SLAMM Forum

General Category => Model Formulation & Parameters => Topic started by: Jonathan S. Clough on November 11, 2009, 02:33:21 PM

Title: Vertical Datums
Post by: Jonathan S. Clough on November 11, 2009, 02:33:21 PM
A Scientist with National Geodetic Survey writes:
QuoteI'm writing to follow up on the question of how the tidal and geodetic datum offsets are applied in 6.0. In a nutshell, the issue is that in 5.0 the tidal / geodetic datum offset value for a single tide station was applied over large geographic areas where that relationship does not necessarily hold constant, therefore adding increasing sources of error/uncertainty in areas farther away from where the tidal geodetic relationship was established.

Jonathan Responds:

I have to clarify a few points about SLAMM 5--

The tidal / geodetic datum could be and usually was defined within smaller polygons or "sub-sites" as we call them in SLAMM jargon.  This allowed us to utilize a gradient based on multiple NOAA tide stations or the VDATUM product when it was available.  In recent model applications we have defined sub-sites as regions in which differences in the vertical datum correction would be minor.  Given that the vertical resolution of our best elevation (LiDAR) is in the range of 5-10 cm we usually would define these areas as having differences of 1 cm or less.  Given the vertical resolution of our input elevation data, the tidal / geodetic datum uncertainty was almost always an order of magnitude or so less. 

Perhaps more importantly, there is an "unpublished" capability in SLAMM 5 to read a raster map of tidal / geodetic datum offset values which we produced using the NOAA VDATUM product.  We have used this capability in many of our SLAMM applications where VDATUM coverages were available.

When there was no VDATUM coverage and considerable distances between NOAA stations, we generally used a spatial interpolation between stations.  In a recent workshop, Allison L. Allen of Center for Operational Oceanographic Products and Services stated that, absent VDATUM results, interpolation of "MTL to NAVD88" between stations is the best manner of determining the "MTL to NAVD88" correction.   

In SLAMM 6 the only difference will be that the VDATUM Raster import capability will be "published" that is to say, written up in the technical documentation and users manual.  Given the fact that VDATUM is now nearly ubiquitous throughout the continental US, I think this should solve the problem for the majority of our users.