Interpretation-based land cover mapping. Possibilities and challenges for rural land management
MetadataVis full innførsel
- Department of Geography 
Land cover/use maps are often seen as a prerequisite for spatial decision-making processes. Those who have the power to settle land cover categories have an implicit power to imprint certain management regimes on the land. Although aspects of power are well known in cartography and geographic information science (GI science), an increased focus on participatory practices and legitimacy in spatial decision-making processes makes it relevant to draw attention to the power of knowledge inherent in the process of land cover/use map production. This thesis builds on theory that establishes maps as knowledge-producing practices. Within such a framework, the focus is on how maps are made and remade in various ways (technically, socially and politically) as solutions to spatial problems by people within particular contexts and cultures. Elements from planning theory are used to underline the importance of maps and geographic information systems (GIS) as an essential foundation for spatial decision-making processes, public participation and legitimacy in land management and planning. Quantitative and qualitative methods are used to explore how technical, social and individual aspects influence and direct the knowledge-producing practice of map production, and how this influence creates implications for participatory practices in spatial decision-making processes. The investigation was conducted with a Norwegian land cover dataset named AR5. AR5 forms an important basis for planning and farmland management in Norway. It consists of a discrete polygon coverage presented as an area class map. The dataset is subject to both a periodic and a continuous updating regime carried out by the national mapping agency and local municipalities, respectively. The study was mainly carried out in different municipalities of the county of Hordaland in the western part of Norway, but for comparison, a study area from the county of Vestfold in the eastern part of the country was included. The thesis consists of three papers: Paper 1 is based on a quantitative GIS change analysis of the AR5 dataset. It explores and discusses technological and human factors that create challenges and opportunities for comparability and change detection. Results show that periodic updating provides unique possibilities for performing change analysis through GIS technology, but boundary uncertainty and closely related area classes challenge the interpretation and can lead to differences in mapping practice. Paper 2 is based on an ethnographic observation of a land cover mapping process, explores the social construction of a land cover map and discusses its implications for the use of GIS as a land management tool. Results show that interpreters from different social contexts classify land areas differently even if the same area class definitions are used. Different interests and needs contribute to this divergence of area classes and result in different mapping regimes. Dissimilarities are most evident when categories are ambiguous and transitions between categories are gradual. Paper 3 is based on a quantitative GIS analysis that compares land cover map products covering the same area produced by five different skilled interpreters. It explores the variation in land cover classification due to individual interpreter assessment. The investigation shows that even if only one mapping regime is involved in a mapping process, the still needed individual assessments can challenge the comparability and consistency of the map product. The results of the papers underline the need to create consciousness about the different interests in the map-producing process, the multiple purposes of a map product and the importance of who has the power to define land cover/use. Mapping reflects the values and judgements of individuals who construct the maps. It also creates knowledge about the land through selective stories. Mapping can therefore always be considered as being political, and access to spatial knowledge production creates power when mapping is used for land management and planning. The main argument of this thesis is that access to spatial knowledge production is required for increased participatory practices in land management and planning. According to a communication theory of power, dialogue and debate among the public are considered necessary to create conditions for legitimate spatial knowledge production representing the values and interests of citizens, thereby ensuring access to spatial knowledge production. Increased public access to spatial knowledge production can be criticised as challenging the map product’s consistency and verifiability. However, if increased public participation is achieved at a fundamental level in land cover/use map production, then increased participation can be achieved without challenging the consistency and verifiability of the map product. This increased participation in spatial knowledge production should concern the legitimacy of map standards, category definitions and the aim of the map product rather than the agreement over local land covers and borderlines.
Består avPaper I: Straume, K. (2013). Monitoring Norwegian farmland loss through periodically updated land cover map data. Norsk Geografisk Tidsskrift - Norwegian Journal of Geography, 67(1), 36–48. This article is not available in BORA. The published version is available at: 10.1080/00291951.2012.759616
Paper II: Straume, K. (2014). The social construction of a land cover map and its implications for geographical information systems (GIS) as a management tool. Land Use Policy, 39, 44–53. This article is not available in BORA. The published version is available at: 10.1016/j.landusepol.2014.03.007
Paper III: Straume, K. (2014) Variation in land cover classification due to individual interpreter assessment: A case study of farmland mapping in Norway. Manuscript submitted for publication. The article is not available in BORA.