Transform coordinate systems

A new data frame has no coordinate system. When you add your first layer to it, the data frame takes on that layer's coordinate system. For any subsequent layers that you add, one of two things happens. If the new layer's coordinate system already matches the data frame, the layer is added without question. If the new layer's coordinate system is different, you get a GCS warning and ArcMap changes the coordinate system of the new layer to match the existing one. This operation is called a datum transformation (or coordinate system transformation). For all new layers that you add, ArcMap does these transformations automatically so that layers with different GCSs can be displayed together. (The transformations are done "on the fly," meaning they are applied only inside the data frame; the GCSs of the data sets on disk are not changed.)

 

What is the exception to this rule?

If the first layer you add to a data frame has latitude-longitude coordinates but no coordinate system information, the data frame becomes GCS_Assumed_Geographic_1 (based on the NAD27 datum). If you then add a second layer with a coordinate system other than NAD27, you'll get a GCS warning message, but no datum transformation will be applied.

 

It's great that ArcMap does datum transformations for you, but complications arise because there may be multiple versions of a transformation. For example, ArcMap has seven formulas that convert from NAD27 to NAD83. Each works best for a particular region.

 

Available transformations

Optimal transformation

Location of data

NAD_1927_To_NAD_1983_6

Canada - Quebec

NAD_1927_To_NAD_1983_Alaska

Alaska

NAD_1927_To_NAD_1983_CSRS98_1

Canada - Quebec

NAD_1927_To_NAD_1983_CSRS98_2

Canada - Saskatchewan

NAD_1927_To_NAD_1983_NADCON

Conterminous US (not Alaska or Hawaii)

NAD_1927_To_NAD_1983_NTv2_Canada

Canada

NAD_1927_To_NAD_1983_PR_VI

Puerto Rico - Virgin Islands

 

 

Similarly, there are 21 different transformations from NAD27 to WGS84.

 

The different transformations from NAD27 to WGS84

Optimal transformation

Location of data

NAD_1927_To_WGS_1984_1

Mean for Antigua, Barbados, Barbuda, Caicos Islands, Cuba, Dominican Republic, Grand Cayman, Jamaica, and Turks Islands

NAD_1927_To_WGS_1984_2

Mean for Belize, Costa Rica, El Salvador, Guatemala, Honduras, and Nicaragua

NAD_1927_To_WGS_1984_3

Mean for Canada

NAD_1927_To_WGS_1984_4

Mean for United States (CONUS)

NAD_1927_To_WGS_1984_5

Mean for United States (CONUS East of Mississippi River and MN, MO, and LA)

NAD_1927_To_WGS_1984_6

Mean for United States (CONUS West of Mississippi River)

NAD_1927_To_WGS_1984_7

United States (Alaska)

NAD_1927_To_WGS_1984_8

Bahamas (except San Salvador Island)

NAD_1927_To_WGS_1984_9

Bahamas (San Salvador Island)

NAD_1927_To_WGS_1984_10

Canada (Alberta, British Columbia)

NAD_1927_To_WGS_1984_11

Canada (Manitoba, Ontario)

NAD_1927_To_WGS_1984_12

Canada (New Brunswick, Newfoundland, Nova Scotia, and Quebec)

NAD_1927_To_WGS_1984_13

Canada (Northwest Territories, Saskatchewan)

NAD_1927_To_WGS_1984_14

Canada (Yukon)

NAD_1927_To_WGS_1984_15

Panama (Canal Zone)

NAD_1927_To_WGS_1984_16

Cuba

NAD_1927_To_WGS_1984_17

Greenland (Hayes Peninsula)

NAD_1927_To_WGS_1984_18

Mexico

NAD_1927_To_WGS_1984_21

United States (Alaska - Aleutians East of 180E)

NAD_1927_To_WGS_1984_22

United States (Alaska - Aleutians West of 180E)

NAD_1927_To_WGS_1984_30

Cuba

 

For each datum transformation, ArcMap has to pick a default formula. For the NAD27 to NAD83 transformation, it uses NAD_1927_To_NAD_1983_NADCON, which is optimized for the continental United States. If your data happens to lie in Alaska or Canada, your features may still not line up properly until you override the default transformation and pick the optimal one.

The default NAD27 to WGS84 transformation is NAD_1927_To_WGS_1984_1, which is the optimal transformation only for Caribbean data.

In this exercise, you will see how ArcMap applies its default datum transformations and you will learn how to change these settings.