Case Study: Upper Midwest Science Center (UMESC) Legacy Data Inventory

The Data at Risk Project’s legacy data inventory investigation at the Upper Midwest Environmental Science Center (UMESC) gave insights into how the structure and organization of a science center can affect how legacy data is managed, inventoried, and preserved.

Over the course of their 50-year history, the name and structure of UMESC has changed several times. Despite this change the Center’s mission has remained clear: provide the scientific information needed by resource managers, researchers, decision makers and the public; to protect, enhance, and restore the ecosystems in the Upper Mississippi River Basin, the Midwest, and the world. The ecosystem sciences in this region were originally conducted out of separate field stations; each station focused on different projects and operated under different funding sources. Eventually these field stations were combined into what is now the Upper Midwest Environmental Science Center (UMESC). While the four branches of UMESC are housed under the same roof, the original funding and ownership for many of the projects within the branches still exist.

The four branches at UMESC and their main funding sources are as follows:

• Geospatial Sciences (GS)—Funding from Fish and Wildlife Service (FWS), the Army Corps of Engineers (ACE), and the National Park Service (NPS)

• Long Term Resource Monitoring (LTRM)—Funding from ACE

• Aquatic Ecosystem Health (AEH)—Funding and regulations from the Food and Drug Administration (FDA) and the Environmental Protection Agency (EPA)

• Ecosystems—Roughly 50% of Ecosystem funds come from USGS headquarters, the other half from other entities. This branch has the largest proportion of funds coming from headquarters out of all four branches.

The structure and history of the center ultimately dictated how our legacy data inventory was conducted at UMESC.

The definitions below are relative terms that are used frequently throughout this case study. They are provided for clarity in the context of this case study and are not necessarily universal.

• LDIRS: The Legacy Data Inventory Reporting System (LDIRS) is an online application designed to catalog the USGS legacy data inventory and score data sets based on their significance, risk factors and spatiotemporal extent. The highest scored datasets will receive preservation funds.

• Inventory: The DaR team’s process of assessing what legacy data a science center has and recording them in LDIRS.

• UMESC Archival System: A system that was designed by the UMESC data manager and archivist to keep track of the center’s archives. The system has been implemented over the past 6 years to document and organize the scientific records at UMESC.

• Record: Any paper or digital file in the science center related to a scientific study, correspondence, or administrative details.  

• Study: the collection of documents that all contribute to the same research question conducted by a scientist or group of scientists. A study typically includes a publication, data set, iterations of the data, protocols, cruise reports, background information on the study, etc. A ‘study’ can be split different ways, as many research programs occur over many years or in many places. In most cases, one study is contained in one file folder but in some cases many file folders can make up one study.

• Physical Record: Non-digital records, usually in paper form.

• Identified Record: A record that is accounted for in the UMESC archival system.

• Unidentified Record: A record that is not accounted for in the UMESC archival system.

• Personal Records and Branch Archives: Records in either physical or digital form that are in the possession of the center’s staff or specific branch, usually stored in personal office spaces or hallways, and are not accounted for in the center’s archival organization system.

• Study Number: Every identified study has a designated study number, a naming convention that indicates scientific branch, year of study, topic, and study number. This naming system is consistent for all numbered studies (e.g., AEH-13-LAMPTFM-02).

To carry out the legacy data inventory investigation the DaR data steward spent 4 days (32 hours) at UMESC planning the inventory process, cataloging valuable legacy data sets from the center into LDIRS and in doing so assessing the challenges and successes of the methods. The general process involved learning about the center’s records management and organization, assessing what physical and digital records were held in the center, deciding which area(s) of the center were the most efficient spaces to spend our inventory time, and actually sifting through the data and writing up LDIRS entries about them. We used this initial inventory trial to project the potential time needed to complete an inventory of the entire science center. The efficiency and success of our process was greatly influenced by the center’s existing archival system and institutional knowledge.

At UMESC there is one person in charge of data management, one archivist, and one quality assurance officer (QAO). Between these three staff members there has been notable progress in the center’s archival organization. Over the course of about 6 years they’ve identified about 62% of the centrally stored physical records in the center and organized them using the UMESC archival system. This internal archival system allows them to operate efficiently within their center, with tasks such as identifying and locating studies of interest, tracking records management schedules, and preserving legacy data.

