Environmental Engineering Reference
In-Depth Information
of concern. Most biotic inventories in natural areas are woefully incomplete
(Stohlgren
et al
. 1995). In most areas, only one or a few species or genotypes might
have been surveyed (low taxonomic completeness; Crall
et al
. 2006). Within a
landscape, county, or region, large areas may be poorly surveyed for target species
(low geographic completeness). In addition, large areas of a site or region may have
been surveyed only once or very few times (low temporal completeness). Maps of
the distribution or abundance of invasive species should be accompanied by infor-
mation on the various levels of completeness throughout the study area so that
managers would understand the limitations of the maps for risk assessment.
Estimating the potential distribution and abundance of an invasive species requires
the information needs outlined above (Table 2.1; needs 1-5), integrated with a geo-
graphic information system (GIS), remote sensing, and fairly sophisticated math-
ematical models. In many cases, the resulting product is a 'habitat suitability map',
a map or model that describes the vulnerability of habitat to target species' invasion
(e.g. Chong
et al
. 2001; Schnase
et al
. 2002b; Venevski and Veneskaia 2003). h ese
models are generally based on a few climate, topographic, or soil variables. Maps
and models of target species' abundance are rare because abiotic and biotic factors
must be carefully quantifi ed relative to species population estimates, and because
of the plasticity and adaptive potential of the target species and genotypes. Patterns
of habitat invasibility have been slow to come, let alone mechanisms explaining
these patterns (Mack
et al
. 2000), and the complexity of this task should not be
underestimated.
It is well recognized that 'potential distribution' models have several limitations.
h ey are based on only a few predictive factors and are aff ected by the selection
of the factors, scale, resolution, and accuracy of spatial data inputs (Morisette
et al
. 2006). h e models do not include information on more than one biological
species (the target organism), so they do not include the complex of interspecies
interactions (e.g. competition, herbivory, predation). h e environmental factors
in the models are all held constant, and the local disturbances (e.g. fi res, fl oods)
and processes such as grazing and mortality of competing species are generally
presumed constant, which is rarely or never the case. Species-habitat relationships
and species mapping (except for humans) is in its infancy. Yet, developing these
capabilities is paramount to the next di cult challenge in risk assessments of inva-
sive organisms—predicting rates of spread of invasive species.
Mathematical models predicting the spread of invasive species are essential.
Managers may want to set priorities for control based partly on the priority path-
ways of spread, the area of potential habitat, the eff ects of the species throughout its
range, and rate at which the species could spread from its current distribution to its
full potential distribution (Pemberton and Cordo 2001). h ere are many models
