<![CDATA[Accurate rainfall forecasting is crucial in Indonesia, where climate change exacerbates the risks of droughts and floods. This study conducts a comparative analysis of Seasonal Autoregressive Integrated Moving Average (SARIMA) and its extension with exogenous variables (SARIMAX) to evaluate the impact of incorporating air humidity in rainfall prediction for Bandung City. Unlike SARIMA, which relies solely on univariate data, SARIMAX integrates external climatic factors, potentially enhancing predictive accuracy. This study analyzed monthly rainfall and air humidity data from January 2014 to December 2023. The modeling procedure included stationarity testing, seasonal decomposition, model identification using ACF and PACF diagnostics, parameter estimation via Maximum Likelihood Estimation (MLE), and residual diagnostic checks. Forecasting performance was comparatively evaluated using Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), and Mean Absolute Scaled Error (MASE). The findings indicate that SARIMAX consistently outperforms SARIMA, yielding lower AIC and BIC values and achieving a MASE of 0.690 compared to 0.840 for SARIMA. This demonstrates that exogenous climatic variables play a crucial role in reducing forecasting error, particularly for seasonal and climate-sensitive time series. Beyond methodological contributions, the findings offer practical implications: incorporating humidity into forecasting models provides policymakers and disaster management authorities with more precise information for climate adaptation and risk mitigation strategies.]]>
