Abstract
The assessment of land use–land cover (LULC) and normalized difference vegetation index (NDVI) changes on hydrology is essential for the development of sustainable water resource strategies. This study focuses on Sperchios river catchment, a complex and very heterogeneous area made up of numerous land covers that are difficult to map due to spectral similarities. Difficulties such as multi-seasonal spectral variables corresponding to different stages of land cover phenological development were addressed by choosing an unsupervised classification algorithm called k-means in combination with the Corine land cover database and past scientific studies. Available remote sensing data included three Landsat images (MSS, TM, OLI) all of which were resampled to a common 30-m resolution. Final LULC classifications in the period 1972–2013 separated forest, agricultural land and heath land and detected unified bare with urban land and pastures with areas of low vegetation. The recorded results include a significant reduction (–80%) of vegetation (NDVI values), 28 and 26% reduction rates in forests and pastures-low vegetation, respectively, a strong increase (114%) in heath land and an 47% increase, as concerns the agricultural land. Using of random points estimated the classification accuracy to be 75.4 and 85.2% for LULC of 1972–2013, respectively. This particular study constitutes a preliminary stage of an attempted integrated basin management of Sperchios river intending to contribute to overall ecological quality assessment of Sperchios river.
Aknowledgements
This work have been funded by the Hellenic General Secretariat of Research and Technology, under the KRIPIS framework program with title: “Development of an integrated management scheme combining both catchment and coastal/sea zones”, ESPA 2007–2013.
Notes
Presented at the 5th International Conference on Environmental Management, Engineering, Planning and Economics (CEMEPE) 14–18 June 2015, Mykonos, Greece