Applying multiple statistical approaches to predict the spatial distribution of soil contaminants in the Chao Phraya Watershed | |
| Author | Chor Pangara |
| Call Number | AIT Thesis no.EV-21-03 |
| Subject(s) | Spatial ecology--Environmental aspects--Thailand--Chao Phraya Watershed Multiple comparisons (Statistics) Environmental monitoring--Geographic information systems |
| Note | A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering in Environmental Engineering and Management |
| Publisher | Asian Institute of Technology |
| Abstract | A study using inverse distance weighted, inverse distance weighted with land use, ordinary kriging, ordinary kriging with land use and geographically weighted regression is done to compare their performances for soil contaminants (total carbon, total nitrogen, total phosphorus, lead, mercury, arsenic, cadmium, chromium, nickel, copper and zinc) prediction in the Chao Phraya watershed. Ordinary kriging with land use has the best performance with root mean square error at 16.22 with inverse distance weighted having the highest root mean square error for total carbon. Geographically weighted regression underperforms with its root mean square error at 1.38 while ordinary kriging with land use has 0,99 which is the best result for total nitrogen. For total phosphorus, inverse distance weighted completely underperformed with its root mean square error at 0.77 when compared to the other four methods and ordinary kriging with land use with root mean square at 0.53 still perform better than the rest. As for lead, cadmium and zinc, geographically weighted regression performance improves with its root mean square error at 41.35, 0.64 and 252.43, respectively when compared to other methods. For total carbon, geographically weighted regression prediction map shows high concentration at more than 80 grams per kilogram to the northern most part of the basin while the other four methods show that such high concentration only occur at the north-eastern part of the basin. The lower part where it is mostly plain shows consistent concentration of less than 20 grams per kilogram. For total nitrogen, geographically weighted regression maps follow the same trend as total carbon geographically weighted regression map with the northern most part having high concentration at more than 4.5 grams per kilogram. The plain area where paddy land resides has consistent concentration of less than 1.5 grams per kilogram. For total phosphorus, ordinary kriging with land use show that large variation occurs between urban land and other type of land use. The concentration can go as high as 1 gram per kilogram at the most populated and urbanized area. Geographically weighted regression model can generate a spatially varied pollutants distribution due to its nature of correlating different factors including human factors to the pollutant’s nature. However, the prediction accuracy while acceptable cannot outperform ordinary kriging with land use method. Since land use is very important, converting land use to binary maps for each land use type cannot properly explain the spatial distribution of pollutants. Land use stratified method produce the best results of all the methods. This is due to the importance of land use types as it is the main indicator of pollutants source. The assumption that pollutants will be varied in different types of land use hold true. Among the five methods, the ordinary kriging with land use performs the best followed by geographically weighted regression, inverse distance weighted with land use, ordinary kriging, and inverse distance weighted, respectively. |
| Year | 2021 |
| Type | Thesis |
| School | School of Environment, Resources, and Development (SERD) |
| Department | Department of Energy and Climate Change (Former title: Department of Energy, Environment, and Climate Change (DEECC)) |
| Academic Program/FoS | Environmental Engineering and Management (EV) |
| Chairperson(s) | Xue, Wenchao |
| Examination Committee(s) | Ekbordin Winijkul;Shrestha, Sangam |
| Scholarship Donor(s) | Loom Nam Khong Pijai (Greater Mekong Subregion) Scholarships |
| Degree | Thesis (M. Eng.) - Asian Institute of Technology, 2021 |