The integration of data analysis and simulations has become increasingly important in the design of physical systems. By developing a virtual replica that targets specific functions, a digital twin helps us control the key responses of the system and allows better insight into decision-making processes. Our application of interest consists of thermal drills that access remote ice sheets and retrieve geophysical data, e.g., in Antarctica. Future exploration missions, however, will cover the icy moons of our Solar System, which requires extrapolating the performance of the cryobots to extreme environments and underlines the need for a virtual testbed. Forward modeling is a fundamental component of the testbed. To this end, we derived a hierarchy of models to predict the cryobot’s trajectory [1] and melting velocity [2]. Simulations are then integrated with the Ice Data Hub [3], a GUI-based interface that stores and provides Cryosphere measurement data for reproducible preprocessing. However, when performing live-response predictions during operations, computational times present a difficult challenge. Alternatively, we can employ surrogate models, e.g., Gaussian process emulation, to replace high-fidelity simulations. By lowering the computational costs, real-time analysis can be paired with information acquired by the cryobot’s sensors. In this contribution, we present the cryotwin structure and discuss its potential applications as a digital twin for thermal drills. Additionally, we show results for the prediction of the heat distribution in the melting head and its integration with surrogate models and physical experiments. REFERENCES [1] F. Baader et al., Field-test performance of an ice-melting probe in a terrestrial ana- logue environment. Icarus 409 (2024) 115852. [2] L. Boledi et al., A scale-coupled numerical method for transient close-contact melting. Computers & Mathematics with Applications 143 (2023), pp. 277–288. [3] A. Simson et al., Enriched metadata for hybrid data compilations with applications to cryosphere research, Helmholtz Metadata Collaboration Conference, online, October 5-6, doi: 10.5281/ZENODO.7185422 (2022).