News:

Due to SPAM attacks, new members must be approved before posting.  Please email jclough@warrenpinnacle.com when registering and your account will be approved.

Main Menu

subsites

Started by Dean, November 16, 2009, 01:51:30 PM

Previous topic - Next topic

Dean

Is there a better, more accurate way to define the subsites by creating subsites from polygons,etc?  For example, using a USGS quad as the delineating lines for the different NWI dates and NED dates would be much more accurate than hand drawing approximate boundaries.  Am I missing something about how to do this in slamm 5?  Will slamm 6 have the capability to handle imported polygons for these subsites (or at least a grid file with the correct dates)? I did notice in the site file that there's some reference to this grid-derived date file, but did I overlook the technical notes on how to do it?

Thanks for the help!
Dean

Jonathan S. Clough

#1
Thanks for the feedback.  This capability (to import shapefiles) is not something that we currently have planned for SLAMM6 unfortunately.  I agree that it would be quite useful.  It might not take too much effort to undertake such a capability but it has not been funded at this time. 

The use of a raster input site to designate NWI photo date was a short-lived capability in the model but using the hand-drawn polygons proved to be more useful ultimately so it was dropped from the model.

Hand drawn input polygons are a bit crude but I have found them to meet our needs without too much difficulty in the majority of cases.

-- Jonathan

Dean

Hmm, but then that leaves the areas along these hand drawn regions with higher error, regarding nwi and ned dates. This would matter more too, the larger the difference in the datasets, for the datum adjustments, right?  What are the coordinates of the corners in the subsites txt file? UTM? If they were in the same coordinate system of the input data files, then one could obtain the exact locations from the shapefiles in ArcMap, and manually correct these in the SLAMM output subsite txt file. Would that be possible, to lessen the error in the subsite "border" regions regarding datum corrections?

-Dean

Jonathan S. Clough

Unless there is a dramatic difference in terms of NWI dates, the model may not be particularly sensitive to the way that the sub-sites are drawn at the boundaries.  If the NWI map seems fairly contiguous between quads (as most of them do, even given varying NWI dates) not much may have changed in the interim. 

Additionally, by zooming in to 200% when drawing boundaries, the quad markers may be drawn fairly precisely.

The situation is similar with DEM dates.  SLAMM tries to use the DEM date and the estimated rate of land movement (based on historical SLR) to create a land elevation estimate at the time of the NWI photo.  Unless land is dramatically vertically moving the model is not likely to be sensitive to differences in the DEM date, especially if we're only talking about a "pixel or two" of misspecification in the boundaries. 

Let me put it this way:  I've never seen artifacts in model results as a result of mis-specifying these boundaries by a pixel or two, though we try to be as precise as possible.

The coordinates of the subsites.txt file are "pixels relative to the upper left corner" so they are not georeferenced.  If you are concerned about errors in the border regions you could do some precise calculations as to pixel locations and edit the subsites.txt file but I would be surprised if you would see any change in model results.   

Please let me know the results of your ongoing investigation and think that this is an important capability to add to the model.  In my experience, it really might not be very important, so has not made the "to do" list.