The first two days of the week were mostly spent working with the data manager and the archivist to assess the quantity, layout, and security of the records at UMESC so we could determine which area(s) of the center would be most effective for a legacy data inventory.

The general layout of the center’s records is as follows:

• Identified Physical Records:

  • The majority of the identified physical records at UMESC are stored in the center’s archive room. The archive room contains 38 file cabinets for scientific records (datasets, publications, protocols, etc.) and 12 file cabinets for administrative records (time sheets, budgets, order forms, etc.). There are 5 file cabinets outside that are accounted for in the archival system, but wouldn’t fit in the archive room.

  • All scientific records in the archive room (plus the 5 outside the room) are accounted for in the center’s archival system, meaning that each study has an identifiable study number and an index sheet that lists all components in that study’s folder. The studies are organized by branch and can be located in the archive room by using a spreadsheet that lists the study numbers found in each numbered file cabinet.

• Unidentified Physical Records:

  • There are about 10 file cabinets stored in the stairwells of the building that hold physical records that have not been identified in the archival system. In most cases these records have been lying untouched for several years and there is little information about them. Work has been started by the archivist to sift through these unidentified physical records and archive anything of value.

• Personal Records and Branch Archives:

  • There is a large volume of materials that exist in the possession of individual scientists. The content, amount and format of these materials is largely unknown to anyone other than the scientists who possess them. The recovery, organization and preservation of these materials typically occurs only if the individual strives to do so or during the 9-090 Exit Interview Process.

  • The LTRM branch and the Wildlife Ecology branch keep their physical records in their own file cabinets (LTRM has about 10 file cabinets and Wildlife Ecology has an unknown number) . These cabinets are stored in different parts of the building and are not accounted for in the center’s archival system.

• Security and Risk:

  • There is a sprinkler system in the newest part of the building (bldg D), which is where the UMESC archive room is located. The security of the file cabinets in the center vary depending on the branch or individual responsible for storing the records; certain regulated studies are required to be stored in locking, fireproof cabinets while others are up to their own discretion. The risk factors of the records vary depending on where they are located and how they are stored.

After assessing the layout of the center’s records, we decided to focus our efforts in the center’s archive room. The organization and documentation of the studies in the archive room made the next step of sifting through records and identifying valuable datasets much more efficient. We spent the final 2 days conducting the following inventorying process:

1. identifying studies of interest (complete datasets with significant study topics and/or large spatiotemporal scale) using the archive room spreadsheet for reference,

2. locating the study’s dataset and protocol; and

3. inventorying the dataset and writing an LDIRS entry that included information about the data and contextual information from the study’s protocol.

Results

Over the course of the week we inventoried 9 datasets and subsequently created 9 new entries in LDIRS. The most complete and valuable datasets we selected from the archive room were those dealing with chemical development for aquatic invasive species control--all 6 of these studies were conducted under the Aquatic Ecosystem Health Branch. After gathering more details on the studies, we found that all 6 of these datasets were technically owned by an outside entity (usually the EPA or the FDA) and, therefore, have restrictions on sharing the data. As a result, some of these datasets had to be removed from LDIRS, although others we have looked further into and have discussed at the end of this case study.

The other 3 datasets were inventoried by talking to various UMESC staff who knew about the data. These datasets are data from the long term ecological monitoring branch, aerial photographs, and hard drives from retired scientists.

A breakdown of the time allocations and estimates:

• Time spent during the week’s investigation:

  • 13 hours spent inventorying 9 datasets. Each dataset required an average of 75 minutes to inventory.

  • 8 hours spent planning the inventory.

  • 11 hours spent interviewing UMESC staff to contextualize the data and clarify any rights and restrictions that might be associated with it.

• Science Center Estimates

  • The estimated total number of file cabinets in the science center is 85 (this takes into account the archive room (50 cabinets), the identified records outside the archive room (15 cabinets), the LTRM branch’s records (10 cabinets), and the unidentified records in the stairwells (10 cabinets). Digital records, personal and other branches records that are not accounted for in the archival system are not considered in this total.

