I’m new to this list and new to geoserver and like what I see so far, however I’m having a little bit of a problem. I’m trying to load storm track data from NOAA and it doesn’t seem to be projecting correctly. I am able to load the shapefiles, symbolize them, etc., my problem is that everything ends up 8 miles north of where they should be.
File I’m Downloading: http://www.nhc.noaa.gov/gis/forecast/archive/al072008.latest_5day.zip
This file has the latest data on Gustav and I’m trying to load it into geoserver. The file contains 4 separate shapefiles, all of which have the same issue. All files have their own .prj with the following:
GEOGCS[“GCS_Sphere”,DATUM[“D_Sphere”,SPHEROID[“Sphere”,6371200.0,0.0]],PRIMEM[“Greenwich”,0.0],UNIT[“Degree”,0.0174532925199433]]
Initially when I tried to lookup the SRS it would come up UNKNOWN so I attempted to reporject the file using the following command:
ogr2ogr -t_srs EPSG:4326 out.shp in.shp
That results in a new shapefile and lookup SRS comes up with 4326 as expected. When I use the KML output option to load the track into google earth the results are approximately 8 miles due north of where they should be in reality. (For example, the _pts.shp files includes fields for lat/lon, when plugging those lat/lon values into google earth the point is ~8 miles south of where the point from the KML is).
I have also tried adding my own SRS by adding the following line to epsg.properties under user_projections:
999999=GEOGCS[“GCS_Sphere”,DATUM[“D_Sphere”,SPHEROID[“Sphere”,6371200.0,0.0]],PRIMEM[“Greenwich”,0.0],UNIT[“Degree”,0.0174532925199433], AUTHORITY[“EPSG”,“999999”]]
After doing that it correctly (I believe) finds 999999 when I lookup SRS but I end up with exactly the same shift in the results.
My next troubleshooting step was to load the original shapefile from NOAA into ArcGIS 9.2 SP6 and project it. When I do that it works perfectly and everything comes out exactly where it should. I can project it from GCS_Sphere (as detected by ArcGIS from the .prj) to either GCS NAD 83 or WGS 84 and Geoserver will detect and display the results perfectly in the KML. I am wanting to automate the process of downloading the latest fileand setting it up in geoserver which is why I don’t want to dedicate an ArcGIS machine to it. I’ve already automated the process and all I need to do now is fix the apparent projection issue. Has anyone else had this problem, or does anyone have suggestions on where to go from here?
Thank you,
Josh Keller
