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

321Re: [beemonitoring] The Soap Collecting Jar

Expand Messages
  • Sam Droege
    Jun 5, 2008
      Jerry:

      Very interesting comparison.  There is quite a large literature on inter-observer differences now in bird counts, particularly the common point count system.  The results demonstrate that, no surprise, there are quite large differences in abilities to detect birds among observers.  Due to all of that we almost always model in an observer co-variable when looking at trend data.  A similar thing is generally to be expected in any skill-based system of counting or catching wildlife as your turtle work clearly shows (very interesting there wasn't a clear association with training!).  I think the main feature in the case of bowls is the more interesting notion of whether people somehow affect capture rates due to a way they may be laying out the bowls and also whether bowl captures are more variable due to differential capture rates under different conditions.

      In the later regard we have 2 projects this summer that hopefully will bear on the topic of variability of captures due to extrinsic factors.

      1.  Patty String will be working in Northern Virginia on stormwater ponds.  These ubiquitous ponds are mown once a year through a county contract and she has been collecting permissions from the various landowners to do a before/after/control/treatment experiment.  The general idea is to see if capture rates change (generally thought to increase) after mowing due to a loss of nectar resources and perhaps an increased visibility.

      2.  Leo Shapiro will be working with Region 5 (Virginia to New England) National Wildlife Refuges to look at variability of bowl captures across and among refuges on a set of fields on a fixed day and in some cases across several days.  This will give us some idea about how bowl captures of bees vary from field to field which will give us some sense of sample size allocation, biodiversity estimation, uniformity (or lack there of) of regional populations of bees, etc.  Leo emailed today and we will be talking tomorrow about the possibility of adding a separate net component, but since we only got one third of what we requested, we likely won't be able to do that much.

      sam


      Sam Droege  Sam_Droege@...                      
      w 301-497-5840 h 301-390-7759 fax 301-497-5624
      USGS Patuxent Wildlife Research Center
      BARC-EAST, BLDG 308, RM 124 10300 Balt. Ave., Beltsville, MD  20705
      Http://www.pwrc.usgs.gov


      Excuse my wandering.
      How can one be orderly with this?
      It's like counting leaves in a garden,
      along with the sound notes of partridges,
      and crows.
      Sometimes organization
      and computation become absurd.
                         Rumi
      P Please don't print this e-mail unless really needed.


      Jerry_Freilich@...
      Sent by: beemonitoring@yahoogroups.com

      06/05/2008 11:43 AM

      Please respond to
      beemonitoring@yahoogroups.com

      To
      beemonitoring@yahoogroups.com
      cc
      Subject
      Re: [beemonitoring] The Soap Collecting Jar





      I've been following this thread with some interest. I've always felt
      sympathy for E.O Wilson who turned to ants because of a vision problem
      resulting from an accident. I have no excuse as good as Uncle Ed's, but
      I've always felt that I don't have very good hand/eye coordination in
      netting small, fast-moving bees. When I'm with other bee folks, I always
      miss the things they catch. I've tended to use pan collectors and other
      bulk traps simply because I'm not good at the other stuff despite having
      had excellent teachers. I've always assumed that the pan collecting may be
      more repeatable but that perhaps I'm missing the "good stuff" I'd only be
      able to get by being an expert, and by being fast.

      T'ai's comments on teaching novices and inter-observer differences rang a
      bell. I worked on a similar problem years ago when screening people to
      search for desert tortoises. In a controlled experiment on replicated
      plots, I found that certain people were markedly better than others at
      finding tortoises, but that previous experience at doing this work was NOT
      a predictor of that ability. In other words, there were pronounced
      differences among people in developing a search image, staying focused, or
      discerning patterns, but that previous experience did not necessarily help
      with this. These results suggested that spending some time to find those
      talented searchers would be worthwhile. I also tried using a page from
      "Where's Waldo?" as a quick surrogate to see if it was any predictor. I.e.,
      to see if people with better Waldo finding skills would be better tortoise
      finders. That didn't work out, but I still think that people vary widely in
      their ability to pick up on things and that that talent should be sought
      out and rewarded. The skill sets used with bees and tortoises may be very
      different but I bet that the underlying principles are likely the same.
      While researching the paper I found that there were relatively few studies
      on inter-observer bias. (Freilich, J.E. and E.L. LaRue. 1998. Importance of
      observer experience in finding desert tortoises. Journal of Wildlife
      Management 62 (2): 590-596.)

      Jerry
      __________________________
      Jerry Freilich, Ph.D.
      Research Coordinator, Olympic National Park
      Coordinator, North Coast & Cascades Research Learning Network
      Olympic National Park
      600 E. Park Ave.
      Port Angeles, WA 98362

      Phone: 360-565-3082
      Fax: 360-565-3070
      Cell: 360-477-3338

      Jerry_Freilich@...

