Friday, February 16, 2007

Global Climate Change as a Complex Problem of a Complex System

Conventional sciences tend to focus on isolated systems under controlled circumstances. Decision support sciences must often deal with real world problems, the complexity of which far exceeds the complexity of the problems typically studied in conventional sciences. This realisation has lead to the designation of a new class of problems, complex problems, which includes issues such as global climate change, ... and stratospheric ozone depletion.
Complex problems are characterized by one or more of the following properties
(adapted from NRC, 1988; van Asselt, 2000; Funtowicz et al., 1999; Holling, 2001;
UNESCO, 2005):

  • There is not one problem, but a tangled web of related problems;
  • The underlying processes interact with one another within some sort of hierarchy;
  • The dynamics of the systems studied are not necessarily regular, but are characterized by synergistic and/or antagonistic relationships, indirect relationships, long delay periods between cause and effect, thresholds or non-linear behaviours;
  • The issue lies across or at the intersection of many disciplines, i.e., it has economic, environmental, socio-cultural and political dimensions;
  • There are a number of different, equally legitimate and plausible, perspectives on how the problem should be conceived.

The hierarchical relationships encountered in complex systems may be hierarchies of inclusion and scale, as in a watershed that includes streams, ponds, rivers, lakes and the sea, at ascending levels. Alternatively, they may be hierarchies of function, as in an organism that is comprised of a number of separate organs, each performing a function subordinate to the overall function of the organism, which itself may be sub-ordinate to the overall function of an ecosystem.

Environmental systems may also include human and institutional sub-systems, which are themselves systems. In complex systems, causes and effects are not always obviously related. Pushing on a complex system "here" often has effects "over there" because the parts are indirectly dependent of one another. Similarly, the future conditions in a complex system may not always follow closely on the conditions in the past.

When a particular threshold is exceeded, the system can abruptly shift away from a period of relative stability in one state, to another, fundamentally different state. An example is that of the impacts of climate change on thermohaline circulation, the large-scale ocean circulation that currently transports heat from the mid-latitudes to the high latitudes. Geological analysis and model experiments suggest that these currents can be on or off, and that the two states are characterized by drastically different environmental conditions in Western Europe (Broeker, 1997; Cusbasch et al., 2001). Similar dramatic regime shifts have now been documented for a wide range of environmental systems (Scheffer et al., 2001; Scheffer & Carpenter, 2003).

Because of the hierarchical, indirect, synergistic and non-linear relationships that can characterize complex systems, any attempts at reductionist analysis will be inherently incomplete. The concepts used to represent the functionings of the system will necessarily be rough approximations. The empirical data required may not be available, or may only be available in a form that requires interpreting or massaging to make it relevant to the problem at hand.

An approach inherent to conventional sciences is that of reductionism, whereby an overall system is understood as an assembly of sub-systems. By studying and understanding each of the sub-systems, an understanding of the overall system is achieved.

While the reductionist approach has led to many great achievements in Western science, the properties of complex problems, ... greatly reduce the effectiveness of the [reductionist] approach. Systems that are complex are not merely complicated; by their nature they involve deep uncertainties ...

MR2005-202.pdf (application/pdf Object)

No comments: