Management of Cultural Landscapes: What Does this Mean in the Former Soviet Union? A Case Study from Latvia

Abstract Concern about changing cultural landscapes has increased recently, with the advent of the European Landscape Convention placing signatory countries in a position of having to develop action for protecting and managing cultural landscapes. In countries of the former Soviet Union the landscape underwent many changes as a result of agricultural collectivisation and its aftermath. This situation has been analysed for six sample rural municipalities (pagasts) in Latvia, one of the three former Soviet countries to join the European Union (EU), using maps from the period 1901 to 1927 (to represent the ‘traditional landscape’) and 1997 orthophotographs updated to 2000 (to represent the ‘post-Soviet landscape’), abandoned agricultural land to 2007 to show the continuing aftermath of the Soviet system and field assessment of their character. It was found that all sampled pagasts had experienced significant landscape change during the Soviet times that replaced the pre-Soviet, traditional character with a new ‘ideological landscape’. The implications for the protection and conservation of such landscapes created by a previous foreign occupying power are many, raising questions of what landscapes or elements to conserve under the requirements of the Convention.


Introduction
Concern about vanishing cultural landscapes and new emerging landscapes has increased over recent years (Antrop, 2005). Many people see landscape change as a threat for a number of reasons, for example, the loss of identity of the existing landscape. However, as Antrop also notes, landscapes are always subject to change and as such they express the way that natural and cultural processes interact with one another. The protection, management and conservation of cultural landscapes has also become more important as a result of the European Landscape Convention coming into force (Council of Europe 2000), with signatory countries committed to the development of concomitant programmes. The definition of the landscape within the Convention is ''. . . an area, as perceived by people, whose character is the result of the action and interaction of natural and/or human factors' '. Under Resolution 128 (2002) of the Council of Europe, on the problems of Europe's countryside, the Council noted, amongst other things, that: Europe's countryside, and the people who live in it, are a highly valued and varied asset for the whole population of the continent: the largest part of rural Europe is covered by agricultural land and forests, which have a strong influence on the character of European landscapes; the great diversity of nature and of human culture enriches the quality of life for all Europeans. It is our duty to understand, protect and enhance this heritage; at present, in many parts of Europe, the rural heritage is being rapidly eroded and even destroyed by social or technological changes, modern agriculture, urban growth, neglect and other forces. (Council of Europe, 2002) As noted by Nikodemus et al. (2005), Europe experienced rapid social transformations over the course of the twentieth century, which present a range of threats to the continuing presence of what might be described as the traditional European cultural and historical rural landscape. The European Environment report (European Environmental Agency, 1995) identities seven factors currently having an impact on the rural landscape, though to different degrees in different places: . Intensification of agriculture; . Overgrowth of agricultural lands; . Urbanisation and development of infrastructure; . Standardisation of building materials and designs; . Tourism and recreation; . Excavation of mineral resources and establishment of land-fill sites; . Disappearance of natural biotopes, habitats and ecosystems.
It is the social and economic impacts that often determine the types of land use within a given region and which in turn affect environmental aspects (Ojima et al., 1993;Mander & Palang, 1994;Melluma, 1994) including the landscape in the sense of the European Landscape Convention definition quoted above. In the former Soviet Union, and other Central and Eastern European (CEE) countries, centralised land planning systems were more important than in other countries and have had a marked effect on the way that the countryside has developed since, as a result of the legacy of land ownership, land use, infrastructure and communications. The three countries of Estonia, Latvia and Lithuania are the only former countries of the Soviet Union to join the EU, although the Council of Europe (responsible for the European Landscape Convention) also covers other former Soviet Union Countries including Belarus, Ukraine, the Russian Federation (the European part) and the Caucasus states. This paper will explore the landscape that resulted from changes that took place between the period 1901-1927, from when baseline data are available, representing the 'traditional landscape', and 2000, concentrating on the Soviet period, 1945Soviet period, -1991 representing the 'post-Soviet landscape', in a case study of selected rural municipalities (known as 'pagasts') from Latvia. While land abandonment is a trend that started after Latvia regained independence in 1991 (see below), in terms of the overall landscape structure and the implications for what can be termed the cultural landscape of Latvia, it is the impact of the period of the so-called Soviet Occupation which in many ways remains a driver of the pattern and structure of the cultural landscape of Latvia's countryside. The resulting landscape and its continuing change presents an interesting challenge for implementation of the European Landscape Convention, to which Latvia is a signatory.
During the twentieth century Latvia experienced several major transformations in land use. In 1935, agricultural land covered 57.3% and forests occupied 26.6% of the land area. By 2000, the agricultural area was only 38.5%, at the expense of which woodland had almost doubled in area, then occupying 44.4% of the territory (Ministry of Agriculture, 2001;Latvian Environmental Agency, 2002). This period covers several major upheavals, including the Second World War, when Latvia was invaded several times and fought over with damaging consequences for the landscape and population (around a third of the population died, were exiled or fled abroad during the war period and many houses were destroyed in the battles). Thus, the current changes, visible and ongoing as they are, have to be seen in the context of the much more considerable changes which happened, more in some areas than others, as a result of pre-war activities, wartime depredations and Soviet land use policies, the last of which in particular continue to affect the character and appearance of the landscape. The broad statistics stated above do not reflect the diversity of landscape, nor the different degrees to which agriculture changed over the whole period, including the major era of the Soviet period, in different places and as a result affected and in turn helped to shape the current character of the cultural landscapes.
For the purposes of this study, landscape character is defined as a ''distinct and recognisable pattern of elements that occur consistently in a particular type of landscape'' (Countryside Agency and Scottish Natural Heritage, 2002). Character makes each part of the landscape distinct, and gives each its particular sense of place. In Latvia no comprehensive assessment of landscape character yet exists, although a simplified version based on dividing the landscape into zones on the basis of topography and land cover has been adapted for guiding forestry planning (Bell & Nikodemus, 2000). As well as topography (including geology), land use pattern (land use types and configuration, ecology, etc.), landscape scale, settlement pattern (traditional house types and location patterns), the cultural landscape character is informed by historical aspects, communication patterns and key features such as churches, castles, unique landform or historical events and persons. The condition of the landscape is also important because this can often determine trends for the future and the need for conservation. This definition forms the basis for an analysis of the Latvian landscape of the sample pagasts in this paper.
The main ways that the Latvian landscape changed under the Soviet system are as follows: . Farms were nationalised and amalgamated into one of two varieties of collective farms (kolkhoz or sovkhoz, terms being short for kollektivnoe khoziaistvo or

