Is something wrong with the calculate geometry function in ArcGIS Pro version 3.5.1?

[CaseyLott1]

Hello,

I have a polygon-based fishnet grid (in Albers Equal Area Conic USGS) where each polygon is within 1 meter of 10,000,000 square meters. This means that each polygon should be ~exactly 1,000 hectares or 2471.053815 Acres. 

The grid has nearly 1 million squares. It is an irregularly shaped polygon, based on ecoregion boundaries, that covers about ~392,000,000 acres of the eastern US from Minnesota to Maine and then south through the Appalachians. 

Today, when I used the calculate (geodesic) option in Calculate Geometry to calculate Hectares and Acres for this grid, each square was reported to be 495 hectares and 1,222 acres. This is wrong. Again, the result should be 1,000 hectares and 2471.053815 acres.

My fishnet grid layer is contained within a Project where the first layer (and therefore, the whole map) used the Web Mercator Aux. Sphere projection. If this were messing up the hectare/acreage calculations, it seems like squares would vary in size due to the inherent distortions of Mercator. Instead, hectare/acreage values for each square are all the same, they are just wrong. 

I was under the impression, which may be my own misinterpretation, that the Area (geodesic) calculation should produce the same result REGARDLESS of projection, since it calculates area using a standard ellipsoid.

Can anyone explain why my hectare and acre calculations ended up incorrect here?

Thank you

[CaseyLott1]

Update: I added my Albers Equal Area Conic version of the file into a new Project, repeated the same Area (geodesic) calculations and got the exact same wrong results. 

[DanPatterson]

I would have this check because there is an un explained paragraph in

Calculate Geometry Attributes (Data Management)—ArcGIS Pro | Documentation

The geodesic length and area properties use a shape-preserving algorithm. This produces highly accurate results that are not biased by an assumption that the input line or polygon features are constructed with geodesic arcs between the vertices, which is the assumption regarding traditional geodesic length and area.

It might be worth reaching out to tech support to see if there is any support documentation explaining the difference

[AyanPalit]

@CaseyLott1 Assuming usage of the Calculate Geometry Attributes with Area (geodesic) , Area Unit (Hectares OR International Acres). Since you specify Pro version 3.5.1 - have you run this same calculation in a previous version and results were in line with your expectations?

[CaseyLott1]

@AyanPalit Yes, you are correct. I did a number of tests back in February of this year and I had convinced myself that no matter what projection my data were in, the geodesic area calculation provided the same results.

This would be a game-changer for me, since it would allow me to make accurate area measurements in Projects that rely on web services that insist on using the idiotic, yet somehow omnipresent, Web Mercator projection. 

My projects usually cover most of the US and accuracy of acreage calculations always matters to me. This tends to keep me away from Web Mercator, but I always have collaborators that want to use web services in that projection. 

In a perfect world, I could bring web services in Web Mercator into a project, project all of my ancillary data into Web Mercator, and then accurately calculate acreage for any given location using calculate (geodesic). Like I said before, I had convinced myself this worked back in February but when I retested it yesterday my acreage estimates of the same-sized objects covering a north-south/east-west gradient were either variable or biased low. 

I don't remember exactly what version of Pro I was using when I did my tests in February. I usually stay pretty up to date. I can't remember what version was current this past February.

I hope that this additional information is helpful. If anyone wants to weigh in here, I'd be grateful.

Until I persuade myself that accurate acreage estimates are possible in Projects with Web Mercator based web services, I will probably still set up projects to use a continental equal area projection. This means that I will always be managing enormous sets of imagery in equal area projections instead of enjoying the easy use of a web service. 

[JoshuaBixby]

Although it could or is likely unrelated, what data format is the fishnet grid layer stored in?  File geodatabase, mobile geodatabase, enterprise geodatabase, etc...?  It is always good to state the data storage format.

In what projection was the fishnet originally created in?

Can you share the fishnet grid layer by itself?