Quantifying the added value of climate information in a spatio-temporal dengue model

摘要: Dengue is the world’s most important vector-borne viral disease. The dengue mosquito and virus are sensitive to climate variability change. Temperature, humidity precipitation influence biology, abundance habitat, replication speed. In this study, we develop a modelling procedure quantify added value of including information in model for 76 provinces Thailand, from 1982–2013. We first developed seasonal-spatial model, account dependency structures 1 month next between provinces. then tested temperature variables at varying time lags, using linear nonlinear functional forms, determine an optimum combination lags describe relative risk. Model parameters were estimated integrated nested Laplace approximation. This approach provides novel opportunity perform selection Bayesian framework, while accounting underlying spatial temporal or forms. quantified additional variation explained by interannual variations, above that provided model. Overall, 8 % variance risk can be variations previous month. inclusion functions framework improved 79 % Therefore, forecast could significantly contribute national early warning system Thailand.