Management of Cultural Landscapes 427
sovietskoe khoziaistvo). A kolkhoz was collectively run by the members, whereas a sovkhoz was centrally administered and the workers were not members. In practice these were bureaucratic differences and made no difference to the way they were managed or the resulting landscape. They were run as a business and all the former landowners (who were not deported to Siberia as 'kulaks') became members (kolkhoz) or workers (sovkhoz). . The previous dispersed settlement pattern was changed as people were moved into blocks of flats constructed in the new village centres. . Large production facilities were constructed in the centres, such as barns, heating plants for the houses and flats, grain silos, intensive pig sheds, dairy facilities, machine tractor stations and storage units. . Land capable of being drained or improved ('ameliorated') by drainage or levelling to allow large machines to operate was brought into production in large contiguous fields that ignored the original field patterns or ownership boundaries. . Land deemed marginal and inefficient for mechanised agriculture was left uncultivated and allowed to become colonised with forest. . Old houses of former land owners were in many cases left empty, some were demolished to make way for large fields while others remained in use, perhaps as storage or where some people continued to live, sometimes several families being moved into what were originally single family houses. . Some forest areas were also drained and improved in terms of productivity.
This can be clearly seen as a large-scale and pervasive restructuring of the land and landscape in order to meet an ideological requirement, centrally planned and carried out by an administrative process (Melluma, 1994).
Subsequent to the collapse of Communism and the restitution of land to the original landowners, many of whom are old, non-resident or not interested in agriculture, land abandonment started and has continued to the present day. This is manifested by scrub growing on many fields, a sign of their eventual transformation into forest. Of the seven factors named in the European Environment report noted above, the main one affecting Latvia is overgrowth of agricultural lands. This is in marked contrast to many other countries in Western Europe, for example, although it also occurs in Eastern European countries and some, such as Portugal, in Western Europe (Nikodemus et al., in press). Urbanisation and the development of infrastructure are also occurring but are comparatively limited in extent to date. This land abandonment is one of the main aftermaths of the Soviet era and in many ways is one of its enduring legacies to the landscape.
Furthermore, in relation to one specific cultural aspect of landscape character, that of settlement pattern, a study in the 1980s of the Latvian settlement pattern types (Sˇteins, 1986) showed that there were originally a number of distinct traditional regional settlement patterns which became altered as a result of the Soviet changes. These can be broadly classified as follows: 1. Homogeneous dispersed farmsteads, found in open and mainly arable plains in Zemgale and Kurzeme; 2. Areas of dispersed farmsteads with occasional linear concentrations along valleys, for example, found locally around the town of Al uksne in northern Latvia; 428 S. Bell et al. 3. Most farmsteads in small hamlets with a few scattered ones in between, found in the more forested landscapes of northwest Vidzeme; 4. Almost all farmsteads in small hamlets, found in several locations; 5. Clusters of houses in hamlets, found along the coastal areas; 6. Linear settlements along roads.
The implications of these changes to settlement pattern will also be explored in the paper.
The European Landscape Convention requires signatories to protect and manage the cultural landscape, which means conserving and keeping up the characteristic features as justified by its heritage value. However, what exactly does this mean in the context of landscapes that were mainly formed during the time of occupation by another power? If many of the built structures from such a period are derelict or destroyed, and if many of the landscape elements which are, arguably, more traditional and of more importance culturally are also in a poor condition or were destroyed, what does conservation and management mean? Moreover, in situations where the population do not identify with the Soviet landscape and where a sense of national identity is associated with the older, traditional elements of the landscape (Bell et al., 2008), what are the options for landscape policies? Furthermore, since the Soviet era is a historical reality and part of the cultural history of the country and landscape, regardless of how people perceive it, how should it be taken into account?
In order to explore the range of aspects described in the introduction and to assess the deeper aspects of these landscape changes and the processes and influences that are most significant, research using several complementary approaches was carried out and is described in this paper. In order to delve more deeply into the subject and to compare how the changes have manifested themselves in different ways in the regions, research into the changes that have taken place in the Latvian landscape and its character was conducted in six pagasts. The research questions are as follows: 1. What has been the impact of changes to the land use and landscape character of Latvia as a result of the Soviet land use planning system? 2. How does the aftermath of these changes continue to affect the development of the landscape? 3. How does the understanding of this situation help in defining goals for landscape protection and management under the requirements of the European Landscape Convention?

