Using of Climate Information in Early Warning Model (EWM) of Dengue Haemorrhagic Fever (DHF) Incidence
(Case Study in Mataram 2010)
Sandro Wellyanto Lubisa* Daniel Naekab Rahmi Ariania Diana Ra Fitrie Aa Rini Hidayatic
a Department of Geophysics and Meteorology, Bogor Agricultural University, Bogor, 16680, Indonesia
b Coordinator of Indonesian Climate Student Forum (ICSF), Bogor, 16680, Indonesia
c Head of Department of Geophysics and Meteorology, Bogor Agricultural University, Bogor, 16680, Indonesia
Incidence rate of Dengue Hemorrhagic Fever (DHF) is associated to the rain season and tropical lowland areas (Hidayati 2009). Researches which have been working before show that there are significantly relationship between climatic factors and DHF incidence rate. Thus, It is expected that climate elements are able to be used accurately as predictors in arrangement of the DHF incidence prediction model. Models have been arranged based on the method of Least Square Multiple Linear Regression (MLR) and statistical stochastic approach. In the model, predictor variables are only consisting of climate information data and incidence rate (IR) of DHF per total population. Validation is analysed to determine the relationship between the accuracy of model output and actual observations data. Models arrange by using climate information data from Mataram (case study 2010) and results some best prediction models. The first models are arranged by using an IR value a week before prediction point (IRn-1). The prediction models are IRn = 0.889 IRn-1 + 0.0188 CH3n-4, where the value of P<0.05, R-square (adj) more than 85 %, S = 2.48617, Durbin-Watson statistic = 2.10305 and if temperature data is available another model which can be used is IRn = 0.040 + 0.878 IRn-1 + 0.0206 SD4n-4 where the value of P<0.05, R-square (adj) more than 85 % , S = 2.49535 and Durbin-Watson statistic = 2.10632. The second prediction models, arrange by using an IR value two weeks earlier (IRn-2). The prediction models are IRn = 0.789 IRn-2 + 0.0353 CH4n-4 where the value of P<0.05, R-square (adj) more than 75 % , S = 3.23325, Durbin-Watson statistic = 1.09148 and another one is IR = 0.813 IRn-2 + 0.0317 CH3n-4 where the value of P<0.05, R-square (adj) more than 75 %, S = 3.54146, by Durbin-Watson statistic = 1.01878. Validation of the models shows a good value and can be consistently applied in Early Warning Model (EWM), results of the validation of these models respectively are 78%, 79%, 77%, and 77%. EWM is designed by using a visual language program and stochastic spread sheet in a crystal ball software. The output models are expected to be useful to anticipate and mitigate DHF incidence rate through the implementation of mosquito eradication and determination of optimum time for fogging activity.
Keywords: Climate information, Prediction Model, Early Warning Model (EWM)
* Correspondence to: Sandro Wellyanto Lubis, Department of Geophysics and Meteorology, Bogor Agricultural University, Bogor, 16680, Indonesia. E-mail: Sandro.email@example.com. +6281385644350.
Results and Discussion (to be continued)