to enclosed forested landscapes (Ihse 1995; Pykälä 2000). The traditional grazing system carried nu- trients from meadows to arable land and created unique nutrient-scarce semi-natural grasslands associated with high species richness and a great number of specialized species (Alanen 1997). A revolution in Finnish agriculture took place at the end of the 19th century. Resulting from the rapid population growth, continuous crop failures and the decreasing world market prices of cere- als, the basis of agriculture was changed from crop production to stock-raising. As a major renewal, hay-making in arable land became common and sustained higher cattle numbers (Soininen 1974). In addition to forest clearances, new areas for fod- der production were obtained by including the most productive meadows in the cultivation cycle. As a consequence, the area of semi-natural grass- lands started to decrease rapidly (Linkola 1922; Valle 1951; Oksanen 2001). The modernization of Finnish agriculture start- ing in the 1950s almost completely terminated traditional farming. Agricultural production in- tensified as a consequence of new machines, fertilizers and techniques. Agricultural policies fa- voured large farms specialized in crop cultivation, whereas traditional small holdings were paid com- pensation for fallowing and livestock slaughter. A marked change was the regional specialization of agriculture in southern and western Finland into crop production and in eastern Finland into dairy farming (Granberg 1989; Luostarinen 1997). The role of cultivated, artificially fertilized grasslands in dairy farming increased at the expense of semi- natural grasslands. Semi-natural grasslands situ- ated in the most favourable agricultural areas were converted to arable land, and meadows of poorer regions became wooded by natural succession, or were afforested (Alanen 1997; Pykälä 2001). In general, the total number of dairy cows and the area under hay and grasses has declined drasti- cally since the 1960’s (Tiainen 2004). Further, only the remnants of semi-natural grasslands created by traditional practices of animal husbandry are left in Finland (Vainio et al. 2001). During the time of Finland’s EU membership, the agri-environmen- tal support of EU has provided opportunities for biodiversity and landscape management on farms (Luoto et al. 2003b). Management of the remain- ing semi-natural grasslands is included in the na- tional EU agri-environment program. Moreover, furthering and management of other open non- cropped rural habitats are considered as important measures sustaining agricultural biodiversity (MAF 2004a). Material and methods Creation of the grassland database The grassland data was derived from the land use database of SLICES (Separated Land Use/Land Cover Information System) (Mikkola et al. 1999). SLICES is based on the joint use of geographic in- formation of several organizations and the source data for the land use classes has been derived from various existing land use registers. The raster for- mat SLICES land use database with 10 m resolution covering the whole of Finland was completed in 2000. The combination of various source datasets into a single database had required reclassification of data into SLICES land use classes, coordinate transformations and conversion of polygon format data into raster format (Mikkola et al. 1999). Over- lapping problems of the land use classes resulting from the disagreement of different source data sets had generally been resolved so that geometrically more accurate data overrode more inaccurate data and newer data overrode older data. The grassland database of this study was based on two SLICES classes, ‘perennial grasses and meadows’ and ‘long-term fallows’ derived from the Finnish Land Parcel Identification System and the Topographic Database, updated in 1998–1999. Perennial grasses and meadows consist of over five years old grasses, meadows and pastures used for agricultural purposes. Long-term fallows are classi- fied in the system as arable land that has not been cultivated or utilized in other ways for at least five years and has not become wooded. The grassland database was built using the ArcInfo 8.1 software (Esri 1992) and it covered mainland Finland and the inner archipelago south of 67° latitude (Fig. 1a). Northernmost regions and Åland were excluded from the analysis due to de- ficiencies in the data. In order to facilitate faster data processing, the data was generalized by elim- inating all grassland patches less than 0.1 hectares. The grassland database was compared to digital 1:20,000 base maps and distinctively misclassified patches, such as forest patches classified as grass- land were removed. In order to distinguish ecologically the most valuable semi-natural grasslands, an existing spa- FENNIA 183: 1 (2005) 41Regional distribution and biodiversity perspectives of Finnish… tial database containing traditional rural biotopes found in a nationwide inventory in 1992−1998 (Vainio et al. 2001) was used to complete the SLICES grassland data. Inventoried traditional ru- ral biotopes consist mainly of various meadows, wooded pastures and grazed forests. Wooded pas- tures are biotopes characterized by a mosaic of tree groups and meadow patches. Grazed forests, excluded here from the analysis, were differenti- ated from wooded pastures in the inventory on the basis of ground vegetation characterized by forest species and denser tree cover (>35%) (Pykälä et al. 1994). Traditional rural biotopes had been classified in the inventory on the basis of their recent status into nationally, regionally and locally valuable tradi- tional rural biotopes (Pykälä et al. 1994) and these value classes were used in this study. In general, nationally valuable biotopes have been managed continuously mainly by mowing or grazing at least for the last 50 years and have particularly diverse species assemblages, often including nationally endangered and rare plant species. Regionally val- uable biotopes are characterized by long-contin- ued traditional or near-traditional land use, typical vegetation created by traditional agricultural prac- tices and diverse species assemblage. In locally valuable biotopes, land use has ceased to be tra- ditional and typical vegetation for traditional rural biotopes occurs often in small patches. National, regional and local biotopes can also include non- managed, partially overgrown areas where diverse vegetation has still been preserved. In order to obtain a picture of the validity of SLICES grassland data, a field check of 401 grass- land patches (other than inventoried traditional rural biotopes) was carried out in the summer of 2004. Grassland patches were selected randomly within southern, western, eastern and northern parts of Finland. Grasslands were checked for ac- tual classification errors in SLICES data (for exam- ple misclassified forest and field areas). Moreover, the proportion of bush cover in grassland patches was estimated. Other datasets A spatial database of arable land was built similarly to the grassland database from SLICES land use data with 10 m resolution. The source data for arable land was based on the Finnish Land Parcel Identifi- cation System. Arable land is defined in the SLICES system as continually cultivated ground, including fields and not more than five years old fallows and cultivated grasslands (Mikkola et al. 1999). Further- more, general land cover information was derived from a nationwide CORINE land cover database in 100 m resolution. The CORINE land cover data is based on satellite image interpretation completed with digital databases (FEI 2003). Studied species The reflections of grassland and arable land abun- dances on species distributions were studied us- ing two butterfly species and four bird species. The species data were derived from the butterfly atlas (Huldén et al. 2000) and the bird atlas (Väisänen et al. 1998), in which the occurrence of both species groups were presented in 100 km2 squares. Spe- cies data were digitized using ArcView software (Version 