Detecting recent disturbance on Montane blanket bogs in the Wicklow Mountains, Ireland using the MODIS enhanced vegetation index

Abstract

Irish peat soils are extensive, covering approximately 14–20% of the national land area. They contain between 53% and 62% of the national soil organic carbon stock. Montane blanket bog covers approximately 25% or 242 650 ha of the total peatland area in Ireland and is the dominant peatland type covering the upland area of Wicklow. Blanket bogs are very sensitive systems and have experienced much disturbance in Ireland due to overgrazing, burning, drainage, forestry and turf cutting. It has been estimated that disturbance of blanket bog, on a national area basis, ranges from 74% to 82% and in Wicklow is 57%. Disturbance can be detrimental to stocks of soil organic carbon in peatlands. Monitoring disturbance in peatlands, which tend to cover large, remote areas, is difficult and expensive using conventional surveying methods. Satellite remote sensing offers a way to gather data for these areas. In this paper a method of determining the probability of disturbance is presented. This method uses the Moderate Resolution Imaging Spectroradiometer (MODIS) Enhanced Vegetation Index (EVI) in combination with univariate image differencing along with thresholding and binary logistic regression. A probability map was produced depicting the geospatial patterns and pressures on the peatland soil organic carbon stock in Wicklow. Peat soils in higher and steeper areas were more disturbed and the primary disturbance in between 2000 and 2005 was fire. Lower, flatter areas did not experience as much disturbance probably because they are wetter. The consumer's and producer's accuracy for the map was 76% and 42%, respectively.

Acknowledgements

This work was supported by the Environmental Protection Agency under the Environmental Research, Technological Development and Innovation programme and was funded by the ERTDI National Development Plan (2000–2006). We would like to thank Enterprise Ireland and Dr Shane Colgan of the Environmental Protection Agency (Ireland) for supporting this work through an International Collaboration Travel Support grant and a Short Term Research Mission grant. We would also like to extend thanks to Professor Nigel Roulet for facilitating John Connolly during a 6-month research visit to the Geography Department at McGill University. Finally, we would like to thank the anonymous reviewers whose comments enhanced this paper.