electronic access to sensitive locality data

Bob Makinson rom at ANBG.GOV.AU
Tue May 24 11:22:55 CDT 1994

Julian Humphries writes:

>>      I seriously doubt the accuracy of the story about rare plants being collected
 based on Internet Accessible plant databases (at the current time).  Why,
 because there aren't any!
and further:
>>      I have not found a single USA based botanical database with non-type specimen
 data available on the net.  (the only non-USA one is ANBG).

Indeed and perhaps so - this week!  But the issue is absolutely real and will
 rapidly become more so as more collections come on line.  It is a certainty
 that some of the the more astute pirate collectors are already alert to the
 eventual possibilities for easy access to detailed locations of organisms that
 are valuable on the black-market or the open market.

I agree with Julian's view that the default position (whether via Gopher or via
 in-person enquiries to herbaria and other biological collections) should be
 open access to data.  We are after all public institutions (or at least we are
 in Australia).  Nevertheless, it is the the real-time nature of internet access
 that is the nub of the issue here, and the fact that it is without even the
 cursory opportunity to vet enquirers that is offered by the traditional front

We have been discussing the issue here at ANBG, as far as rare plant locality
 information is concerned, since early last year, and have had a reasonably
 comprehensive discussion that has also drawn in personnel from the
 Environmental Resources & Information Network, Australian Nature Conservation
 Agency, CSIRO, and the Rare Or Threatened Australian Plants (ROTAP) listing and
 assessment project.

The discussions arose from concerns expressed from various quarters about
 security of locality data, especially for rare and endangered plants, given
 that the Gopher system allowed relatively unchecked outside access to IBIS
 herbarium records, i.e. real-time access to specimen label data held on the
 IBIS at ANBG Herbarium.

The point was made repeatedly that there are no guarantees of data security
 under present (pre-Gopher) systems, and that black market collectors and others
 could, with peristence, usually find out where a plant occurs.

This was agreed by all participants, but the counter-point was made that one had
 the advantage of being able to check bona fides of herbarium visitors, and that
 an unauthorized collector seeking data from other sources had at least some
 chance of being detected by people or organisations sympathetic to protection
 prior to getting the data.  Locality information contained in Specimens Seen
 lists in publications could be (and is sometimes) selectively censored, with a
 referral to the appropriate wildlife management authority for those wishing
 more detail.

The problem with Gopher is that it provides instantaneous access for outside
 enquiries, whether bona fide or not, and there is no chance of policing data
 flow before the event.  The job with Gopher is therefore to seek appropriate
 trip-wire procedures that have a good chance of screening out at least some
 pirates.  It was accepted by all that a sophisticated computer operator could
 find avenues of database interrogation that might circumvent these trip-wires,
 and that this could not be totally prevented without either highly complex
 programming or banning public access altogether.  The latter alternative was
 not acceptable to anyone.

ERIN does not supply point localities for any taxa; enquirers wanting these are
 referred to the appropriate herbarium or state conservation agency as the data
 custodians and providers.  ERIN supplies "fuzzed" geocode data for 1o or 0.5o
 grid cells (or, in the case of Cape York, 6 km grid cells).  This is deemed
 broad enough that a black market collector is unlikely to gain much assistance
 in locating a vulnerable taxon (unless its preferred habitat is known and is
 very highly distinctive within the grid cell).

One problem with keeping totally open network access for all data, particularly
 with listed Rare/Vulnerable/Endangered taxa,  is that over the years
 duplicates, direct site access, and other forms of locality information has
 often been supplied by other herbaria or state management authorities on the
 basis (explicit or implicit) that it would be subject to the consensus rule
 governing herbarium access and enquiries, i.e. bona fide researchers only, with
 referral to the management body offered if access is refused.  Enforcement of
 this, and exercise of the necessary discretion, has always been left to the
 institution holding the specimens.  There are known cases where we could expect
 a partial or complete cut-off of information and assistance from wildlife
 management officers if rare taxon information that they had supplied was on
 open network access.

