I think I'm having a similar issue. The view changes to -180 to 180 (meaning when I create a grid it is Pacific-centered at 0* (IDL). However, the data extent is still left 0 to right 360, so it starts in the center of the grid (the left edge of my map is at 0* (IDL) and the eastern half extends 180* eastward beyond 180E of the grid).

If I switch the Meridian of the data file (ArcCatalog-SpatialReference-Edit-Modify to -180, the map is in the "right place", but now it is Atlantic-centered, splitting Alaska at the eastern and western edges, which is what I'm trying to avoid.

This seems like it should have an easy fix, but I can't think of it! Thanks.

We can't quite tell what you're doing from your description. We think you should try to leave the data's coordinate system as geographic, using Greenwich prime meridian. This should work whether the raster was built as -180 to +180 or 0 to 360. Then use a projected coordinate system (if you wants a look-alike pseudo-plate carree, use plate carree) with a central meridian of 180. It could be that you're using an old enough version of the software that it wasn't handling 0-360 rasters correctly when projecting/transforming them, but that is fixed now.

We reread the original question here and I think we may have originally misinterpreted it. Yes, it's working properly. By changing the prime meridian of the geographic coordinate system, the longitude values are changed. So what was originally +180, is now 0. In essence, this changes the "break" in the longitude values. Rather than occur at the +/-180 longitude line, it occurs at 0 longitude (Greenwich).

Yes, that's correct. It is working, i.e. what was originally +180 is now 0, which is exactly what I wanted it to do.

Now I realize that what I need is simply a grid to match that extent. I created a grid (under Layers) with the same projection, and defined my own origin changing the X from -180 to 0. But, the grid still plots -180 to +180 rather than 0 to 360.

Sounds like your grid is reflecting the coordinate system of the data frame, which is as it should be.

Thank you, Aileens for your help. A few things I learned:

1) I created a new data frame and re-added my file with the correct projection. That allowed me to change the grid origin to 0,-90. However, though the decimal degrees are 0 to 360, the grid labels go from 0E to 180 (in the center) to 0W.

2) At one point, my file would not reproject to Pacific-centered, even though I was following the same steps. It is because the file had a .prj that was being carried with it. Once I erased the associated .prj it worked.

So glad this is now working for you! I am sure you learned a lot though this exercise. :) Thanks for sharing what you learned with us.