Land cover in Upper Egypt assessed using regional and global land-cover products derived from MODIS imagery

Abstract

Irrigation along the Nile River has resulted in dramatic changes in the biophysical environment of Upper Egypt. In this study we used a combination of Moderate Resolution Imaging Spectroradiometer (MODIS) 250 m Normalized Difference Vegetation Index data and Landsat imagery to identify areas that changed from 2001 to 2008 as a result of irrigation and water-level fluctuations in the Nile River and nearby waterbodies. We used two different methods of time series analysis – principal components analysis (PCA) and harmonic decomposition – applied to the MODIS 250 m Normalized Difference Vegetation Index images to derive simple three-class land-cover maps and then assessed their accuracy using a set of reference polygons derived from 30 m Landsat 5 and 7 imagery. We analysed our MODIS 250 m maps against a new MODIS global land-cover product (MOD12Q1 collection 5) to assess whether regionally specific mapping approaches are superior to a standard global product. Results showed that the accuracy of the PCA-based product was greater than the accuracy of either the harmonic decomposition or MOD12Q1 products for the years 2001, 2003 and 2008. However, the accuracy of the PCA product was only slightly better than the MOD12Q1 for 2001 and 2003. Overall, the results suggest that our PCA-based approach produces a high level of user and producer accuracies, although the MOD12Q1 product also showed consistently high accuracy. Overlay of 2001–2008 PCA-based maps showed a net increase of 12,129 ha of irrigated vegetation, with the largest increase found from 2006 to 2008 around the districts of Edfu and Kom Ombo. This result was unexpected in light of ambitious government plans to develop 336,000 ha of irrigated agriculture around the Toshka Lakes.

Acknowledgements

This research is supported by a grant from the National Institutes of Health (award 1R01GM093345-01 to the University of Miami, FL, USA, and Misr University for Science and Technology (MUST), Six of October Governorate, Egypt). The authors also thank the anonymous referees and the editor for many helpful comments and corrections.