Materials and Methods
A small sample of Latvian pagasts was selected for study. These were located in each of the historical divisions of the territory that is now Latvia (Kurzeme, in the west, Vidzeme in the north and centre, Zemgale, in the south and Latgale in the east), and selected to reflect the main differences in land use and soil quality, topography and landscape character. The selection criteria were limited by the availability of data on historical land use, which is not available comprehensively across the country. The locations were, therefore, chosen because good data from the period 1901 to around 1927 were available (topographic maps and archival materials still extant in the Management of Cultural Landscapes 429 municipalities) showing the land use before the land reforms of the first period of independence, before numerous new farmsteads were constructed and before the Second World War and the subsequent Soviet Occupation. Six locations were chosen, two in Vidzeme (Dzerbene and Vecpiebalga), one in Zemgale (Vecsaule), two in Kurzeme (Barta and Priekule) and one in Latgale (Nautreni). Table 1 summarises the main features of the samples. Figure 1 shows the location of the study areas.
The study of landscape involves not only the physical aspects of the land use and settlement which can be identified from maps but also the underlying processes that created it (and continue to affect it) and its perceptual aspects (see the definition of 'landscape' from the European Landscape Convention as noted in the Introduction). Therefore the research used both map-based and visual-description-based approaches to analyse the landscapes under investigation. The understanding of landscape character flows logically from the analysis of changes to land use and settlement patterns-how these changed landscapes are experienced on the ground.
Using the ARCView Geographical Information System (GIS), datasets were compiled from two sources: historic maps from the early twentieth century (supplemented by archival material such as local maps and survey reports) to provide a picture of the landscape before the First World War and aerial orthophotographs from the late twentieth century to show the post-Soviet land use and settlement (updated by field survey to show the current extent of the process of land abandonment).
The historic maps were produced by the then Russian Imperial cartography service with some updating in the time of the First Latvian Independence period. This area is the second example from Kurzeme and is different because it has a town, Priekule, located in the middle of the pagast. The natural conditions are similar to those as Barta, except that the very eastern part falls into the so-called Rietumkursa upland. Nautreni Lies in the east of Latvia, the area known as Latgale. This area traditionally has a different character owing to its history which included a period under Polish rule and the scale of land use has always been smaller. The southern part of the area is hilly with poorer soils, lying in the Latgale upland while the northern part is flatter and more fertile and lying in the Austrumlatvijas lowlands.
They were of a scale of 1:75 000 dating from 1901 to 1927, depending on the mapping date (there is only one set of maps for each area for the early twentieth century, with a wide range of survey dates), supplemented with some local data available for each pagast, mainly local maps and some survey reports) to provide more detailed information about land use at that time. Table 2 shows the dates of the maps used for each of the sample pagasts. Frequently, the maps were only updated to show additional roads in the 1920s but the basic survey was of the early 1900s, except for Latgale, where a new survey was undertaken in the early 1920s. Although new  1916 and 1925 1927 The northern section is the newer map, the southern section was updated to show additional roads.

