Using Multiple Methods: ADx

Submitted by Ben Smith | published 19th Jul 2011 | last updated 13th Jan 2020

Forecasting, scenarios, and hope (Postma, Liebl 2005 p. 166)

Decision making models can be categorised by many dimensions such as qualitative-quantitative, participatory, etc. One of the key category is probabilistic or non-probabilistic. Although, these two types of approaches use similar types of inputs such as 1) consistent, predetermined (risk), and uncertainties, the two types of models have different requirement, purposes, and trends due to the amount of input types they can handle (Figure x).

Probability based decision making is generally used to predict future scenarios based on the information available at the present time. The likeliness is usually assigned to the alternative futures in the form of probability. If decision maker can set many probabilities or predetermined based on the frequencies in previous events or mathematical reasoning, e.g. a dice, the model will shows likelihoods of particular events in the future. However, this approach has severe weakness where probability is not possibly estimated by empirical studies. Strategic level adaptation planning to climate change is one of these situations. Climate change brings too much uncertainties for its strategic decision making processes; therefore, it is not possible to carry out a predictive decision making analysis unless many assumptions are included (Lempert, Nakicenovic et al. 2004 p. 2, Rayner, Malone 1998 p. 15).

On the other hand, non probabilistic approaches focus on drawing the plausibility or possibility of future states. In other words, these approaches help to generate possible future scenarios for decision making processes (Postma, Liebl 2005). The scenario approaches are more torrent to uncertainties than predictive approaches. However, this approach is still has some limitations if users are try to "know" future through one of these approaches. Scenario approaches do not tell you, which decision you should take. Instead, these approaches only show possible future states and its potential ranges, and then, a decision maker has to make her/his decision on her/his own, i.e. the purpose of scenario approaches is different from that of predictive approaches. Moreover, there are "unknowable unknown information" (Type 3 of known knowns) and the type of information will not possible to find as it can be ignorant (Ansoff 1976). Therefore, if the situation has too many ignorant, even scenario approaches cannot contribute much on a decision making process.

There are some more shortcomings in scenario approach. The scenario approach cannot capture diverse view of future scenarios. Also, most scenario approaches including IPCC's SRES scenarios failed to suggest some distributions attached to scenarios. Although uncertainties like climate change are not possible to be presented as probabilities, but likelihoods will remain as a centre of discussion (Parson, