# Propose a new location for Watch Tower using Viewshed Analysis

I want to suggest location for new watch tower in a forest area, but i am confused , on what criteria i can propose a best location. If i am taking DEM , slope and distance from existing tower , then how can i propose a best location in terms of GIS. I don't want to select points mannualy and check viewsheds , again a new point and its viewshed, until I get a maximum visible area.I want to utilize gis tools. I tried with dem ,slope, proximity from existing tower and there reclassification into 10 classes and then weightage overlay with more percentage to dem, then distance from existing tower and least to slope.But I couldn't get desired results. Please suggest me a solution, I will be very thankful to you friends.

Friends today I have proceeded with given problem as follows: I want to locate a watch tower to monitor forest fire:---- 1) First Criteria: Distance From Roads and Drainage ( 200 meter and 500 meter Buffer respectively), and Intersect them to find suitable areas . 2) Then I have made slope file from SRTM DEM and taken less then 6 degree as criteria (using Raster Calulator) , and I got almost 85 % study area comes under this range. DEM ranges from 400 to 763 for my study area, and using Raster Calculater I got new layer with elevation > 500 (ROUGHLY). And Converted both slope layer and DEM to polygon. 3) ROAD_DRAINAGE_BUFFER is now INTERSECTED with SLOPE_DEM Layer. And I got fewer Polygons, but Still many. Now if I am trying to get Viewshed at highest possible point , then again viewshed from this point, is not giving me maximum visible area. Still I am getting big area to analyse. How to get reduce my area. Still a problem. I don't want to make you uncomfortable , asking same problem. Please If you still give some clues, I will be thankful. Thank You Friends.,....for giving most precious .thing......(TIME)........

-
What are the most important factors for this watch tower? Maximum viewshed for a location? Does it need to see a certain canyon? Suitable soil, accessibility, environment, etc. for site locations? I.e., do some sites provide excellent viewsheds, but are poor locations to build a tower for other reasons? If you can answer some of these questions, you may be able to reduce the areas you need to calculate viewsheds on. Those details may also help us to better formulate an automated solution for you. –  Baltok Apr 15 '13 at 20:06
Thanks for replying, Baltok. My watch tower visibity constraint is OFFSETA 50 feet or 15.24 meter and RADIUS2 is 6000 meter. No soil, or environment factors are under consideration. I don't know exactly which way I have to proceed, or if there any alternative tool to Viewshed. Or there is possible reverse viewshed, that is from area as input and it give us points of interest. –  Engineer_Yudh Apr 16 '13 at 3:26
As @Baltok said: First answer this question: What are the most important factors for this watch tower? (1) I do not understand why a (tourism ?) watch tower should within 500 m of a river. (2) What do you mean with ...it will not giving me desired visible area? Is there a place which must be seen from the tower? –  Jens Apr 17 '13 at 13:47

One of the activities (Activity 3) published recently by Esri discusses step by step details about locating a wildfire observation tower using freely available public data. Here is the Scribd Link

For ready reference, the criteria they have used there:

• Within 100 meters of a roadway, for easy access
• At least 380 meters in elevation, for the best visibility of the surrounding terrain
• On a slope of less than five degrees, to minimize construction costs and maximize ground stability
• On pasture/hay fields, since these are the grassy areas most susceptible to wildfires in this area
-

Since high elevations are more likely to have large visible areas than low elevations, you could start with a suitability analysis that combines elevation and distance from existing tower(s) through a weighted linear combination (or similar). Once you can have a limited number of suggested locations, it's much simpler to run viewshed on those points and find how much area each one can see. This isn't fully automated, but at least narrows down the possibilities more than random guesswork, and provides you with a quantified reason for why you selected that set of finalist sites.

I am not aware of any "reverse" viewshed, at least not in the terms you're describing. It is possible to select "possible fire" points and run the viewshed there to evaluate which "possible tower" area can see it, but that seems to be even further away from what you really want to do.

-
The approach of @Erica is good. Here are a few additions: (1) instead of using high elevations you can also use ridgelines. To find the ridgelines you can compute a flow direction grid. From the flow direction grid compute a flow accumulation grid. Use the pixel where the flow accumulation is zero – these pixel are the ridgelines. (2) if you have a dataset from which you can estimate the tree height: ignore all areas as location for a tower where height of tree is more than height of tower (50 feet or 15.24 meter). –  Jens Apr 17 '13 at 10:48
ERICA>> Thanks For Your Suggestion. I am getting path...........:) –  Engineer_Yudh Apr 17 '13 at 12:41
Thanks Jens>> I will definitely try with your concept of ridgeline. –  Engineer_Yudh Apr 17 '13 at 12:43
Jens>> I am using SRTM as SURFACE ELEVATION DATA. I think , here tree height doesn't make so much difference, Will you please elaborate the concept of Ignoring Tree height for such an instance. Thanks –  Engineer_Yudh Apr 17 '13 at 12:51
the idea of to considering tree height is: if the trees are higher than your tower you will see noting. If you have a good potential place (= height elevation or a ridge) but there is a forest with large trees (e.g. 70 feet) at this place it makes no sense to build a tower of 50 feet. If the tower is lower than the trees at the same place you will only see branches and leaves. –  Jens Apr 17 '13 at 13:21