| 研究課題/領域番号 |
17F17372
|
|---|
| 研究機関 | 東京工業大学 |
|---|
研究代表者 |
鼎 信次郎 東京工業大学, 環境・社会理工学院, 教授 (20313108)
|
|---|
| 研究分担者 |
DHANAPALA ARACHCHIGE SACHINDRA 東京工業大学, 環境・社会理工学院, 外国人特別研究員
|
|---|
| 研究期間 (年度) |
2017-11-10 – 2020-03-31
|
| キーワード | climate change impact / downscaling / environmental flow |
|---|
| 研究実績の概要 |
This JSPS project investigates producing improved projections of future likely climate at a daily time step and hence quantification of impacts of changing climate on the availability of flows to the environment considering a case study of the Sagami River basin. Existing statistical downscaling approaches should be improved to produce more reliable climatic information at catchment scales and at daily time steps. Also, with the changing climate the changes induced in the hydrological characteristics of streams in relation to timing, magnitude, duration and frequency can have a profound impact on the aquatic ecosystem. Following is this year's achievement for that purpose projections of catchment scale hydroclimatic variables at a daily time step. Reanalysis and AmeDAS data were compiled for the study area. Matlab codes for downscaling Reanalysis data to catchment scale climate variables were prepared. Downscaling models for daily maximum and minimum temperature were developed using multi-population genetic programming for stations located in the Sagami River basin. Potential of 4 input selection procedures for downscaling models was assessed. Model performance was assessed for the calibration and validation periods.
|
| 現在までの達成度 (区分) |
現在までの達成度 (区分)
2: おおむね順調に進展している
理由
In the research is being conducted as planned. However, the research so far has given rise to several new issues (e.g. limited performance of several input selection approached, outlier generation) which have to be further investigated.
|
| 今後の研究の推進方策 |
Downscaling models for daily precipitation and streamflows will be developed using multi-population genetic programming for stations located in the Sagami River basin. Performance comparison of genetic programming and multi-population genetic programming will be conducted. A hydrological model for the Sagami Basin will be developed using SWAT (Soil Water Assessment Tool). A performance comparison between direct downscaling of streamflows and simulation of streamflows with physics-based hydrological modelling will be performed. 3 Journal papers and a few conference papers are expected to be written along the following themes (note that these are not the exact titles of papers)
1. Potential of employing multiple populations in genetic programming in statistical downscaling, 2. Impact of migration policy in multi-population genetic programming on downscaling model outputs., 3. Comparison of direct downscaling with multi-population genetic programming against physics-based hydrological modelling.
|
