Problem:

There is a discrepancy in the actual location of an uploaded shape and how it is displayed. When I run into this issue it is usually a small discrepancy, maybe 5 meters, but in this example the difference is ~95 meters.


Short answer: 

This is probably because the shapefile is using a projection based on NAD_1927, which is notoriously difficult to transform to NAD 83/WGS 84. There's no single right equation which will always work in an accurate manner. If they're able to convert any uploaded shapefiles to a projection based on the NAD 83/WGS 84 datum, then they'll probably have much more accurate results within the Biotics map viewer. This is probably worth adding to the help desk solutions.

See https://desktop.arcgis.com/en/arcmap/latest/tools/coverage-toolbox/transforming-between-nad27-and-nad83.htm for some light reading.


Longer answer: 

The problem is caused because the identify tool attempts to transform the clicked coordinates (which are in Web mercator, based on WGS84) to the coordinates of the uploaded shapefile. It's taking its best guess at the transformations it needs to apply based upon what is returned by ArcGIS Server's findTransformations web service available through the the GeometryServer utility. I believe it uses the first option returned - which might not necessarily be the best.

The projection for this shapefile is "NAS_1927_UTM_Zone_11N". 

PROJCS["NAD_1927_UTM_Zone_11N",GEOGCS["GCS_North_American_1927",DATUM["D_North_American_1927",SPHEROID["Clarke_1866",6378206.4,294.9786982]],PRIMEM["Greenwich",0.0],UNIT["Degree",0.0174532925199433]],PROJECTION["Transverse_Mercator"],PARAMETER["False_Easting",500000.0],PARAMETER["False_Northing",0.0],PARAMETER["Central_Meridian",-117.0],PARAMETER["Scale_Factor",0.9996],PARAMETER["Latitude_Of_Origin",0.0],UNIT["Meter",1.0]]

In this case, ArcGIS Server returns the following transformations:

{  "transformations": [    {      "geoTransforms": [        {          "wkid": 108190,          "latestWkid": 108190,          "transformForward": true,          "name": "WGS_1984_(ITRF00)_To_NAD_1983"        },        {          "wkid": 108001,          "latestWkid": 1241,          "transformForward": false,          "name": "NAD_1927_To_NAD_1983_NADCON"        }      ]    },    {      "geoTransforms": [        {          "wkid": 8494,          "latestWkid": 1515,          "transformForward": false,          "name": "NAD_1983_To_WGS_1984_5"        },        {          "wkid": 108001,          "latestWkid": 1241,          "transformForward": false,          "name": "NAD_1927_To_NAD_1983_NADCON"        }      ]    },    {      "geoTransforms": [        {          "wkid": 1737,          "latestWkid": 1737,          "transformForward": false,          "name": "NAD_1983_To_WGS_1984_OR_WA_41"        },        {          "wkid": 108001,          "latestWkid": 1241,          "transformForward": false,          "name": "NAD_1927_To_NAD_1983_NADCON"        }      ]    },    {      "geoTransforms": [        {          "wkid": 8232,          "latestWkid": 1308,          "transformForward": false,          "name": "NAD_1983_To_WGS_1984_4"        },        {          "wkid": 108001,          "latestWkid": 1241,          "transformForward": false,          "name": "NAD_1927_To_NAD_1983_NADCON"        }      ]    },    {      "geoTransforms": [        {          "wkid": 8088,          "latestWkid": 1188,          "transformForward": false,          "name": "NAD_1983_To_WGS_1984_1"        },        {          "wkid": 108001,          "latestWkid": 1241,          "transformForward": false,          "name": "NAD_1927_To_NAD_1983_NADCON"        }      ]    }  ] }

The map viewer uses the first one. The result doesn't seem to align with how the coordinates are being transformed by ArcGIS Server to display on the map, and are shifted by a fairly significant margin.

Due to the nuances of converting NAD 27 to NAD 83, I suspect that ArcGIS Server's visualization engine is using a more sophisticated approach to determine the transformation which should be used, not just the first one it lists. I noticed the the third transformation is named "NAD_1983_To_WGS_1984_OR_WA_41", which I'm speculating might end up being more appropriate for this shapefile since the data is in Washington state.

However, without a lot of work, I don't think it's likely that we can significantly improve the map viewer's ability to determine which of the possible transformations are likely to produce the best results. My recommendation is to avoid using NAD 27 data whenever possible (i.e. if the dataset uses NAD27, project it yourself to make sure the resulting data is in the correct location, and then upload the NAD83 version into Biotics).