13. Integrating a SARIMAX model and climate scenarios to project dengue incidence in An Giang province by the late 21st century
Abstract
Climate change is a critical driver of dengue transmission in tropical regions, yet long-term local-scale projections in Vietnam remain scarce. This study applies a Seasonal Autoregressive Integrated Moving Average with Exogenous Variables (SARIMAX) model (2,0,1)(0,0,1,12) to quantify the relationship between climate variables and dengue incidence and to project future incidence in An Giang province throughout the 21st century. Monthly dengue case counts (2001 - 2021) were integrated with outputs from 20 global climate models under Shared Socioeconomic Pathway (SSP) scenarios. Climatic predictors were reduced using Principal Component Analysis (PCA) to address multicollinearity, with lagged associations incorporated to improve model performance. Temperature exhibited the strongest association with dengue incidence at a lag of 2 - 3 months, whereas precipitation showed significant effects at lags of 0 - 2 months. Projections indicate a consistent upward trend, with dengue incidence increasing by approximately 25 - 30 % under high emission SSP scenarios by the late 21st century. These findings highlight the importance of integrating climate information into local dengue early warning and forecasting systems.
