Correcting perturbed tracks for landfall

[WPringle]

I think as highlighted by NHC team before, the tracks ideally should be corrected for landfall.
Once the storm makes landfall it should have decay in intensity, so when the TC is perturbed (e.g., along_track), the weaker intensity after landfall could fall into the ocean if the TC is made slower, or the last more intense ocean point could fall on land if the TC is made faster. We can expect similar changes due to cross_track perturbations when the coastline is curved not exactly perpendicular to the track causing the TC to make landfall earlier or later.

The fix would likely involve working out where the coastline is and the time when it crosses the coastline for the original track and each perturbed track. Then interpolating between the two times before and after landfall to get an estimate of the landfall intensity. Then using this information to adjust the perturbed track intensities (I guess this must happen before the intensity perturbation). Need to think about the logic at bit more.