U.S. Interagency Elevation Inventory: Changes

Introduction

Over the last few months, we’ve been rebuilding the U.S. Interagency Elevation Inventory (USIEI) to add new features. While I’ll touch a little bit on those features, what I really want to talk about is the restructuring of the data. That’s going to be relevant to you if you plan to pull down the inventory geodatabase or geopackage into your own GIS software.

New Features

First, a little on the new features targeted to roll out February 8, 2021. The old USIEI only allowed you to drop a point to do a search. Now you’ll be able to search by text, drop a point, or draw a polygon. The text search allows a wide range of inputs, including addresses, states, GeoJSON and WKT geometries, and hydrologic unit codes (8, 10, or 12 digit HUCS). There are now filters to narrow your search results, including filtering by text in the title. The filters are also where you can restrict the sources from the search, which can be handy when searching along the coast where you’ll see both topographic and bathymetric returns but might only want one or the other. I’ll write another post to review new features in more detail, but here’s the pretty picture:

Screenshot of the new US Interagency Elevation Inventory application.

New Structure

The old data structure was just a flat table and it had some problems. Some fields in the table were made to hold multiple items and surrounded by html to allow them to show up as links in the viewer. There were also fields, particularly the vertical accuracy, where there might be multiple estimates but only one value could be stored. Now there are separate tables to allow multiple pieces of similar information for a given data set. The following figure of the new database structure should make the descriptions below a bit clearer.

Database diagram of the new US Interagency Elevation Inventory public download.

Links

There are multiple types of links to be represented for a given dataset. These include the link(s) to access the data online, to the metadata, to data reports, and maybe even to map services. Each of those might have more than one link (e.g. a dataset could be available on both The National Map and on Digital Coast). Instead of trying to cram all that into a few fields in the original table, there is now a DataLinks table to hold them. As shown in the database diagram above, the DataLinks table contains rows with a link, a label for the link, and an identifier of what type of link it is. It has a key, InvUUID, the matches the UUID field in the main NationalElevationInventory table to associate the link with the right data set. The LinkTypeID is just an integer which is further defined in the LinkType table. It’s simply providing the information that if LinkTypeID is 1, it’s a data access link. If it’s 2, it’s a metadata link (3 = report or resource, 4 = map service). Unfortunately, we didn’t build a way to tell which data access link goes with which metadata link when there are multiple. It’s usually not hard to tell as they tend to have the same domain. For example, the usgs.gov links go together.

Vertical Accuracy

Over time we came to realize that a single vertical accuracy value was limiting our ability to provide information. This is probably most obvious for topobathy datasets where there could be one accuracy for the bathymetry and another accuracy for the topography. We also started to find that there might be an accuracy estimate done by whomever acquired the data, as part of their evidence that they’d met the specification, and another accuracy done by the receiver of the data. As with the links, the vertical accuracy is now moved to it’s own table to allow multiple values per dataset and includes an ID field to indicate the basis for the accuracy estimate (BasisID). You’ll note that the table has both a field for the vertical root mean square error in cm (VertRMSEcm) and a description field that also tends to have a number. While the description field often has the same number as the VertRMSEcm field, sometimes the original reports are in a different metric, such as a 95% confidence.

In the future, we may also add metrics like the vegetated vertical accuracy, but those aren’t in there now. They’d also likely make a mess of the USIEI application, so I think we’ll hold off until there is a known demand. That information should be in the metadata anyway.

New Fields

There are a few new pieces of information that we’ve added. Some of them are only place holders and won’t have much in them until we fill in the information over time. New additions will get the information first. How far we’ll be able to go back-filling information remains to be seen.

LAS Format

This field should really be just format, but the vast majority of the data in the inventory is lidar in LAS. The contents of this field should be the format and version of the data, such as LAS 1.2 or LAS 1.4. Since data can be transformed from one version to another, this should be the format the data was originally stored as.

Point Classes

This field assumes you’re working with LAS data. It’s a comma separated list of the classes present in the dataset. If you’re not familiar with LAS classes, typical classes are 1 = unclassified, 2 = ground, 7 = noise, and 9 = water. The data should follow the ASPRS LAS class table, but it doesn’t always.

Leaf On or Off

The leaf on or off field was requested by the National Forest Service. It currently says unknown for all entries, but we’ll start filling this in with better information. For many data set, there may not be a documented answer or even a correct one.

Info Contact

The info contact was added to provide a more direct way to find out more information. Since completed datasets already have metadata available, this is most applicable to those data sets that are still in progress. Usually people want to know when it will be available and in the past they’ve sent an email to us and we’ve had to reroute. This is intended to skip a step in that process. For many data sets this is an email, but it might also have a phone number or other information to a secondary information source.

Contract Specification

While this could be applicable to any contracted data collection, its target use is to record which USGS Lidar Base Specification was used. This is another one that may take awhile to fill in for old records.

Conclusion

We do hope you’ll find the new inventory application and database more useful, though it may take a little getting used to if you aren’t used to relational tables. The geodatabase has relationship classes already built to make it a bit easier to work with in ArcGIS. The geopackage doesn’t have anything to explicitly connect the tables, so use the information here to do the appropriate joins in your GIS. The zip file will probably have both the geopackage and the geodatabase initially, but that may be split in the future. There will also be a data dictionary to explain all the fields.

As you might guess, the map services to support the USIEI application have also been updated. Another post describes the changes to the topo and topobathy services.