Detecting fluctuation of rice cultivated areas in semi-arid regions by combined use of MODIS and Landsat imageries

DOI Web Site 参考文献8件 オープンアクセス

抄録

We developed an improved method to detect rice cultivated areas in semi-arid regions by combining the usage of Landsat imagery for detecting rice fields and MODIS for spatial and temporal upscaling of rice cultivation areas. We selected Haryana State in northwestern India as a case study area, where average farm plot size is small (~4,000 m2) and cultivated areas largely fluctuate from year to year due to water availability. Firstly, rice cultivated areas were detected by unsupervised classification of Landsat images of a specific year. Secondly, two conditional parameters for time-series EVI and LSWI were optimized by Powell’s method to best match the rice cultivated areas detected by MODIS to the areas detected by Landsat. Thirdly, calculated rice cultivated areas were compared with statistical data. Until results showed reasonable agreement, rice cultivated areas were reclassified in Landsat images and the following procedures were repeated. A good coefficient of determination (R2 = 0.82) was obtained between the estimated rice area and the statistical data at the district (sub-state) level in the study area. This demonstrates the potential and increased accuracy of the developed methodology to detect and map harvested rice areas in water scarce arid and semi-arid regions.

収録刊行物

  • Hydrological Research Letters

    Hydrological Research Letters 9 (4), 107-112, 2015

    水文・水資源学会／日本地下水学会／日本水文科学会／陸水物理研究会

参考文献 (8)*注記

もっと見る

関連プロジェクト

もっと見る

キーワード

詳細情報 詳細情報について

問題の指摘

ページトップへ