Ahmed M Razu, Rahaman Khan Rubayet, Kok Aaron, Hassan Quazi K
Department of Geomatics Engineering, University of Calgary, 2500 University Dr. NW, Calgary, AB T2N 1N4, Canada.
Sensors (Basel). 2017 Oct 14;17(10):2347. doi: 10.3390/s17102347.
The northeastern region of Bangladesh often experiences flash flooding during the pre-harvesting period of the rice crop, which is the major cereal crop in the country. In this study, our objective was to delineate the impact of the 2017 flash flood (that initiated on 27 March 2017) on rice using multi-temporal Landsat-8 OLI and MODIS data. Initially, we opted to use Landsat-8 OLI data for mapping the damages; however, during and after the flooding event the acquisition of cloud free images were challenging. Thus, we used this data to map the cultivated rice acreage considering the planting to mature stages of the crop. Also, in order to map the extent of the damaged area, we utilized MODIS data as their 16-day composites provided cloud free information. Our results indicated that both the cultivated and damaged area estimates based on satellite data had strong relationships while compared to the ground-based estimates (i.e., ² values approximately 0.92 for both cases, and RMSE of 18,374 and 9380 ha for cultivated and damaged areas, respectively). Finally, we believe that our study would be critical for planning and ensuring food security for the country.
孟加拉国东北部地区在该国主要谷物作物水稻的收获前时期经常遭遇突发洪水。在本研究中，我们的目标是利用多时相陆地卫星8号运营陆地成像仪（Landsat-8 OLI）和中分辨率成像光谱仪（MODIS）数据，描绘2017年突发洪水（于2017年3月27日开始）对水稻的影响。最初，我们选择使用陆地卫星8号OLI数据来绘制受灾情况；然而，在洪水事件期间及之后，获取无云图像具有挑战性。因此，我们利用这些数据绘制了从作物种植到成熟阶段的水稻种植面积。此外，为了绘制受灾区域的范围，我们利用了MODIS数据，因为其16天合成数据提供了无云信息。我们的结果表明，与地面估计值相比，基于卫星数据的种植面积和受灾面积估计值都具有很强的相关性（即两种情况的R²值均约为0.92，种植面积和受灾面积的均方根误差分别为18374公顷和9380公顷）。最后，我们认为我们的研究对于该国的规划和确保粮食安全至关重要。
