Station elevation of 0 for dry nodes can be misleading - should output a sentinel value instead

[PhilMiller]

In testing and discussion with @jduckerOWP we noted that if SCHISM is asked for water surface elevation at a station point in a dry element, the output value is 0. In a modeling scenario around sea level, this non-value could be misinterpreted as an actual value.

I'd suggest using NaN (strongly preferred) or a sentinel like -9999 for the surface elevation of dry station points instead, so that it can't possibly be misinterpreted.

[josephzhang8]

Actually the elev output near wet/dry can be non-0 as well. In dry instances, it can be any value below the local bottom. This is important information for developers so we cannot change that.

Therefore, to discern wet/dry at a node, use dryFlagNode in out2d*

[PhilMiller]

Interesting. What does the value at a dry node represent?

And, to be clear, I'm talking about sta_out, which is MPI_Reduce'd to sta_out_gb, and exposed in BMI as TROUTE_ETA2, not the general water surface elevation variables.

[josephzhang8]

Same applies to station outputs for elev. The elevation at wet or dry nodes is part of finite-element budget.
So for station outputs, need to check elev again local bottom. I can help with that if needed.