Data Worth Analysis for a Fluvial Geothermal System using an Ensemble Smoother with Multiple Data Assimilation

Abstract

Long-term geothermal production is subject to considerable uncertainty due to limited data availability and inherent geological heterogeneity. While observation and data acquisition improve our understanding of the reservoir, they also contribute significantly to project costs. It is essential to identify the most informative observation strategy. In this study, we apply a previously developed scenario-based data assimilation framework that integrates rapid geological modelling, efficient numerical simulation, and Ensemble Smoother with Multiple Data Assimilation (ESMDA) to constrain uncertainties in reservoir properties and production forecasts to a synthetic but geologically realistic fluvial geothermal system and conduct a data worth analysis to evaluate the impact of different observations (production temperature and injection pressure, well temperature and pressure profiles, etc.) on uncertainty reduction. Results show that production temperature and injection pressure alone, though cost-effective, are insufficient to significantly reduce uncertainties in reservoir performance forecasts. In contrast, well temperature and pressure profiles exhibit substantially higher data worth, leading to much better-constrained predictions. Moreover, incorporating a monitoring borehole further constrains uncertainty by capturing subsurface dynamics between the injector and producer. These findings underscore the importance of monitoring pressure and temperature profiles in the wells of a geothermal doublet.