Difference of areas with the Obstruction Identification Surfaces tool from ArcGIS for Aviation

1170
4
04-15-2019 01:15 PM
IngJuanMaSuarez
Esri Contributor

Hi!

I'm using the Obstruction Identification Surfaces tool from ArcGIS for Aviation, but there is a difference of centimeters in the length at each threshold of approach to the track, at the end there is a difference of 1 meter of area between the result of the tool and the mathematically calculated result with calculator, paper and pencil. I have ArcGIS Desktop 10.6.1.

Is it possible to know the algorithm that ArcGIS for Aviation uses to calculate the Obstruction Identification Surfaces? I don't know about the aviation, the question is asked by a user and although the difference is minimal, if we want to know why this happens.

Thanks for all.

0 Kudos
4 Replies
DavidWickliffe
Esri Contributor

Hi Juan,

   I am going to contact you directly for more details.  Hopefully we can find out the cause of the differences and can later post the answer.

David

DavidWickliffe
Esri Contributor

Hi Juan,

   I looked at your test data and found that if you use an XY coordinate system focused on your runway, the Obstruction Identification Surface tool will produce results better matching your CAD data.  I tried this by making my own projection from the Transverse_Mercator one used by your dataset.  You can do this in the ArcMap Data Frame Properties> Coordinate System tab.   Select the existing coordinate system and right-click> copy and Modify.   Give the new coordinate system a new, more local name,  and edit the values for Central_Meridian and for Latitude_Of_Origin.  I gave Lat/Long coordinates corresponding to the midpoint of the RunwayCenterline feature.  Ensure you set your ArcMap data frame to your new local coordinate system.

    Now You can use the  Create Aviation Geodatabase GP tool to create a new Airports 18B geodatabase and set the XY coordinate system to your new local projection (and the Z coordinate system to the same used by your test data).     Now when you run the Obstruction Identification Surface tool,  output to the Airspace\ObstructionIdSurface.   You can then use the Polygon To Line GP tool and look at the SHAPE_Length values in the output feature class, in particular see the 280 meter wide Runway Strip width that you reported.   You can also use the Feature Vertices To Points GP tool and do some Measure tool measuring on the points.

 

There are distortions in a projection the farther away from its origin- in this case the intersection of the Central Meridian and Latitude of Origin.  Your runway location is about 617 Km away from the coordinate system’s origin and I see about 1.2 meter difference in the Runway Strip width within this coordinate system.   So it is best to make use of a more locally focused coordinate system.

David

DavidWickliffe
Esri Contributor

Hang on there David!    This tool is supposed to compensate for this issue by projecting the data into a local azimuthal projection in a scratch workspace to generate output with the precise dimensions before projecting back.    We are looking into this further!

DavidWickliffe
Esri Contributor

Hi Juan,

  Looking into this further,  I found that I had made a mistake in telling you that the runway in question was ~617Km from the coordinate system's origin.  I was testing with the country's East coordinate system, but your data uses the West coordinate system.  I mixed up oeste with este!  Lo siento!   But it only further illustrates how that can be a factor.  The runway in question is ~ 55 Km from the Central Meridian of the projection, and distortions occur as you get farther away from it since it is a Transverse Mercator projection.   So, things scale slightly worse at 55 km away- hence 280.0147 instead of 280.0000.

The tool reprojects the runway into an azimuthal equidistant projection centered on the runway midpoint to construct the surfaces.  Then it reprojects the surfaces back into your target feature class.   The farther away the features are from the Central meridian, there greater the scale, distance, and area distortions.  It is just a compromise of using a projection over a wide area.  The coordinates are more accurate in the azimuthal equidistant projection.  The accuracy can be better preserved by reprojecting into a more accurate target location.   So, the suggestion is still to keep your airport data in a projection that is centered on the airport to minimize distortions.  Note:  Other projections like Lambert Conformal Conic are most accurate along the standard parallels, so it really depends on the type of projection used in regards to how you customize those projection parameters. 

0 Kudos