Geoscience Reference
In-Depth Information
0.5
Exponential
Power
Linear
Arrthenius
Lloyd-Taylor
0.4
0.3
0.2
0.1
0
0
5
10
15
Temperature (
°
C)
20
25
30
35
40
Figure 25.3
Variation in temperature sensitivity of soil respiration, with curves based
on fi ve different soil respiration models. Taken from Qi et al. (2002).
higher temperatures, and having a variable response along the temperature gradient
(fi gure 25.3). This indicates that simple empirical models generally used to calculate
soil respiration are ineffective, as they do not account for this non-linearity of
response, and therefore, lead to substantial errors in predicting carbon budgets and
fl ux within ecosystems at broad spatial scales. This is just a simple example of how
a small non-linear variation in response at fi ne scales can potentially lead to inac-
curate predictions at broader scales.
The recent suggested prevalence or dominance of chaotic or non-linear processes
in ecosystems does not necessarily mean that all ecosystem processes are complex
or non-linear; indeed ecosystems do have some characteristics of linearity, organisa-
tion, and predictability (e.g., Brown et al., 2002). Nevertheless, most ecosystems
are complex and will have elements of non-linearity, which need to be accounted
for in management.
Ecosystem individuality
For ease of interpretation and management, ecosystems have traditionally been
categorised according to dominant physical or biological characteristics (e.g.,
Rodwell, 1991). Table 25.1 implies that ecosystems within the same category func-
tion in very similar ways, but in reality, each ecosystem and ecosystem component
has notable individuality due to its historic context (e.g., sequence of past distur-
bances, management regimes, etc.), spatial variations in the characteristics and
arrangements of biotic and abiotic components (from landscape confi guration to
genetics), and stochastic processes, among other reasons (e.g., Jørgensen et al.,
2005). Consequently, although ecosystems may be categorized by type, and indeed
show superfi cial similarities in functional processes, no two ecosystems will function
identically. This context dependence frustrates generalisations and confounds man-
agement regimes that rely on predictive precision.
