ArcGIS Path Distance: Tobler's Hiking Function

[BrooksLawler]

Hi All,

I am definitely no expert in ArcMap and could really use some help.

I have been working on calculating the cost to travel from raw material toolstone sources in Interior, Alaska using Tobler's Hiking Function. I followed Kaitlin Yanchar's tutorial: ArcGIS Tutorial: Tobler’s Hiking Function (Anisotropic Distance) | Kaitlin Yanchar, MA, RPA in a version of ArcMap 10. I used a 100m resolution DEM, 32 bit floating point raster. I clipped the DEM to only include the analysis extent which is about a 46,000 square mile area, and then followed the steps of the tutorial. Slope is the only friction surface I used. The cost raster and resulting values of the cost raster appear to be anisotropic and correct in relation to one another, but I am having a problem similar to Mathew Schmidtlein. My values are not underestimated, they seem to be significantly overestimated, such that the time it would take to travel several hundred miles, when the cost is converted from minutes to years, would take several years. I followed the tutorial twice and came up with the same values. There is steep terrain in the area, but I looked at the slope values and there are none that are over 70 degrees, so I don't think ArcMap is getting hung up on terrain that may be impossible to traverse. Does anyone have any ideas why I may be getting such huge cost values? Is it possible that the values of the cost raster are not in minutes?

Thank you for your help,

Brooks

[ChrisDonohue__GISP]

One guess from a quick look at the tutorial - the tutorial mentions this:

30m DEM: A Digital Elevation Model with at least 30 meter resolution. These are raster images that contain elevation information and can easily be found online.

The 100 meter DEM you have is very coarse compared to what is suggested, so that may be what is causing things to go awry.

EDIT - forgot to italicize the text from the tutorial to distinguish it from my comment (now fixed),

Chris Donohue, GISP

[DanPatterson_Retired]

I suspect that the cell size and elevation values are not in the same units.  I have seen cases reported when a 30m DEM isn't actually 30m but a fraction of a decimal degree which, if converted to planar units, would be about 30 m (ish).  Check the extents and cell size of the raster, confirm that they are in planar coordinates (ie metric) and that the Z values are in the same units.  If the cell size is actually in decimal degrees, then you would have wildly weird slopes etc since the

z values are metric a the cell size/width are not