An additional field data were gathered from 760 locations in February 2007 for validation of the results. MODIS 250m NDVI datasets were found to have sufficient spatial, spectral, and temporal resolutions to detect unique multi-temporal signatures for rice crops grown in the basin. Rice crop's multi-temporal NDVI signature was consistent with its general phenological characteristics and rice crops were spectrally separable at some point during the growing season. The maps of irrigated rice crops for all three irrigation areas have been compared with the records of available crop data with the irrigation authorities.Similarly, time series of MODIS satellite images have been used to estimate actual evapotranspiration (ET) using SAM-ET (spatial algorithm for mapping ET) algorithm, developed for an Australian agro-ecosystem, for the southern MDB over the time period of 2000-2007 and the results being further validated with the Eddy Covariance Systems. Finally, the volume of actual water use from the mapped irrigated rice crops have been estimated for the irrigation areas of Coleambally, Murrumbidgee and Murray. The results of actual water use are also compared with the records available with the irrigation authorities and show that remote sensing based volume estimates of actual water use are much more accurate in closing the water balance for irrigated rice system than standard methods.Due to continuous and prolonged drought, many irrigation areas in Murray Darling Basin (MDB) of Australia have been seriously affected to varying degrees, this has resulted in a significant reduction of water allocation for growing cereal crops such as Rice. Traditionally, rice crop is grown mainly in the in Southern part of MDB including New South Wales Riverina regions, such as Murrumbidgee, Coleambally, and the Murray Irrigation Areas. Less water availability forces the farmers not to grow rice crop in the MDB.The main aim is to map the irrigated rice areas using MODIS 250m NDVI time series data for 2000-2007 and then quantify the spatial volume of actual water use from MODIS optical-thermal satellite data using SAM-ET algorithm.Time-series MODIS 250m NDVI datasets hold considerable promise for irrigated rice area mapping in an agriculturally intensive region due to their global coverage, intermediate spatial resolution, high temporal resolution (16-day composite period), and cost-free status. The Savitzky'Golay filter was applied to smooth out noise in NDVI time-series that caused primarily due to cloud contamination and atmospheric variability. Our results indicate that this filtering method is more effective in obtaining high-quality NDVI time-series. A combination of unsupervised ISODATA and decision tree classification were performed on 12-month time-series of MODIS NDVI data for 8 years to map irrigated rice crop across the basin.
|Title of host publication||Third International Rice Congress|
|Subtitle of host publication||Rice for Future Generations|
|Place of Publication||Vietnam|
|Number of pages||1|
|Publication status||Published - 2010|
|Event||International Rice Congress - Hanoi, Vietnam, Viet Nam|
Duration: 08 Nov 2010 → 12 Nov 2010
|Conference||International Rice Congress|
|Period||08/11/10 → 12/11/10|
Hafeez, M., Song, K., Rabbani, U., Chemin, Y., & Sixsmith, J. (2010). Spatial Mapping of Actual Water Use from Irrigated Rice Areas in the Southern Murray Darling Basin of Australia. In Third International Rice Congress: Rice for Future Generations (pp. 2). IRRC.