# ArcMap Area Calculations vs MultiNet FEATAREA Field Discrepancy

I'm finding a huge discrepancy between the values supplied in the `FEATAREA` field in my MultiNet data and the area calculation results from ArcMap v10.2.2.

In this case, I'm looking at the area of U.S. counties in the `*__A8` MultiNet files. A spot check of the values supplied in the `FEATAREA` field suggests that the values are correct (based on Google searches for county area).

However, when I project the layer to `WGS 1984 Web Mercator` in ArcMap, then add a new data field to the attribute table and do a "Calculate Geometry" on that field, the resulting value is off by a factor of 50 to 100 percent - it varies from one row to the next and isn't a constant scale factor error. I've verified that I'm asking for the same unit of measure (square meters in this case).

Obviously this leads me to question both my methods and the accuracy of the polygon data supplied by MultiNet. It seems more likely that I'm doing something wrong or omitting a required step, but I can't figure out what that might be.

• All projections warp reality to some extent; a global projection like Mercator isn't very well-suited to accurate measurements of area. Try projecting to a Projected Coordinate System such as USA Albers Equal Area Conic, and see if the Calculate Geometry results are more what you expect. Sep 15, 2014 at 19:19
• Erica, you are correct. My concern is with this part, "it varies from one row to the next and isn't a constant scale factor error." Sep 15, 2014 at 19:27

## 1 Answer

Rather than doing a "Calculate Geometry" on the new field, I needed to do a "Calculate Field" and specify the Geodesic Area function (via Python 9.3):

``````   !shape.geodesicarea@SQUAREMETERS!
``````

There was no need to do a projection at all.

For my purposes, calculating the Geodesic Area yields values that area extremely close (within roughly 0.005% of the value provided by MultiNet).