A great discussion of the DUHHHHH factor. From MIT.
Why is global warming a flat earth proffer!
Natural Variability. Even a chaotic system may have a well defined attractor, i.e., a well defined distribution describing the probability of occurrence of various climates. Such a distribution would describe the natural variability of climate which arises from the climate system's chaotic elements, and this distribution may itself be predictable beyond the limit of deterministic predictability. The natural variability of the climate system on time scales of decades and longer is very much dependent on oceanic and cryospheric processes. Paleoclimatic evidence indicates that it may be very large. However this natural variability is not yet well determined or understood.
Errors in Initial Conditions. Even in a non-chaotic system, the ability to predict future behavior depends on the accuracy of the knowledge of the system's initial state. For example, we do not yet know how accurately we need to know the temperatures of the deep ocean today in order to make a prediction of climate of a given accuracy, say, fifty years in the future. Given the long time scales of the deep ocean circulations, the climate trends of the coming decades may be crucially dependent on the history of climate during the last few centuries. The extent of this dependence is not known.
Model Errors. Even if the above factors did not introduce limits on our ability to predict climate, our climate predictions would be hampered by the imperfect state of current climate models. That these models have major errors is demonstrated by the fact that no coupled atmosphere-ocean general circulation model has been able to reproduce a climate equilibrium like the current climate without introducing artificial sources of heat, moisture, and/or momentum at the interface between the atmosphere and the oceans (i.e., "flux corrections"). The source or sources of these errors is not always known, but there are many potential sources that need to be investigated. Some examples are inadequate resolution of oceanic currents and eddies, and inadequate representations of subgrid scale processes such as moist convection and clouds in the atmosphere. The need to parameterize subgrid scale processes may introduce severe limitations on our ability to represent these processes and their impact on climate change, and we need to understand these limitations and the resulting uncertainty.
From web.mit.edu
tom watson tosiwmee |