3.2, Esri, Redlands, CA, USA). Both studied butterflies, Coenonympha pam- philus (Small Heath) and Lycaena hippothoe (Purple-edged Copper) are meadow species. C. pamphilus is a grassland generalist, living in all kinds of grassland areas, such as dry and mesic meadows, road verges and clear fellings. Larvae feed on grasses (Poaceae). L. hippothoe occurs in forest-sheltered, especially dry flower-rich mead- ows and also in river shore meadows and grassy slopes in the north. Larvae feed on Rumex acetosa and Rumex acetosella. Both species are common in southern and central Finland and occur more scarcely in northern Finland. Species abundances vary annually, but C. pamphilus appears to have declined especially in parts of western and eastern Finland and L. hippothoe in parts of southern and western Finland (Marttila et al. 1990). Acrocephalus palustris (Marsh Warbler) and Lo- custella naevia (Grasshopper Warbler) represent farmland bird species feeding and breeding in suc- cessional semi-open agricultural land. The habitat selection of A. palustris contains lush bush vegeta- tion in meadows, abandoned land and neglected gardens, and particularly moist brookside bushes. L. naevia breeds in meadows and abandoned ar- able land in bushes with dense grass undergrowth and is particularly abundant in cultivated areas located near coastal bays and lakes. A. palustris and L. naevia occur in southern and central Fin- land. L. naevia has bred in Finland already since the 19th century. A. palustris arrived in Finland in the mid-20th century from the south and south- east and has become more abundant and spread northwards ever since (Väisänen et al. 1998). 42 FENNIA 183: 1 (2005)Sonja Kivinen The occurrence of two bird species, Crex crex (Corn crake) and Perdix perdix (Gray Partridge) living in open agricultural land was also studied in relation to grassland abundance. C. crex lives in open hay fields, fallow fields and dry shore meadows in southern and central Finland. P. per- dix breeds in bushy and weedy edges of large field areas and midfield islets in southern Finland and along the western coast. Intensification of agricul- ture has led to rapid decline of the populations of both species throughout Europe. C. crex has been particularly affected by nest-destroying harvest ma- chines and the decline of meadows surrounding cereal fields. P. perdix is also affected by modern cropping practices as well as the by the decreased habitat quality of production fields (Väisänen et al. 1998; Meyer-Aurich et al. 2003). Both species have been classified as near-threatened species in Finland (Rassi et al. 2001). Spatial analyses on the national scale The spatial distribution of grasslands was studied in 100 km2 grid squares and within drainage basin units. Grid squares were created using the ArcIn- fo generate function. Drainage basin units were based on the first level partition of Finland’s water system areas (Ekholm 1993) and this GIS data was taken from a database of the Finnish Environment Institute. The proportion of grassland in grid squares and drainage basin units was calculated using the ArcInfo intersect function. The propor- tion of arable land derived from the SLICES data- base was also calculated in the grid squares. The amounts of grassland and arable land in 100 km2 squares were compared to the occur- rences of butterfly and bird species. The coordi- nates of land use grids coincided with the grids of animals observed. The quantity and quality of butterfly observations collected in different parts of Finland varied considerably and there were no proper estimates of the survey activity in atlas squares. Therefore only observations after the year 1988 and observation squares included in the But- terfly monitoring scheme (see Huldén et al. 2000) were taken into account in order to reduce inac- curacies of data. By contrast, survey activity and nesting probabilities in squares were available for the bird atlas data. Well and satisfactorily surveyed atlas squares and the squares with certain, prob- able, or possible nesting were chosen for analyses. The northern limits of species distributions were taken into account in the calculations. The spatial analyses were carried out using ArcInfo intersect functions and the statistical relationships between species and habitat distributions were analysed us- ing the SPSS-program (SPSS 2002). Spatial analyses on the regional scale The number, connectivity and type of traditional rural biotopes as well as land cover in four grass- land-rich and two grassland-poor drainage basins were studied. CORINE land cover data was reclas- sified as 1) grassland, 2) arable land, 3) urban area, 4) forest, 5) mire, 6) bare ground and 7) water and was further complemented by the SLICES grass- land database. Land cover proportions for drain- age basins were calculated using ArcInfo intersect functions. Landscape connectivity is determined by inter- actions between landscape spatial structure and the movement behaviour of species (e.g. the re- view of Tischendorf & Fahrig 2000). For connectiv- ity analyses, a total of 11 traditional rural biotopes (meadows and wooded pastures), including 1 nationally, 3 regionally and 7 locally valuable biotopes were selected within drainage basins. Nationally valuable biotopes were not found in grassland-poor drainage basins, and therefore the total number of biotopes was 10 in these regions. Connectivity for selected traditional rural biotopes was measured using a distance-weighting scheme, in which the amount of grassland area was cal- culated within a user-specified threshold distance from a biotope patch. Following Hanski (1994), connectivity S is given by the sum of contributions of all grassland patches: S i = ∑exp(–αd ij )A j where A j is the area of grassland (in ares), d ij is the distance between grassland patches i and j (in km*10). Connectivity was calculated for each tra- ditional rural biotope patch up to five kilometres from the centre point using the Si program (Moi- lanen 2000). Because no specific species was studied, the parameter α was given a value 1. The exponential function gives more weight for the patches, which are close to the focal patch. Results The validity of grassland data Field checks showed that classification of the SLICES data was relatively accurate. Out of 401 FENNIA 183: 1 (2005) 43Regional distribution and biodiversity perspectives of Finnish… grassland patches 1.7% turned out to be urban green cover, 1.6% grassy clear cuttings and 0.7% former grazed forests. Moreover, 2.7% of grassland patches carried planted young trees and 0.5% of patches were ploughed. Otherwise, field checked grassland patches consisted mainly of old, culti- vated grasslands and long-term fallows in accord- ance with SLICES classification. Over half of the visited grassland patches were open habitats. One third of the patches had bush cover up to ten percent and approximately one tenth of patches had bush cover between 11 and 60% of the total area. Bushes were often found by the side of ditches, whereas the core of the patch was typically open habitat. The characteristics of distribution patterns of grasslands The proportion of grassland varied noticeably in different parts of Finland, and several grassland belts could be distinguished in 100 km2 squares (Fig. 2a). Grasslands formed an almost continu- ous zone running from western Lapland through coastal areas to south-eastern Finland. Another wide grassland belt extended from northern Fin- land southwards along the eastern border. Further- more, a noticeable grassland zone covered parts of southern and central Finland, and a diagonal grassland belt extended from eastern Finland to the western coast. Fig. 2. The distribution of grasslands in Finland in 100 km2 squares. a) The amount of grassland in hectares. b) The mean size of grassland patches in hectares. c) The proportion of grassland of all agricultural land in per- centages. d) The occurrence of nationally valuable tradi- tional rural biotopes (grazed forests excluded) (see Vainio et al. 2001) and the amount of arable land in hectares. 