Michael Karteris

Burned area mapping using logistic regression modeling of a single post-fire Landsat-5 Thematic Mapper image

Logistic regression modeling was applied, as an alternative classification procedure, to a single post-fire Landsat-5 Thematic Mapper image for burned land mapping. The nature of the classification problem in this case allowed the structure and application of logistic regression models, since the dependent variable could be expressed in a dichotomous way. The two logistic regression models consisted of the TM 4, TM 7, TM 1 and TM 4, TM 7, TM 2 presented an overall accuracy of 97.37% and 97.30%, respectively and proved to be the most well performing three-channel color composites. The discriminator ability in respect to burned area mapping of each one of the six spectral channels of Thematic Mapper, which was achieved by applying six logistic regression models, agreed with the results taken from the separability indices Jeffries-Matusita and Transformed Divergence.

Logistic regression modelling of multitemporal Thematic Mapper data for burned area mapping

This study focused on the development of a logistic regression model for burned area mapping using two Landsat-5 Thematic Mapper (TM) images. Logistic regression models were structured using the spectral channels of the two images as explanatory variables. The overall accuracy of the results and other statistical indications denote that logisticregression modelling can be usedsuccessfully for burned area mapping. The model that consisted of the spectral channels TM4, TM7 and TM1 and had an overall accuracy of 97.62%, proved to be the most suitable. Moreover, the study concluded that the spectral channel TM4 was the most sensitive to alterations of the spectral response of the burned category pixels, followed by TM7.

Classification analyses of vegetation for delineating forest fire fuel complexes in a Mediterranean test site using satellite remote sensing and GIS

If fuel, weather and topography are considered to be the most important determinants of wildfire occurrence, it is evident that only fuel can be kept under human control and modified to reduce fire potential. In the present study, forest fuel mapping is considered from a remote sensing perspective by the assessment and mapping of general vegetation complexes. The purpose is to delineate forest types which present a particular fire behaviour and to explore the use of Landsat TM data for their mapping. The spectral classes were derived by considering as key elements of the classification scheme the main species that prevail in the overstory layer, as well as meaningful mixtures of them, discriminated by their degree of density as indicated from vegetation indices. The study area, Halkidiki, Greece, which has strong spatial heterogeneity in both the composition and structure of its ecosystems, as well as of their spatial distribution and arrangement, is a characteristic area and representative of the majority of landscape types found across Greece. The overall classification accuracy of the original Landsat TM image (85.30%) was not improved significantly when other synthetic spectral channels or the digital elevation model were integrated with the satellite data, possibly because the detailed classification scheme adopted was determined using the overall spectral discrimination offered by the original satellite data.

Local-Scale Fuel-Type Mapping and Fire Behavior Prediction by Employing High-Resolution Satellite Imagery

Judicial wildland fire prevention and management requires precise information on fuel characteristics and spatial distribution of the various vegetation types present in an area. The aim of this study was to present an integrated approach to forest fire management, combining local-scale fuel-type mapping from very high spatial resolution imagery with fire behavior simulation. The specific objectives were (i) to develop a detail site-specific fuel model in a Mediterranean area that is suitable for fire behavior prediction; (ii) to produce a detailed local-scale fuel-type map with an object-based approach; and (iii) to generate accurate fire behavior maps. The spatial extent of the different fuel types of a forested landscape in northern Greece characterized by heterogeneous vegetation and topography was determined using a Quickbird image. Site-specific fuel models were created by measuring fuel parameters in representative natural fuel complexes. Following necessary preprocessing of the image to compensate for geometric errors, multiscale components of the scene were delineated through a segmentation algorithm. The resulting image objects were assigned to respective fuel types using a CART statistical model with an overall accuracy over 80%. The FARSITE fire simulation model was applied to simulate potential wildland fire growth and behavior. Utilizing the spatial database capabilities of geographic information systems, FARSITE allows the user to simulate the spatial and temporal spread and behavior of a fire burning in heterogeneous terrain, fuels, and weather.

Design of forest management planning DSS for wildfire risk reduction

Abstract Forest management planning is generally a complicated task. The amount of data, information and knowledge involved in the management process is often overwhelming. Decision support systems can help forest managers make well documented decisions concerning forest management planning. These systems include a wide variety of components, depending on the management goals of the forested land. Although an increased growth of decision support systems in specific domains of forest management planning exists, there is no special design model for the deployment of forest management planning. To this direction, this paper has the following objectives: Firstly, to propose a conceptual design model for developing goal-driven forest management planning decision support systems. Secondly, to apply the design model for a particular case of these systems, the wildfire risk reduction decision support systems. Thirdly, to present the deployment of a wildfire risk reduction decision support system as well as its results for a specific forest area.

Development of a nationwide approach for large scale estimation of green roof retrofitting areas and roof-top solar energy potential using VHR natural colour orthoimagery and DSM data over Thessaloniki, Greece

Buildings, and other engineered structures that form cities, are responsible for a significant portion of the local and global impacts of climate change. Therefore, there is a need for the increase and large-scale implementation of energy-efficient and renewable energy-generation technologies. In this paper, we investigate a remotely sensed approach for assessing the potential capacity of installing rooftop photovoltaic systems and/or implementing green roof retrofitting measures in an urbanized area. In particular, we analyse the ways in which freely available, no-cost, very high spatial resolution orthoimagery and Digital Surfaces Models, along with geospatial vector data, can assist in delineating available roof area at the level of individual buildings across a Mediterranean city. The available data-sets are integrated and processed within a Geographical Object Based Image Analysis framework. The method was developed on the basis of a test area consisting of the town of Thessaloniki in Northern Greece. The correlation between the automated extracted roof area and the manual digitized reference data had a value of up to 0.9. The described approach can be applied nationwide across Greece, considering the national availability coverage of the data-sets, while the semantics can readily be transferred to neighbouring Mediterranean countries.

An object based approach for the implementation of forest legislation in Greece using very high resolution satellite data

The possibility to extract forest areas according to the criteria included in a legal forest definition was evaluated, in a mountainous area in the Northern-central part of Greece, following an object based image analysis approach. While a lot of studies have focused on the estimation of forest cover at regional scale, no particular emphasis has been given so far to the delineation of forest boundary line at local scale.