Management of Cultural Landscapes 431
farmsteads were built after the land reforms of 1920 none were included in the updated surveys as few if any would have been constructed by the time the survey field work was carried out. Thus, for all practical purposes the maps show the situation of land use and settlements before the First World War, if not earlier, and will be referred to as the '1900s' era maps. Aerial orthophotographs from 1997 were used to create the maps of land use in the post-Soviet era. Some of the changes in the landscape since 1991, such as some land abandonment, could be identified from the aerial photographs or from pagasts surveys, by the age of the vegetation structure; for example, scrub and young trees growing on fields show land that was abandoned 10 years ago at the most. Therefore, the picture is more accurately that of the landscape emerging from the Soviet era, from which it may be possible to interpret trends resulting from its aftermath. Field surveys in 2007 allowed the state of land abandonment to be updated and recent statistics verify that the process continues (Nikodemus et al., in press). The updating of the land abandonment data helps to show how the trends of land use change are continuing, rather than presenting a study that is already historical. It was not possible to identify from these maps what changes had occurred before or during the Second World War. However, from the knowledge of the type of changes undertaken during the Soviet era (as noted above) and their scale, it was fair to assume that most were from this time, apart from the land abandonment evidenced by the scrub growth. The exception is the housing pattern, as some of the farmsteads built in the 1920s and 1930s survived the war and Soviet era and therefore appear on the later maps. This fact presented difficulties for analysis of the changes to settlement (see Results section below). The maps therefore represent the basic land use pattern as it emerged from the Soviet era with land abandonment representing the main changes that have occurred since. They will be referred to as the '2000s era maps'.
The spatial data recorded in the GIS included forest cover (no information on forest type or age was available from the 1900s), fens (which may also include bogs and mires), houses/farmsteads, roads, rivers and streams, ditches (recording land amelioration) and features such as collective farm centres, quarry pits and other elements. The remaining land was identified as agricultural land but not divided into pasture, meadow or arable types as no specific data from the pre-First World War period were available. Two comparative datasets were therefore prepared at the same scale and using the same classifications, one for the 1900s era and one for the 2000s era. The maps for the 2000s era also include the abandoned land surveyed in 2007. The areas are identified as percentages of the agricultural land, a class to which they still technically belong rather than being identified as a separate class. This enables the 2000 land cover and settlement pattern to be seen and analysed separately from the issues of abandonment (Figures 2-7).
Although both sets of maps show houses/farmsteads (divided into active farms and smaller farmsteads in the 2000s era maps, in order to indicate the extent of commercial, as opposed to semi-subsistence farming and its correlation to land abandonment) this is more indicative of the pattern because the earlier maps do not clearly show the function and status of all buildings marked on them. Therefore, the farmsteads and houses marked on the maps are mainly used to show the changes in the pattern and distribution of settlement rather than an analysis of numerical

Management of Cultural Landscapes 433
changes. The Encyclopaedia of Pagasts (Anon, 2001(Anon, , 2002 describes the main factors that affected the settlement of each area and have been used to help understand the character and significance of the changes visible on the maps.

Management of Cultural Landscapes 435
In order to collect data on the landscape character and the condition of the elements of the landscape each of the sample pagasts was visited by car to obtain a sense of its character, and a number of photographs and notes taken of the main

Management of Cultural Landscapes 437
features of each, in relation to the evidence presented by the 2000s era maps. As the visits took place in 2003 the contemporary features observed on the ground reflect those extant at that date. However, the key aspects noted above were readily visible and many of the major changes such as abandonment of the collective farm centres had already taken place well before 2003, shortly after the collapse of Communism. Although now some six years old, recent updates of data on land abandonment show that the character remains the same and that the trend for land abandonment continues (see also Nikodemus et al., in press). The landscape character survey recorded how the landscape appeared, using the criteria noted in the introduction: landform, land use pattern, landscape scale, settlement pattern, land use condition and key features. There is clearly a correlation between the mapped information and what is visible on the ground. The landscape character information however, provides more on the perceptual or experiential aspects and reveals more than maps alone.

Results
The results are presented in a number of summary tables that allow comparison between the pre-Soviet and post-Soviet situations in each of the sampled pagasts. This is followed by the results of the landscape character surveys to give a more indepth interpretation of character. Table 3 presents a summary of the changes from one land use class to another, expressed as percentages in the 1900s era and then in the 2000s era, followed by the overall percentage change of each land use type over the period. Figures 2-7 show the maps of land use for each of the periods. This shows a range of scales of land use change, depending on a number of factors, such as the initial landscape conditions. The key reasons for many of the changes are linked to the political changes as well as the agricultural planning of the collectivisation period. All these aspects are intimately connected as part of the larger Soviet political project (see below).

Land Use Change in the Sample Pagasts
The figures presented in Table 3 describe the quantitative aspects of the changes, and summarises the significant aspects of these patterns for each sample area, showing how the land use changes are, in many cases, very dramatic and, as will be seen below, have also changed the landscape character. The maps, Figures 2-7, present the visual pattern of these changes.

