Here are a few notes on things that could be improved in a future workflow: - Filter pixels within bins out of 5 NMAD - Test LOWESS filtering for hypsometric interpolation - Include along-track bias correction