It is known that snowcover properties change rapidly due to effect of weather and radiation, detailed models mapping effect of weather and radiation processes to evolution of snowpack have been developed. These models are capable of accurately simulating entire evolution of snowpack at a specific point if a sufficiently detailed time-series of weather and radiation parameters affecting the point is known. In this study we consider the reverse problem of finding the weather and radiation parameters that lead to changes in snowpack parameters, we have used a simulation approach to study the feasibility of finding this reverse map. We mapped a time-series of snowcover states to their corresponding time-series of weather and radiation states using a machine learning model. The data of snowcover states was generated using a well known and rigorously validated snowcover simulation model (SNOWPACK). The results of our experiments show that snow surface time-series contains important information about the meteorological time-series affecting it. We were able to find the meteorological parameters from the simulated data under certain conditions, we expect these results to generalize with actual data. There maybe important applications of these results in optimization of weather data collection systems, weather interpolation algorithms and downscaling algorithms, combining the snowpack data with weather observations can lead to improvements in these algorithms. This study makes a preliminary feasibility study of the reverse problem, our results are positive and encourage further field work using actual data.