Combining LiDAR and IKONOS data for eco-hydrological classification of an ombrotrophic peatland

This paper presents the results of a study aimed at improving the spatial representation of eco-hydrological communities on ombrotrophic (rain-fed) peatlands. Remote sensing techniques have shown potential for peatland monitoring but most previous work has focussed on spectral approaches. Such methods often result in poor discrimination of cover types and also neglect information on the presence of peatland microtopes (e.g. hummocks and hollows), which show strong links to hydrological condition, biodiversity and carbon sequestration. Spatial information on surface structure is therefore a useful proxy for peatland condition. The paper first demonstrates how airborne LiDAR data can provide such information. Secondly, we apply a combined multispectral-structural approach using IKONOS and airborne LiDAR data to the mapping of peatland condition classes (defined by the lowland raised bog inventory) at Wedholme Flow, Cumbria, UK. LiDAR data were pre-processed so that spatial estimates of minimum and maximum land surface height, variance and semi-variance (from semivariogram analysis) were extracted. These were assimilated into a spectral classification procedure with an IKONOS dataset where thematic outputs from per-pixel maximum likelihood classification were compared. Field ecological survey data were used to validate the results which showed considerable improvements in thematic separation of peatland eco-hydrological classes, when spatially-distributed measurements of LiDAR variance or semivariance at 5 m lag were included. Of key importance was the improved delineation of management classes (Eriophorum bog, active raised bog, Rhynchospora alba communities, and Calluna- and Erica-dominated degraded raised bog). The paper demonstrates how combined textural-optical approaches can offer improvements in land cover mapping products in areas where reliance on purely spectral discrimination approaches would otherwise result in considerable thematic uncertainty