Browsing by Subject "Habitat suitability models"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Publication
    Development of assessment tools for Lake Sevan (Armenia) by the application of remote sensing data and geographic information systems (GIS) techniques
    (2011) Agyemang, Thomas Kwaku; Schmieder, Klaus
    Lake Sevan is the biggest source of water in Armenia. Its littoral zone, in addition to being a food source and a substrate for macrophytes, algae and invertebrates, provide refuge and spawning habitats for both young & old organisms especially fishes. Between 1933 and 1960s, the lake level had been lowered by 20 m below the original level by increasing the lake outflow intermittently for irrigation and electricity generation. This evidently had ecological and economical consequences on the lake ecosystem. The importance of assessing the accuracy of spatial data classifications derived from remote sensing methods and used in geographic information system (GIS) analyses has been regarded as a critical component of many projects. In this project, supervised classified QuickBird satellite imageries of both submersed macrophytes and landcover types (emersed vegetation) of the Gavaraget, Tsovazard and Masrik Regions of the study area were validated in a GIS environment. The results of these assessments were represented by error matrices presenting the overall accuracy, the user and producer accuracies in each category, as well as the kappa coefficients. For submersed macrophytes at the vegetation level, the overall accuracy ranging between 77-88% was achieved in all the investigation years. Alga blooms in the different years impacted on the accuracy of the classification. However, even through severe algal blooms user accuracies between 55% and 95% were achieved. On the other hand, at the growth type level, the overall accuracy was as high as over 70% and as low as below 49%. For emersed vegetation types, predominantly high overall accuracies of more than 70% were obtained in 2 of the investigation years. Above all, in 2008, only slight overall accuracy could be obtained. For reeds areas, high user accuracies of more than 78% could be obtained, while for shrubs, trees, no vegetation and grasses in the different years, very different classification accuracies were attained. Two habitat suitability models (one for fishes and one for birds) were built in a GIS environment in this project. While the Crucian Carp (Carassius auratus Gibelio Bloch) was chosen as lead species for the fish habitat, the Common Coot (Fulica atra) and the Great Crested Grebe (Podiceps cristatus) were chosen for the bird habitat models based on expert knowledge on Lake Sevan. Five fish habitat suitability classes were assigned in the model. There was a similar trend in the fish habitat areas in all the landscapes in Gavaraget, Tsovazard and Masrik regions. The habitat areas increased in 2007 and decreased in 2008. The increases in all the regions were the same (around 43%) while the highest reduction occurred in Gavaraget (47%) followed by Masrik (38%) and Tsovazard (25%) respectively. Apart from the reductions in habitat areas in 2008, there were severe decreases in the quality of the habitat areas in all the regions of interests. The increases and decreases were as a result of interannual fluctuations due to water level fluctuations and algal blooms of Lake Sevan. Also, for the bird habitat model, five classes were assigned. Tsovazard and Masrik had a similar trend in habitat areas with an initial increase in 2007 followed by a decrease in 2008. However, Gavaraget had reductions in 2007 and 2008. Again, in addition to the severe reductions in the habitat areas in 2008, there were severe decreases in the quality of the habitat areas in all the regions of interests. The changes in emersed macrophyte vegetations and the lake water level fluctuations effected the different changes in the bird habitat areas.