| Hainan province has been one of the major malaria epidemic areas in China. Malaria has been one of three most prevalent epidemics in this area. It has greatly threatened the healthy of people and depressed productive capacity. As a vector-borne disease, most malaria epidemics are the result of abnormal environmental condition, such as geographical and meteorological condition. It can provide valuable decision support for malaria preventive and control measures to research the relationship between malaria epidemics and environmental condition and to develop prediction model of malaria epidemics, then to construct GIS-based malaria monitoring and warning system,. In the present research, the data on meteorological variables , NOAA-AVHRR NDVI and malaria incidence in Hainan province were used to explore the possibility of building malaria prediction model and, at the same time, to try to construct the GIS-based malaria monitoring and warning system of Hainan province .The results were indicated as below:1. To explore the relationship between the meteorological factors (including temperature , rainfall and relative humidity) and the epidemic ofmalaria and to model the changes of epidemic of malaria for predicting the incidence of malaria in the future in Hainan Province, Spearman's rank correlation analysis was used to analyze the correlation between meteorological factors and malaria incidence per month for the period 1995-2000, then stepwise regression was used to simulate the changes of malaria incidence with those meteorological factors which were correlated with malaria incidence. Malaria incidence show significant positive correlation with temperature and rainfall. The result of stepwise regression analysis suggested that on the scale of Hainan province, t-02 (the average temperature of the present month and its previous two months) was significant climatic variable in the transmission of malaria (R2=0.59); for southern high epidemic areas in the province, 1-02 and t-oimin (t-o2min is the average minimum temperature of the present month and its previous two months) were significant climatic variables (R2=0.63) ; and for Wanning city, t-o2max (t-ozmax is the average maximum temperature of the present month and its previous months) (R2=0.48). When we introduced the I-1 (the incidence of the last month) to the regression analysis with meteorological variables, we got better fitting effect of the regression functions. For whole province, R2 reached 0.81; for southern high epidemic areas in the province, R2 reached 0.81; and for Wanning City, R2 reached 0.53. When we introduced the I-2 (the incidence of the second last month), R2 were 0.72,0.73 and 0.49 respectively. So we can conclude that the meteorological factors can influence the prevalence of malaria, and we can fit the trends of the epidemic of malaria with meteorological factors and use the fitting model to predict the malaria epidemic in the future.2. For validating the usefulness of the vegetation index obtained by satellite remote sensing for the development of maps of malaria risk and for prediction of malaria epidemics in Hainan province, we analyzed the relationship between the malaria incidene and the NORR-AVHRR NDVI of 19districts in Hainan province for the period 1995/2-1996/1. We found that malaria incidences have positive correlations to mean and maximum values of NDVI , the area proportion of NDVI values of 145+, and have negative correlation to the area of NDVI values of 145-. The malaria epidemic regions are in accordance with which regions with NDVI values of 145+ continue for 9 months or more. Negative binomial regression analysis showed that mean and maximum values of NDVI , area proportion of NDVI values of 145-, months with NDVI values of 140+ were significant variables. And based on the statistically demonstrated relationship between the mean NDVI values and the malaria incidence, NDVI-1 (<125), NDVI-2 (125-140) , NDVI-3 (140-145), and NDVI-4 (>45) were considered to be indicative of areas without risk of malaria transmission(incide...
|
PDF Full Text Request