Environmental Engineering Reference
In-Depth Information
could be defined as the amount of total nitrogen in ppt and independent variables of
interest could be watershed type (WS), month (MO), precipitation measure (PR),
water depth (WD), and vegetation cover (VEG). Such a model could initially be
defined conceptually as
¼
F
WS
½
;
;
;
;
:
Y
MO
PR
WD
VEG
Such a model is considered indeterminate at this stage because the list of
explanatory factors is likely not complete and the response is not predictive (i.e.,
nature of the relationship of each independent variable with the response variable is
not defined). However, such a model is useful in designing a study. Fixed effects
(if control of experimental units is possible) could be watershed type, water depth,
and vegetation cover. Random effects could be month and precipitation. However,
final determination of random and fixed effects depends on the amount of control an
investigator has on the system and variables used to measure the effects.
An important principle in conducting experiments is to hold all factors, except
the one of interest, constant so that any response to treatment can reliably be
attributed to the treatment. Unfortunately, this is rarely possible in wetland studies
and thus, investigators must design studies to limit the variation within all variables
that are not of interest. In this example, one may be only interested in the relation-
ship between watershed type and total nitrogen while recognizing that factors other
than watershed type influence nitrogen levels in wetlands. The proper study design
can remove or partition the influence of independent variables in such a way that
priority treatment effects can be estimated and evaluated. Ideally, the model will be
constructed based on the hypothesis, available data, assumptions, and potential
unknown parameters.
Conceptual models can be specific, similar to the above example, or encompass an
entire wetland type (Fig.
1.1
). A critical step in constructing a conceptual model is
identification of the problem or question. Brooks et al. (
2005
) used existing data to
develop a conceptual model of wetland degradation and restoration in an effort to
improve scenarios for the use of mitigation wetlands to replace lost wetland area and
ecological function. They hypothesized that increasing influence of stressors
homogenizes wetland diversity and variability. Devito and Hill (
1998
) developed a
conceptual model of wetland sulfate (SO
4
) retention and export based on watershed
hydrogeology. An investigator should use conceptual models to develop objectives
and competing hypotheses for experimental testing. Ogden (
2005
) developed a
conceptual ecological model for anthropogenic stressors on an Everglades ridge
and slough system. He identified five major ecosystem stressors (reduced spatial
extent, degraded water quality, reduced water storage capacity, compartmentaliza-
tion, and exotic species) and made predictions on stressor effects. In addition, he
identified a series of biological indicators of wetland restoration success that can be
incorporated into future studies. In addition, conceptual models can be informed by or
produce predictions from ecological theory. Euliss et al. (
2004
) developed a concep-
tual model for prairie pothole wetlands - the wetland continuum. The model allows
for simultaneous consideration of climate and hydrologic setting on wetland
