Workshop: Bases para a Conservação da Biodiversidade do Estado de São Paulo

Comments by Robert Colwell

Dept. Ecology and Evolutionary Biology

University of Connecticut, U-43 Storrs, CT 06269-3043

Tel: (860) 486.4395

Fax: (860) 486.3790

Mail: colwell@uconnvm.uconn.edu

Based on discussions at the Workshop, I see the biodiversity project focused on five distinct objectives. These are my recommendations regarding them:

1. Information Recapture. Existing information on the biodiversity of Sao Paulo exists in several forms. The objective is to recapture or reorganize alll existing information.
   1. Existing information already in electronic form within the State of SP. Databases and data files exist in a variety of formats in many locations and institutions in SP. These databases and data files need to be studied, preferably by those currently responsible for them or by their originators, to provide for the Project a standard set of metadata for each dataset. No effort should be made, in this step, to synthesize the datasets themselves. An impressive first step was taken at the Workshop, in asking participants to fill out survey forms about existing datasets, but not all were prepared to do so, and more complete metadata are required.

   2. b. Existing information already in electronic form in Brazilian institutions outside the State of SP and in other countries. Increasingly, data files exist in a variety of electronic formats in many locations and institutions outside SP. Working out cooperative arrangements for datas-sharing with these institutions can begin with queries to their databases for electronic information on Brazilian specimens, followed by arrangements either to download this information to a central data bank somewhere in SP, or perhaps longo-term agreeements for online access over the Internet, integrated into the SP metadata files.

   3. Existing information on historic specimen labels or field notes, but not yet in electronic form, from specimens deposited in SP collections. The task of bringing this category of information into usable electronic form is of course enormous, but can be focused and scheduled to proceed in parallel with other objectives. Type specimens, for example, could be made top priority, then groups of specimens of particular conservation interest, or priorities could be set by other criteria. Were it my decision, I would attempt to pursuade all teams engaged in this process to use the same software, or at least the same data structure, to carry out the data entry, with the objective of easing the later task of distributed data integration, and to promote a consensus solution to the inevitable decisions regarding how to parse locality descriptions, how to handle non-obligatory taxonomic ranks, and so on.

   4. Existing information on historic specimen labels or field notes, but not yet in electronic form, from specimens deposited in collections in other parts of Brazil and in other countries. This process (pioneered by CONABIO for Mexico) requires either lengthy visits by data-entry personnel from SP, or contracts with the donor institutions to provide data entry and transmittal. Note: the term "data repatriation" should be avoided, as it unnecessarily raises tempers at many otherwise cooperative data-donor institutions. Although the specimens are from Brazil, the information being offered to Brazil under data sharing arrangements was added, usually at considerable expense, by personnel at the data-donor institution. In other words, the information is not being "repatriated", because it did not exist until added outside Brazil, in most cases. "Data sharing " is a much more accurate and less inflammatory term.

2. Information discovery. Plans initiated at Serra Negra to delineate both taxonomic gaps (lacunae) and geographic gaps will guide proposals for the gathering of new data and specimens from the State of SP. In the inevitable cases in which no taxonomic authority exists in SP (or perhaps in Brazil) for an "orphan" group, perhaps foreign specialists can be enticed to participate, through invitations for extended visits to SP collections, combined with intensive, short-term training of SP personnel in collecting, preparation, and rough-sorting for the group in question, allowing the specialist to make the most of his or her time with well-mounted and databased specimens (this process is modeled on our experience with Project ALAS).

3. Information, integration and management. Integrating all the foregoing forms of biodivesity information (and others I have not mentioned) on the distributed (or database federation) model requires careful planning, clever experimentation to adapt system to local needs, and close scrutiny of similar efforts already underway elsewhere. Some examples include ERIN, the Ecological Society of America's Data Archive, and the Organization for Tropical Studies/Smithsonian Tropical Research Institute data sharing project. (I am involved in the last two, and will let you know more about them as time goes by. Both are just getting under way).

4. Information publication. The goal of providing Internet access to biodiversity information for the State of SP will be reached most effectively and efficiently if it is an integral part of 3., above. In addition, public involvement in the conservation of biodiversity may be moved farther by providing appropriate levels of information in Visitor's Centers at natural areas, through local identification groups of organisms, through natural history programs in elementary and secondary schools, and possibly through a television series on the biodiversty of SP.

5. Intensive study of a few archetypal or "reference" sites in SP, with less intensive study in "comparison" sites. The goal of intensive surveys and inventory throughout the State is obviously unrealistic, but I believe the overall level of knowledge will prove greater with a well-designed mix of intensive and less intensive efforts. The future value is very great, both for research and conservation objectives, for a virtually fully known biota in one site in each of the major habitat/ecosystem types in the State.

Strengths of the Program

1. All major stakeholders involved from the start.

2. Commitment to a multipath approach, including all the objectives I have outlined above.

3. A conscious decision, from the outset, to keep the process as decentralized as possible, thus empowering and rewarding those closest to the information to maintain and amplify it. Obviously, this approach also requires continuous and careful attribution of credit where credit is due. If those "in the trenches" come to feel the administrators are receiving all the rewards, the distributed model will not work. Alienation and perceived disenfranchisement of stakeholders and participants would be the worst enemy of the project.

4. The setting at the outset of simple, clear, and minimalist data standards.

5. A genuine effort to learn from the experience (and specially from the mistakes) of others, while maintaining a commitment to find your own path.