Influence of Resolution in Irrigated Area Mapping and Area Estimation

Velpuri, N M and Thenkabail, P S and Gumma, M K and Biradar, C M and Dheeravath, V and Noojipady, P and Yuanjie, L (2009) Influence of Resolution in Irrigated Area Mapping and Area Estimation. Photogrammetric Engineering & Remote Sensing , 75 (12). pp. 1383-1395. ISSN 0099-1112

PDF - Published Version
Download (1MB) | Preview


The overarching goal of this paper was to determine how irrigated areas change with resolution (or scale) of imagery. Specific objectives investigated were to (a) map irrigated areas using four distinct spatial resolutions (or scales), (b) determine how irrigated areas change with resolutions, and (c) establish the causes of differences in resolution-based irrigated areas. The study was conducted in the very large Krishna River basin (India), which has a high degree of formal contiguous, and informal fragmented irrigated areas. The irrigated areas were mapped using satellite sensor data at four distinct resolutions: (a) NOAA AVHRR Pathfinder 10,000 m, (b) Terra MODIS 500 m, (c) Terra MODIS 250 m, and (d) Landsat ETM+ 30 m. The proportion of irrigated areas relative to Landsat 30 m derived irrigated areas (9.36 million hectares for the Krishna basin) were (a) 95 percent using MODIS 250 m, (b) 93 percent using MODIS 500 m, and (c) 86 percent using AVHRR 10,000 m. In this study, it was found that the precise location of the irrigated areas were better established using finer spatial resolution data. A strong relationship (R2 = 0.74 to 0.95) was observed between irrigated areas determined using various resolutions. This study proved the hypotheses that “the finer the spatial resolution of the sensor used, greater was the irrigated area derived,” since at finer spatial resolutions, fragmented areas are detected better. Accuracies and errors were established consistently for three classes (surface water irrigated, ground water/conjunctive use irrigated, and nonirrigated) across the four resolutions mentioned above. The results showed that the Landsat data provided significantly higher overall accuracies (84 percent) when compared to MODIS 500 m (77 percent), MODIS 250 m (79 percent), and AVHRR 10,000 m (63 percent).

Item Type: Article
Subjects: Others > Watershed Management
Others > Agriculture-Farming, Production, Technology, Economics
Depositing User: Mr Sanat Kumar Behera
Date Deposited: 24 Oct 2013 14:39
Last Modified: 18 Dec 2015 06:48
Official URL:
Funders: International Water management Institute
Acknowledgement: Dr Gumma Muralikrishna is presently associated with ICRISAT. This study was supported by the International Water management Institute’s (IWMI) core funding. We want to acknowledge the vision and support shown by Professor Frank Rijsberman, Program Director, Google (formerly Director General of IWMI). We thank three anonymous reviewers for their helpful comments and suggestions. The paper has not been internally reviewed by U.S. Geological Survey (USGS). Hence, the views expressed in the paper are not endorsed by USGS.
View Statistics

Actions (login required)

View Item View Item