44 FENNIA 183: 1 (2005)Sonja Kivinen Over three quarters of the studied grid squares had 50 hectares or less grassland, and 3.5% of squares had no grassland. Grassland areas great- er than 100 hectares were found in 5.7% of the squares and the maximum value was 1044 hec- tares in the northern Bothnian Bay region. In com- parison, three quarters of the squares had arable land more than 100 hectares, and arable land was absent in 7.4% of the squares. Grassland patches were generally larger in northern Finland than in the south (Fig. 2b). Inventoried traditional rural bi- otopes (meadows and wooded pastures) (Vainio et al. 2001) accounted for approximately one tenth of the total grassland area. Grasslands were found both in the main agri- cultural regions and in regions where agriculture has only a minor role. The Spearman rank correla- tion coefficient (r) between grassland area and ar- able land area calculated for 100 km2 grid squares was 0.299 (p < 0.001). The proportion of grassland of all agricultural land was greatest in northern Finland and in coastal areas (Fig. 2c). A general absence of most valuable semi-natural grasslands (traditional rural biotopes of national value) was found in the most intensively cultivated regions of southern and western Finland (Fig. 2d). Traditional rural biotopes and landscape properties The grassland networks in six drainage basins lo- cated in different parts of Finland (Fig. 3) varied considerably in terms of the number and type of traditional rural biotopes, patch sizes, connectivity and surrounding land cover. A northern drainage basin including the lower course of Tornionjoki and Kaakamajoki (Fig. 4a) was generally characterized by extensive forest areas (Table 1). Grasslands covered 1.5% of the to- tal area and were generally found in river valleys. Traditional rural biotopes, with flood meadows as the most common type (Fig. 5a), were rather abundant, but mainly of local value. The studied traditional rural biotopes were located in the well- connected network of generally large grassland patches (Fig. 6a). A drainage basin including the upper courses of Vienan Kemi and Iijoki, also located in northern Finland (Fig. 4b), was dominated by forests, mires and numerous watercourses and grasslands cov- ered 0.8% of the area (Table 1). The characteristic traditional rural biotope type was fen meadows (Fig. 5b). The number of traditional rural biotopes was small compared to the area of the drainage basin. The grassland patches were large, but had a rather scattered location (Fig. 6b). Thus, traditional rural biotopes of the region were significantly iso- lated. The Kyrönjoki drainage basin located in western Finland (Fig. 4c) was characterized by arable land and forested areas (Table 1). Grasslands covered 0.2% of the total area and were located mainly by rivers. Traditional rural biotopes, nearly all of lo- cal value, were rather few compared to the area of drainage basin and no nationally valuable biotopes were found in the region. The most typical biotope types were mesic meadows, shore meadows and wooded pastures (Fig. 5c). Selected traditional ru- ral biotopes were poorly connected with the other, relatively small grassland patches (Fig. 6c). The eastern Koitajoki drainage basin (Fig. 4d) was dominated by forests and mires and grass- Fig. 3. The grassland proportions within drainage basins (in percentages) and the locations of six study areas. a) The lower course of Tornionjoki–Kaakamajoki, b) The upper courses of Vienan Kemi and Iijoki, c) Kyrönjoki, d) Koitajoki, e) Vanajavesi–Pyhäjärvi and f) Uskelanjoki–Halikonjoki (see Ekholm 1993). FENNIA 183: 1 (2005) 45Regional distribution and biodiversity perspectives of Finnish… Fig. 4. The distribution of grasslands presented in 1 km2 grid squares and the occurrence of traditional ru- ral biotopes within six stud- ied drainage basins. a) The lower course of Tornionjoki– Kaakamajoki, b) The upper courses of Vienan Kemi and Iijoki, c) Kyrönjoki, d) Koita- joki, e) Vanajavesi–Pyhäjärvi and f) Uskelanjoki–Halikon- joki. Table 1. Land use statistics for the studied six drainage basins. MPS = mean patch size of grassland, Grass = Grassland, Ar- able = arable land, Urban = urban area, Bare = bare ground. Drainage basin Area km2 MPS ha Grass % Arable % Urban % Forest % Mire % Bare % Water % Total % Tornionjoki–Kaakamajoki 1416 2.1 1.5 7.2 3.8 76.1 7.9 0.2 3.3 100.0 Vienan Kemi–Iijoki 3122 2.0 0.8 1.5 1.8 69.2 10.9 0.0 15.8 100.0 Kyrönjoki 4923 1.2 0.2 24.6 5.1 61.0 7.5 0.2 1.4 100.0 Koitajoki 3741 1.0 0.2 0.9 1.2 71.8 14.0 0.0 11.9 100.0 Vanajavesi–Pyhäjärvi 2759 1.1 0.8 17.7 11.2 55.2 1.2 0.1 13.8 100.0 Uskelanjoki–Halikonjoki 873 1.9 1.6 40.5 8.2 48.0 0.9 0.2 0.6 100.0 46 FENNIA 183: 1 (2005)Sonja Kivinen erate amounts of arable and urban land, forests and large watercourses (Table 1). Grasslands cov- ered 0.8% of the total area and were concentrated near to lakes. Compared to the area, the number of traditional rural biotopes was highest of all stud- ied drainage basins and a considerable number of the biotopes were of national or regional value. The main biotope types were wooded pastures and mesic meadows (Fig. 5e). Traditional rural bi- otopes were located in the dense network of rather small grassland patches (Fig. 6e). A drainage basin including Uskelanjoki and Halikonjoki (Fig. 4f) in south-western Finland was characterized by wide arable land areas (Table 1). Grasslands covered 1.6% of the total area and were located mainly by rivers. A high number of traditional rural biotopes were found in the region, over half of them classified as semi-natural grass- lands of national or regional value. Mesic mead- ows and shore meadows were the most common biotope types (Fig. 5f). Selected traditional rural biotopes were highly connected with the other relatively large grassland patches (Fig. 6f). The area and connectivity of traditional rural biotopes varied notably between drainage basins (Fig. 6g, Table 2), but less distinctively between value classes (national, regional, local) within drainage basins (Fig. 6). Some differences between value classes were found in Tornionjoki–Kaakama- joki and Vanajavesi–Pyhäjärvi, where the most valuable patches were the biggest ones. However, their connectivity was not generally greater than that of less valuable patches. Furthermore, in the Koitajoki drainage basin and partially in Uskelan- joki–Halikonjoki the most valuable traditional ru- ral biotopes were either biggest in size or had the highest connectivity values. Fig. 5. The proportions of traditional rural biotope types in six studied drainage basins (grazed forests excluded). a) The lower course of Tornionjoki–Kaakamajoki, b) The upper courses of Vienan Kemi and Iijoki, c) Kyrönjoki d) Koitajoki, e) Vanajavesi–Pyhäjärvi and f) Uskelanjoki–Halikonjoki. lands covered 0.2% of the total area (Table 1). Few traditional biotopes, with mesic meadows as the dominant type, were found in the region (Fig. 5d). Most of these biotopes were of local value and no nationally valuable traditional rural biotopes oc- curred in the region. The studied traditional rural biotopes were located in the disconnected net- work of small grassland patches (Fig. 6d). The landscape of a southern Vanajavesi–Py- häjärvi drainage basin (Fig. 4e) consisted of mod- Table 2. Connectivity statistics for the selected 10 or 11 traditional rural biotopes. FPS = focal patch size, NN = mean distance between the focal patch and the nearest neighbour- ing grassland patch, Conn = connectivity value of focal patches. Drainage basin FPS (ha) NN (km) Conn mean st dev mean mean st dev Tornionjoki–Kaakamajoki 11.8 23.2 0.4 20.9 25.1 Vienan Kemi–Iijoki 14.3 16.3 1.4 4.5 11.0 Kyrönjoki 2.2 1.5 0.5 7.3 7.5 Koitajoki 1.8 1.8 1.0 6.8 10.4 Vanajavesi–Pyhäjärvi 1.1 0.8 0.4 16.1 23.9 Uskelanjoki–Halikonjoki 4.3 6.5 0.3 29.6 29.3 FENNIA 183: 1 (2005) 47Regional distribution and biodiversity perspectives of Finnish… 806040200 A re a (h a) 80 60 40 20 0 a) 403020100 50 40 30 20 10 0 b) Connectivity 3020100 7 6 5 4 3 2 1 0 d) Connectivity 806040200 