Time series extractions - bug for mesh across the IDL

IDL: International Date Line (+/- 180 meridian)

For how it is implemented in the IN_TRIANGLE routine:

The test loop goes like this:

for every tide gauges
    for the triangles
        check if point is in triangle 

When dealing with mesh that lies across the IDL (like Pacific Ocean model or Global models), this will not work.

At some point the IN_TRIANGLE check will see a triangle, that is at the same latitude, that overlaps the IDL and "contains" the target point, exits the loop and get the triangle id.

graph example, for the first 20 gauges. Stars are the triangles vertices

Fix will follow the creation of this issue.