[Qgis-user] Dumb USGS geotiff questions

[Mike Flannigan]

It may sound like I know what I am talking about,
but you may not want to make that assumption.

Everything is working fine for you.  Older USGS
topo maps (~1980 and earlier) are in NAD27 datum.
Newer maps (~1990 and newer) might be in WGS84,
but I'm not sure about that.

I'm thinking if you change your datum to NAD27
that deviation may go away.

I believe the older files can be changed to WGS84
with GDAL if needed, but I generally just leave them
in NAD27.  In MO, where I normally work, the difference
between NAD27 and WGS84 is pretty small - roughly the
amount shown on your maps.  I would expect it to be
higher where you are working - near Oakland, CA:
http://www.mflan.com/temp/shift.png


Mike



On 4/19/20 4:41 AM, qgis-user-request at lists.osgeo.org wrote:
> Date: Sat, 18 Apr 2020 14:43:03 -0700
> From: Morgan Fletcher<morgan at hahaha.org>
> To:qgis-user at lists.osgeo.org
> Subject: [Qgis-user] Dumb USGS geotiff questions
> Message-ID:
> 	<CAPwdt1Xd78gQOVago1d86KteK0vt0ow2dT_zJ+XrECn4anPNxw at mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> (Resending, with screenshots as google drive links, to cut down message
> size.)
>
> I am running 3.12.1-București on a mac, os-x 10.14.6. My knowledge of GIS
> is rudimentary, I've learned what I need to, to satisfy my curiosity about
> my region, using old maps.
>
> I have downloaded all the USGS maps for my region - Oakland, CA, USA, which
> corresponds to "CA Concord" - from USGS topoView, in geotiff format.
> Opening them in QGIS gives fairly wild offsets; they are not all encoded
> the same way, yet have common corners. Features on the map vary with what
> modern maps show, but getting their corners right would help me a lot.
>
> If I create a new project, set my CRS to WGS 84 / Pseudo-Mercator
> (EPSG:3857) as I think is correct for my base map, Google Maps (EPSG:3857 -
> WGS 84 / Pseudo-Mercator - Projected) then Layer > Add Layer > Add Raster
> Layer > Raster Dataset > CA_Concord_465520_1943_62500_geo.tif > Add, I get
> prompted with:
>
> https://drive.google.com/file/d/1FKVPrHLtXXKK8a-QTTz3IigT9dpj1Hou/view?usp=sharing
>
> I don't know what to choose, here. I chose the first one. I've tried
> others. The tif.prj file does say
> DATUM["D_North_American_1927",SPHEROID["Clarke_1866",6378206.4,294.9786982]],
> which I am guessing conforms to NAD27. If I add an opacity slider to the
> new raster layer, set its opacity, and then use the Zoom to Coordinate
> panel to navigate to 37º45'N 122º15'W, I find that the point is off from
> what's on the map, see:
>
> https://drive.google.com/file/d/10vYtcM1FZ-vdS3z1FPW2EYbrNy2sj65L/view?usp=sharing
>
> This is similarly true at the top left:
>
> https://drive.google.com/file/d/18naBKuwk6rZ9DP0tFOZxnGKlIpRQgDKM/view?usp=sharing
>
> Also, note how the map shows two lines at top left, it's true at top right.
> This is something to do with projection, I think. I've tried to
> georeference these lines, to get them to line up right, but which
> intersection do I use, when there are two vertical lines intersecting with
> the top, horizontal border, to reference 122º 15' 38º 00' ? I've had mixed
> results, trying to use deg-min-ss coordinates to georeference these; I get
> a map off the west coast of Africa. Using points on the map is more
> successful.
>
> All the maps have slightly variant values in their tfw files. Their prj
> files are more uniform, but also show variations. I'd like to normalize
> them all, so that I can view the USGS topo maps for my region over time.
> Advice?
>
> Thanks,
>
> Morgan