Environmental Engineering Reference
In-Depth Information
expected from our models. By examining both long-term data from our Cary Institute sites
and shorter-term data from nymph populations distributed throughout Dutchess County,
we found reasonably strong correlations between the infection prevalences predicted by
our models and those observed in nature (LoGiudice et al. 2008).
Another important observation was that the roles played by individual species change
with the ecological context. For example, adding shrews to a community consisting of
only deer and mice resulted in a reduction of tick infection prevalence, but adding shrews
to a community with many host species tended to increase tick infection prevalence.
Consequently, the order by which species are added to or removed from communities
(community assembly or disassembly rules) becomes crucial in determining natural pat-
terns of change in disease risk with changes in host communities.
In the other approach, we compared nymphal tick abundance and infection prevalence
in forest fragments that differ in size or in species richness. In both cases, we found that
infection prevalence is higher in smaller fragments (
Allan et al. 2003
) and in fragments
with fewer host species (LoGiudice et al. 2008). Results concerning nymphal tick abun-
dance have been mixed. In general, knowledge of the specific identities of the host species
provided considerably more predictive power over a simple accounting of the number of
species in the host community. Interestingly, this same phenomenon, whereby species
composition is a better predictor than species richness alone, has been found in other stud-
ies of the ecosystem functions provided by highly diverse communities compared with
depauperate ones (e.g.,
Loreau et al. 2001
).
It's the Ecosystem
Lyme disease is typical of many zoonotic disease systems. The emergence of the disease
is followed by hot pursuit of the species (pathogen, reservoir, host) responsible, which
allows scientists (largely from health specialties) to produce a sketch of ecological sources
of risk. These sketches are typically insufficient at best and inaccurate at worst. Their fail-
ings are a direct consequence of a reductionist focus on the minimum number of species
that play what are assumed to be the strongest roles in producing a pool of pathogens.
More holistic study of these systems tends to reveal that (1) predicting variable disease
risk requires knowledge of a larger number of host species; (2) these hosts are embedded
with communities and ecosystems in which unrelated processes (e.g., acorn masting or
other resource pulses) can have profound effects; (3) individual host species can interact
with one another, for example, with hosts “competing” for vectors or pathogens; (4) the
aggregate of species that are good hosts/reservoirs and poor hosts/reservoirs can be cru-
cial in determining disease dynamics; and (5) effects of individual host species can be
highly context dependent, with even the qualitative effect of a species (increasing or
decreasing risk) depending on the composition of the remaining community members.
On the basis of combining both relatively thorough case studies, such as Lyme disease,
and basic epidemiological models (
Rudolf and Antonovics 2005; Dobson 2004; Keesing
et al. 2006
), new theory is emerging that will guide the pursuit of general principles that
allow the emergence of infectious diseases to be understood and even predicted. An eco-
system approach is an indispensable part of this pursuit.
