topic Re: Determining the observation point with largest viewshed in Renewable Energy Questions

Determining the observation point with largest viewshed

ToryElmore — Mon, 16 Nov 2015 20:57:52 GMT

I am looking for the optimal location to place a radio repeater on a 25,000 acre preserve. Therefore, I need to determine a point with the largest viewshed of the property that is within a 1/4-mile buffer of a hiking trail or road.I have layers for the preserve boundaries, DEM, trails/roads, and a 1/4-mile buffer all laid out, but I am unsure how to proceed. I know there are viewshed and observer point tools, but I am not certain how to use them to my desired end.

Re: Determining the observation point with largest viewshed

DanPatterson_Retired — Mon, 16 Nov 2015 21:02:21 GMT

are there any fire observation towers in the preserve?

Re: Determining the observation point with largest viewshed

ToryElmore — Mon, 16 Nov 2015 21:04:20 GMT

There are not, unfortunately.

Re: Determining the observation point with largest viewshed

DanPatterson_Retired — Mon, 16 Nov 2015 21:15:30 GMT

I was just trying to think of a shortcut for you to potentially find sites with optimal view. You could isolate areas of maximal elevation within your polygon buffer, however, that still doesn't guarantee that would be the optimal site. In any event, I doubt that a ground based point is going to be appropriate so perhaps you have other options (like a tower) that you are considering. This could broaden the candidate sites since there is going to be more that elevation that needs to be considered.

Re: Determining the observation point with largest viewshed

DarrenWiens2 — Mon, 16 Nov 2015 21:24:38 GMT

I would start by identifying ridgelines and sampling those pixels as potential observation points. You can get at ridgelines by finding pixels where flow accumulation = 0.

Re: Determining the observation point with largest viewshed

ChrisDonohue__GISP — Mon, 16 Nov 2015 21:59:51 GMT

I'd start with the suggestion by Darren Wiens, and then consider using Modelbuilder (or Python) to iterate a Viewshed on each potential point (this may take several hours of run time), then compare the results.

Some thoughts/questions:

If you have not done a Viewshed before, I would do some trial runs to get a feel for it.
If you can employ a tower, you will need to know the height for the OFFSETA (attribute field)
It will probably be worthwhile to employ a RADIUS2 value (attribute field) to exclude distant cells outside your area of interest (this will help reduce processing time). For the same reason, it may be useful to employ a AZIMUTH1 and AZIMUTH2 (attribute fields) if your observation point is to one side of your buffered area.
The Curvature and Refraction corrections could get interesting, as they are designed for light. How does your radio signal propagate? Is it line-of-sight or does it go further?
Does the Repeater receive signals from another radio source? Would it be limited in placement by where it could receive the signals to repeat?
For best results a complete DEM will be needed. I mention this as your example seems to be missing some portions of the DEM where there is a trail buffer.

ArcGIS Help (10.2, 10.2.1, and 10.2.2)

Chris Donohue, GISP

Re: Determining the observation point with largest viewshed

ToryElmore — Mon, 16 Nov 2015 22:07:33 GMT

Ok, these all seem like good suggestions. I need to try out the Viewshed tool a bit before I determine how to proceed. The DEM is clipped to our property boundaries, but I see now that the inholdings need to be represented regardless.

Re: Determining the observation point with largest viewshed

ChrisDonohue__GISP — Mon, 16 Nov 2015 22:13:18 GMT

Another idea:

Besides Viewshed, look into Visibility (Spatial Analyst). I have not used this tool, but it seems to have much of the functionality of Viewshed but with some added functions. Using an analysis type set to Frequency may be the way to go.

ArcGIS Help (10.2, 10.2.1, and 10.2.2)

Chris Donohue, GISP

Re: Determining the observation point with largest viewshed

DanPatterson_Retired — Mon, 16 Nov 2015 22:24:01 GMT

Iteractive visibility add-in http://www.arcgis.com/home/item.html?id=c6d74b84ab37454d9ae8cae6bbe4955f

Viewshed along a path (just add tower height to the extracted path elevation:

http://www.arcgis.com/home/item.html?id=3528bd72847c439f88190a137a1d0e67

Re: Determining the observation point with largest viewshed

ToryElmore — Mon, 16 Nov 2015 23:21:56 GMT

I decided to use Flow Direction & Accumulation to ID ridgelines. I converted those pixels to a point shapefile, then split the shapefile into individual points using Dan Patterson's Split Layer By Attributes tool. I used Model Builder to iterate Viewshed on each point. Need some time to let it process, but I am hopeful.

If you have suggestions on what I should do differently, I'd still like to hear them. Thank you Dan Patterson, Darren Wiens, and Chris Donohue, GISP

Re: Determining the observation point with largest viewshed

DanPatterson_Retired — Mon, 16 Nov 2015 23:24:44 GMT

please post your results...I would be interested in seeing the product.

Re: Determining the observation point with largest viewshed

ToryElmore — Tue, 24 Nov 2015 17:44:32 GMT

First, I used Flow Direction and then Flow Accumulation to identify ridgelines. I converted all pixels where flow accumulation = 0 to a point shapefile, then split the shapefile into individual points using Split Layer By Attributes. I built a model to iterate each point shapefile through the Viewshed tool, then to convert to the resulting rasters into polygon shapefiles. I merged all of the polygon shapefiles and selected for only the viewshed polygons. I clipped the polygon shapefile to my preserve boundary layer and added an Area_Acres field to the attribute table and calculated the geometry for each polygon. I selected for the 6 polygons with an area larger than 8,000 acres, and send the resulting polygons and corresponding observer points to the Preserve Manager. Success!

Re: Determining the observation point with largest viewshed

ToryElmore — Tue, 24 Nov 2015 18:32:40 GMT

Here is a picture of the sample output