Landscape-scale Remote Sensing and Classification of Pond and Lake Habitat Diversity in a Tropical City

Abstract Ponds and lakes are common freshwater habitats in urban landscapes, and often have a high biodiversity and conservation value. The importance of landscape-scale conservation of pond networks has recently been recognised, yet the categorisation and classification of pond network spatial structures is missing. Developing spatial methods and tools to characterise and understand pond networks is a critical first step to accurately conserve pond habitats and inhabiting species. This paper presents an inventory of ponds and lakes in Greater Kuala Lumpur, Malaysia, characterising their distribution, abundance and type. Remote sensing was first employed to map and characterise these habitats, followed by multivariate cluster analysis to classify and develop a typology of the ponds identified. Physicochemical data was collected from a sample (n=60) of ponds to compare with the remotely sensed pond classification. Results demonstrated that multi-source remote sensing can be highly accurate and effective in inventorying ponds and lakes of varying sizes. A total of 1013 ponds and lakes were identified within the Greater Kuala Lumpur region and were found to be highly environmentally heterogeneous. Typology clusters were driven by land cover rather than local physicochemical variables demonstrating that specific remotely-sensed variables may be sufficient proxies for certain chemical variables. Landscape-scale conservation and management of pond networks should utilise remote sensing tools, to establish pond network structure, and to maintain wide environmental heterogeneity among pond habitats. Incorporating remote sensing tools into pond conservation will ensure that effective pond conservation can be achieved and biodiversity protection can be maximised.