Biology Reference
In-Depth Information
to the protocol for data collectors to record a location every 10 min, regardless of
whether the target species is present. If not present, the location should be recorded
as an absence location. This addition would create a record of what areas have been
surveyed for the species, providing a better temporal view of a species' distribution.
Additionally, returning to at least a subset of the sampled present and absent loca-
tions each year would provide information on local spread, such as how quickly a
particular patch of an invader is growing. Adding a “professional” layer of plot data
to a larger set of weed mapping data as demonstrated by Barnett et al (2007) can
also help with these issues.
Typically weed mapping is only focused on a single or a small subset of species.
However, when people are out searching for species, they may find a previously
unreported species that has established in the area of which managers were una-
ware. Although it is unlikely that these types of data collectors would distinguish a
cryptic invader, they might notice a showy species or one that is suddenly abundant
in the area. Additionally, they could be provided with a short list of species that
have a high probability of invading the area based on distribution predictions. Data
collected by weed mapping over time can then be used to forecast invasions.
6.3.3 Statistical Techniques
Using statistics to forecast the distribution of a particular species and its rate of
spread involves developing a relationship between data collected in the field and
other predictor variables available across the full extent of the area of interest such
as climate data or satellite imagery. Statistical methods can be used with weed map-
ping data or with research plots that capture information on a suite of species rather
than a single or a few individual species.
Gilbert et al. (2005) offer one of the few examples of a published modeling effort
that actually predicts the spread of a species through time over a large area, and
their products provide useful information for the control of the species. The
researchers followed the spread of the horse chestnut leafminer (
Cameraria
ohridella)
from its initial invasion from 2002 to 2004 in the United Kingdom and
were able to use this information to develop models to predict the further spread of
the species in the next four subsequent years (2005, 2006, 2007, and 2008). So, they
produced a map of the United Kingdom showing the distribution in each of the
seven years (Fig. 6.1).
However, models like this one are not easy to develop. A major limitation of
these statistical techniques is the data needed to parameterize them. More detailed
predictions such as those including rates of spread require very specific data.
Modeling the spread of an organism that moves quickly like an insect or disease
may be hampered by the inability to gather needed datasets or by not having the
time to fully understand the organism's ecology and dynamics. In contrast, the chal-
lenge may be quite different when attempting to model rates of spread of a plant
species that may take decades to move across a landscape. When looking at this
