11/21 Moorea Experimental Garden Array (Contact: Neil Davies, neiltahiti [at] gmail.com)
Big ecological questions, diverse data, new methods.
Friday October 31, 2014. 9am - 5:30pm.
Additional details: Andy Rominger, rominger [at] berkeley.edu
Museum of Vertebrate Zoology, Grinnell Miller Library, 3101 VLSB.
Discovering universal mechanisms that drive ecological patterns will increasingly come from analyses that bridge scales of space, time and phylogeny. Gaining insights across these scales requires the development and application of analyses that are able to utilize data from varied sources (e.g., standardized surveys, haphazard specimen data, citizen science initiatives). Analyses of such data require methods that are able to account for non-standardized collection methods. In the context of community ecology, evolutionary ecology, and biogeography, new statistical methods provide the ability to fit mechanistic models rooted in biological hypotheses to disparate data spanning spatial scales from plots to continents and time scales stretching deep into the fossil record. Invited speakers will highlight one or multiple of the key themes underlying this workshop: big unanswered questions in ecology and evolution, diverse data that could be brought to bare on these questions, and the methods that will connect data with questions.
PI: Andrew Rominger (Graduate Student, ESPM)
Co-PIs: Leithen M'Gonigle (Postdoc, ESPM), Karthik Ram (Assistant Researcher, BiGCB), Giovanni Rapacciuolo (Postdoc, BiGCB), Jenn Weaver (Postdoc, ESPM and BiGCB), Adam Zeilinger (Postdoc, ESPM and BiGCB).
|09:00 – 09:45||ARRIVAL & breakfast|
|09:45 – 10:00||Introductory comments on goals for workshop|
|10:00 – 10:20||Elise Zipkin: Multi-state population models using count data|
|10:20 – 10:40||Brian Maurer: On the scaling of abundance and the hierarchical structure of geographic ranges|
|10:40 – 11:00||Adam Wilson: TBA|
|11:00 – 11:20||[tentative] Naia Morueta Holme|
|11:20 – 11:40||Carl Boettiger: Big data, ecological forecasting and the parameter problem|
|11:40 – 1:20||LUNCH|
|1:20 – 1:40||Perry de Valpine: NIMBLE: New algorithms for hierarchical models in R|
|1:40 – 2:00||Maggi Kelly: Rescuing and sharing historic vegetation data for ecological analysis: the California Vegetation Type Mapping project|
|2:00 – 2:20||[tentative] Student lightning talks|
|2:20 – 3:00||COFFEE break and open discussion|
|3:00 – 3:20||Rosemary Gillespie: TBA|
|3:20 – 3:40||Seth Finnegan: TBA|
|3:40 – 4:00||Cindy Looy: TBA|
|4:00 – 4:20||COFFEE break and open discussion|
|4:20 – 4:40||Cory Merow: TBA|
|4:40 – 5:00||Margaret Evans: Process-based range models|
|5:00 – 5:20||David Ackerly: TBA|
|5:20 – 5:30||CLOSING remarks|
|Evening plans to be determined|
Understanding Taxon Ranges in Space and Time.
November 7th, 2014. 1-5pm.
Additional details: Brent Mishler, bmishler [at] berkeley.edu
University and Jepson Herbaria Seminar Room, 1002 VLSB.
All studies of past and future changes in the distribution of biological taxa require an accurate knowledge of their geographic ranges. Many sources of data are used to infer ranges, including expert opinion, citizen-scientist reports, ecological plot data, models using climate data, floristic and faunistic lists, and museum specimens. There are many problems with interpreting all such sources of data, including taxonomic misidentification, and even if "correctly identified," uncertainty with respect to which taxon concept was used. Specimen-based data have an advantage over all sources of data with respect to solving these identification issues, since specimens have the diagnostic morphological and (in many cases) molecular features needed to clarify their correct classification under sometimes conflicting taxonomies.
While specimen-based data have the potential to be the "gold standard" for determining accurate taxon ranges, the usefulness of specimen data—particularly its usefulness to non-taxonomists or to large-scale meta-analyses—may be compromised by data errors. Due to the high demand of specimen data from the collective resources of the Berkeley Natural History Museums (BNHM) and their collaborators, and the sensitive nature of the research being conducted (e.g., climate change), high quality specimen data are needed. Theoretical and technical advances in the general field of biodiversity informatics are also needed, e.g., to better link specimens with their "correct identification" under different taxonomic concepts (the taxon concept problem).
The BNHM collections total over 12 million specimens of living and extinct flora and fauna. Each of the museums serve specimen data online and participate in large-scale, multi-institutional collaborations through which tens of millions of specimen records can be accessed. We have taken the lead in many advances in biodiversity informatics and are uniquely poised to help tackle the taxon concept problem, "the most difficult problem of them all" (John Wieczorek, pers. com.).
We will hold two workshops: (1) an initial, one-day meeting for just the on-campus collaborators and nearby data-users to clarify the issues and specific problems, and to discuss the most productive agenda for the following workshop; (2) a later two-day workshop for both on-campus and national participants (with an associated set of general talks open to all interested parties in the Bay Area, to broaden participation). The goal of this pair of workshops is to share examples of recent projects that requested large-scale downloads of electronic specimen data, receive feedback from users of the data, and discuss possible improvements in data curation that can make specimen data more valuable to researchers engaged in climate modeling, modeling current and future distributions of species under climate change, tracking biodiversity, preparing biological inventories, and correlating geographic range with other variables (such as ploidy and ecological traits).
Workshop participants will tackle three important questions that span all types of biodiversity data (1) assessing confidence in the range of a taxon in time and space (both spatial and taxonomic certainty), (2) how to apply taxonomic concepts to specimen data and other electronically available data that use taxon name as a reference point (such as photographs), and (3) what is needed to apply such taxon concepts to data sets, including how to capture annotations by researchers who use and clean specimen data for analysis.
Collaborators will include domain scientists, database managers, and IT specialists from each of the BMHMs serving specimen data, the Berkeley Ecoinformatics Engine, outside users who have been provided with large downloads of specimen data from the BNHM for their research, and a few key individuals from national and international projects (listed below).
Resulting products of the workshop will include developing an "white-paper" about the use of specimen data and what researchers need to be mindful of, constructing a list of potential improvements that can be made to specimen databases so that they can better serve both the research community and the public, and appointment of a core working group to develop a grant proposal that will seek external funding to pursue some of the more difficult issues related to specimen data (e.g. tracking taxon concepts).