From these profiles we can compute climate averages, contour maps, and wind fields — and watch the atmosphere expand as it warms. We note that in the lowest few km where moisture can become prominent the contributions of temperature and moisture to the signal delay become ambiguous. However, since we can model (project downward from ~5 km altitude) the temperature based on climatological patterns, we can then recover moisture in the lower atmosphere with high accuracy. Actual weather models generally don't bother with this. They simply ingest the pure refractivity or bending angle measurement, which is not ambiguous, and let the detailed physics built into the model, combined with information from other sensors, sort out the relative contributions of temperature and moisture at different altitudes, a strategy that has proven highly effective.