The emergence of digital twin amongst the scientific and engineering applications has paved the way for integration of data and classical simulations towards prediction of system behaviour. The virtual replica of a physical system can be evaluated in numerous scenarios, e.g. for performance, that otherwise would require substantial amount of time and capital. Our application of digital twin concerns with the melting robots, i.e. cryobots, which are designed to access the remote ice selves for retrieving the geo-physical data e.g. in Antarctica. Future exploration missions however would target the icy moons of our solar system, like Europa, which in way necessitates the extrapolation of cryobots performance to extreme environment conditions. Cryotwin, a digital twin for cryobots, provides the necessary virtual infrastructure for evaluating the efficiency and trajectory of cryobots pertaining to a given environmental condition, such as in Europa. Physics based models, numerical and semi-analytical, form the backbone of Cryotwin and is extended by the inclusion of measured data. Given a particular input power, the efficiency model helps in predicting the actual usable power for propagation along with melting velocity and this is then used for trajectory calculation. To end we have developed a hierarchy of efficiency models [1,3] and they are used for calculating trajectories for different cryobots [2,3]. In this contribution we present trajectory calculations, performed in the Cryotwin, using yet another efficiency model, which extends the previous. The unknown environment, such as in Europa, brings us the challenge of uncertainty in conditions and thus require sensitivity analysis of design parameters, like maximum melting depth. The numerical and semi-analytical solution strategies do not comply well with these requirements as they are computationally expensive. The Cryotwin is equipped with gaussian emulator that assists in sensitivity analysis and in real-time solution strategies. We also present the trajectory predictions by the gaussian emulator and verify them against the calculated results.