Arctic sea ice acts as a crucial barrier, separating the warm ocean air from the colder atmosphere, thereby playing a significant role in regulating the climate. Heat transfer primarily occurs through narrow linear openings in the ice, known as leads. This study introduces a spatiotemporal clustering method to identify and estimate leads from trajectories derived by the RADARSAT Geophysical Processing System (RGPS). The extracted features will be instrumental in calibrating the MAPS-MPM model. To fully quantify model parameter uncertainties, we employ the Bayesian emulation and calibration method, and are in the process of developing an emulator for the high-dimensional spatiotemporal model outputs.