Currently I work as an engineer dealing with water resource issues for the St. Johns River Water Management District.
I initially setup the District's Beowulf cluster (in 2001) running Scyld, then OpenPBS (2002), and now Grid Engine (since 2004). The cluster's initial purpose was to run 3-D hydrodynamic models of estuaries and lakes. It is starting to be used for hydrologic and groundwater modeling and advanced optimization runs. It is a modest cluster consisting (in 2007) of 4 single-core Opteron and 4 dual-core Opteron nodes for a total of 24 slave processors. It has 3 TB of RAID disk space.
I built a parameter space, model data set creation program that might be useful to other Grid Engine users called Pyslice.
Also developed Tidal Analysis Program in Python in case there are any water levels you need to analyze. :-)