Digitization in US Herbaria, 2012

These are the results as of Oct. 18, 2012.

NEW: Histograms of the data by number and percent of total. Not updated since Sept 7]

Introduction

The goal of the US Virtual Herbarium project is to enable development of a single access point to the information associated with all the herbarium specimens in all US herbaria. This requires thate the specimens be databased, georeferenced, and, when appropriate, imaged. The project conducted its first survey of herbaria in 2011. The 2012 survey asked a few more questions, one concerning georeferencing and others asking for more information than in the previous survey on where the data were made available.

Procedures

The survey was initially posted online. During the 4 weeks it was open, 87 people responded, including 5 representing herbaria not listed in Index herbariorum (IH) The total number of specimens reported was 15,682,050 of which 41.7% were databased, 5.9% were imaged, and 15.5% were georeferenced.

Because of the low response rate, a paper version was sent out to 570 herbaria that had not responded to the online survey and were, so far as one could tell from IH, still active. On June 18, the paper version was made available on the web and another plea for returns posted to the herbaria and taxacom listserves. On July 14, the format of the posted survey paper survey was modified to address criticisms made at Botany 2012; fhe questions were not changed. In mid-August, regional representatives were asked to encourage people in their regions to respond and Barkworth sent out a few individual appeals.

Results

Eleven envelopes were returned as undeliverable. In some instances, a search of the web revealed that there had been a change of address; others appear to have closed. Somebody had written on one envelope "This college closed 15 years ago". I have sent Dr. Thiers, who maintains IH, a summary of the information concerning the undeliverable addresses. She tells me that there has been an unusually high number of corrections coming in from US herbaria. We both thank you for helping her maintain its value.

On Sept. 3, 2012, there were 736 entries in Index herbariorum for U.S. herbaria. The USVH database has 744 records because it includes a few that are not listed in IH. Of those listed in IH, 142 are no longer in existence (this number includes those reported closed in this year's survey. Most (111) of these herbaria transferred their specimens to other herbaria before closing. For the purposed of this report, "transferred" includes herbaria that have been combined administratively even if they are being maintained separately and those who have sent their specimens on long term loan to another herbarium.

It is not always clear what happened to the other 31 herbaria. In some instances it seems they were simply thrown out.

Returns

I have received 330 responses so far; a few more have been promised. Some respondents reported totals for two or more herbaria being administered jointly, e.g., RM and RMS; BRIT, SMU, and VDB; A, ECON, FH and GH; UC, JEPS, and GOD. If the total number of responses, such jointly reported herbaria count as a single herbarium. In 2011, some of these herbaria sent individual responses so, overall, there have been several more returns this year than last.

Specimen data.

The responding herbaria are estimated to hold a total of 70,783,776 specimens of which 22,672,899 (32.0%) have been databased; 3,878,581 (5.5%) imaged, and 7,433,748 (10.5%) georeferenced. The georeferencing figure is probably an underestimate; many respondents answered the question with "unknown" which was counted as none, others said something like "<5%" so I entered a number that was less than 5% of their total. Among the responding herbaria, 238 (72.1%) have begun databasing, 121 (36.7%) imaging, and 163 (49.4%) provided estimates of how many of their specimens were georeferenced. Eleven herbaria have more specimens imaged than databased.

in 2011 survey, the number of respondents was 292. In combination they held 55,465,574 specimens of which 16,947,122 (30.6%) and 152,042 (0.3%) were imaged. There were no questions about georereferencing last year.

Comparing the data for the two years, it appears that, since last year, the number of specimens in US herbaria has increased, that progress is being made in databasing and even greater progress in imaging. Interpretation is complicated by the fact that the herbaria responding in the two years were not the same.

The data suggest that 81 of the responding herbaria have shrunk since last year. Some have had some or all of their specimens transferred to another herbarium. Other reductions probably reflect a change in the way the estimate was made or the respondents' frame of mind when filling in the form rather than real shrinkage. Another possibility is that, as digitization has progressed, the respondent realized previous estimates were high.

There are no specifications as to how the estimate should be made; some herbaria include only accessioned specimens, some include their backlog of unaccessioned specimens. The reason for the absence of any guidelines is that Index herbariorium has none and the intent is to keep the data obtained from these surveys compatible with its data.

The good news is that some of the returns were from newly registered herbaria and some from herbaria that are not registered. The bad news is that some respondents (I will evenutally go back and count how many) said that he or she was retiring - or was already retired - and that the fate of the herbarium was uncertain.

Internet Access

The majority of the responding herbaria (200/329) provide no online access to their data they have databased 3,789,058 specimens, imaged 846,183 specimens and georeferenced 691,575 specimens. A few stated that they would be providing data to a network in the future, often the near future, but this reports reflects the situation at the time the survey was completed.

Local portals: Of the responding herbaria, 89 provide searchable access to their data via a local portal, i.e., a portal that provides access only to their own data. Of the herbaria providing local access, 50 are not contributing records to a regional or taxonomic network. Herbaria which expect to offer such access soon are not included in the above counts.

Regional portals: Several herbaria contribute records to a regional nework and some contribute to more than one. The networks and the number of responding herbaria contributing to them are: Flora of Alabama (5), Consortium of California Herbaria (17), Consortium of Northeastern Herbaria (6), Florida (2), Great Plains (2), New Mexico Biodiversity Network (1), Consortium of Pacific island Herbaria (1), Consortium of Pacific Northwest Herbaria (12), Great Plains (2), Oklahoma Plant Atlas (1), Oregon Atlas (3), Pennsylvania (1), SEINet/Intermountain (21), SERNEC (2), Flora of Texas (1), New Mexico Biodiversity (1), Wisconsin Flora (4).

Fifteen respondents provide access through the Global Biodiversity Information Facility. Four have provided data to CONABIO, NY provides data to a network for plants of Brazil, ASU to one for American Myrtaceae.

Taxonomic portals: Each of the three NSF-funded taxonomic networks was represented among the respondents. The numbers contributing records to each of the three networks are: Bryophytes (7), Lichens (14), Macrofungi (5). The tritrophic network also includes plants but there is, as yet, no access to the records being acquired.

All the regional portals and most of the local portals have mechanisms for searching by taxon, user-defined area, and collector. Most also provide access to images of living plants.

Interest: Several respondents stated that they would like to digitize their collection but did not have the time, staff, or funding to do so. Some individuals stated that they would like to join a network. In some cases, I was able to place them in contact with the leader of an existing network. All the leaders concerned were happy to be told of the interest even though it undoubtedly means more work for them. Two respondents from small herbaria said that digitization was simply not a high priority for them.

Caveat

The major problem with this report is the number of herbaria that have not responded. This is, of course, always a problem in surveys. The problem with this survey it that the population being sampled cannot be described as uniform. The impact of large herbaria on the specimen numbers is significant. I shall, like last year, eventually compare the frequency of returns in different size classes.

FEEL FREE TO TELL A NON-RESPONDING COLLEAGUE TO SEND THEIR INFORMATION IN! The survey (a pdf verson) can be downloaded from http://herbarium.usu.edu.

I appreciate that I sent the survey out when people may have left for the field so I shall keep the file open through August.

The list of respondents is now sorted by state.