Loading ...
Sorry, an error occurred while loading the content.

396Re: [beemonitoring] Re: Standardized Sampling Methodologies and a Common Database

Expand Messages
  • John S. Ascher
    Aug 15, 2008
    • 0 Attachment
      This sounds good Sam. I have a few minor additions as follows:

      " 1. Standardized vs. Opportunistic samples or surveys"

      I'm not sure that these can be broken down so simply. My sense is that a
      wide array of sampling techniques are appropriate depending on the
      questions of interest and the circumstances. Much "Opportunistic" or
      taxonomically-focused sampling can be standardized to some degree, but
      using methods appropriate to descriptive and historical science (e.g.,
      historical biogeography) and therefore quite different from those applied
      to experimental studies such as those designed by statistically savvy bee
      ecologists.

      "3. Everyone does their own thing and keeps data in whatever
      database/spreadsheet they like and periodically contributes a text file
      with column headers to a central repository. Each database is owned by the
      contributor and is maintained (and included or excluded) by that group.
      Another body provides a service for extraction or
      display of these datasets...Discoverlife is a good example of this."

      A useful model, already implemented at Discoverlife, is for small
      contributors and those lacking computer resources to periodically send
      static data (e.g., from a spreadsheet) whereas larger and/or more
      computer savvy contributors can set up dynamic, continuously updating
      links (e.g. to a relational database) between their servers and the
      community resource.

      Many groups have already been developing useful standards for sharing
      pollinator data and we can usefully consult these and suggest that people
      adopt them. If people nonetheless persist in doing their own thing for
      whatever reason much of their data may still be rendered useful to all if
      a clever computer scientist can extract these.

      It is extremely important to note that there are already multiple linked
      central repositories in place. All data sent to one central repository can
      and should be shared dynamically with other collaborating repositories.
      Local repositories can enhance centralized (global) data by providing
      additional more particular services (e.g., customizable dynamic local maps
      and potentially analyses based on these) and by sending corrections
      discovered locally back to the general repositories.

      As a specific example, note that bee specimen records sent to GBIF can
      also be sent to other centralized data sources. This map of Bombus
      includes 135,000+ GBIF records and many others, all error-checked by the
      Global Mapper:

      http://www.discoverlife.org/mp/20m?kind=Bombus

      This example shows how the community can and should take advantage of
      multiple central repositories, as these have different strengths and can
      usefully link to each other to collectively display and error-check data.

      When planning this or any other project we should try to take full
      advantage of existing tools. Of these, web-based collaborative tools are
      already very powerful and are being improved every day.

      Images in particular can have a very wide array of uses once copyright
      issues can be addressed.

      In summary I suggest that we as a community assemble globally relevant
      data, which can of course easily be repackaged for local use, and
      establish dynamic links among central repositories (plural) and between
      these and local repositories.

      John

      P.S. On the subject of sampling oligolectic bees, these are not
      efficiently sampled using single-site/ecological protocols designed to
      obtain an unbiased cross-section of the community from an unbiased sample
      of floral resources. However these can be found very effectively using
      taxonomically-oriented methods, such as targeted collecting at sites were
      the particular taxa of interest have been recorded historically or at
      biogeographically similar sites. In this case sampling bias in favor of
      the oligolectic species of interest is a very good thing.





