| Summary: | The sea ice community seeks an integrated-instrument approach to measure sea ice thickness from its components of draft, freeboard, and surface elevation (including snow loads). Current bias error estimates are included in archives. However, many issues are not yet accounted for. Hence, it is very timely to ask the science question: What are the different types of uncertainty which impact the measurement accuracy of sea ice thickness, its distribution, and resulting volume? This project developed a collection of "demonstration papers" to identify and support solutions for the integration of data sets across multiple scales. The essence of the problem comes down to the interpretation of data collected by one instrument with one footprint size versus another instrument with a different footprint. Plots of sea ice thickness frequency distribution serve as a “Rosetta Stone” within the sea ice community to communicate sea ice thickness information between scales. Spatial aliasing distorts this communication across the scales by introducing false peaks, like garbled language, into our Rosetta-Stone communicator. The problem introduces a resolution error as a function of the size (length scale) and shape (waveform) of an instrument’s footprint. We are finding that instruments with measurement footprints which exceed sea ice features incorporate spatial aliasing into their thickness records. Aliasing occurs whenever deep narrow features are smoothened into wider shallower features. The integrated volume is not initially impacted and hence the problem is accepted as a small second-order effect for individual instruments. But the conserved ice volume is not an archived data value. The archived data values are thickness and associated frequency distributions which are important model parameters for heat fluxes at the air-sea interface. The problem is only now becoming relevant as geolocation errors diminish to the benefit of coincident measurement campaigns. Specifically, aliasing increases when we combine measurements from different instruments at different scales. This is a problem that needs to be addressed because sea ice scientists are being asked by non-scientists and policy makers for higher data quality so that they can make critical decisions about global human activities. Spatial aliasing is a ubiquitous problem that is not limited to sea ice thickness nor is it limited to instrument measurements. The problem becomes most pronounced for strongly non-Gaussian (skewed) distributions such as sea ice and snow thickness. In particular, cases with strong bi-modal distribution can incur resolution errors which impact large-scale sea ice volume estimates.
|
|---|