Vertical/Horizontal Streaks in Model Output

[Jonathan S. Clough]

FAQ:  What are these vertical/horizontal streaks in model output:

Answer

Soil Saturation: 

Maps of SLAMM results occasionally include green linear stripes of fresh marsh moving through what was previously dry land (red).  These stripes are a result of the SLAMM soil saturation algorithm which is geographically quite simple.  The water table for each cell is estimated by moving unidirectionally in the off-shore direction and finding the nearest wetland cells.  This water table is then adjusted for the estimated increase in the fresh water table due to sea level rise.  If the estimated water table is greater than the elevation of the dry land, saturation takes place and the dry land is predicted to be converted to wetland.
 
Because of the unidirectional search algorithm, horizontal streaks may appear on model results as shown above.  Despite this minor loss of geographic realism on output maps, the algorithm does provide useful information about the susceptibility of dry land to water saturation due to changes in the water table.

[Dean]

Jonathan,  thanks for this. I was just thinking I needed to ask about that.

-Dean

[Jonathan S. Clough]

You're welcome.  I'm thinking I should include an "include soil saturation as a process" flag in SLAMM6 so that it can be turned on and off...  -- J

[Dean]

Yeah, I think that would be great idea. I was presenting some preliminary results and the GIS staff I was working with thought the public would be skeptical of the output if it didn't look "real." I agree, and so we were thinking about how to take the lines out of the final outputs for public presentation. Having the option to turn soil sat. off would be good.

-Dean

[Jonathan S. Clough]

Turning off soil saturation  is now an option in SLAMM 6.0.1.  -- J

[Dean]

Awesome! Thanks, Jonathan.  We'll let you know how our year 2 results come out next month.  Looking forward to getting back into this after a long time out.

-Dean

[Anna]

Hi,
I'm getting the striping problem with other land use categories such as tidal flats and estuarine beach.  This is occurring with the soil saturation function turned on and off.  Any ideas?
Thanks

[Jonathan S. Clough]

The other model capability that could cause this is overwash, especially when run at cell sizes of under 30M.  We don't recommend running overwash at resolutions of under 30M.

If this is not causing the problem, look to ensure that the streaks are not artifacts within the elevation layer. 

If neither of these cause the problem, write back and we'll continue to help you search for the cause.

We are currently looking for funding to improve the geographic realism behind soil saturation and overwash when run at smaller cell sizes.  We also would like to refine these portions of the model with an updated literature search and potentially improved model formulations.  -- J

[Anna]

Thanks!  That helped substantially.

[RRunting]

Hi,

I'm having some problems with horizontal streaking in my model output. I am using a 5m resolution DEM and a sea level rise of 1.8m. I find there is no to minimal streaking if I use a lower sea level rise (eg A1F1) or a coarse (e.g. 30m) DEM (but I would like to use the finer resolution DEM and higher sea level rise).
I have tried checking/unchecking overwash, but this didn’t help (nor did changing all the overwash parameters to 0).
I've also tried checking/unchecking soil saturation, which also didn’t help (although checking it made it even streakier).
I've experimented with altering many of the other parameters (e.g. use elevation pre-processor, putting 'false' next to accretion parameters) but this also didn’t help.
I couldn’t find any artifacts in the elevation dataset on a visual inspection.
Is there a solution to this using my current data resolution and sea level rise? Or should I scale up the elevation data?

Thankyou,
Rebecca.

[Jonathan S. Clough]

We haven't seen this type of streaking in our other 5M analyses, though I guess it is possible that erosion predictions could cause such streaking.  If you'd like to send us a screen capture of your maps we'd be willing to look at it and try to discern what may be happening.   Best regards -- Jonathan