Handbook for Atlasing
Planning a Biological Atlas:
Ontario Atlas Project
Recently, a number of people have been discussing the possibility of starting a new collection of field data for an atlas of the plants or the herpetofauna of the Province of Ontario. These ideas flow naturally out of the success of the Ontario Breeding Bird Atlas.
Much has been learned about the operation of an atlas project from the Bird Atlas project, which might be of assistance to others who are considering the planning and operation of a biological atlas project. It is essential that preplanning of an atlas be as thorough as possible. There were two years of methodological preparation put into the Bird Atlas (1979 and 1980), but this was insufficient in several ways which we will discuss below.
The major items to be considered are: scale, field methodology, institutional structure, administration, staffing, funding, data handling, and analysis.
Decisions must be made early on the scale of the project: all of the province or only a portion thereof? We are quite thankful now that some people pushed hard for all of Ontario to be included in the Bird Atlas, but the size of the project dictated that different sizes of data collection unit would have to be used in different parts of the province. For example, in southern Ontario, comprehensive coverage was the goal, using the 10 X 10 km UTM grid, while in northern Ontario data were collected from individual 10 km squares within 100 X 100 km blocks. This system is quite feasible for a large area such as Ontario (over 1 million km2), but it does add complexity to the system and requires the use of the two-map format used in the Atlas: one for southern Ontario by 10 km2, and one for all Ontario by 100 km block.
The necessity of keeping the data collection unit consistent across the entire province was not recognized in the initial stages of the project, and as a result we have some data for 100 km blocks without any idea of what 10 km2 it was collected in. This problem was largely solved part way through the project.
In Ontario and in Canada as a whole, it is probable that all future atlases will utilize the 10 X 10 UTM grid system for data collection. The UTM grid is printed on all Canadian topographic maps and is being used for atlas projects in Ontario, Quebec, the Maritimes, and Alberta. This helps ensure that data from provincial atlases can be combined into a national atlas data base as it becomes available. It would also be of tremendous research potential to be able to undertake correlations between the data for different life forms, and with physiographic data.
In remote areas individual 10 X 10 squares could be "sampled" within some larger unit of measurement. In the case of rare species the specific 6 digit number UTM code could be used to locate the exact location of specific records.
At the same time it is reasonable to consider the coverage of the entire province in any future atlas. It has been proven that it can be done and the resultant data base is more valuable than if it was only a portion of the province. The resolution of the data in northern Ontario is obviously very coarse, but this must be measured against the vastness of the area and the paucity of data which existed previously.
It is essential that the details of field methodology be worked out well in advance of any field data collection. The important factors that need to be considered here are:
We assume that all future atlas projects will use largely volunteer labour backed up by certain levels of technical expertise of a more professional nature.
The institutional structure of the organization for the operation of the project is of critical importance. Our experience has confirmed our belief that all major groups involved in a particular field must be involved in the planning and operation of a project from the very start. This would include all government agencies, museums, university and college departments, NGOs and major corporations that would have interest in the field. There are many diverse aspects of running a project which require effective management: Who will run the operation? Who raises the money, hires the staff, encourages the volunteers, pays the bills, owns the data, receives requests, answers questions, prints and distributes materials?
The Ontario experience suggests that a non-governmental organization was an excellent choice as administrative home for an atlas project. The Federation of Ontario Naturalists, with the assistance of the Long Point Bird Observatory, played a key role in the Bird Atlas project and would be ideal candidates for any future projects. Governments and universities, as institutional homes for atlas projects, suffer from a number of drawbacks which volunteer groups do not. Usually, the most important advantage of a volunteer group is that of independence. Such a group can act quickly over a wide area and with flexibility. Such factors are often limited in large government or quasi-government agencies. A number of bird atlases in other countries are faltering because they are being operated on too narrow an institutional base with the concurrent lack of potential, profile, and administrative capability.
The Bird Atlas was basically governed by two committees - the management committee and the technical committee. The former dealt with such items as intergroup cooperation, project structure, funding, staffing, budgeting and project development. The latter dealt with methodological issues such as data handling, interpretation of field data, field card and instruction booklet design, field work design and error checking. These committees served two vital purposes: project Supervision, and communication back to the institutions represented by each committee member. The committees were very carefully designed so that representatives from a large number of agencies and groups were present.