      > OK, I can see Matt's original message if I look on the listserv's web
      site...it was somehow corrupted by my email browser originally...
      >
      > For future reference all these messages are archived at:
      >
      > http://tech.groups.yahoo.com/group/beemonitoring/
      >
      > I believe that anyone can see these.
      >
      > So, this will be another important set of topics at any meeting.
      >
      > 1. Standardized vs. Opportunistic samples or surveys
      >
      > 2. Databasing and datasharing.
      >
      > In regards to topic one...Both general approaches are very useful, in
      their places and there is no reason not to develope systems for both.
      >
      > A survey or set of surveys can be established (likely at several
      geographic scales) that is systematic, standardized, and repeatable that
      will provide the most statistically rigorous means of looking at change
      and another complementary system can be established that
      > compiles unstandarized studies, data collections, museum information,
      general collecting etc.
      >
      > In regards to topic number 2. Sharing data and databasing are often big
      bottlenecks in collaborative projects. I have seen a number of ways for
      the NOT to work in the past, but only 3 that seem to work well.
      >
      > 1. One agency or group pays for, collects, analyzes, databases ALL the
      data (relatively unrealistic in this case). North American
      > Waterfowl Surveys or the Breeding Bird Survey are good examples of these.
      >
      > 2. One group maintains a data entry web site in which everyone
      > shares and produces reports and dataset of equal value to the
      > stakeholders. The North American Amphibian Monitoring program and
      FrogwatchUSA are good examples.
      >
      > 3. Everyone does their own thing and keeps data in whatever
      > database/spreadsheet they like and periodically contributes a text file
      with column headers to a central repository. Each database is owned by
      the contributor and is maintained (and included or excluded) by that
      group. Another body provides a service for extraction or display of
      these datasets...Discoverlife is a good example of this.
      >
      > sam
      >
      >
      >
      >
      >
      >
      > --- In beemonitoring@yahoogroups.com, "Matthew Sarver" <mjsarver@...>
      wrote:
      >> All -
      >> Clearly, we each have different opinions on this topic, biased by
      > our own
      >> interests and specializations. Such is the challenge of
      > collaborative work
      >> in the age of academic globalization! The common ground, as I read
      > it, is
      >> threefold:
      >> (1) A desire for some level of standardization in methods of
      > inventorying
      >> bees for the specific purpose of monitoring long-term population and
      distributional trends (Sam's original point, and the goal of his
      > work, if I
      >> understand it correctly)
      >> (2) A way to incorporate and make available the massive amount of
      non-standardized data already available in museums, and that will
      > continue
      >> to be generated by taxonomists and ecological field workers. This
      > data, as
      >> John points out, is of tremendous importance in natural history,
      > taxonomy,
      >> and biogeography, and can add to the standardized data in (1), and
      > may
      >> supersede it in many cases of rare or infrequently collected
      > species.
      >> (3) Following from the first two points, and as has been alluded to
      > by John
      >> and others, the need for a collaborative and
      > accessible "clearinghouse" for
      >> the resultant data from both standardized and non-standardized
      > origins.
      >> As a bit of an outsider (I often find myself walking a tightrope
      > between
      >> academia, government, non-profits, etc) perhaps I can offer a start. It
      seems to me that the standardization of protocols is only useful
      > if that
      >> data ends up in a common database for analysis and sharing. If we
      > are to
      >> build a common database for bee records, it would be foolish not to
      > include
      >> all of the records from non-standardized methodology, including
      > museum
      >> specimens, expert-identified photographs, etc.
      >> While the georeferenced specimen mapping tools in the Discover Life
      > guides
      >> are a good start, I would argue that an expanded version of that
      > database,
      >> with a much fuller feature set and search functions, and including
      > more
      >> fields, would be highly desirable. This North American Bee
      > Database (or
      >> whatever it might be called) could become the standard location for
      > storage
      >> of all bee specimen and photo records for the continent, and could
      > be made
      >> accessible on the web.
      >> Issues of standardization could be dealt with by populating, for
      > each import
      >> of records, a selection of fields indicating the type of record, the
      collection methods used, etc. This would hopefully not be as hard
      > as it
      >> might seem. Most bee specimens could be assigned to one of the
      > following
      >> collection methods: malaise, net/hand, bowl, vane trap, photograph
      > only, or
      >> unknown method (for museum specimens). Another field could ask for
      > the
      >> specific protocol used. Still more linked fields would hold floral
      association, habitat data, etc
      >> In this way, all relevant data could be compiled in a centralized
      > clearing
      >> house. Researchers interested in monitoring trends could simply
      > filter the
      >> database and view only specimens from standardized methods, while
      > those
      >> interested in floral associations or distributions could make use
      > of the
      >> complete data set.
      >> Several challenges come to mind here:
      >> (1) Funding / Personnel - such a project would require full time
      > attention
      >> from at least a few people building and managing the database, in
      > addition
      >> to much time from taxonomists (who, as John points out, are already
      overextended).
      >> (2) Academic intellectual property - Regrettably, this is a major
      > issue when
      >> dealing with such an endeavor, but that is the nature of our field,
      > and
      >> everyone should get due credit for their contributions. Perhaps
      > this could
      >> be overcome by a lock that contributors could place on data of
      > their own
      >> specimens. This "lock" would allow the data to show up in certain
      > contexts
      >> (e.g. state species list queries), but not in full detail until any
      > relevant
      >> publications were completed.
      >> (3) Data accuracy - a database such as this would require much
      > effort from
      >> competent individuals to ensure the accuracty of determinations,
      > etc.
      >> Including det. codes and dates in the database would be a minimal
      > step to
      >> help ensure the validity of records.
      >> (4) Accessibility. Difficult decisions would need to be made about
      > use of
      >> the contributed data. I am in the open data-sharing camp, but many
      > are not,
      >> and I understand the reasons for that. If full funding could be
      > found to
      >> support the efforts of staff and taxonomists, it would compel open
      > access to
      >> the compiled data.
      >> I feel that this is the direction that we should be going in this
      information age. We should all strive to overcome our own self-
      > interests
      >> and work toward a true collaborative effort!
      >> Sam, I apologize if I have hijacked your original intention, but it
      > seems to
      >> me that standardized methodologies are closely intertwined with
      > this idea.
      >> My two cents
      >> Matt Sarver
      >
      >
      >


      --
      John S. Ascher, Ph.D.
      Bee Database Project Manager
      Division of Invertebrate Zoology
      American Museum of Natural History
      Central Park West @ 79th St.
      New York, NY 10024-5192
      work phone: 212-496-3447
      mobile phone: 917-407-0378
    • Show all 28 messages in this topic