Settlement Pattern Changes
Settlement patterns are a key aspect of the cultural landscape, representing many social and cultural aspects of an inhabited place. One of the major changes to occur during the Soviet period, as noted in the introduction, resulted from collectivisation and was the rationalisation and centralisation of dwellings into blocks of flats in new village centres. The influence of this can be seen quite clearly. There were also other reasons for the loss of houses and farmsteads in some areas, particularly war damage, deportation of the population and the large-scale land amelioration programmes.
From the analysis of the land use changes summarised in Table 3 above, it will be seen that each area has fared differently under the impact of the Soviet planning system at the time of the collective farms. The reasons for this are partly because while all planning was done centrally according to pre-established targets (Melluma, 1994), other factors also influenced the degree to which standardisation was possible. Table 4 compares some of the major influences on agriculture and the possible

73.76
A much smaller area of forestry in the 1900s than for other sample areas, and much smaller increase in forest area. Significant areas of fen were drained, mainly for more farmland but also for more forest in places. In the 1900s forest was mainly in concentrated areas. By 2000 these areas had expanded with some additional areas around the fen margins. Forest had more than doubled in area, reduction in agriculture from land changing to forest being, in part, compensated by the drainage of the fens into highly productive arable land. Example of major land rationalisation under the collective farm system.  reasons that led to the changes in land use. It takes the main aspects that influence agricultural production and sets out the scale and nature of the changes relating to a range of factors. The results from this comparison tend to suggest that the driving force for increased productivity was mainly based on the capacity for production of the soil, the suitability of the topography for large-scale operations, and the potential to increase available agricultural land through large-scale amelioration. Thus, those flatter areas with potentially fertile fens, such as Vecsaule and Nautreni, saw the largest scale amelioration based largely on the drainage of these fens. Soil does not vary much between sample areas and it is noticeable that those areas that saw largescale drainage of fens also saw some of the poorer soils converted to forest and the newly drained areas replacing and substituting for the land that went to forest. No afforestation programme was carried out in Soviet times, so that the increase in forest is all as a result of abandonment, which was clearly a force for change during that period as well as at the present time. In some places wet forest was also drained to increase timber productivity. What is also noticeable is that forest expanded everywhere but it did so most in the areas with little scope for amelioration and large-scale mechanised agriculture, although topography is not an especially significant factor (Barta is quite flat but saw a similar drop in agricultural land to Dzerbene, which is hilly).

Management of Cultural Landscapes 441
This degree of land use change is only possible at the scale demonstrated here if the land can be treated as a single unit. It is not so easy when the land is owned and managed by many small farmers-cooperation is more difficult. The same applies to rationalisation and movement of settlements, afforestation and reorganisation of roads. The starting point of the land use pattern also affected subsequent patterns: for example, Dzerbene was already heavily forested and Barta already had a large area under forest, restricting to some extent the amount of change that was possible. Once the land was handed back to the original owners and a smaller-scale farm unit pattern was re-established, the centrally organised elements such as drainage systems, which require maintenance, started to break down and individual land use decisions such as maintaining cultivation or allowing abandonment became more important, as seen in the patterns of abandonment on the maps from the 2000s, which are in small scattered patches, though often more associated with already heavily forested and enclosed areas than large-scale open ones. Another study (Nikodemus et al., in press) shows that the reasons for abandonment are not related to soil types but are largely due to social factors, such as the age of the landowners. The 2000s maps showing 'active farms' demonstrate that these are not in places where abandonment has been significant. There are large numbers of farms in Latvia classed as 'semi-subsistence' and which do not produce anything for the market.
No clear conclusions can be made concerning the picture gained from the changes to settlements in the sample areas and such conclusions only tell part of the story as far as the effects on landscape character are concerned. First, as already noted, the number of dwellings (whether active farms, farmsteads, houses or flats) does not tell us anything about the type of building, whether significant numbers of older traditional houses survived, how many new farms were built in the 1920s and 1930s or how many were replaced with newer versions on the same site. The known pattern of change to settlements was broadly that, during the 1950s, farming activities were moved to the centres of the newly established collective farms (Nikodemus et al., 2005). This process was intensified after the 'Resolution of 1961' on the displacement of inhabitants to villages and multi-storied blocks of flats and manufacturing units. Urban building types such as blocks of flats began to appear. The construction of central villages and the elimination of individual farmsteads continued until the 1980s (L use & J akobsone, 1990;Gr ave & L use, 1990;Riekstin sˇ, 2001).
The building of new residential areas comprising multi-storied blocks of flats and so-called 'L ıv ani-type' detached houses (a form of standard design) tended to take place only in the centre of each pagast and on the bigger collective farms, if there was more than one per pagast (they tended to be amalgamated over time). To assess the stock of older, traditional farmsteads and other building remains that might be worth conserving it would be necessary to conduct site surveys. Such surveys have not so far been undertaken. However, the sample areas were visited and some information is available for them (see below). As a result of the changes to the settlements it can be observed that the traditional types no longer exist in any of the studied pagasts. Table 5 shows how the settlement patterns have changed with respect to the types categorised by Sˇteins (see the Introduction). Table 6 presents a summary comparison of the main features observed during the surveys to assess landscape character. Figures 8a-f show typical views of the landscapes of these study areas.