A re a (h a) 25 20 15 10 5 0 f) Connectivity 35302520151050 A re a (h a) 16 14 12 10 8 6 4 2 0 d c e f b a g) National Regional Local Value classes of traditional rural biotopes C onnectivity 20151050 A re a (h a) 5 4 3 2 1 0 c) Connectivity 100806040200-20 A re a (h a) 3.5 3.0 2.5 2.0 1.5 1.0 .5 0.0 e) Fig. 6. The patch size and connectivity of traditional rural biotopes in six stud- ied drainage basins. a) The lower course of Tornionjoki– Kaakamajoki, b) The upper courses of Vienan Kemi and Iijoki, c) Kyrönjoki d) Koita- joki, e) Vanajavesi–Pyhäjärvi and f) Uskelanjoki–Halikon- joki and g) the mean values of traditional rural biotopes for drainage basins. Note that the axes have different scales. Studied species in relation to grassland and arable land Butterflies C. pamphilus and L. hippothoe were present in 32.3% and 28.4% of 776 investigated grid squares, respectively. Occupied squares had significantly higher amounts of grassland than unoccupied squares (Fig. 7a & b, Table 3). The occurrence of L. hippothoe was also significantly positively related with arable land area, in contrast to C. pamphilus. Both species occurred densely in grassland-rich areas of southern, central and eastern Finland. In the north, C. pamphilus occurred in the regions of Bothnian Bay with abundant grasslands. Both species were absent in several squares with high proportions of arable land in southern and western Finland (Fig. 8a, c & d). 48 FENNIA 183: 1 (2005)Sonja Kivinen Fig. 7. The relationship between species presence and ab- sence and grassland area in 100 km2 squares. a) C. pam- philus, b) L. hippothoe, c) A. palustris, d) L. naevia, e) C. crex and f) P. perdix. The box represents the inter-quartile range, the line within the box shows the median, and the whiskers extend to the smallest and largest observations that are not suspected outliers. Table 3. Statistics for the relationship between butterfly and bird species occurrences and agricultural land use based on 100 km2 grid-square analysis. p-values derived from the Mann-Whitney U-test designate the significance of the differences between presence (1) and absence (0) squares. p-values were corrected using the Bonferroni method. n.s. = not significant, ** p < 0.01, *** p < 0.001. Species Grassland (ha) Arable land (ha) n mean st dev p mean st dev p C. pamphilus 0 525 36.8 60.9 *** 1046.6 1105.1 n.s. 1 251 54.1 57.6 1124.1 1027.9 L. hippothoe 0 556 40.9 65.4 ** 1007.7 1080.3 *** 1 220 46.1 45.3 1233.4 1066.9 A. palustris 0 1008 32.9 60.0 *** 933.4 1117.1 *** 1 449 51.2 52.1 1389.4 1202.2 L. naevia 0 1106 32.6 41.1 *** 717.1 1029.4 *** 1 547 54.9 80.6 1446.7 1193.9 C. crex 0 1087 35.0 59.6 *** 1021.1 1163.9 *** 1 314 47.0 46.2 1428.9 1145.7 P. perdix 0 1383 36.5 43.8 n.s. 848.2 975.5 *** 1 156 56.3 119.3 2599.4 1416.9 Birds The farmland edge species A. palustris and L. nae- via were the most common of the studied bird spe- cies found in 30.8% and 33.1% of the grid squares, respectively. Both grassland and arable land were significantly more abundant in occupied than in unoccupied squares (Fig. 7c & d, Table 3). A. palustris and L. naevia occurred densely in south- ern Finland and in coastal regions with abundant grasslands. The densest occurrences of the species in central and eastern Finland coincided largely with grassland patterns. The species occurred spo- radically or were absent in forest-dominated re- gions of western and eastern Finland. Moreover, they occurred scarcely in intensively cultivated regions of western Finland as well as in parts of agricultural areas in southern Finland. In northern Finland, species were found mainly in grassland- rich regions of the Bothnian Bay (Fig. 8b, e & f). The distributions of C. crex and P. perdix, the species breeding in fields, differed greatly from each other. C. crex was present in 22.4% and P. perdix only in 10.1% of the studied squares. The occurrence of C. crex was significantly positively related to the amount of grasslands and arable land, whereas the occurrence of P. perdix was sig- nificantly positively associated only with arable land (Fig. 7e & f, Table 3). The spatial distribution pattern of C. crex in relation to grasslands and ar- able land resembled largely that of A. palustris and L. naevia. P. perdix was mainly present in squares with a great amount of arable land located in the main agricultural regions of southern and western FENNIA 183: 1 (2005) 49Regional distribution and biodiversity perspectives of Finnish… Fig. 8. The spatial relation- ships between studied spe- cies and agricultural land use patterns. a) Observation squares for butterfly species and b) bird species. The oc- currence of c) C. pamphilus, d) L. hippothoe, e) A. palus- tris, f) L. naevia and g) C. crex in relation to grassland area (ha/100 km2) and h) P. perdix in relation to arable land area (ha/100 km2). Finland and in the Bothnian Bay region (Fig. 8b, g & h). Discussion Regional distribution of grasslands in the Finnish landscape The abundance of grasslands varied greatly in dif- ferent parts of Finland. Grasslands were generally scarce in parts of southern and western Finland. In these regions, favourable physical conditions, such as mild climate, flat topography and fine-sediment soils have promoted intensification of agriculture (see Alalammi 1994; Palomäki & Mikkonen 1999). Regions with high cover of grasslands were gener- ally characterized by varied topography, which has led to smaller field sizes and more mosaic agri- cultural landscapes compared to easily cultivated fine-sediment plains (e.g. Luoto 2000). As a special case, land uplift processes have created a unique 50 FENNIA 183: 1 (2005)Sonja Kivinen grassland pattern in western coastal areas. Land uplift is particularly rapid in the Bothnian Bay re- gion, as much as 8 mm per year (Kakkuri 1990). As the land rises, the bottoms of long, sheltered reedy bays gradually turn into coastal grasslands. How- ever, the expansion of reeds due to eutrophication and natural succession makes these grasslands rather instable habitats (Vainio et al. 2001). Physical conditions have traditionally deter- mined the interactions between nature and ag- ricultural land and have created the traditional landscape structure of Europe (Jongman 2002). However, agricultural land use today, and thus the present distribution of grasslands, is determined to a growing degree by economic conditions, cultur- al and social aspects and agricultural policies (Ihse 1995; Jongman 2002). Agricultural landscapes of southern and western Finland, characterized by relatively monotonous field plains, share similar biodiversity problems with Western European agri- cultural regions. Large crop cultivation areas often have only few grasslands. Moreover, in the west- ern regions characterized by effective dairy farm- ing, grasslands are continually renewed. In other words, rural habitats of different age and structure are generally lacking. Agricultural subsidies direct- ed to such favourable cultivation regions further intensify the development (Jongman 2002; Robin- son & Sutherland 2002; Waldhart & Otte 2003). The number of dairy farms grows eastwards and northwards, which is naturally reflected in the oc- currence of various grassland habitats. Further, the occurrence of non-renewed grasslands may partially reflect the decrease of agricultural activi- ties. Small-scale dairy farming is diminishing to a growing degree as production is concentrated on economically more profitable larger farms and as the current farmer population ages (e.g. Pyykkönen 2001). For example, between 1990 and 2000 the number of dairy farms decreased by about 50% (TIKE 2002). Agricultural marginaliza- tion, affecting in a larger context mainly northern and eastern Europe, does not have a positive ef- fect on biodiversity in the long run (Pärtel et al 1999; Norderhaug et al. 2000). Although the area of non-renewed grasslands first increases in the landscape, the grasslands will eventually revert to forest if they are not adequately managed (Luoto et al. 2003a; Waldhart et al. 2003). In other words, semi-open