  • In our week’s inventory we spent 13 hours inventorying 9 datasets. The 9 datasets take up about 25% of a file cabinet. Based on that, it would take about 52 hours to inventory one file cabinet and about 4,420 hours to inventory the physical records at UMESC.

The main takeaway is that previous record management and inventory efforts enabled the DaR team to go in and easily identify which datasets we wanted to inventory and exactly where to find them.

The main factors we found that influenced the management, inventorying and preservation of legacy data were the organization of the center, the presence of outside entities in funding and ownership, and the utilization of institutional knowledge. 

Organization:

The existing archival organization system at UMESC helped to make the legacy data inventory process more efficient and more focused. The time the archivist spent going through and identifying records in a way that allowed for easy identification and location of particular studies saved the DaR team time and allowed us to go directly to studies of a particular topic. This ease of access has also proved to be beneficial for UMESC scientists who want to go back and find a particular dataset for further study or reference.

As tool, the master index sheet is useful for quick assessments of what’s contained in a study/file folder, as well as an efficient way to request project records from a Federal Records Center (FRC). In many cases, UMESC has sent physical records, master index sheet included, to a FRC. Keeping a copy of the master index sheet enables UMESC staff  to identify a study or piece of a study of interest and request the records from the specific FRC if needed. Locally managing the master index sheets saves time and storage space by minimizing the need to digitize and store on-site every record from a study that may or may not be used again. Sending  records to FRC or the National Archives (NARA) on schedule is not only good records management practice but it's also an effective way to make space available in the center and keep track of the constant flux of data.

While their system isn’t fully developed, it’s important to recognize the advantages of UMESC’s systematic organization and control they have over their science center’s scientific records, specifically legacy data.

Outside Ownership:

With different outside entities providing funding for all or most of a branch’s projects, there are direct effects to how the data can be handled once the study is done. While conducting the UMESC inventory, we found that there were some advantages and disadvantages to data being owned by someone other than USGS.

Having studies tightly regulated by outside entities proved to be advantageous for legacy data preservation by enforcing excellent records management and detailed documentation throughout the studies. One example of this at UMESC is the Aquatic Ecosystem Health Branch that’s funded and regulated by the EPA. This inter-agency relationship was established to regulate studies dealing with chemicals (mostly pesticides for invasive aquatic species) that needed to be accepted by the EPA and FDA before becoming available for use. These tight regulations caused studies from the AEH branch to be well organized and well documented, compared to other branches who weren’t held to such strict regulations. In addition to being well organized, the EPA and FDA regulated studies required all records to be stored in fire-protected file cabinets.

The main disadvantage in having outside ownership of the data was that it was often restricted for uses we were interested in, such as releasing to the public. The restrictions varied based on the project, funding entity, data agreements, etc., but in all cases if an outside entity owned the data, permissions would have to be given to us before we could release the data.

We delved into these policies a bit more with some of the datasets we inventoried from the AEH Branch. One example is a collection of studies on TFM, a chemical toxic to sea lampreys. The establishment of Great Lakes Fisheries Commission (GLFC) in the 1960’s spurred the regional effort for sea lamprey control to combat the detrimental invasive species across the Great Lakes and Midwestern waterways. The GLFC remains in control of the studies conducted on sea lamprey control out of the Midwest, therefore, in this case the Data at Risk project will follow up with the UMESC-GLFC connection to determine whether or not we can release these high profile legacy data.

Institutional Knowledge:

One of the most valuable resources we used in our inventory process was the center’s institutional knowledge. Communication is a key part of legacy data preservation. In general, a study is best understood by the scientists who conducted the work and the information managers that assist them. The UMESC data manager and archivist provided a deep working knowledge of who best to talk to about specific studies and how to effectively get information from them. Directly engaging UMESC scientists provided project-level context and a practical  understanding of the legal restrictions of the work. Without that input, legacy data preservation would not be possible in many cases. Utilizing current institutional knowledge is, therefore, a crucial part of preserving USGS legacy data. Without this knowledge the data loses much, if not all, of its value.