Hi -

Thanks to those who sent me comments and guidance, this helped me focus on an approach that I think will work well.  And David, thanks for your thoughts and guidance, I respond within your note below, in red.

After considering your guidance and talking with the GIS staff on the project, here's how I think NGMDB will proceed:

  *   We'll georeference as many geologic maps as possible, for including in our mapView interface.  Georeferencing will enable us to compute the fractional scale.
  *   For the many maps that aren't georeferenced, either because we run out of available staff time or because the maps simply can't be georeferenced for any of several reasons, we will:
     *   determine the fractional scale from the USGS Library's paper copy, or
     *   for maps that aren't available in our Library, I'll try using the scale determination methods recommended by Kate Cramer and(or) David Medeiros.

As I mentioned, I address David's comments within his note below, but there's one point that's long concerned me for USGS publications, which I raise here in a more general context (but I admit to not having the experience or skills that you do, so it might be unrealistic).  That is, ideally, there could be a "scan status list", in which every part (or "object") of a publication would be separately tracked, with the goal of identifying who has scanned it and where the "best" scan resides.  Perhaps not practical to build or maintain, but if it were possible, that'd help us all, I think...

Thanks again.

Dave


________________________________
From: David Medeiros <[log in to unmask]>
Sent: Wednesday, March 29, 2023 1:38 PM
To: [log in to unmask] <[log in to unmask]>; Soller, David R <[log in to unmask]>
Subject: Re: [EXTERNAL] RE: Determining map scale, online?

Hi David,

I was in the middle of forming a more technical reply to your map scale problem when I finally saw your note on these being served as PDFs. I'll leave my other info copied below, it may still be helpful. But I also wanted to touch on (rant over) the bigger use-case issue here. I know this is out of your/our control, we're usually using older systems designed by people we never meet and almost never have any input into how these maps are digitized, preserved, and served to our users. But... for if/when we do get a chance to be involved in redesigning some of these services it's worth having a bigger picture of what these maps are for and how best to share them. And PDF's are not how best to share them!  Most users are familiar PDF, but I've found many are confused by tifs, and so NGMDB offers PDF, TIF, and JPG (and KMZ if it's a map we've georeffed).  But perhaps you're not referring to sharing with end-users, but with cooperators such as other libraries?  Related to the technical aspects of determining scale, you can get the image out of the PDF, but that's adding yet another layer of resampling and resizing to the original scan that has already been shoved into a file format that's mostly unusable in other tools. Looking at the bigger use case picture for these scans, are they just finding aids and previews for the paper maps, or is the scan to be a resource itself to be used directly?  Our scans are intended to be used directly, not just as previews or thumbnails -- they're 300 dpi RGB tifs.  My experience is mostly that once a decent scan has been made from a paper map, users will almost always want the digital version over the physical one - for image manipulation or geospatial work. Neither the paper map nor the PDF best serve those use cases. Most of my end users are often unsure of how to deal with a PDF image in the first place. Relatedly the RF scale of the original paper map is also not usually that useful to a digital user, except in understanding the broad category of small vs large scale the map may fall into which often has to do with data-resolution, but they can get that info faster from the image preview. I know none of this helps you now and probably none of it is in our control. But it's worth keeping in mind for later when someone whose building a new system for dealing with these maps and scans casually asks for our input on their new design. I think the goal of cataloging digital materials like they were physical objects is one that gets us in trouble later as you're dealing with now.

Best,
David

Original reply:

Are these maps high resolution scans for download or just digital thumbnails for the map viewer as finding aid to the paper map records?  They're relatively high res (300 dpi) tifs, for download and use in GIS and other applications.  We store the tifs uncompressed, but serve compressed tifs (source tifs available upon request).

If knowing the scans full size helps, you can determine that straight from the downloaded image files embedded metadata. Both the DPI and the pixel dimensions should be listed in your file viewer info window, or just open it in PS.  But that's the scan print size, not the original paper map size so not sure that helps.  I agree, it doesn't help much, and is only readily available for some (not all) of the maps I'll be dealing with.

For maps you know the original paper dimensions of, you can derive the RF scale by opening that scan in PS or AI, measuring the scanned length of a side and rescaling by the required proportion to match the original dimensions. Measure the scale bar in the scan and determine the RF from there. This would be the original paper map RF.

Or, for maps of known locations you can always compare measured distances between known points in the scanned map and the same points in a GIS or Google Earth Web to again derive the RF of the scan.  This is similar to approach we'll be taking, by georeferencing the maps...



David Medeiros

Geospatial Reference & Instruction Specialist

Stanford Geospatial Center

650.561.5294

@mapbliss


SGC website: gis.stanford.edu

GIS cartography: bit.ly/giscart<https://gcc02.safelinks.protection.outlook.com/?url=http%3A%2F%2Fbit.ly%2Fgiscart&data=05%7C01%7Cdrsoller%40USGS.GOV%7C74fb60f973c3482deac408db307c713d%7C0693b5ba4b184d7b9341f32f400a5494%7C0%7C0%7C638157083269437686%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Bs4hbyo24oSLO6oxvZDNkju291yEVa5y6B9bchnNKTk%3D&reserved=0><https://gcc02.safelinks.protection.outlook.com/?url=http%3A%2F%2Fbit.ly%2Fgiscart&data=05%7C01%7Cdrsoller%40USGS.GOV%7C74fb60f973c3482deac408db307c713d%7C0693b5ba4b184d7b9341f32f400a5494%7C0%7C0%7C638157083269437686%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Bs4hbyo24oSLO6oxvZDNkju291yEVa5y6B9bchnNKTk%3D&reserved=0>

GIS email list: bit.ly/GISlist


________________________________
From: Maps-L: Map Librarians, etc. <[log in to unmask]> on behalf of Soller, David R <[log in to unmask]>
Sent: Wednesday, March 29, 2023 7:47 AM
To: [log in to unmask] <[log in to unmask]>
Subject: Re: [EXTERNAL] RE: Determining map scale, online?

