Elevation and tide modelling improvements

[robbibt]

Some important fixes to both the elevation code and ensemble tide modelling code:

Elevation:

• Added a min_count param that masks out any rolling median windows with less than N valid satellite observations. This greatly improves noise in the lower intertidal
• Added min_periods param and set sensible default to ensure the rolling mean smoother has enough data points for reliable results (also improves noise on lower intertidal)
• Moved all elevation param logic out of pixel_dem_debug and into pixel_dem so it only has to be maintained in one place, and made some assorted cleanups to pixel_dem_debug

(before vs. after)

Tide modelling:

• Rather than a big if/else that runs either pixel_tides or pixel_tides_ensemble, the pixel_tides_ensemble func has been updated to support both ensemble tide modelling and individual models (acting as a thin wrapper around pixel_tides if a single model is provided). This greatly simplifies code and makes it harder for silly mistakes to creep in.
• pixel_tides_ensemble has been updated to make sure important params like reduce_mean and resolution are actually passed through - this was silently failing before and we had no idea.
• Exposure and elevation code has been cleaned up to use the updated function

Other:

• All notebooks updated
• In-progress commented code temporarily removed from exposure script - this content is still available in the exposure_filters branch, and removing it from main will make merging those changes much easier in the future.

[github-actions]

For full integration test results, refer to the Tests directory README.

[vnewey]

Nice work Robbi!