retain both VRM and rugosity measures

[scw]

After discussion with Lisa Wedding, she suggested we retain the old rugosity calculation, but include the new VRM calculation under the name 'VRM' to distinguish between the two. That way, we'll keep continuity with the older release (someone could reproduce an existing analysis), but provide a better way at 'getting at' what rugosity is meant to represent via the VRM.

[patiam]

Just a quick clarification here- The surface area : planar area ratio metric (ala Jenness) that was included in the original BTM is no more or less a "rugosity" metric than VRM; both are measures of rugosity, and using the name for only the former is not exactly right in my opinion. In fact, I don't believe Jenness ever even called his algorithm or metrics "rugosity"; I think I'm the one that stuck that name on it way back when we began using it in our lab and I mentioned it in some presentations and papers. But I've run into more than one situation recently where the usage seems to be tending towards "rugosity" meaning specifically the Jensess algorithm, as opposed to VRM, and I think we should steer away from that wherever possible. Sorry if this seems like a semantic nitpick, but I do think it matters.

[scw]

Pat,

Great point. I think we do want to use some sort of labeling that will guide users of older versions of BTM to the equivalent tool in this release, but should try to use appropriate names where possible. Jenness (2004) does include the labels 'rugosity', 'terrain ruggedness', and 'topographic roughness' among the keywords for the paper, and states "... [the] ratio of this surface area to planimetric area provides a useful measure of topographic roughness of the landscape." However, I agree that we should be clear in our labeling. The historical chain measurement technique of reef ecologists (straight chain length to contoured length) maps well onto the approach used by Jenness, but as you mentioned it is just one possible label. Perhaps we could have a 'surface roughness' tool group, and then call each tool by a more appropriate name? So 'Surface Area to Planar Area' for Jenness, VRM for the Sappington approach, and others if we added them (fractal dimension, roughness, etc).

It is tricky, because as you mentioned their is the connection in literature between the Jenness technique and rugosity (as for, in example, Wilson 2007). So we do need some way of letting folks know that it is referred to by that label, but simultaneously try to educate. Perhaps some text on the nature of roughness measures (scale dependent; effects of resolution and neighborhood size) would help clarify the issue.

For reference, here's the help documentation topic on the tool from the older release.

Rugosity can be best defined as the ratio of surface area to planar area. Basically, rugosity is a measure of terrain complexity or the "bumpiness" of the terrain. In the benthic environment, rugosity can be used to aid in the identification of areas with high biodiversity, depending on the scale of the input bathymetry.

The BTM uses a process developed by Jeff Jenness to derive rugosity from an input bathymetric data set. This novel methodology creates an output that is similar to a Trangulated Irregular Network (TIN), but doesnt require the ArcGIS 3D Analyst extension for its creation (figure 1). Similar to BPI data set creation, rugosity derivation relies, in part, on a neighborhood analysis using a 3 grid cell by 3 grid cell neighborhood. An algorithm is passed through the Raster Map Algebra Operation object within Spatial Analyst that calculates the planar distance between the center point of the center cell and of each of the eight surrounding cells in the neighborhood. Next, using the Pythagorean Theorem, the surface distance is calculated for each planar distance using the difference in elevation between the cells. The result of this function is sixteen separate grid data sets with each cell value equal to this surface distance. The next step in the process is to calculate the area formed by three adjacent sides. The result is eight triangular surface area grids. These grid datasets are combined to obtain a surface area data set for the input bathymetric data set. The final step in the process is to create a data set that represents the ratio of surface area to planar area. This final data set represents rugosity for the study area.

Jenness, Jeff S. "Calculating landscape surface area from digital elevation models." Wildlife Society Bulletin 32.3 (2004): 829-839.

Wilson, Margaret FJ, et al. "Multiscale terrain analysis of multibeam bathymetry data for habitat mapping on the continental slope." Marine Geodesy 30.1-2 (2007): 3-35.

[patiam]

Good points, Shaun. The analogous nature of the SA:PA method to the diver chain method is definitely something that led me to begin exploring its use way back when. At the risk of belaboring the point, here's a timeline:

2002 - Jenness introduces his Surface Areas & Ratios Tool for Av 3.x http://www.jennessent.com/arcview/surface_areas.htm Note the lack of the use of the term "rugosity" in the description page, the user's manual, and the original poster.

Later in 2002- I began hyping the idea of using the Jenness algorithm as one type of rugosity metric in posters and presented papers at a number of meetings. Also began beating the TPI drum, stealing the idea from Andy Weiss's poster I saw at the ESRI UC the previous year.

2004- Jeff Jenness begins using the term rugosity, including it as a keyword in his WSB paper. Presumably this is influenced by other folks using the term to refer to what his algorithm calculates.

2005- Original BTM released by Dawn's lab & NOAA CSC, making it easier for folks to calculate these metrics (Yay Dawn!) and further associating the term "rugosity" with the Jenness algorithm.

2006-2007 - Lundblad and others (incl. Wilson et al.) are publishing papers using the Jenness algorithm for "rugosity", as well as TPI/BPI.

2008-present - Usage in lit & elsewhere varies; some think of "rugosity" as a generic type of metric, which can be calculated various ways, while others associate the term specifically with the Jenness SA:PA method. I'm in the former camp, and am only picking this nit b/c I want to encourage that usage, while acknowledging my unintentional part in initially encouraging the "rugosity = SA:PA" definition (the latter).

