Problem with misaligned spatial projection (.prj) – British National Grid

[Igloo_79]

I’m new to HEC-RAS and having problems with projected coordinate systems of my HEC-RAS (5.0.3) geometry and the WMS maps in RAS Mapper. I imported georeferenced cross sections via a .CSV file (i.e. not using GIS), however, they are offset by ~120m from the river seen on WMS maps (Google, Open Streetmap etc).

I’ve used .prj files (27700: OSGB 1936/British National Grid) I downloaded (http://spatialreference.org/ref/epsg/27700/) as well as those generated by QGIS, leading to the same results.

I found a workaround by manually editing the Easting & Northing in the .prj file which aligned my geometry and the WMS correctly – however, I’m wondering if this may affect 2D modelling e.g. generating misprojected flood maps. I’d be grateful for any suggestions please?

.prj file
PROJCS[“OSGB 1936 / British National Grid”,GEOGCS[“OSGB 1936”,DATUM[“D_OSGB_1936”,SPHEROID[“Airy_1830”,6377563.396,299.3249646]],PRIMEM[“Greenwich”,0],UNIT[“Degree”,0.017453292519943295]],PROJECTION[“Transverse_Mercator”],PARAMETER[“latitude_of_origin”,49],PARAMETER[“central_meridian”,-2],PARAMETER[“scale_factor”,0.9996012717],PARAMETER[“false_easting”,400000],PARAMETER[“false_northing”,-100000],UNIT[“Meter”,1]]

[Anonymous]

I had the exact same issue, never found a solution…

[Anonymous]

I ended up editing http://spatialreference.org/ref/epsg/27700/ as follows:

PARAMETER[“false_easting”,400107],PARAMETER[“false_northing”,-100056]

Are these the same values you got?

[Igloo_79]

Mine were very similar. Have you experienced offset issues when loading terrain (e.g. DTM with the standard 27700 UCS), or when performing 2D floodplain modelling?

I am concerned that whilst the 1D network may appear to line up neatly with the WMS imagery, the offset could affect 2D flood mapping.

[lkhoury]

I had the same issue. It turned out that the survey data was adjusted by a correction factor (eg: 1.006). Are you sure your survey data point do not have any issues?

[Igloo_79]

Ok, so it seems that the WMS maps are projected to different coordinate systems e.g. Google Maps Global Mercator (Ras Mapper>Map Layers>Image Info).

I’ve gone back to my original .prj (British National Grid, 27700) to avoid potential issues with misaligned 2D mapping (e.g. my terrain layer is offset from my 1D geometry due to different projection systems), so have opted for using terrain and/or OS maps as my backdrop.

[benp]

I had the same offset with the ESPG:3857 (Spherical Mercator WGS84 projection used for OSM et Google map).
I solved the problem by reprojecting with QGIS my orginal geotiff/rectangular pixels 5.89×8.47 m² ESPG:3857 in square pixel 5x5m² ESPG:3857
I tried ESPG:4326 (Google earth geographic spherical coordinate) with no success: there is no offset but the scale is approximatively 0.01 m for 1 km.

[LorenAmelang]

So in my post at
http://hec-ras-help.1091112.n5.nabble.com/Terrain-from-EPSG-3310-file-is-475-south-of-Web-Imagery-td6257.html

Is the key:

Here is what RAS Mapper GDAL sees in the 3310 file it imported incorrectly:
—
Pixel Size = (1.030469526568122,-0.969388829487901)
—

vs:

RAS Mapper importing original USGS DEM:
—
Pixel Size = (1.000000000000000,-1.000000000000000)
—

It must know about non-square pixels, since the features aligned perfectly once manually displaced to matching positions. I guess it just calculates the Origin without considering pixel shape?

[Author]

Posts