Methods need to suit both purpose and available resources (money, manpower,
and skill levels). To improve the rate of data generation, methods need to be
acceptable to observers. In countries well endowed with amateur and profes-
sional ornithologists, it may not matter that most do not contribute to surveys
because there are plenty who will. In places with far fewer ornithologists, it
becomes more important that methods are simple enough to be acceptable to
them all.
Overall, there is a convergence on the value of species tick lists that are really
the simplest kind of recording that could be imagined. They are also the most
similar to the kinds of notes that almost all birdwatchers routinely collect. Such
lists improve in value with narrower time boundaries. One month, 1 day, 1 h, and
20 min have all been used to good effect. They also improve in value with tighter
area constraint. Areas of several hundred square kilometres (quarter degree
grids), 10 km^2 , 2 km^2 down to 2 ha have again all been used to good effect.
Smaller time boundaries are appropriate for smaller area boundaries. If data
are collected at a finer resolution of time and space, there are more analytical
possibilities. Such records can always be aggregated to a coarser resolution if
needed. The converse cannot be done if the original recording was at a coarse
scale. An example might be the fact that a record of presence in a quarter degree
cell cannot necessarily be attributed to a major habitat type. Nor can it even be
relocated although this might be required if it is an uncommon species.
Tick lists are not an efficient recording method for congregatory species such
as breeding seabirds or non-breeding waders or wildfowl. Confronted with
a sight of tens or even thousands of individuals, it is wasteful not to record
a number even if it is only an order of magnitude.
To date, there has been something of a separation of design of studies looking
at spatial or temporal patterns. Atlas studies have attempted to achieve complete
coverage of grid cells. They tend to last for several years—five is typical—and to
be repeated after tens of years. Atlases generally use simple methods. Temporal or
monitoring studies have tended to use more sophisticated methods, such as
mapping, transects, or point counts. Plots are distributed as observers choose
although increasingly with an element of sampling design. Plots are visited annu-
ally for as long a run as the individual observers can sustain. It is clear from those
atlases that have been repeated that quite large changes in range (and presumably
numbers) can be found over periods as short as 20 years. It is also clear, from the
American bird atlas or the African mammal atlas, that valuable distribution maps
can be drawn from studies not based on uniform coverage of a grid. There seems
to be an as yet under-exploited potential to design studies that measure spatial
and temporal patterns within the same design.
Conclusion| 13