mixed farm landscape will become more enclosed, forested landscape at the expense of the diversity of agricultural habitats and species (Ihse 1995). Traditional rural biotopes in the agricultural landscape mosaic The occurrence of traditional rural biotopes as a part of grassland networks varied considerably in different parts of Finland. Traditional rural biotopes can be considered as diversity “hot spots” in the grassland network (see Vainio et al. 2001). They contribute regional biodiversity, providing essen- tial habitats for specialised species and act as es- sential source populations in depleted agricultural landscapes (Duelli & Obrist 2003). Further, other grasslands enlarge the non-crop habitat selection and can provide important dispersal routes in the landscape (Steffan-Dewenter & Tscharntke 1997). Well-connected grassland networks containing abundant traditional rural biotopes were found in three studied drainage basins located in northern and southern Finland. In Tornionjoki–Kaakamajoki previously widely mowed and grazed flood mead- ows, once cleared from bushes and grassy flood forests (Cajander 1909; Kalliola 1973), have been partially preserved in the landscape along with cultivated grasslands. The Vanajavesi–Pyhäjärvi region has long traditions of successful, versatile agriculture (Maisema-aluetyöryhmä 1993), and rather numerous wooded pastures and meadows are still found in the landscape. The Uskelanjoki– Halikonjoki drainage basin represents a special case in Finland. Agri-environmental support and successful restoration projects have increased the area of various semi-natural grasslands located in topographically steep, economically marginal river valleys surrounded by intensively cultivated fine sediment plains (Luoto et al. 2003b). From the metapopulation point of view, these regions rep- resent landscape mosaics with high biodiversity values (Hanski 1998). Disconnected grassland networks including few traditional rural biotopes occurred in three drain- age basins in northern, eastern and western Fin- land. Vienan Kemi–Iijoki and Koitajoki were both forested, sparsely settled drainage basins affected by agricultural marginalization. In Vienan Kemi–Ii- joki, half of the traditional rural biotopes were fen meadows that are typically less species-rich than drier meadows, and the relatively high proportion of grasslands in the region results from the large and scattered non-renewed cultivated grasslands (see Maisema-aluetyöryhmä 1993). In Koitajoki, scarce, but species-rich dry and mesic meadows have been preserved in the rather enclosed for- ested landscape. In the Kyrönjoki drainage basin FENNIA 183: 1 (2005) 51Regional distribution and biodiversity perspectives of Finnish… the intensification of agriculture in fine-sediment river valleys has resulted in a scarcity of traditional rural biotopes and grassland habitats in general (see Alalammi 1994). On the basis of metapopula- tion theory (Hanski 1998) these landscapes can- not maintain a diversity of agricultural nature as high as landscapes with well-connected grassland networks including several patches of semi-natural grasslands. The number of traditional rural biotopes is rap- idly decreasing and their quality has largely de- clined. The most valuable semi-natural grasslands are typically located in small-holdings owned by elderly farmers and are thus threatened by the end of management in the near future (Pykälä 2001). The maintenance of biodiversity in Finland re- quires the management and restoring of traditional rural biotopes on a large scale. However, this is naturally cost-demanding and can be difficult to accomplish. As mosaic agricultural landscapes with diverse habitat selection are crucial for the maintenance of biodiversity, it would be more pref- erable to direct the limited resources to traditional rural biotopes located in well-connected grassland networks than to isolated grassland patches (e.g. Dunning et al. 1992). Grassland patterns and the distribution of species The distributions of the studied butterfly and bird species were generally related to the grassland pat- terns. Moreover, a general absence of the studied species excluding P. perdix favouring large field areas was observed in the most intensively cul- tivated regions in southern and western Finland. These results are in accordance with other studies that have pointed out the impoverishing impact of agricultural intensification on diversity (Robinson & Sutherland 2002; Jeanneret et al. 2003a; Steiner & Köhler 2003). The location and dispersal of butterfly popula- tions strongly depend on flowering plants, and various meadows have been found to be the rich- est habitats for butterflies (Kuussaari et al. 2001). Further, butterflies need an appropriate landscape structure in order to be able to move between the habitat patches, and often require several habi- tats to complete their life-cycles (Jeanneret et al. 2003b). In general, the widespread decline of but- terfly species in Finland and throughout Europe has been associated with the decrease of open, uncultivated land (Pitkänen et al. 2001). It has been noted that old, abandoned fields may par- tially compensate the scarcity of meadows for but- terflies (e.g. Heliölä et al. 2000). Farmland birds have in many cases been affect- ed by the loss of the mosaic of pastures and field in agricultural landscapes (Tiainen & Pakkala 2001). The habitat requirements of birds are wider com- pared to several other species groups and birds use various open and semi-open grasslands as feeding and breeding habitats (Söderström et al. 2001; Bar- nett et al. 2004; Peach et al. 2004; Virkkala et al. 2004). Farmland birds breeding in bushes, such as A. palustris and L. naevia studied here, have bene- fited from the increase of abandoned successional agricultural land (Väisänen et al. 1998). Further, C. crex has recently recovered slightly in Finland, possibly due to the decreased intensity of agri- culture in the neighbouring Baltic countries and in Russia and the establishment of non-cropped shelter zones on Finnish fields (Tiainen & Pakkala 2001). It has also been observed that C. crex is nowadays found more often on low-productive grasslands than on fields (Kunttu & Laine 2002). In contrast, P. perdix has been noted to benefit partic- ularly from short-term rotated set-aside (Tiainen & Pakkala 2001), which was included in arable land in the classification system of this study. The distribution of plant species was not exam- ined in this study. The highest diversity of plant species is generally found in traditionally man- aged dry and mesic meadows in boreal agricul- tural landscapes (Pykälä 2000). A large part of grasslands studied here had earlier been subject to a rotation cycle and the development of vegetation in these areas strongly depends on the intensity of former arable management, which in turn affects the nutrient balance and seed bank (Hansson & Fogelfors 1998; Pärtel et al. 1999; Perner & Malt 2003; Waldhart & Otte 2003). The signs of former cultivation can be seen in a grassland patch even one century after cultivation has ended (Skånes 1991; Gibson & Brown 1992). Common plant species for meadows may disperse rapidly on grassland patches, but rarer ones are often miss- ing. The proximity of well-preserved species-rich meadows enhances species dispersal (Dunning et al. 1992; Norderhaug et al. 2000; Pykälä 2001; Duelli & Obrist 2003). Maps of the studied agri- cultural landscapes reflect the different dispersal potentials of plant species from species-rich “hot spots” of semi-natural grasslands in the network of grassland patches (Fig. 4). 