Landscape Character Surveys
The analysis shows that the current landscape character is dominated by the changes that took place during Soviet times that have in most cases created a different pattern and scale from the landscape of the 1900s. This landscape is in some cases much more enclosed, with fewer open views than would have been possible before. Most of the wet fen areas have been drained, removing a significant element The amelioration moved houses from the clusters and concentrated them along the roads Barta 1 6 The dispersed pattern with no centre was replaced by linear clusters and new villages Priekule 1 6 As for Barta but the town also has a lot of influence on the pattern Nautreni 3 2 The clusters became more linear in pattern and more dispersed farmsteads emerged  from the landscape of those areas where they were extensive in the 1900s era, such as Vecsaule and Nautreni. The landscape also contains the buildings of the collective farms, in some cases located in dominant positions in the landscape, such as at Vecpiebalga (Figure 8b). In some ways the landscape has become more diverse, with more patches of forest in places such as at Vecpiebalga or Nautreni. In terms of condition, in all areas there are derelict collective farm buildings and houses in varying states of repair. This all suggests that there is no such thing remaining as a Latvian traditional landscape in these areas, assuming that the scene pre-war still represented a traditional type as a result of the land management and tenure system of manor estates with small tenant farms that had lasted for several centuries, allowing for the addition of the new farmsteads in the 1920s and 1930s. The period between 1991 and 2007 shows that land abandonment is also occurring as evidenced by the area of scrubby bushes recorded in the data and visible in the photographs. Some of these scrub areas are also concentrated along the lines of the drainage ditches, suggesting that maintenance of the drainage systems may not be kept up, leading to eventual water-logging and loss of the land to production. The condition of the remaining older landscape elements is also poor, suggesting that the losses of the remaining traditional elements are continuing.
This character as observed on the ground is the reality of the landscape today and there will probably be few people alive who remember what it was really like in prewar years. The changes wrought by the Soviet era are a reality and many of the elements have been there for several decades. The implications of this for landscape management and conservation, especially in the context of the European Landscape Convention is discussed below.

