Abstract
Climatic variability and rise in temperature are considered as the key determinants to the transmission of malaria. In the present study, the trends in the cases of malaria caused by Plasmodium falciparum and Plasmodium vivax were investigated by using the nonparametric MannโKendall test after removing the effect of significant lagโ1 serial correlation from the time series of cases of malaria incidence by preโwhitening in annual, seasonal, and monthly time scales at Bikaner, located in the Thar Desert of Rajasthan, in northwest India. Multiโcollinearity within the datasets under consideration was investigated by means of correlation matrix, the Bartlett sphericity test, and the KaiserโMeyerโOlkin measure of sampling adequacy, subsequent to which it was removed by using principal component analysis. Finally, artificial neural network models were employed to predict cases of malaria incidence caused by P. falciparum and P. vivax at various scales. During the last 34 years from 1975 to 2008, P. falciparum malaria incidence cases have been found to increase significantly corresponding to monthly (April and September) and seasonal (monsoon) time scales over Bikaner. On the other hand, no significant trends were observed in P. vivax malaria cases at Bikaner. Concomitant increases in P. falciparum cases of malaria incidence and observed temperature increases at Bikaner hint that P. falciparum malaria may have grown significantly under the warming climate of the Thar Desert. Copyright ยฉ 2012 Royal Meteorological Society