-1

I'm using ESRI's Path Distance Analysis based on Toblers hiking function to estimate the walking time from an archaeological site to establish it's context. Each color represents an hour of walking time, the red line is river which I made 100 meters wide and re-classed to NoData. I think the first four hours of walking make a lot of sense.

But what is going on with when the model hits the coast line?

Especially with the middle blue value (the 7th hour of travel) - Tobler's estimates travel at roughly 5km per hour optimal speed but along the coast it is estimating travel at 16km an hour.

Is this edge distortion?

Toblers attempt

  • Check your table and units. And no travel possible over the river based on your scenario – FelixIP Jun 6 '17 at 7:52
  • The river can be crossed at the four locations where there is archaeological or historical records of bridges. All of them are nearby the location of the site. They are represented in the raster as a break of a few pixels in the river NoData line. – 9spades Jun 6 '17 at 12:54
0

My best guess is that you did something wrong when preparing the DEM and SLOPE rasters for the analysis. It seems that the "presence" of the river is actually disregarded by the analysis, so ArcMAP is not "realizing" that the river is set as NoData. Therefore, I am wondering if, once you buffered the river and reclassified it as NoData, you actually "subtracted" it from the DEM and related SLOPE rasters. In order to do that you may want to add the reclassified river raster to the DEM and SLOPE rasters via RASTER CALCULATOR. The rationale being that, in Map Algebra, data+NoData=NoData. That would allow the river to show up as NoData in the DEM and SLOPE rasters that you use to implement Tobler's hiking function in Path Distance.

  • I should have provided more details in my initial question. I have a few breaks in the river - where there are bridges (or were bridges) contemporary with the archaeological site. That looks as though it is working properly. Along the coast though it seems as if there is are impossible walking speeds (circa 15km per hour). – 9spades Jun 6 '17 at 13:05
  • Sorry but I do not get your point. By "coast" do you mean the left-most part of your raster? If that holds true, what I see is areas with darker colours, relative to the core zone. My best guess would be that those colours indicate values smaller than your core zone. But, not knowing the colour ramp you used, that remains just my impression. Secondly, at the best of my knowledge, in archaeology what it is used is the inverse of the Tobler's function, that is the function solved for time. So, in my experience, the countours indicate walking time-zones (isochrones). – NewAtGis Jun 6 '17 at 18:13
  • If my latter comment holds true, your raster could be read off as indicating that the outermost blueish zones can be reached with a longer (time-wise) walk. I am wondering what procedure did you use to get that raster. – NewAtGis Jun 6 '17 at 18:15
  • The coastline is the far left - I should have included the shapefile for the coastline in my clip. That was an oversight on my part. So the dot at the center is the archaeological site (it is just outside Rome) the river is the Tiber - you intuited correctly that the far left was the coast. I was informed by this lab. – 9spades Jun 7 '17 at 4:27
  • continued...The darkest blue would be about 9 hours travel on foot based on tobler's which suggests around 5kmph max on level ground. According to virtually all pedestrian mobility scholarship is the average comfortable walking speed (it does vary according to many factors but standard is 5km). I think I identified the problem - along the coast most of the elevation was zero or even a negative elevation value. I think it was resulting in a zero cost to travel in some areas. I converted the water from 0 to ND, and added 4 meters in elevation and had much more reasonable results. – 9spades Jun 7 '17 at 4:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.