Landscape Ecological Analysis and Assessment in an Urbanising Environment - forest birds as biodiversity indicators

University dissertation from Stockholm : KTH Royal Institute of Technology

Abstract: To achieve a sustainable development, impacts onbiodiversity of urbanisation, infrastructure, land use changesand other developments must be considered on a landscape andregional scale. Landscape ecology can provide a conceptualframework for the assessment of consequences of long-termdevelopment processes like urbanisation on biodiversity on alandscape scale, and for evaluating the impacts of alternativeplanning scenarios. The aim of this study was to explore theeffects of habitat quality, quantity and connectivity on forestbird diversity in an urban-rural gradient. The purpose of theanalyses was to develop knowledge and methods for integratingbiodiversity issues in planning and assessments in anurbanising environment, on landscape and regional scales.The study area was situated in and around Stockholm, thecapital of Sweden, covering the city centre, suburbs andperi-urban areas. Data on breeding forest birds were collectedthrough bird censuses in an urban-suburban gradient. In orderto embrace also the peri-urban areas for a more completeurban-rural gradient, data on two fragmentation-sensitiveforest grouse species were obtained through a questionnaire tohunters in the whole study area. Response variables in theanalyses were forest bird species richness and diversity,relative species richness and occurrence of single sensitivespecies like selected sedentary forest birds, including theforest grouse species, and red list species. Habitat quality,quantity and connectivity were analysed using available data onabiotic conditions, including urban disturbances, andvegetation in geographical information systems. In addition, afield study on vegetation structure and composition wasperformed in a subset of the smaller sample sites.Relationships between the response variables and habitatquality, quantity and connectivity were explored usingstatistical methods like multivariate statistics and regressionmodelling. Further, for some models, spatial dependencies werequantified and accounted for. When habitat models wereretrieved, they were used for spatial predictions of habitatsuitability. They were also applied on future planningscenarios in order to predict and assess the impacts onsensitive species. In the urban-rural gradient, the foreststructure and composition changed, so that in more urban areas,coniferous forest on rich soils, wet forests and wetlandsbecame less abundant and more scattered. Sensitive birdspecies, tied to these habitat types, were shown to besensitive to habitat fragmentation caused by urbanisation.Large, well-connected habitat patches and aggregations ofsuitable habitat in the landscape had a higher probability ofoccupancy when compared to other patches. For the forest grousespecies, effects of car traffic added to the explanation oftheir distribution. By contrast, deciduous forest was stillquite common in predominantly urban areas, due to both latechanges in land use and a history of human preferences. Certainred listed bird species tied to deciduous forest did not seemto be affected by isolation, and also occurred in suitablehabitats in some highly urbanised areas. Furthermore, relativespecies richness in the urban-suburban gradient was related tomulti-layered deciduous forest habitats with a large amount ofdead wood. Such habitats were associated with natural shorelineand with old pastures and parks. From the derived statisticalmodels, describing the relationships between sensitive speciesand environmental variables, predictive habitat maps could becreated for the present situation and for planning scenarios.The predictions of the impacts on habitats of sensitive speciesmade it possible to quantify, integrate and visualise theeffects of urbanisation scenarios on aspects of biodiversity ona landscape scale.