Comparative Analysis of Machine Learning Models for Groundwater Level Forecasting: The Impact of Contextual Data

This paper presents a comparative evaluation of three distinct categories of models applied to groundwater level data: traditional batch learning methods, time series deep learning methods, and time series foundation models. By enriching the water level data with weather-related features, we significantly improved the effectiveness of simpler models. The results demonstrate that, despite their state-of-the-art performance on univariate datasets and the corresponding publicity, advanced models without contextual feature support are still surpassed by traditional methods trained on enriched datasets.