52 FENNIA 183: 1 (2005)Sonja Kivinen The main limitations of the species data were the large number of uninvestigated squares of the butterfly atlas and the lack of estimation of spe- cies abundance in the atlas squares. Furthermore, species distribution is driven by multiple factors acting on multiple spatial and temporal scales (Wiens 1989; Levin 1992). For example, the im- pact of climate on species distribution is particu- larly pronounced in large scale studies (Thuiller et al. 2004). Thus, the multivariate analysis including both climatic and habitat variables, as well as spe- cies abundances instead of presence/absence data would provide interesting information concerning the relationship between species distributions and environmental factors. Advantages and limitations in using large spatial databases Technical advantages and recent production of nationwide spatial databases have made signifi- cant contributions to biogeographical research (Johnson 1990; Goodchild 1994; Burnett & Kal- liola 2000; Weiers et al. 2004). In this study, new information on the distribution of grasslands was produced by combining large databases with rela- tively high spatial resolution and detailed land use information. From a biodiversity point of view, these methods can provide effective tools and new insights for management and monitoring of Finn- ish agricultural landscapes. However, the use of large databases does cause some problems. Because nationwide databases are usually created by combining different kinds of source data, they can be variable by quality and age, and even include classification errors. Thus, GIS users should be aware of these potential pit- falls and their influence on the studied distribu- tions (e.g. Berry 1987). Here, the SLICES grassland data was based on two continually updated vector databases, the Finnish Land Parcel Identification System administered by the Ministry of Agricul- ture and Forestry and the Topographic Database produced by the National Land Survey of Finland. These databases include the most accurate infor- mation of the Finnish agricultural land use and ter- rain. However, the production of SLICES database naturally required compromises in resolving spa- tially overlapping land use classes of source data (see Mikkola et al. 1999) and these decisions are largely out of reach of the ordinary database user. The field check of the grassland database car- ried out here showed that actual classification errors (e.g. clear-cuttings classified as grassland) were rather few. Thus, the data can be consid- ered to represent rather reliably the distribution of grasslands in Finland. A more crucial potential problem in extracting habitats from land use da- tabases is the disagreement between the desired habitat properties and the classification criteria of land use in the database. In this study, the SLICES classification system was suitable for the mapping of low-intensity grassy habitats, i.e. old grasslands and long-term fallows. However, in order to dis- tinguish ecologically the most valuable semi-natu- ral grasslands from other grassland habitats, it was necessary to use the database containing inven- toried traditional rural biotopes. Furthermore, low intensity grasslands habitats are continually threat- ened by bush encroachment if not adequately managed and the information of open grassland habitats in the database may become obsolete rather rapidly especially in agriculturally marginal regions. In this study the main part of the grass- lands were open or nearly open habitats. How- ever, clear signs of bush encroachment were also observed, and so the correspondence between the information in the database and the real state of grasslands is continuously changing. Conclusions The distribution of grasslands in Finland was ef- fectively mapped using a nationwide land use da- tabase. The increased production and availability of large digital spatial databases provide powerful possibilities for research and biodiversity manage- ment. However, classification errors, the age of data and particularly classification criteria of land use classes compared to aspired habitat properties may limit the usability of the databases and must be taken into account in habitat mapping. The results showed great variation in the re- gional abundance of grasslands, due to physical conditions and various social-economical and his- torical as well as political factors. The results also indicated the positive effect of grassland abun- dance and the negative effect of intensified land use on the occurrence of the studied species living in various grassland habitats. Agricultural intensi- fication in the most advantageous farming regions and the end of dairy farming in small-holdings in the near future threaten the existence of versatile agricultural landscapes with a diverse selection of grassland habitats and associated species. Thus, FENNIA 183: 1 (2005) 53Regional distribution and biodiversity perspectives of Finnish… the maintenance of well-connected grassland net- works including diversity “hot spots” of semi-natu- ral grasslands is an essential, but demanding chal- lenge in the management of Finnish agricultural biodiversity. ACKNOWLEDGEMENTS I thank Dr. M. Luoto, Prof. R. Kalliola and two ref- erees for their valuable comments on the manuscript. I am grateful to Dr. R. Heikkinen for contributions to the collection of field data. I also thank M. Bai- ley for language checking of the English manuscript. This study was supported by the LUMOTTU program funded by the Ministry of Agriculture and Forestry, the Ministry of the Environment and the Finnish Cul- tural Foundation. REFERENCES Ahti T, L Hämet-Ahti & J Jalas (1968). Vegetation zones and their sections in northwestern Europe. Annales Botanici Fennici 5, 169−210. Alalammi P (ed) (1994). Finland’s landscapes, and ur- ban and rural milieus. Atlas of Finland 350. 234 p. National Land Survey of Finland & Geographical Society of Finland, Helsinki. Alanen A (1997). Perinnemaisemat. In Luostarinen M & A Yli-Viikari (eds). Maaseudun kulttuurimaise- mat. The Finnish Environment 87, 71−78. Barnett PR, MJ Whittingham, RB Bradbury & JD Wil- son (2004). Use of unimproved and improved lowland grassland by wintering birds in the UK. Agriculture, Ecosystems and Environment 102, 49−60. Bender DJ, L Tischendorf & L Fahrig (2003). Using patch isolation metrics to predict animal binary movement in binary landscapes. Landscape Ecol- ogy 18, 17−39. Benton TG, JA Vickery & JD Wilson (2003). Farm- land biodiversity: is habitat heterogeneity the key? Trends in Ecology and Evolution 18, 182−188. Berry JK (1987). Computer-assisted map analysis: po- tential and pitfalls. Photogrammetric Engineering and Remote Sensing 53, 1405−1410. Briggs D & F Courtney (1989). Agriculture and en- vironment. The physical geography of temperate agricultural systems. 442 p. Longman Scientific and Technical, Singapore. Burnett C & R Kalliola (2000). Maps in the informa- tion society. Fennia 178, 81−96. Cajander AK (1909). Niityt ja viljelysmaat. In Palmén EG, E Hjelt, J Gummerrus, G Melander, JA Pal- men, K Krohn & J Jantti (eds). Oma maa IV. Tieto- kirja Suomen kodeille, 41−50. WSOY, Porvoo. Carpelan C (1999). Käännekohtia Suomen esihisto- riassa aikavälillä 5100−1000 eKr. In Fogelberg P (ed). Pohjan poluilla. Suomalaisten juuret nykytut- kimuksen mukaan, 249−280. Suomen tiedeseura, Helsinki. Corbet SA (1995). Insects, plants and succession: ad- vantages of long-term set-aside. Agriculture, Eco- systems and Environment 53, 201−217. Cousins SAO, S Lavorel & I Davies (2003). Modelling the effects of landscape pattern and grazing re- gimes on the persistence of plant species with high conservation value in grasslands in south-eastern Sweden. Landscape Ecology 18, 315−332. Dauber J, M Hirsch, D Simmering, R Waldhardt, A Otte & V Wolters (2003). Landscape structure as an indicator of biodiversity: matrix effects on spe- cies richness. Agriculture, Ecosystems and Envi- ronment 98, 321−329. de la Peña NM, A Butet, Y Delettre, G Paillat, P Mo- rant, L Le Du & F Burel (2003). Response of the small mammal community to changes in western French agricultural landscapes. Landscape Ecol- ogy 18, 265−278. di Giulio M, PJ Edwards & E Meister (2001). Enhanc- ing insect diversity in agricultural grasslands: the roles of management and landscape structure. Journal of Applied Ecology 38, 310−319. Duelli P & MK Obrist (2003). Regional biodiversity in an agricultural landscape: the contribution of seminatural habitat islands. Basic and Applied Ecology 4, 129−138. Dunning JB, BJ Danielson & HR Pulliam (1992). Eco- logical processes that affect populations in com- plex landscapes. Oikos 65, 169−175. Ekholm M (1993). Suomen vesistöalueet. Publica- tions of the Water and Environment Administra- tion A 126. 