Study Area

The concept of “breadth versus depth” comes into consideration in this discussion. The smaller area that is covered the more time that we will have for incorporating site-specific knowledge from experts and to search for site-specific parameters for each site. Alternatively, a fairly simple model such as SLAMM may be applied to fairly broad areas due to its relatively limited data requirements.

Related correspondence:

** It was suggested by Virginia Burkett that we not include barrier islands in the model simulation. The barrier island simulation portion of SLAMM is subject to considerable uncertainty due to the simple nature of the overwash model and also uncertainty as to the size and frequency of storms that promote such overwash and resultant migration. We are soliciting additional thoughts on this advice.

** Patty Glick, 4/23/09 -- “The identification of the our specific study site(s) is definitely something we need to do soon. I do think it may be worth trading off having a broader study area but at lower resolution (given computational constraints) for more rigorously modeled sub-sites, if those of you who are working "on the ground" think the latter would be the most valuable output. On the other hand, I wouldn't want the study sites to be so small that we might miss some areas for potential habitat migration, etc. I'll put it to the group for further discussion.”

** Jonathan Clough: Regarding the question of model resolution and computational constraints of the model: The largest site that we have modeled to this point had 90 million land-cover cells (Chesapeake Bay.) This is certainly within the scope of this project and, if modeling at a 30 meter cell-size represents 8 million hectares. If we reduce the cell size to 5 meters, though, the coverage area reduces by a factor of 36 to 0.2 million ha. A 64 bit implementation is coming along which may reduce the memory limitations.

Our concern regarding coverage is primarily one of ground-truthing, time spent working with results, time spent talking to local experts, time spent finding site-specific parameters. As the study area becomes more limited we have more time to work with each area being modeled.

** James Pahl writes:
“Is it possible, instead of running the entire coast at 30m resolution for all scenarios, to run component basins of the coast at a higher resolution and with fewer scenarios? The reason for this is that the State is putting into place a set of three Regional Planning Teams for future project planning for the coast. Those RPTs are going to be roughly organized as such:

RPT 1: East of the Mississippi River (i.e. Lake Pontchartrain Basin and Breton Sound)
RPT 2: Mississippi to Atchafalaya River
RPT 3: West of the Atchafalaya River (Chenier Plain and Acadiana Bays)

Higher resolution SLAMM runs on this type of regional basis would provide a very valuable data source to and be directly applicable to future project and program planning of the State. This also might help on the resolution question that we didn't get to in today's conference call, regarding such issues as canals and protection levees.

Also, it might help with comparability with the output being generated by Rafael Calderon and Chris Zganjar at TNC. They seem to be restricting their runs in the Gulf to the county or sub-basin level (e.g. Jefferson County, TX). I don't know what level of resolution US F&WS has planned.” – James Pahl, 4/23/09