Discussion
Latvian people have a strong sense of what elements constitute a typical or traditional Latvian landscape (Bell et al., 2008). Separate farmsteads, orchards, lines of oak or lime trees, ponds and bath houses (saunas) and storks nests feature in this list. It is perfectly possible to find many examples of these elements but from the analysis of landscape change it is difficult to conclude that they constitute a traditional landscape. These elements have to be set into the context of fields and forests and settlement patterns which, as we have seen, are now very different from how they appeared in the 1900s era. So, while there may be small-scale locations where the landscape has changed little and where it is possible to find the archetypal farmstead with its associated features, extensive cultural landscapes tend to reflect the Soviet character overlaid by recent abandonment.
The same study found that rural inhabitants also identified with the current reality of the Latvian landscape, naming abandoned fields and derelict buildings among other aspects they saw as being associated with living in the country. However, it was also clear that only the oldest people could claim to remember the landscape from before the Second World War and that the younger generations who only really remember the post-Soviet times accept the landscape as it is and do not perceive it as Soviet, apart perhaps from the sight of ruined collective farm buildings and the flats where many of them live. The lower intensity of management has also tended to allow a more natural-appearing landscape to develop, although, in the 2007 study noted above, many older people preferred the Soviet era landscape because it was tidier and better managed. The sight of abandonment is not liked by most people, especially other farmers living in the countryside.
According to Antrop (2005), history is full of examples of major landscape changes, most of which were the result of initiatives by landowners or other powers at the time. A prime example of this is the enclosure movement of England and parts of northwest Europe in the eighteenth and nineteenth centuries that transformed large-scale open landscapes into geometric hedged fields. In so doing a large proportion of the rural population were displaced, causing many people to move to the towns and cities (Hoskins, 2005;Rackham, 2000). Thus, a dramatic change such as witnessed by collectivisation is another example of major landscape change undertaken by the ruling power of the time. Antrop (2005) goes on to identify three main periods of landscape development in Europe. The first of these is the traditional landscape, consisting of pre-eighteenthcentury remnants going back to a remote past. The second type is the landscape of the revolutionary age, mainly the nineteenth-century agricultural revolution and industrialisation leading up to the Second World War. The third type is the postmodern landscape of the post-Second World War period to the present day that results from globalisation and urbanisation. This classification begs the question: where does the Soviet collectivisation era fit?
It is possible to see some elements of the pre-Soviet Latvian landscape as falling into the type of the revolutionary age landscape, with the influence of the landowners on agricultural practices, at least in the areas of better land (Anon., 2001(Anon., , 2002. The land abandonment and recent changes from 1991 clearly fall into the classification of post-modern landscapes, with the farmers coming under the influence, first, of the open markets once the Soviet Union collapsed and, more recently (post-2004), within the Common Agricultural Policy (CAP) of the EU (identified by Antrop as an example of the forces of globalisation). However, the collectivisation, rationalisation and amelioration processes analysed in this paper are not revolutionary, in the sense defined by Antrop, nor post-modern. It is necessary, for this part of Europe, to introduce a new category of 'Ideological Landscapes', defined by the combination of state planning control, nationalisation of land, social control and classification of the population (into kulaks and non-kulaks) driven more by ideology than a need to improve agricultural output (collective farms were notoriously inefficient across most of the Soviet Union).
Such a type of landscape can be found in other countries which were subjected to the Soviet collectivisation system, such as East Germany, the former Czechoslovakia, though not to Poland, which retained its small-scale farms, at least in the original territory; collective farms were established in the lands taken from Germany after the Second World War. Nor are ideological landscapes necessarily a feature solely of communist systems. They have arguably existed for many centuries (Mitchell, 2002;Warnke & McLintock, 1994). In Germany in the 1930s and the early part of the Second World War it was the aim of the Nazi regime to redesign the conquered territories of Poland and other lands into model landscapes suitable for Aryan Nazi settlers (Lekan & Zeller, 2005). It is even possible to consider the clearances of people from the Scottish Highlands in the eighteenth and nineteenth centuries (Prebble, 1965) to make way for economic sheep production as a product of an ideology (capitalism perhaps?) and for the survey grid of the American West (the so-called 'Jeffersonian Grid') and its use as a vehicle for giving land to settlers as part of the ideology of 'Manifest Destiny' promoted by President Andrew Jackson Management of Cultural Landscapes 451 (Cronin, 1983). It can be argued that these landscapes are as much a part of historical reality as traditional vernacular landscapes which, in many places, were not produced by free peasants but were part of a feudal or estate landscape affected by factors such as serfdom, agricultural improvements and the power of the estate owner or landlord. This is the case for many landscapes held in high regard today, such as the hedgerow landscapes of England noted above, which were also the product of an ideology of land-owning capitalism and the abolition of common land in the interests of private property (Rackham, 2000).
If the post-1991 situation falls under the type of post-modern landscape, then it is a kind that has started to develop over 40 years later than in Western Europe. However, it fits into one of the four trends identified by Klijn and Vos (2000). These are: intensification and at the same time an increase in the scale of agriculture (which may occur in the large-scale arable areas such as at Vecsaule if the land is bought by entrepreneurial farmers, but is otherwise unlikely); urban sprawl (not an issue in the areas studied but possible around the capital, Riga); tourism development based on the quiet, natural landscape, wildlife, and charming though not dramatic scenery (country cottages and saunas by lakes, for example); and land abandonment and extensification in remote rural areas (such as in Barta and Priekule, where the abandonment is most significant). It could be argued that intensification and increase in scale of agriculture was what took place in the collectivisation period, and that this has gone into reverse with the restitution of land to the previous owners or their descendents and the restoration of small-scale semi-subsistence farming (seen in the maps of the 2000s era by the small numbers of 'active farms' compared with the farmstead numbers as a whole).
It is not necessarily only the loss of the traditional landscape that is an issue. Many of the traditions and practices have gone with it. In a study in Estonia, Latvia's northern neighbour, where the same Soviet system was applied, Kaur et al. (2004) noted that the 50 years of the landscape of the collective farms put pressure on traditional practices that resulted in 'uninformed' landscape patterns and has tended to undermine rural identity. They found that many farmers who reclaimed the land after the restoration of the Estonian Republic failed to meet the economic challenges that led many to cease farming and then abandon the land. This abandonment not only caused traditional practices to disappear, such as cattle husbandry or haymaking, but has caused an identity crisis among rural people (Kaur et al., 2004). This research has been limited by data availability and only looked at a small number of pagasts in Latvia. A time step of only two periods, approximately 90 years apart is also a potential problem for analysis. Studies of landscape change are always dependent on the availability of data. In a study from Switzerland, Schneeberger et al. (2007) were able to look at 11 time steps for an area between 1885 and 1996. The availability of such data allows for the identification of the precise timing of step changes in the landscape. In another study of rural France, however, Eetvelde and Antrop (2004) looked at two time steps of sample sections from two sets of aerial photographs, although these were only 30 years apart. Frustratingly, for Latvia a third set of data exists which would bridge the wide gap between the 1900s and 2000 with a snapshot of around 1970-old Soviet army maps. These maps are highly detailed (1:10 000 scale) but resources were not available to digitise them at the time of this research project. They would however be very useful additional layers in any further studies of the Latvian landscape.
Methodological and data issues notwithstanding, some significant aspects of the Latvian cultural landscape have been revealed. The results suggest that there is no longer any extensive traditional cultural landscape left from before the Soviet era and that, whilst smaller relicts may survive, the dominant character is formed from the activities of the collectivisation period. Furthermore, the aftermath of this, including the difficulties faced by the landowners after restitution of the land and the resulting continuing process of abandonment in many areas, is not only part of the effects of globalisation but also a direct consequence of the collectivisation era. This leads to the scenario of extensification and abandonment identified by Klijn and Vos above which appears to be a continuing trend according to the most recent data (Nikodemus et al., in press). Antrop (2005) considers how to control changes in the landscape. He notes that this is difficult because the landscape consists of many individual pieces of property owned by people with different interests. Landscape therefore includes many parcels owned by different people but is, itself, a common asset not limited by property boundaries. If each owner has a degree of freedom over how they manage their land, this can lead to chaotic development. In the countryside (as opposed to urban areas, where there is usually more control over the main aspects of built development) planning and management of the landscape is mainly possible through the use of instruments to promote different land management practices, such as agrienvironment schemes under the CAP. The current instruments are the Single Area Payment and the Less-Favoured Area Payment schemes, which are paid to farmers who keep land in a cultivated condition (including cutting the grass on meadows and pasture to prevent scrub forming). A recent study addressed the success of these schemes in halting or reversing land abandonment in Latvia, using a selection of pagasts, including Vecsaule and Nautreni. It found that so far the schemes, introduced in 2004 after Latvia joined the EU, are having little effect, largely due to the attitudes of and level of support given to older, smaller land owners (also mostly non-active farmers) whose land is most likely to be undergoing abandonment (Nikodemus et al., in press).

