| Ref ID: | 36991 |
|---|---|
| Ref Type: | Journal Article |
| Authors: | Thanaphattarapornchai, Montri Sarapirome, Sunya |
| Title: | Ancient salt-making site prediction using spatial modeling: A case study in Nakhon Ratchasima Province |
| Date: | 2021 |
| Source: | Suranaree Journal of Science and Technology |
| Abstract: |
The study aims at predicting ancient salt-making potential areas, prior to the 19th century, in Nakhon Ratchasima province, Thailand, using spatial modeling of the frequency ratio (FR) and logistic regression (LR). Input data are physical and cultural characteristics and their relationship extracted from samples of existing ancient salt-making sites. These include locating on the Maha Sarakham Formation (KTms) and its related units, distance from geologic structures, locating on salt crust area and saline wet belt, Normalized Difference Salinity Index ( NDSI ), distance from ancient settlements, and distance from water bodies. The coefficients of input variables from the LR model could explain how much they influence the chance to discover the new ancient sites. The most influential factor was the distance from the ancient settlement, followed by locating on salt crust area and saline wet belt and distance from water bodies. Frequency ratios strongly supported these findings. Final results from the models were raster-based probability maps of ancient salt-making site occurrence. Relative Operating Characteristics (ROC) were used to validate the FR and LR probability maps and resulted in areas under curves (AUC) of 0.810 and 0.908, respectively. It revealed that LR model could conclusively provide a better result. The study results fruitfully serve the objectives of the study. |
| Volume: | 28: 030047 |
| Number: | 3 |
| Page Start: | 1 |
| Page End: | 11 |