      "This is the most beautiful place on earth,
      there are many such places..."
      Edward Abbey
      ___________________________



      T'ai Roulston
      <
      thr8z@virginia.e
      du> To
      Sent by:
      beemonitoring@yahoogroups.com
      beemonitoring@yah cc
      oogroups.com
      Subject
      Re: [beemonitoring] The Soap
      06/05/2008 09:01 Collecting Jar
      AM AST


      Please respond to
      beemonitoring@yah
      oogroups.com











      Sam:



      Your basic design is reasonable and it would be great to have someone do something
      similar sometime. An hour of netting is definitely not sufficient and there is no
      doubt that bowls would win hands-down for repeatability in most circumstances with
      such limited net collecting. The hour that people use to accompany bowls now is
      not for a real faunal survey but rather to get some things that bowls miss. In my
      current community surveys with nets, I use something on the order of 10 collecting
      hours per day per site (with a substantial crew, total duration depending on the
      number of plant species of interest). Obviously, only people with substantial
      resources (human and otherwise) and a need for data of faunal associations can
      deploy that kind of effort. What is really needed is a comparison of sampling
      efforts (how many bowls are necessary --data you probably have in spades-- for
      species accumulation curves or other metrics to flatten) and how many net hours of
      sampling by single collectors are required for the same saturation estimates. When
      that is established, the variability of collectors can be thrown in to understand
      the relationship between sampling effort and method bias. I train novel net
      collectors every year for intensive field work. I don't have systematic data for
      comparisons of individual collectors (I make sure that everyone collects on all
      plants to avoid systematic person x plant collection biases), but I am impressed
      with how quickly people get good at collecting. I think that most but not all
      biases from individual collectors can be controlled by sampling protocols.
      Selection of method should always depend on the ultimate goal for the data and
      which types of biases and limitations are most tolerable.

      T'ai

      On Jun 5, 2008, at 8:26 AM, Sam Droege wrote:




      T'ai:

      True, I am generalizing based on experience rather than on data. Other than
      some rather small references in papers to differences in observers in
      netting capture rates I can't think of anything published that looks at
      comparisons in variability in any set of different techniques. Maybe some
      design like this:

      Perhaps use 6 collectors (3 experienced, 3 neophyte) and 18 Study Sites
      (Vernal bottomlands, transmission lines, or fields, 1 Hectare but perhaps
      larger so their is a larger collection of bees to sample from)


      Day 1. Observers are assigned to the 18 sites (3 each) in which they set
      out bowls in the early a.m. and then net for 1 hour during the middle of the
      day.
      Day 2. Observers do an additional 3 sites allocated in such as way that all
      pairs of observers have been matched. (note that there would be 3 extra
      sites on day 2)

      It would be nice to repeat this at another 18 sites so that the number of
      bees wouldn't be exhausted at a site.,.... or perhaps better... simply wait
      several weeks and do it again on the same sites

      That's my initial idea. I have a feeling there is a more parsimonious
      design out there somewhere but can't think of it at this point. I am also
      concerned that because we are dealing with day, site, observer, and
      technique factors here that our degrees of freedom might be eaten up. This
      would be a grand summer student project. .... How about doing it at Blandy
      this summer? I would be glad to come down to participate.

      sam


      Sam Droege
      Sam_Droege@...
      w 301-497-5840 h 301-390-7759 fax 301-497-5624
      USGS Patuxent Wildlife Research Center
      BARC-EAST, BLDG 308, RM 124 10300 Balt. Ave., Beltsville, MD 20705

      Http://www.pwrc.usgs.gov




      Albert Einstein
      “Not everything that can be counted counts, and not everything that counts
      can be counted.”

      P Please don't print this e-mail unless really needed.




      T'ai Roulston <
      thr8z@...>
      Sent by:
      beemonitoring@yahoogroups.com


      06/04/2008 09:35 PM





      Please respond to

      beemonitoring@yahoogroups.com



      To

      beemonitoring@yahoogroups.com
      cc

      Subject
      Re: [beemonitoring] The Soap
      Collecting Jar










      The advantage of bowls for general surveys is that they are more replicable
      than following the ramblings, skill, and proclivities of a person with a
      net.




      Sam:

      A lot of people are repeating statements similar to this, casually and in
      submitted manuscripts. Certainly, any standardized collecting is more
      repeatable than non-standardized collecting (the rambling, skill, and
      proclivity part). Is pan-trapping more repeatable than standardized
      intensive netting? It probably is, but I haven't seen the data yet. There
      can certainly be big differences among collectors in catching small or
      speedy insects, but the variability of pan trap catch with placement can be
      extraordinary as well. Repeatability is likely to be strongly related to
      sampling effort (number of collectors/collecting time or number of pan
      traps) plus variation related to biases in method. None of these factors are
      well documented or, to my knowledge, easily generalized at this point.

      T'ai

      T'ai Roulston
      Associate Director Blandy Experimental Farm
      Research Assoc. Prof. Environmental Sciences
      University of Virginia
      400 Blandy Farm Lane
      Boyce, VA 22620
      540 837-1758 ext 276

      thr8z@...







      T'ai Roulston
      Associate Director, Blandy Experimental Farm
      Research Assoc. Professor, Dept Envi Sci. University of Virginia
      400 Blandy Farm Lane
      Boyce, VA 22620



    • Show all 14 messages in this topic