163 p. Esri (1992). Understanding GIS. The ARC/INFO Meth- od. Environmental Systems Research Institute, Inc., Redlands, CA, USA. Fahrig L (2001). How much habitat is enough? Bio- logical Conservation 100, 65−74. FEI (2003). Final report of the first part of the project Corine Land Cover 2000 Finland (CLC2000). Finn- ish Environment Institute. Unpublished report. Freemark KE, C Boutin & CJ Keddy (2002). Importance of farmland habitats for conservation of plant spe- cies. Conservation Biology 16, 399−412. Gibson CWD & VK Brown (1992). Grazing and veg- etation change: deflected or modified succession? Journal of Applied Ecology 29, 120−131. Goodchild MF (1994). Integrating GIS and remote sensing for vegetation analysis and modeling: methodological issues. Journal of Vegetation Sci- ence 5, 615−626. Granberg L (1989). Valtio maataloustulojen tasaajana ja takaajana. Bidrag till Kännedom av Finlands Natur och Folk 138, 1−214. Grotenfelt G (1922). Alkuperäiset viljelystavat Suo- messa. In Krohn K, KO Lindeqvist, G Melander & K Grotenfelt (eds). Oma maa III. Tietokirja Suo- men kodeille, 24−41. WSOY, Porvoo. Hämet-Ahti L (1988). Suomen kasvillisuuden pääpiir- teet. In Alalammi P (ed). Atlas of Finland, Folio 54 FENNIA 183: 1 (2005)Sonja Kivinen 141: Vegetation and flora, 1−2. National Board of Survey and Geographical Society of Finland, Helsinki. Hanski I (1994). A practical model of metapopula- tion dynamics. Journal of Animal Ecology 63, 151−162. Hanski I (1998). Metapopulation dynamics. Nature 396, 41−49. Hanski I & M Gilpin (1991). Metapopulation dynam- ics: brief history and conceptual domain. Biologi- cal Journal of the Linnean Society 42, 3−16. Hansson M & H Fogelfors (1998). Management of permanent set-aside on arable land in Sweden. Journal of Applied Ecology 35, 758−771. Haysom KA, DI McCracken, GN Foster & NW Sotherton (2004). Developing grassland conserva- tion headlands: response of carabid assemblage to different cutting regimes in a silage field edge. Agriculture, Ecosystems & Environment 102, 263−277. Heliölä J, O-P Liinalaakso, R Martikainen & T Schultz (2000). Tummaverkkoperhonen Pirkanmaalla. Pirkanmaan ympäristökeskuksen monistesarja 6. 39 p. Helminen VA (1987). Lämpöolot. In Alalammi P (ed). Atlas of Finland, Folio 131: Climate, 4−10. Na- tional Board of Survey & Geographical Society of Finland, Helsinki. Huldén L, A Albrecht, J Itämies, P Malinen & J Wet- tenhovi (2000). Suomen suurperhosatlas. 328 p. Suomen Perhostutkijain Seura & Luonnontieteel- linen keskusmuseo, Helsinki. Huurre M (1995). 9000 vuotta Suomen esihistoriaa. 271 p. Otava, Helsinki. Ihse M (1995). Swedish agricultural landscapes – pat- terns and changes during the last 50 years, studied by aerial photos. Landscape and Urban Planning 31, 21−37. Ihse M & C Lindahl (2000). A holistic model for land- scape ecology in practice: the Swedish survey and management of ancient meadows and pastures. Landscape and Urban Planning 50, 59−84. Jeanneret Ph, B Schüpbach & H Luka (2003a). Quan- tifying the impact of landscape and habitat fea- tures on biodiversity in cultivated landscapes. Agriculture, Ecosystems and Environment 98, 311−320. Jeanneret Ph, B Schüpbach, L Pfiffner & Th Walter (2003b). Arthropod reaction to landscape and habitat features in agricultural landscapes. Land- scape Ecology 18, 253−263. Johnson LB (1990). Analyzing spatial and temporal phenomena using geographical information sys- tem. A review of ecological applications. Land- scape Ecology 4, 31−43. Jongman RHG (2002). Homogenisation and fragmen- tation of the European landscape: ecological con- sequences and solutions. Landscape and Urban Planning 58, 211−221. Jordán F, A Báldi, K-M Orci, I Rácz & Z Varga (2003). Characterizing the importance of habitat patches and corridors in maintaining the landscape con- nectivity of a Pholidoptera transsylvanica (Or- thoptera) metapopulation. Landscape Ecology 18, 83−92. Kakkuri J (1990). Fennoskandian maankohoaminen. In Alalammi P (ed). Atlas of Finland, Folio 125: Geophysics of the solid earth crust, 35−36. Na- tional Board of Survey & Geographical Society of Finland, Helsinki. Kalliola R (1973). Suomen kasvimaantiede. 308 p. WSOY, Porvoo. Kiviniemi K & O Eriksson (2002). Size-related dete- rioration of semi-natural grassland fragments in Sweden. Diversity and Distributions 8, 21−29. Krebs JR, JD Wilson, RB Bradbury & GM Siriwardena (1999). The second Silent Spring? Nature 400, 611−612. Kujansuu R & J Niemelä (1990). Maaperämuodostu- mat. In Alalammi P (ed). Atlas of Finland, Folio 124: Surficial deposits, 9−11. National Board of Survey and Geographical Society of Finland, Hel- sinki. Kunttu P & J Laine (2002). Turun pesimälinnuston muutokset vuosina 1951−2001. 60 p. Turun kau- pungin ympäristönsuojelutoimisto, Turku. Kuussaari M, J Heliölä, J Salminen & I Niininen (2001). Results of the butterfly monitoring scheme in Finnish agricultural landscapes for the year 2000. Baptria 26, 69−80. Lee JC, FD Menalled & DA Landis (2001). Refuge habitats modify impact of insecticide disturbance on carabid beetle communities. Journal of Ap- plied Ecology 38, 472−483. Levin SA (1992). The problem of pattern and scale in ecology. Ecology 73, 1943−1967. Levins R (1969). Some demographic and genetic consequences of environmental heterogeneity for biological control. Bulletin of the Entomological Society America 15, 237−240. Linkola K (1922). Niityt ja viljelysmaat. In Krohn K, KO Lindeqvist, G Melander & K Grotenfelt (eds). Oma maa III. Tietokirja Suomen kodeille, 1012−1032. WSOY, Porvoo. Luostarinen M (1997). Maaseudun suuri historia. In Luostarinen M & A Yli-Viikari (eds). Maaseudun kulttuurimaisemat. The Finnish Environment 87, 59−63. Luoto M (2000). Spatial analysis of landscape eco- logical characteristics of five agricultural areas in Finland by GIS. Fennia 178, 15−54. Luoto M, M Kuussaari, H Rita, J Salminen & T von Bonsdorff (2001). Determinants of distribution and abundance in the clouded apollo butterfly: a landscape ecological approach. Ecography 24, 601−617. Luoto M, J Pykälä & M Kuussaari (2003a). Decline of landscape-scale habitat and species diversity after the end of cattle grazing. Journal for Nature Con- servation 11, 171−178. Luoto M, S Rekolainen, J Aakkula & J Pykälä (2003b). Loss of plant species richness and habitat con- FENNIA 183: 1 (2005) 55Regional distribution and biodiversity perspectives of Finnish… nectivity in grasslands associated with agricultural change in Finland. Ambio 32, 447−451. MacArthur RH & EO Wilson (1967). The theory of island biogeography. 203 p. Princeton University Press, Princeton. MAF (2004a). Maisemanhoito. Luonnon monimuo- toisuus. Perinnebiotoopit. Maatalouden ympäris- tötuen erityistuet v. 2000−2006. 20 p. Ministry of Agriculture and Forestry. MAF (2004b). Matilda. Information Service of Ag- ricultural Statistics in Finland. Ministry of Agri- culture and Forestry. . 17.5.2004. Maisema-aluetyöryhmä (1993). Arvokkaat maise- ma-alueet. Maisema-aluetyöryhmän mietintö II. 204 p. Ympäristöministeriö, mietintö 66/1992. Marttila O, T Haahtela, H Aarnio & P Ojalainen (1990). Suomen päiväperhoset. 362 p. Karisto Oy, Hämeenlinna. Matson PA, WJ Parton, AG Power & MJ Swift (1997). Agricultural intensification and ecosystem proper- ties. Science 277, 504−509. Meyer-Aurich A, P Zander & M Hermann (2003). Consideration of biotic nature conservation tar- gets in agricultural land use – a case study from the Biosphere Reserve Schorfheide-Chorin. Agriculture, Ecosystems and Environment 98, 529−539. Mikkola A, O Jaakkola & Y Sucksdorff (1999). Valtakunnallisten maankäyttö-, peitteisyys- ja maaperäaineistojen muodostaminen. The Finnish Environment 342. 