A worthwhile variation on this type of structure (that of having each project sponsored and administered on an ad hoc basis) is that which has developed in Great Britain, where the national government has established an agency, the Biological Records Centre, to serve as the central repository for all of their various atlas projects. Their staff serve as advisors on all major aspects of new and ongoing projects. We have no similar group in Canada, so if an atlas is to run to its fullest potential it must raise money to hire a staff.
Any atlas project without a permanent staff person will simply not work at all effectively. There is only so much that volunteers can be asked to do and the drudgery of project administration is not one of them. The responsibilities of the hired staff would include:
Our experience with the Ontario atlas was that a minimum of two, and as many as six, full time staff members were required at various stages in the project, with the greatest effort being required in data review and in the production of the book.
Very little can be done without some funding. The Ontario Breeding Bird Atlas made a decision that a certain level of base funding was necessary for such items as two staff, printing, data handling, office expenses, travel, telephone, mailing and computer time. This base budget amounted to approximately $60,000 per year. Other monies were raised for specific projects such as data collection in remote areas, computer software development and computer equipment. In total, $70,000 of funds and equipment were utilized in the 8 years of the Ontario Breeding Bird Atlas.
If sufficient funding is not available for a basic level of administration, it would be better to hold off on a field project until a later date. This might be preferable to the operation of a sputtering effort which operates intermittently according to the availability of funds or the interest of a volunteer coordinator.
The Bird Atlas was able to raise such excellent levels of support because it was well thought through at a professional level. A hastily conceived project would not have been able to obtain the approval of those in government and industry who are well versed in project administration, especially in these days of fierce competition for funds.
The handling of the data for the Bird Atlas was not at all well thought through when the project started in 1981. But since that time, very hard work by a number of people, including a full-time programmer for the last four years, has resulted in a sophisticated computer-based data handling system. All the data was stored on the IBM mainframe computers at the University of Waterloo. Two PCs and three output devices (two printers and a plotter) were dedicated exclusively for atlas use. In addition, large printers and plotters were available at the University when necessary. The computer system was used for:
With over 400,000 records there is no other way of handling the information. Any new atlas must consider its data handling needs well before any field collection should take place. For example, the design of the field card can influence to a large degree the efficiency of data input, and determine the information which can be extracted later. The data card used by the Ontario atlas did not, for instance, allow for the recording of the exact date of each sighting, and hence it is often impossible to say just when a suspected migrant was seen.
Another primary consideration is the hardware and software required. The Ontario atlas used SAS as the software package for handling the data base, BASIC for driving a desktop plotter, and custom software for driving the laser scanner-plotter used to produce the publication-quality maps in the Atlas. Since the Ontario project got underway in the early 1980s, more sophisticated relational data base management systems have become available, including some good packages for PCs. The value of SAS on the other hand, is that it allows for program development for almost any conceivable purpose, is widely used and well supported, and is available in compatible mainframe and PC versions.
Certainly any project beginning now would want to consider using the very powerful PCs which are now available. For some applications, it would be worthwhile to have such a PC connected to a mainframe computer. This would allow access to the powerful data processing capabilities of the larger machine. It would also allow access to the many output devices that are usually attached to mainframes in a large computer establishment. As a relevant example, this paper was composed on a Digital PC in Paul Eagles' office, then edited on a similar machine that was located in Dave Balser's office and connected to the Eagles' machine via a computer network. The finished product then was transferred to a VAX mainframe computer, then to an IBM mainframe computer, both at The University of Waterloo, in Waterloo, Ontario. The paper was then electronically transferred via an inter-university computer network to Charlie Smith's computer account at the Cornell Laboratory of Ornithology at Ithaca, New York.
One final cautionary note concerns software development. It might be assumed that most of the programming required would be to develop or implement a mapping system. This is a major task, of course, but at least as much effort was devoted to producing summaries, lists, and specialized reports for regional coordinators, authors, editors and those scrutinizing the data.
One important aspect of any database is its use. The Bird Atlas database is now in active use by a variety of people, and it is hoped that the number of researchers using the data will increase as demonstrations of its utility and potential are provided. For a database to be utilized efficiently there needs to be a basic administrative structure in place. The following items need to be addressed:
In summary, we have learned a lot from our efforts in regards to the Bird Atlas. This information could be of use to others who are considering the establishment of new atlas projects. We would like to encourage such initiatives and are willing to provide assistance.
The NORAC website is hosted and maintained by Bird Studies Canada