| 研究実績の概要 |
In this fiscal year, I have first processed the two types of datasets collected in the previous year. Since the datasets are in various spatial resolutions, I converted all the datasets into 5-arc minutes.
Then, the statistical land use and land cover change model was estimated with multinominal logit regression methods. In the modeling framework, the world was divided into 17 AIM regions. Therefore, one model was fitted for each region. The data for the years 1995, 2000, 2005, 2010, and 2015 was used for econometric estimation.
After that, the estimated coefficient was used to project the land use and land cover in the future in a recursive manner, together with the independent variables for the future.
Last, the projected land use and land cover for future years were adjusted using the cross-entropy method, with the objective function of minimizing the distance between the posterior probabilities and the priors.
The results could provide future land use information for the long-term. It could also provide information for the assessment of land-based climate mitigation potential.
|
