Fix fort.19 I/O to allow multiple ADCIRC open boundaries

[gajanan-choudhary]

Currently, only 1 "open boundary", i.e., an ADCIRC boundary with water elevation specified, works correctly. This is because ADCIRC reads in data from fort.19 as it marches in time. I had planned to replace fort.19 with a fort.19.new file which would contain correct values. However, the values in fort.19.new are currently only correct for the coupled boundary. We need to either read in correct values from fort.19 and write them in fort.19.new, or read in values from fort.19 inside Python, and then replace the values at coupled boundary with the correct values and increment the BC timestamp so that ADCIRC skips reading fort.19.