I agree that the best way to address this in the BTM would be to have a category of roughness, complexity, or rugosity tools, within which we can have SA:PA, VRM, and perhaps other measures such as slope of slope, variability of flow direction, etc. Probably more than $0.02, but those are my thoughts...

[scw]

@patiam just to let you know, I've added back the SA:PA tool, and will be releasing a new version with this and a slew of other fixes in the next few days. Once that release is out, would you be able to look over the terminology I've used to make sure its consistent with what we've discussed? Thanks!

[patiam]

@scw sounds good, i'm a bit scattered with field work and teaching, but give me a heads up when it's ready & i'll try to look at it ASAP. keep up the good work, and thanks!

[scw]

Sounds good @patiam. I'm going to post to the BTM list and ask folks to bang away at the current release, so hopefully we can get the documentation issues and labeling worked out in the next month or so.

[patiam]

Looks like this issue has been closed but here are my initial comments regarding the 3.0 release documentation & functionalit(using 10.1):

• SA:PA "normal" range mentioned in help should be more like 1 to 4
• Add the standard Jenness image w/ the grid & triangles to the tool help

• SA:PA tool threw an error during intermediate data deletion because it is looking for a shift_x.tif.tif file (extension is doubled); final output is still created despite fail message

  Deleting intermediate data... ExecuteError: Failed to execute. Parameters are not valid. ERROR 000732: Input Data Element: Dataset D:\test\test_BTM\shift_1.tif.tif does not exist or is not supported Failed to execute (Delete). Executing: Delete D:\test\test_BTM\shift_1.tif.tif File Start Time: Tue Mar 04 19:12:06 2014 Failed to execute. Parameters are not valid. ERROR 000732: Input Data Element: Dataset D:\test\test_BTM\shift_1.tif.tif does not exist or is not supported Failed to execute (Delete). Failed at Tue Mar 04 19:12:06 2014 (Elapsed Time: 0.00 seconds) Python Error: File "C:\Documents and Settings\peej\Local Settings\Application Data\ESRI\Desktop10.1\AssemblyCache{E3209541-0C78-CF2E-7CCE-261C8208CB1E}\toolbox\scripts\surface_area_to_planar_area.py", line 163, in main arcpy.Delete_management(rel_path) Completed script surfacetoplanar... Failed to execute (surfacetoplanar). ## Failed at Tue Mar 04 19:12:06 2014 (Elapsed Time: 17.00 seconds)
• Toolbar does not automatically come up (or appear as a choice to be added manually) after installing Add-in until ArcMap is closed and re-opened
• Help and perhaps parameter entry descriptions for Statistical Aspect need to explain the cos(Aspect) = "northness" and sin(Aspect) = "eastness". A citation would be good too, like Roberts, D. W. 1986. Ordination on the basis of fuzzy set theory. Vegetatio 66:123-31.
• Standardize BPI Grids still requires exactly 2 (rather than n) input rasters (sorry, this is a sore spot)
• Might be good to explain the slope calculation method in help

If I think of or find more, I'll add them. Thanks for all the great work! Cheers, -pat

On Thu, Feb 20, 2014 at 11:38 AM, Shaun Walbridge notifications@github.comwrote:

Sounds good @patiam https://github.com/patiam. I'm going to post to the BTM list and ask folks to bang away at the current release, so hopefully we can get the documentation issues and labeling worked out in the next month or so.

Reply to this email directly or view it on GitHubhttps://github.com/EsriOceans/btm/issues/6#issuecomment-35660358 .

Pat J. Iampietro Research Associate Seafloor Mapping Lab Institute for Science & Environmental Policy California State University, Monterey Bay 100 Campus Center Seaside, CA 93955 (831) 582-4214 (831) 582-3073 FAX piampietro@csumb.edu http://seafloor.csumb.edu

[scw]

These are all great Pat, thanks! I'll add separate issues for these, and make sure they are addressed before we produce a 3.0 final release.

[scw]

Note that on the issue "Toolbar does not automatically come up (or appear as a choice to be added manually) after installing Add-in until ArcMap is closed and re-opened" this is a limitation of how ArcGIS handles Add-ins -- they only will show up after restarting ArcMap. Should I include a note about this in the installation instructions?

[patiam]

I see. The simply adding a sentence to the install instructions should do the trick. As it is, the user kind of expects to see it appear in their current session. Thanks, -pat

On Thu, Mar 6, 2014 at 10:34 PM, Shaun Walbridge notifications@github.comwrote:

Note that on the issue "Toolbar does not automatically come up (or appear as a choice to be added manually) after installing Add-in until ArcMap is closed and re-opened" this is a limitation of how ArcGIS handles Add-ins -- they only will show up after restarting ArcMap. Should I include a note about this in the installation instructions?

Reply to this email directly or view it on GitHubhttps://github.com/EsriOceans/btm/issues/6#issuecomment-36971605 .

Pat J. Iampietro Research Associate Seafloor Mapping Lab Institute for Science & Environmental Policy California State University, Monterey Bay 100 Campus Center Seaside, CA 93955 (831) 582-4214 (831) 582-3073 FAX piampietro@csumb.edu http://seafloor.csumb.edu

[scw]

Pat, I've added new documentation to the README.md which includes the step of closing ArcMap, and a short GIF clip to show the process. It'll be included in the next release in the documentation, and is now included on GitHub and the ArcGIS Online download site.