2020 Fiscal Year Research-status Report
Global inundation area estimation by assimilating multi-sensor satellite observations into a hydrodynamic model
| Project/Area Number |
20K22428
|
|---|
| Research Institution | The University of Tokyo |
|---|
Principal Investigator |
ZHOU XUDONG 東京大学, 生産技術研究所, 特任助教 (20876239)
|
|---|
| Project Period (FY) |
2020-09-11 – 2022-03-31
|
| Keywords | general agreement / explainable mismatches |
|---|
| Outline of Annual Research Achievements |
The historical water surface area has been simulated by CaMa-Flood. Comparisons show a good agreement of CaMa-Flood with results derived from Landsat in major rivers, lakes and open-to-sky floodplains. While modeled water surface area is lower than satellite-derived results in the northern high latitudes and coastal area. In contrast, model results are larger than Landsat in forests and croplands. The differences are related to the limitations in model structures and bias in the model inputs and parameters. The above results are summarized and submitted to international journals.
Additionally, the probability analysis of the water surface is finished. Despite of the permanent water surface, transitional water surface is associated with irrigated area and wetlands, indicating a higher probability for water identification in these land cover types. There are large variations between different satellite results and the time-variant water surface area is systematic biased. This needs to be solved when selecting appropriate satellite for data assimilation.
Various flood areas with different runoff inputs are prepared to account for the uncertainties resulted from the model inputs. The population and economy exposure to the flood is analyzed with uncertainties considered over different continents. Asia is suffering the largest exposure while the uncertainties are the largest in the Africa, indicating a necessity to improve the performance of different hydrological models especially for data-limited regions.
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
The plan for the first year is to understand the characteristics of the water surface area estimated by river model and derived from the satellite observations. This is generally accomplished, and the results have been summarized. One relevant paper is published, one has been accepted for publication.
What is not expected is that the limitation of satellite-derived water is strong in regions such as the high latitudes and the forests. Model also has large shortages in representing local depressions and small rivers. The water surface area derived from different satellites also varies in specific regions. These mismatches must be well treated before future application of the Bayesian and data assimilation.
|
| Strategy for Future Research Activity |
The special resolution is one of the largest reasons for the variations between results from models and different satellites. The next step is to assess the impact of spatial resolution on the water surface area and exposure. In terms of the mismatches, filtering masking can be applied, so that we can test our methods where no significant shortcomings of the model and satellites are found. On the other hand, the results can be pre-processed with help of data from other sources.
The other procedures of the project will be continued as planned.
|
| Causes of Carryover |
The traveling is paused due to the pandemic. Collaboration with other research teams and attendance to domestic or international workshops will be restarted in the next fiscal year. Some additional items will be prepared for research and traveling.
|