Conclusions
This paper aims to address three research questions: . What has been the impact of changes to the land use and landscape of Latvia as a result of the Soviet land use planning system? From the research on six sample pagasts the evidence suggests that land use, in terms of the proportions of farmland and forest as well as other elements such as fens and settlements, has changed greatly in all areas, although more in some than others. The landscape, in terms of the character of the areas, has also changed considerably, with many features having been lost (old houses) or added (collective farm buildings). The landscape is largely a product of this era, though traditional elements of a small scale remain, such as old houses with orchards and storks nests.

Management of Cultural Landscapes 453
. How does the aftermath of these changes continue to affect the development of the landscape? The aftermath includes the upheavals caused, in part, by restitution of land to the former owners, and a delayed exposure of the agricultural sector to the forces of post-modernism (globalisation and EU membership) caused by the isolation of the Soviet system from the global economy until 1991. This has already led to the dereliction of many of the common assets of the collective farms, which remain as rusting monuments to the collective farm era, and is also resulting in land abandonment and marginalisation of remoter areas. This continues to the present day. . How does the understanding of this situation help in defining goals for landscape protection and management under the requirements of the European Landscape Convention? If the landscape is a product of the Soviet system, which, as noted in the introduction, was a system imposed by a foreign occupying power, then apart from this era being understood and recognised as a historical reality, is it appropriate to conserve it in the same way that hedgerow landscapes in England, for example, might be conserved, restored or managed? As part of the historical reality and an inescapable part of the layers that make up the landscape this probably cannot be avoided. Indeed, the conservation of at least some representative examples of collective farm building complexes should also be considered. As argued above, the landscape is similar in some respects to the other ideological landscapes of which the hedgerow landscape of parts of England is arguably also an example. However, no systematic inventory of any aspect of the cultural landscape of Latvia has yet been undertaken and no criteria for what aspects of the landscape or built heritage of whatever era should be conserved have been established.
Under the European Landscape Convention some kind of landscape character assessment is probably necessary if the conditions are to be fulfilled. In the case of the more valued traditional landscapes, as part of the palimpsest of the broader Latvian landscape, a different approach is probably needed, based on a broader survey of character and an identification of where and to what extent traditional landscapes remain and can be conserved or restored. Certainly, the broad pattern of land use and the scale of the landscape cannot be put back to what it was before the Second World War, so that it has to be accepted as being the current cultural landscape. The rate of change through abandonment is perhaps a different issue, where it might be feasible to reduce the rate of further losses due to land abandonment, through the more targeted and supported use of agri-environment programmes such as the Single Area Payment scheme, and to help rural inhabitants to conserve and restore traditional farmsteads and associated features, especially the oldest examples.