Good morning -

Thanks to Kate Cramer, Katie Lage, and Paige Andrew for their replies.  I'll attempt to answer each of you in this response...

First, I apologize for not being clear -- the maps for which a scale needs to be determined are not by USGS, they're non-quadrangle maps mostly published by State geological surveys. Because the NGMDB focuses on maps and their display through our mapView interface (https://ngmdb.usgs.gov/mapview/), I really need to know the scale so that each map can be displayed in the proper scale "bin". I've spot-checked some of these maps in their State GS publications catalog, and in Worldcat (thank you for that hint!), but have found no indication of scale -- the bibliographic records that I've checked indicate the number of plates, and occasionally the overall physical dimension, but not the fractional scale. As an added challenge, because these images are PDFs of scans that I've obtained from the States, I don't know the specs used by their staff (or their contractors) to create the scan (e.g., dpi, file format, whether they were compressed before the PDF was made).

When I referred to "guesstimating" the fractional scale on some maps, I did it in Photoshop by setting the display to 100%, and judging whether that appears to be the "print size".  But for the image I'm staring at right now, that's quite incorrect -- for this one, 50% seems more likely, but that's only a guess. The only time I'm really confident is when I know the map is page-size (8.5 x 11"), which generally means it's not a map that I want to georeference anyway!

I think my best strategy might be to: (1) pull some maps from the USGS Library, from a given Publisher, Series, and time period (e.g., maps from the mid-20th century, in the Bulletin series of the predecessor of the California Geological Survey); (2) determine the fractional scale by applying my handy transparent Map Scale Indicator to the bar scale; and then (3) consider applying that scale to maps in other CGS Bulletins of the same vintage that seem to present the same cartographic appearance and similar-sized map area.  Not an ideal situation, so I'll attempt to find each publication if I can...

Thank you for guidance and questions, they were quite helpful, and I welcome any further opinions!

Dave


________________________________
From: Cramer, Kate <[log in to unmask]>
Sent: Tuesday, March 28, 2023 5:23 PM
To: Maps-L: Map Librarians, etc. <[log in to unmask]>; Soller, David R <[log in to unmask]>
Subject: [EXTERNAL] RE: Determining map scale, online?




 This email has been received from outside of DOI - Use caution before clicking on links, opening attachments, or responding.



If you are dealing directly with the original scan and you know the DPI (dots per inch) they were scanned at, you can work it through from there.  Can be fussy indeed, but once you do a few you’ll likely be able to recognize trends.

Some image viewers have properties displays that will do the pixel calculations for you.  If you are just in windows, you should be able to right-click on the file and look at the properties there (details tab)
If you are dealing with a PDF, you can sometimes get similar information from inside your pdf viewer.

Another option you may have, depending on the map, is they will sometimes say 1cm=5km or the like.  (My guess is you don’t have this, but it would nice if you just missed it and could now save a lot of time!)

If they were born digital, it’s a different can of worms.  At that point there may not be a “print size” per se.  Some digital products are rated in detail recorded at a particular ground-distance-interval.  I think.  I’m a bit shaky on that aspect.  (I mostly deal with paper and scans.)



If you are dealing with a picture someone posted on the web and have no resources to determine it’s digital handling or history, then you may have a different issue.  If the image quality has been degraded, it may no longer be accurate to whatever the scale was originally was due to data loss.  At that point you may want to punt…

  *   Maybe call out observed resolution and a best guess of quality of information?
  *   Possibly determine a pixels per area covered ratio if you can get a bounding box?
     *   Like, if it’s 10 degrees by 10 degrees, and the distance from one edge of the mapped are to the other is 1000 pixels, that does give you a number to work with that could have value?



Another course of action may be to ask the people who will be using the maps what information will be most useful to them.



While I know this was short on ‘proper’ answers, I hope it gives you a few ideas to pursue.  I wish you the best of luck and a Bucket of Perseverance!



-Kate Cramer

East View Geospatial

Geodata Project Lead



From: Maps-L: Map Librarians, etc. <[log in to unmask]> On Behalf Of Soller, David R
Sent: Tuesday, March 28, 2023 3:23 PM
To: [log in to unmask]
Subject: Determining map scale, online?



Hi -



I need to determine the map scale for hundreds of scanned geologic maps that have only a bar scale (i.e., no fractional scale is provided).  I have very limited access to the USGS Library to inspect the printed map, so I was hoping for guidance -- is it possible to accurately determine the scale from an online view of the map?  If not, how do you address the issue of map scale in your citation records?



I've made map scale guesstimates for a few maps in the past, by viewing the map online at what I think is the actual size of the printed map.  But this current batch is too important to proceed without experienced guidance from the Library community.

Thanks.



Dave Soller

National Geologic Map Database

U.S. Geological Survey