Climate change is a phenomenon with enormous want of certainty about its extent and effects. This is not only due to a lack of scientific knowledge. Indeed, the basic physics of climate change are well understood, and the consequences of CO2 emissions for the atmosphere's energy balance is very precisely predictable on a global level. However, this is also where certainty ends. Unfortunately, an increase in the atmosphere's energy level and mean temperature does not mean that climate will change evenly across the globe. Warming on a global scale can also mean cooling in some places, or a more frequent emergence of extreme events such as storms and droughts. Predicting these local impacts of climate change is a major task for climate science today.
On the other side, the further development of greenhouse gas emissions is nearly unpredictable and depends on the global population, economic development, and available technology. There is similar uncertainty about the exact ways in which changes in temperature or precipitation actually influence ecological and human systems. Additionally, the high complexity of the global climate system and lack of knowledge about some ecosystem dynamics means that the impacts on the natural environment are not fully predictable. For example, the effect of forests on local precipitation is not exactly known yet, such that a forest which is shrinking due to decreasing rainfall could amplify as well as dampen the local decrease.
The bottom line is that, given the data and scientific knowledge available, there is a multitude of possible outcomes when trying to estimate the effects of climate change decades or even centuries in the future. Because each step of analysis, starting from the global atmosphere to local socio-economic impacts, adds another layer of uncertainty, scientists speak of a "cascade of uncertainty" that increases along the chain of impacts. Specifically for water and agriculture, climate variability creates the risk of crop failure and endangers food security as well as the livelihoods of rural communities.
Scenario analysis is a common tool used to lower uncertainties. Scenarios are models of possible outcomes dependent on the progress of various factors. Moreover, a multidisciplinary approach within science with a connection to relevant political, private and civil institutions should be achieved. Broad-based and transparent communication of changes and threats is important to gain understanding and action by the people. Important capabilities herein are flexibility to react, visionary thinking for innovative ideas, directivity as well as thinking in a long-term perspective.
Changes in complex systems and the important input factors are often only partly sizeable. Scenarios can assist in projecting possible outcomes by starting on the current base and giving various pathways on the base of most plausible estimations. The margin of discretion is displayed by the scope of several scenarios. Information for scenario development can be quantitative as well as qualitative data.
The selection of scenarios for an analysis should pay regard to the following principles:
- Limited number: the set of scenarios should be as small as possible
- Comprehensive: the framework needs to cover sufficiently different future development to represent a plausible range of assumptions and thus represent relevant uncertainties
- Comparability: the scenario set should make it possible for some research knowledge generated in one community to be compared with information generated in another
- Multi-scale: the storylines should provide enough explicit information on the aggregated scale to be clearly distinguishable also at finer scales. Similarly, storylines and scenarios should embrace near-term and long-term future conditions; the former providing linkages to ongoing trends and planning horizons, and the latter accommodating plausible large-scale divergences in key driving factors
- Structured but flexible: The scenario set should provide enough structure to facilitate consistency, but also offer flexibility for defining relevant details
- ↑ Stainforth et al. (2007): Confidence, uncertainty and decision-support relevance in climate predictions. Phil. Trans. R. Soc. A 15 August 2007 vol. 365 no. 1857 2145-2161. http://rsta.royalsocietypublishing.org/content/365/1857/2145.full