5

I am very new to ArcGIS and trying to build a series of layers of data in ArcGIS 9.3.1 for a river catchment in Leicestershire, using data from several sources. All data covers at least the area of interest.

I have maps using British_National_Grid and NAD_1983_Albers as the projected co-ordinate system. When projected together one British National Grid layer doesn't all appear with the others. The Albers layers match the majority of the National Grid layers.

As far as I can tell the National Grid layers data (see below) is the same for all.

Data Type: Shapefile Feature Class Shapefile: C:\luke-work\Analysis 2 data\aquatic landscapes\AquaticLandscapes.shp Geometry Type: Polygon

Projected Coordinate System: British_National_Grid Projection: Transverse_Mercator False_Easting: 400000.00000000 False_Northing: -100000.00000000 Central_Meridian: -2.00000000 Scale_Factor: 0.99960127 Latitude_Of_Origin: 49.00000000 Linear Unit: Meter

Geographic Coordinate System: GCS_OSGB_1936 Datum: D_OSGB_1936 Prime Meridian: Greenwich Angular Unit: Degree

The other two projections data are

Data Type: Personal Geodatabase Feature Class Location: C:\luke-work\Analysis 2 data\catchment\catchment.mdb Feature Dataset: Layers Feature Class: DrainageLine Feature Type: Simple Geometry Type: Line

Projected Coordinate System: NAD_1983_Albers Projection: Albers False_Easting: 1500000.00000000 False_Northing: 6000000.00000000 Central_Meridian: -100.00000000 Standard_Parallel_1: 27.50000000 Standard_Parallel_2: 35.00000000 Latitude_Of_Origin: 18.00000000 Linear Unit: Meter

Geographic Coordinate System: GCS_North_American_1983 Datum: D_North_American_1983 Prime Meridian: Greenwich Angular Unit: Degree

Any suggestions on what I am doing wrong?

Thanks for your answer. I have tried out what was suggested and answer each comment below.

2). I only have GIS data for the UK that I am looking at currently, so have had to assume (seperately) that each data set is OK, and it is the other that is mis-behaving (i.e. everything has been done twice). Is this even a safe approach.

4). No, data for catchment is still projected about 100 miles south west of where it should be.

6). No. data remains in the wrong place.

Subsequent comments: Data seems to be scaled appropriately - UK is larger than catchment!

Further investigation of the data I was given also reveals that there is a dataset which is projected significanly south west of the data in the above problem (now referred to as "Drain"). I am guessing somehere in Brazil! This appears (to me) to be using the same co-ordinate system as the catchment. This is backed up (in my mind) by the fact that is has previously been used in a modelling exercise for the catchment. That work was carried out on a different computer at a different site (would this even be significant?).

Catchment data:

Data Type: Personal Geodatabase Feature Class Location: C:\luke-work\Analysis 2 data\catchment\catchment.mdb Feature Dataset: Layers Feature Class: Catchment Feature Type: Simple Geometry Type: Polygon

Projected Coordinate System: NAD_1983_Albers Projection: Albers False_Easting: 1500000.00000000 False_Northing: 6000000.00000000 Central_Meridian: -100.00000000 Standard_Parallel_1: 27.50000000 Standard_Parallel_2: 35.00000000 Latitude_Of_Origin: 18.00000000 Linear Unit: Meter

Geographic Coordinate System: GCS_North_American_1983 Datum: D_North_American_1983 Prime Meridian: Greenwich Angular Unit: Degree

Drainage:

Data Type: Personal Geodatabase Feature Class Location: C:\luke-work\Analysis 2 data\catchment\catchment.mdb Feature Dataset: Layers Feature Class: DrainagePoint Feature Type: Simple Geometry Type: Point

Projected Coordinate System: NAD_1983_Albers Projection: Albers False_Easting: 1500000.00000000 False_Northing: 6000000.00000000 Central_Meridian: -100.00000000 Standard_Parallel_1: 27.50000000 Standard_Parallel_2: 35.00000000 Latitude_Of_Origin: 18.00000000 Linear Unit: Meter

Geographic Coordinate System: GCS_North_American_1983 Datum: D_North_American_1983 Prime Meridian: Greenwich Angular Unit: Degree

Improve this question
1
  • Luke, I think it is more appropriate to write comments under an answer or edit your original question with further information.
    – Brad Nesom
    Commented Dec 20, 2010 at 5:23

1 Answer 1

Reset to default
4

The method i use to determine errors or no information in the projection is...
1. load one data set.
2. find another source for data that you can rely on the projection (call this the check datset). This does not need to be the same as the data in question.
3. Note that when you load one dataset the projection of the document is in that data's projection.
4. Load the check dataset. (Does it overlay with data 1?)
5. Now right click on the TOC (table of contents) layers (default frame). All the way at the top of the TOC.
6. Select properties> coordinate system Tab, in the layers list select the check layer and it's coordinate system.

Now does the data overlay each other? If so then you have verified that the data 1 prj file contains the correct coordinate system. Continue to lad each other dataset the same way.

If not you have some searching to do.

Generally I look at the scale and direction of the questionable data in conjuction with the check data. size will tell if it is ll, meters, or foot. direction will help determine if it is out by datum, or zone. if one is VERY tiny and down to the left (US comparison) then it is probably in lat lon or GCS.

If the data 1 is "3" times bigger then one is foot the other meter.

And if one is at the same y value but x is off then it is probably meter but a differrent zone.

These aren't all of what you need to know to work through the problems but it might get you started and able to ask more questions.

Edit: More explanation about crs 9coordinate reference systems)

Don't let it get too confusing.
First, generally speaking you have a cartesion coordinate system in each of your datasets.
Normally systems use (if they can) the positive, positive quadrant of the coordinates.
If I have a road in oklahoma and I assign the state plane system for eastern hew hampshire none of the coordinates change.
However if I look at the data with another layer it will obviously be in the wrong location.
Arcmap does what is called on-the-fly reprojection. This means that all data does not have to be in the same system.

When you open a new document and add the first layer (my reason for instructing to add one layer) the crs of the first layer added is assigned to the document. Subsequent layers added are "reprojected" for the display (only).

If one of the layers does not have a projection assigned then it returns a message that it cannot be reprojected and it is not. whatever the coordinates are in it's data lay where they corespond to on the documents crs.

You said you are restricted to UK data. There is I'm sure a plethora of other data that can be acquired. All you need is something like the country boundary from the esri data probably already on your system.
This will alow you to verify each dataset independently, one at a time.

To clarify. - the projection of a shape file is defined in the prj file. While the projection of a personal gdb is defined in a table in the mdb. These can be changed by using arccatalog and selecting a different projection system. This does not change the coordinate values (and is not a reprojection). It only changes where the data will show up in conjunction with other data.

With this information see if my original step process helps. And you can respond back or comment

Improve this answer
0