86 p. Moilanen A (2000). The Si program. Metapopulation research group, University of Helsinki. Norderhaug A, M Ihse & O Pedersen (2000). Biotope patterns and abundance of meadow plant species in a Norwegian rural landscape. Landscape Ecol- ogy 15, 201−218. Oksanen M (2001). Vihreät laitumet. Laidunyhdistys 1927−1970. 112 p. Karisto Oy, Hämeenlinna. Opdam P, J Verboom & R Pouwels (2003). Landscape cohesion: an index for the conservation potential of landscapes for biodiversity. Landscape Ecology 18, 113−126. Pain DJ, D Hilla & DI McCracken (1997). Impact of agricultural intensification of pastoral systems on bird distributions in Britain 1970−1990. Agricul- ture, Ecosystems and Environment 64, 19−32. Palomäki M & K Mikkonen (1999). Talousalueet ja menestymisen edellytykset. In Westerholm J & P Raento (eds). Suomen kartasto, 68−69. Suomen Maantieteellinen Seura & WSOY, Helsinki. Pärtel M, R Mändla & M Zobel (1999). Landscape history of calcareous (alvar) grasslands in Hanila, western Estonia, during the last three hundred years. Landscape Ecology 14, 187−196. Peach WJ, M Denny, PA Cotton, IF Hill, D Gruar, D Barritt, A Impey & J Mallord (2004). Habitat se- lection by song thrushes in stable and declining farmland populations. Journal of Applied Ecology 41, 275−293. Perner J & S Malt (2003). Assessment of changing agricultural land use: response of vegetation, ground-dwelling spiders and beetles to the con- version of arable land into grassland. Agriculture, Ecosystems and Environment 98, 169−181. Pitkänen M, M Kuussaari & J Pöyry (2001). Butter- flies. In Pitkänen M & J Tiainen (eds). Biodiversity of agricultural landscapes in Finland. Birdlife Fin- land Conservation Series 3, 51−68. Poschlod P & MF WallisDeVries (2002). The histori- cal and socioeconomic perspective of calcareous grasslands − lessons from the distant and recent past. Biological Conservation 104, 361−376. Pöyry J, S Lindgren, J Salminen & M Kuussaari (2004). Restoration of butterfly and moth communities in semi-natural grasslands by cattle grazing. Ecologi- cal Applications 14, 1656−1670. Pykälä J (2000). Mitigating human effects on Euro- pean biodiversity through traditional animal hus- bandry. Conservation Biology 14, 705−712. Pykälä J (2001). Perinteinen karjatalous luonnon mo- nimuotoisuuden ylläpitäjänä. The Finnish Envi- ronment 495. 205 p. Pykälä J, A Alanen, M Vainio & A Leivo (1994). Perin- nemaisemien inventointiohjeet. Vesi- ja ympäris- töhallituksen monistesarja 559. 106 p. Pykälä J & I Lappalainen (1998). Nykyaika näkyy maatalousympäristöissä. In Lappalainen I (ed). Suomen luonnon monimuotoisuus, 184−195. Finnish Environment Institute, Helsinki. Pywell RF, EA Warman, TH Sparks, JN Greatorex- Davies, KJ Walker, WR Meek, C Carvell, S Petit & LG Firbank (2004). Assessing habitat quality for butterflies on intensively managed arable farm- land. Biological Conservation 118, 313−325. Pyykkönen P (2001). Maatalouden rakennemuutos eri alueilla. Pellervon taloudellisen tutkimuslai- toksen raportteja 180. 61 p. Rassi P, A Alanen, T Kanerva & I Mannerkoski (eds) (2001). Suomen lajien uhanalaisuus 2000. 432 p. Ministry of Environment & Finnish Environment Institute, Helsinki. Robinson RA & WJ Sutherland (2002). Post-war chang- es in arable farming and biodiversity in Great Brit- ain. Journal of Applied Ecology 39, 157−176. Ruuhijärvi R (1988). Suomen kasvillisuuden pääpiir- teet. In Alalammi P (ed). Atlas of Finland, Folio 141: Vegetation and flora, 2−6. National Board of Survey and Geographical Society of Finland, Helsinki. Saunders DA, RJ Hobbs & CR Margules (1991). Bio- logical consequences of ecosystem fragmentation: a review. Conservation Biology 5, 18−32. Seppälä M (1986). Maanpinnan muodot. In Alalammi P (ed). Atlas of Finland, Folio 122: Geomorphol- ogy, 1−8. National Board of Survey & Geographi- cal Society of Finland, Helsinki. Simonen A (1990). Suomen kallioperä. In Alalammi P (ed). Atlas of Finland, Folio 123: Bedrock, 1−4. National Board of Survey & Geographical Society of Finland, Helsinki. 56 FENNIA 183: 1 (2005)Sonja Kivinen Skånes H (1991). Förändringar i odlingslandskapet och dess konsekvenser för gräsmarksfloran. Stock- holms Universitet Naturgeografiska Institutionen Rapport 86, 1−70. Söderström B, B Svensson, K Vessby & A Glimskär (2001). Plants, insects and birds in semi-natural pastures in relation to local habitat and land- scape factors. Biodiversity and Conservation 10, 1839−1863. Soininen AM (1974). Vanha maataloutemme. Maa- talous ja maatalousväestö Suomessa perinnäisen maatalouden loppukaudella 1720-luvulta 1870- luvulle. Journal of the Scientific Agricultural Soci- ety in Finland 46, 1−459. Helsinki. Solantie R (1987). Sade- ja lumiolot. In Alalammi P (ed). Atlas of Finland, Folio 131: Climate, 18−22. National Board of Survey & Geographical Society of Finland, Helsinki. SPSS (2002). SPSS. Release 11.5.1. SPSS, Inc. Steffan-Dewenter I & T Tscharntke (1997). Early suc- cession of butterflies and plant communities on set-aside fields. Oecologia 109, 294−302. Steffan-Dewenter I & T Tscharntke (2001). Succes- sion of bee communities on fallows. Ecography 24, 83−93. Steiner NC & W Köhler (2003). Effects of landscape patterns on species richness – a modelling ap- proach. Agriculture, Ecosystems and Environment 98, 353−361. Thuiller W, MB Araújo & S Lavorel (2004). Do we need land-cover data to model species distribu- tions in Europe? Journal of Biogeography 31, 353−361. Tiainen J (2004). Maatalousympäristön historia. In Tiainen J, M Kuussaari, IP Laurila & T Toivonen (eds). Elämää pellossa. Suomen maatalousym- päristön monimuotoisuus, 26−42. Edita, Hel- sinki. Tiainen J & T Pakkala (2001). Birds. In Pitkänen M & J Tiainen (eds). Biodiversity of agricultural land- scapes in Finland. Birdlife Finland Conservation Series 3, 33−50. TIKE (2002). Agricultural census 2000. 275 p. Maa- ja metsätalousministeriön tietopalvelukeskus – In- formation Centre of the Ministry of Agriculture and Forestry, Helsinki. Tischendorf L & L Fahrig (2000). On the use and measurement of landscape connectivity. Oikos 90, 7−19. Tuhkanen S (1984). A circumboreal system of cli- matic-phytogeographical regions. Acta Botanica Fennica 127, 1−50. Vainio M, H Kekäläinen, A Alanen & J Pykälä (2001). Suomen perinnebiotoopit. Perinnemaisemapro- jektin valtakunnallinen loppuraportti. The Finnish Environment 527. 163 p. Väisänen RA, E Lammi & P Koskimies (1998). Muut- tuva pesimälinnustomme. 567 p. Otava, Keuruu. Valle O (1951). Maatalous. In Granö JG, R Jurva, J Ke- ränen, V Kujala, A Laitakari, U Pesonen, P Kalaja & E Kanervo (eds). Suomen maantieteen käsikir- ja, 404−427. Geographical Society of Finland & Otava, Helsinki. van Swaay CAM (2002). The importance of calcare- ous grasslands for butterflies in Europe. Biological Conservation 104, 315−318. Vickery JA, JR Tallowin, R Feber, EJ Asteraki, PW At- kinson, RJ Fuller & VK Brown (2001). The man- agement of lowland neutral grasslands in Britain: effects of agricultural practices on birds and their food resources. Journal of Applied Ecology 38, 647−664. Virkkala R, M Luoto & K Rainio (2004). Effects of landscape composition on farmland and red-list- ed birds in agricultural-forest mosaics. Ecography 27, 1−12. Vuorela I (1999). Viljelytoiminnan alku Suomessa paleoekologisen tutkimuksen kohteena. In Fogel- berg P (ed). Pohjan poluilla. Suomalaisten juuret nykytutkimuksen mukaan, 143−151. Suomen tie- deseura, Helsinki. Waldhart R (2003). Biodiversity and landscape − summary, conclusions and perspectives. Agricul- ture, Ecosystems and Environment 98, 305−309. Waldhart R & A Otte (2003). Indicators of plant species and community diversity in grasslands. Agriculture, Ecosystems and Environment 98, 339−351. Waldhart R, D Simmering & H Albrecht (2003). Flo- ristic diversity at the habitat scale in agricultural landscapes of Central Europe − summary, conclu- sions and perspectives. Agriculture, Ecosystems and Environment 98, 79−85. Weiers S, M Boch, M Wissen & G Rossner (2004). Mapping and indicator approaches for the assess- ment of habitats at different scales using remote sensing and GIS methods. Landscape and Urban Planning 67, 43−65. Wiens JA (1989). Spatial scaling in ecology. Func- tional Ecology 3, 385−397.