Address non-linear dynamics by filtering statistics points by angle variance

[robbibt]

Currently, rates-of-change statistics are calculated by measuring the nearest distance to shorelines from each year in the time series. This works well for linear erosional/progradation dynamics, but less well in environments undergoing highly dynamic, non-linear change through time (e.g. sandbars, river mouths etc).

One way to address this would be to calculate the variance of angles to each nearest shoreline. In the case of linear erosion/progradation, all angles should be similar. However, more complex examples of change are likely to have a larger range of angles. This angle variance could then be used to filter or flag relevant statistics points as lower certainty.

[robbibt]

Done