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

NWI

Started by Loch, March 03, 2010, 10:33:44 PM

Previous topic - Next topic

Loch

G'day. I'm using SLAMM6 for SLR in Queensland Australia. I have a query on (the required SLAMM category) NWI: these don't exist for the region I work with (Non US).
Also, if I was to source aerial imagery for the region, do the Nrows and Ncolumns have to match? The pixel size won't be the same.

Thanks,
Lochran

Jonathan S. Clough

Hello! 

What you need to use the SLAMM crosswalk table provided with the model (SLAMM6_nwi_codes_2009.xls in the same directory as SLAMM.EXE) is a wetlands cover class that uses the Cowardin Classification system.  On the other hand, we have cross-walked several other spatial wetland datasets to SLAMM classes fairly easily.  The SLAMM classification is pretty broad and usually wetland inventories are more precise.  This usually makes for a fairly simple linkage (e.g. "high marsh" or "brackish marsh" --> SLAMM irregularly flooded marsh)  You should read the technical documentation carefully to ensure that you are matching the categories properly and that the SLAMM conceptual model is appropriate for your category assignments.

Yes, all of your raster inputs need to be resampled so that they all have the same numbers of rows and columns and precisely the same lat and long locations.  There is a thread that covers some details regarding this here:  http://warrenpinnacle.com/SLAMMFORUM/index.php?topic=51.0

Good luck and please let us know how it goes. 

-- J

Loch

Thanks Jonathan, we'll re-categorise our wetland classes according to the SLAMM6_nwi_codes. Can we change these codes within the xls - what do you mean by 'crosswalk'? I note that it is read-only.

Also, thanks for answering the query on rows + cols, I got that sorted.

Lochran

Jonathan S. Clough

By "crosswalk," I meant to map a relationship from your existing wetlands classification scheme to the SLAMM categories.  It is a term related to metadata and I may not have used it entirely precisely -- ( http://en.wikipedia.org/wiki/Crosswalk_(metadata) )

I'm not sure why the Excel file is read-only, that must be an artifact of the installation script.  Feel free to copy it into whatever directory you'd like, turn off the read-only, and make changes to that document.   -- Jonathan


Loch

Hi Jonathan,

You don't happen to know what the vertical control datum (VCD) is for Australia, or the Asia-Pacific region? If I use LiDAR data for the Australian region, then will these have a pre-specified VCD?

Lochran

Jonathan S. Clough

The vertical datum for the elevation data should be indicated in the metadata of any elevation data set.  I'm not sure what the standard is for Australia. 

The most important thing is that you convert this to an MTL basis for the use within SLAMM.  A correction may be applied on a sub-site (polygonal) basis using the "MTL minus NAVD88" correction that could also be defined as the "MTL minus [your datum]" correction or it may be applied on a cell-by-cell basis using a VDATUM correction raster.

Loch


Loch

Hi Jonathan,

Re: land-cover categories and the decison tree - are there a minimum number of categories that can be used for a model? Our wetland data comprise just 11 categories and while we may be able to subcategorise some of these, the data just aren't as specific as the US NWI.

Thanks,
Lochran

Loch

Hi Jonathan. Sorry, one more important matter:
In tropical Australia, we have a coastal wetland classification that differs to the US NWI, not only by species type, but also by (species-specific) response to inundation and erosion. We go through the process now of cross-walking our wetland community data, but our communities may not necessarily convert to say open water or tidal flat. One possible option may be for us to alter the (global) source code, so that we derive a new decision tree based on Australian wetland types, with specific response to inundation, accretion and erosion. What are your thoughts on this?
Lochran

Jonathan S. Clough

Hi Loch:

This is a tricky issue.  You noted that one possible option is to alter the source code so that a new decision tree is derived based on Australian wetland types.  Unfortunately, I think that is required for issues raised in both of your last two posts.  If you try to run the model with fewer wetland categories than exist in the model you will, by definition, be collapsing wetland types.  This will likely bring your wetland designations out of phase with the SLAMM conceptual model.

Additionally wetland types or characteristics that are region-specific would need to be added to the model code (or an existing type would need to be modified).

If you are interested in taking these source-code changes on, I suggest that you look carefully at the global.pas, and transfer.inc code.  You should be able to find the decision tree accounting (that matches with "The SLAMM Decision Tree" chapter in the Tech Doc.) about two thirds of the way down transfer.inc.

Perhaps the optimal long-term solution is to bring the land-cover definitions and decision tree flow chart into the graphical interface and make the program completely flexible in this regard  (i.e. the user would be free to designate the conceptual model for each category including category name, elevation range, salinity tolerance, habitat switching assumptions.  Unfortunately, this would be a very time-consuming addition to the code and would require a funding source and then time to implement...

Let me know how you're progressing on this and good luck! -- Jonathan

Loch

Hi Jonathan. Thanks for your last reply.

We've thought carefully about our crosswalk situation and we've managed to transfer most of our wetland types across to ecologically similar (NWI) wetlands. We have one further issue. Some of the tropical Australian wetland communities don't respond to inundation according to the (decision tree) criteria used in SLAMM (Table 3, page 31 Technical documentation). We need to change some of these, otherwise our model will not be relevant. So for example, we require scrub-shrub to change to a community other than salt marsh.
Can we do this without altering the source code? If not, will minor modifications to source code have unforeseen consequences?

Thanks,
Lochran

Jonathan S. Clough

Quote from: Loch on April 25, 2010, 10:41:37 PM
So for example, we require scrub-shrub to change to a community other than salt marsh.
Can we do this without altering the source code? If not, will minor modifications to source code have unforeseen consequences?

This cannot be currently be done without altering the source.  Maybe in a future version! 

This minor modification should not have unforeseen consequences but obviously I can't promise anything about code that others are modifying.  Feel free to post suggested code modifications here for my (eventual) review. 

-- Jonathan