Landslides triggered by degrading mountain permafrost pose an increasing risk to life and infrastructure in mountain regions due to current climate change trends. Because discontinous mountain permafrost is not directly detectable by remote sensing, permafrost molards offer a way of indirectly detecting its location and state. Permafrost molards are cones of loose debris that are associated with landslide deposits. They originate from icecemented blocks of sediment that are transported downslope with the landslide. These ice-cemented blocks are fragmented parts of the permafrost initially located in the landslide source area. Over time, these blocks degrade into conical mounds with diameters typically ranging from decimeters up to 40 m. They indicate the presence of an area of degrading permafrost at the level of the detachment zone. The physical processes that lead to the formation of molards from the ice-cemented blocks have not yet been studied in detail. Therefore, we perform an experimental study to recreate molards in a controlled laboratory environment and investigate the key formation processes. We downscale the molards to an initial cube side length of ~30 cm, and reduced the model's complexity by using two different types of gravel with a defined granulometry and lithology. We quantified the degradation phase by using a novel photogrammetric time-lapse system to detect changes in the digital elevation model between half-hourly time-steps. For the first time, we successfully recreated morphologies resembling molards under controlled laboratory conditions on a decimeter scale. We find the main processes to produce the final molard shape are cascades of grainfall for slightly cohesive sediment, and individual grainfall for noncohesive sediment. Our experiments reveal three possible cross-section shapes (bell-shaped, triangular, trapezoidal) that correspond to molards that can be found in the field. Along with these field observations, we suggest that it may be possible to identify the cohesion of a molard's sediment based on its morphology. Together with future field data and experiments investigating the granulometry, ice and clay contents of the initial icecemented block, we aim to use molards as a record to better understand the landslide dynamics and state of the mountain permafrost.