It developed a robust stream processing in the Google Earth Engine for near-daily re-visit of high-res satellite observations.

When I submit this work to the Journal of Remote Sensing , I was a little bit hesitate because this is currently the top open access journal (Rank 4.1). But to my surprise, the reviewers really like it and feedback very positively. After publication, It even gets further attention from the community as well, so I guess others people also like it and grouped these comments here Just For Record.

Link to the app: https://ndminhhus.users.earthengine.app/view/cropninhthuan2019

Characterization of droughts using satellite-based data and indices in a steep, highly dynamic tropical catchment, like Vu Gia Thu Bon, which is the most important basin in central Vietnam, has remained a challenge for many years. This study examined the six widely used vegetation indices (VIs) to effectively monitor droughts that are based on their sensitivity with precipitation, soil moisture, and their linkage with the impacts on agricultural crop production and forest fires. Six VIs representing the four main groups, including greenness-based VIs (Vegetation Condition Index), water-based VIs (Normalized Difference Water Index, Land Surface Water Index), temperature-based VIs (Temperature Condition Index), and combined VIs (Vegetation Health Index, Normalized Difference Drought Index) were tested using MODIS data from January 2001 to December 2016 with the support of cloud-based Google Earth Engine computational platform. Results showed that droughts happened almost every year, but with different intensity. Vegetation stress was found to be mainly attributed to precipitation in the rice paddy fields and to temperature in the forest areas. Findings revealed that combined vegetation indices were more sensitive drought indicators in the basin, whereas their performance was different by vegetation type. In the rice paddy fields, NDDI was more sensitive to precipitation than other indices; it better captured droughts and their impacts on crop yield. In the forest areas, VHI was more sensitive to temperature, and thus had better performance than other VIs. Accordingly, NDDI and VHI were recommended for monitoring droughts in the agricultural and forest lands, respectively. The findings from this study are crucial to map drought risks and prepare an effective mitigation plan for the basin 

Understanding of reservoir storage is vital for flood management and operation of dams. However, in Vietnam many reservoirs are located in remote regions, which make difficult the deployment and maintenance of in-situ measurement systems. The Srepok River is a major tributary of the Mekong River, flowing from the Central Highlands of Vietnam into northeastern Cambodia. Due to rich hydropower potential, within 125 km on the mainstream, seven hydropower dams has been built and at least two more are under construction by Cambodia counterpart. Therefore, the Srepok River is in need of a monitoring system that could facilitate transboundary basin-scale management of water resources for upcoming challenges. As a means to estimate reservoir storage for sites without reliable in-situ data, this study proposed a combination of freely available satellite based data including all Landsat imagery, Sentinel 2A imagery, SRTM digital elevation model and altimetry radar to achieve more frequency of reservoir storage information in the Srepok River. A set of algorithms for detection and removal of cloud in Landsat 5, 7 and 8 images were developed. Then, Modified Normalized Differenced Water Index (M-NDWI) derived from less than 1 percent cloud in all Landsat imagery over the ROI was used for calculating surface water extent area from 1985 to present time. Radar altimetry (TOPEX/Poseidon, Jason) was used for sensing water level. Hypsometric curves of the reservoirs were drawn using SRTM 30m DEM. When both Altimetry WL and in-situ gauge were unavailable, water area extent calculated from Landsat scenes were used to derive reservoir WL using hypsometric curves. This composite WL dataset increases the frequency and reliability of reservoir capacity monitoring in remote and low income regions. Google Earth Engine cloud platform used in processing these remote sensing data greatly improved the speed and automation process. The study indicated that freely available remote sensing data nowadays produce a great opportunity to support the study and monitoring of water bodies which may not possible before.