A Unified Sea Ice Thickness Data Set for Model Validation

Can we, as a community, do better at using existing ice thickness measurements to more effectively evaluate the changing nature of the Arctic ice pack and to better evaluate the performance of our models? We think we can if we work together. We are trying to create a unified ice thickness data set by combining observations from various ice thickness measurement systems. It is designed to facilitate the intercomparison of different measurements, the evaluation of the state of the ice pack, and the validation of sea ice models. Datasets that might be included are ice draft estimates from various submarine and moored upward looking sonar instruments, ice thickness estimates from airborne electromagnetic instruments, and satellite altimeter freeboard measurements. Three principles for the proposed data set are: 1) Full documentation of data sources and characteristics, 2) Spatial and temporal averaging to approximately common scales, and 3) Common data formats. We would not mix data types and we would not interpolate to locations or times not represented in the observations. The target spatial and temporal scale for the measurements would be 50 lineal km of ice and/or one month. Point measurements are not so useful in this context. Data from both hemispheres and any body of ocean water would be included. Documentation would include locations, times, measurement methods, processing, snow depth assumptions, averaging distance and time, error characteristics, data provider, and more. The cooperation and collaboration of the various data providers is essential to the success of this project and so far we have had a very gratifying response to our overtures. We would like to hear from any who have not heard from us and who have collected sea ice thickness data at the approximate target scales. With potentially thousands of individual samples, much could be learned about the measurement systems, about the changing state of the ice cover, and about ice model performance and errors.