One participant pointed out that the (Australian) legal position governing
 public access to specimen information (if not the actual specimen) was unclear:
  does any herbarium have the right to refuse access?  How does FOI legislation
 affect our legal standing?  It was agreed that in the longer run the legal
 rights and obligations of collection and wildlife managers need to be
 clarified.  However it was also agreed that FOI-type obligations do not
 necessarily require us to supply data in real time, and that delayed
 information was a reasonable minimum step in being able to assess enquirers
 against a likely threat to a taxon by unauthorized collecting.

One taxonomist argued strongly for an embargo on all locality information,
 certainly for listed rare, vulnerabale, and endangered species of orchids, but
 preferrably for all orchids.  He cited cases to show that whole-population
 collecting was a real threat in this family.

He also advanced the view that individual researchers should be able to request
 embargo of whole records of taxa under study if they thought there was a risk
 of taxonomic piracy (e.g. unscrupulous overseas taxonomists searching for "sp.
 nov." records, retrieving information for these, and rushing to unrefereed
 publication; cases are known, although not so far through database access and
 not at CBG).   A solution short of complete embargo (honour system) might be to
 have a "work in progress" flag, with contact details, at the gateway to such
 taxa.  It was agreed by the meeting that this problem and possible solutions
 needed separate argument and consideration.

John Briggs, oner of the compilers of the ROTAP listing over many years,
 favoured some degree of restriction for all ROTAP taxa, of all categories (i.e.
 R as well as V&E), but not necessarily of the whole record.

Briggs made a key point:  that while black-market collecting was a problem for
 some taxa, the greater threat to R&E taxa was through ignorance, of what they
 are and where they occur.   A real estate developer or local authority was much
 more likely  to damage or destroy rare taxa if they did not know that they were
 there!  Hence any information embargoes should be minimised for genuine users,
 and any extra clearances should be subject to minimum necessary time delays.
 It was  acknowledged (and cases were cited) that developers had occasionally
 been known to deliberately and pre-emptively destroy rare plants as soon as
 they became aware of their presence and before any protective action could be
 taken, but this was much less likely than the destruction-through-ignorance

A question remains as to whether whole-record or part-only information needed to
 be embargoed.  One problem with blanking only the plain-language locality field
 (and perhaps generalising the geocode) might be that enquirers could view the
 partial record, identify a distinctive phrase in another field, and retrieve
 the whole record (or an adjacent one) by querying on that phrase (without using
 the taxon name and invoking the embargo message).

Some other creative possibilities for circumventing embargoed information
 (whether whole-record or part-record) were discussed, but are not discussed

Consensus of the meeting was for a solution including the following features:

*       Records for sensitive taxa to be not available for external users without
 further clearance.  This to apply also to newly discovered taxa, not yet
 ROTAP-listed, which are likely to qualify for ROTAP listing; herbarium staff to
 use discretion in designating these.

*       A notice to be inserted at the IBIS gateway stating that records for sensitive
 taxa are not available without further clearance, and giving contact details.
 Also to have an invokable electronic form requesting clearance for such taxa,
 for completion by the enquirer (giving bona fides, maybe with referees); this
 form to be capable of printing/postage by the enquirer, or direct email

*       The embargo would need to be programmed such that queries on the NAME  of an
 embargoed taxon, or on the COLLECTOR NAME/NUMBER combination for a particular
 record of the taxon, or on the REGISTRATION NUMBER for a particular record of
 the taxon, would all invoke the same embargo notice and clearance application
 form (in addition to the IBIS gateway notice).

*       Password access for embargoed taxa to be available for approved external users
 at ANBG discretion (e.g. probably National Parks services, other herbaria,
 etc., with caution re sensitivity of information).

*        Re technical implementation, a two-level Gopher was one possibility; one
 level, including sensitive data, for local network users only (plus external
 password holders); a second level, excluding sensitive data, for open-access
 external enquirers.

Bob Makinson
Curator of the Herbarium
Australian National Botanic Gardens, Canberra
email rom at anbg.gov.au

More information about the Taxacom mailing list