Arsenic is a serious hazard in groundwaters in many parts of the world, including Cambodia. In this study, Geographic Information Systems and geostatistics have been used to develop hazard and risk maps of arsenic in shallow groundwaters of Cambodia. Firstly arsenic concentration in shallow groundwaters was mapped in 500 m x 500 m sized grid by employing a regression kriging (RK) model using 1150 arsenic data points collected by various institutions between 1999 and 2010 from tube wells 16-120 m deep. The dataset was split into a training dataset (85 %) and a testing dataset (15 %). Additionally, some environmental covariates such as elevation, slope, soil types and geological types were converted to 11 principal components and then those components selected for the model by stepwise regression in order to develop the deterministic trend for the RK model. The Leave-one-out cross-validation (LOOCV) result for the RK model shows the amount of variation in arsenic concentration explained by the model is approximately 54 %, and the root of mean squared error (RMSE) is 56. The paired t-test shows that the RK model is good for predicting areas where As > 10 ppb. The results of the RK prediction maps are also compared against the IDW and OK models for evaluation. All three models predict that the high arsenic risk areas are along the Mekong River, from Kratie town to the Vietnamese border, and that Kandal province has the highest risk for arsenic in groundwater in Cambodia. The second step of the study was to generate arsenic risk maps, which was done from the RK model including ratios of arsenicosis, skin cancer, and incidence rates of arsenic induced cancers based on algorithms from Yu et al. (2003). The number of arsenic induced diseases was then estimated using the arsenic hazard map and available Cambodian population density data. The estimated number of total cases of arsenic induced diseases for the whole of Cambodia based on 2008 population statistics is 57,967 cases of arsenicosis (39,137 cases for hyperpigmentation; 18,830 for keratosis), 2,052 cases of skin cancer, and 5,690 cases of internal cancers (5,263 for liver cancer, 391 for lung cancer, and 36 for bladder cancer). The estimated numbers of arsenic-attributable deaths per annum through lung, skin and bladder cancers in 2008 were approximately 300, 0.2 and 8 cases respectively. Comparison to WHO's estimated total deaths caused by lung, skin, and bladder cancer in 2004, shows that arsenic in groundwater contributes 28 %, 0.6 % and 6.1 % respectively to the deaths in Cambodia caused by these cancers.
Date of Award | 1 Aug 2011 |
---|
Original language | English |
---|
Awarding Institution | - The University of Manchester
|
---|
Supervisor | David Polya (Supervisor) |
---|
- Arsenic, GIS, Geostatistics, Hazard Map, Risk Map, Groundwater, Cancer, Risk Assessment
USING GIS AND GEOSTATISTICS TO DEVELOP HAZARD AND RISK MAPS OF ARSENIC IN SHALLOW GROUNDWATERS OF CAMBODIA
Sovann, C. (Author). 1 Aug 2011
Student thesis: Master of Philosophy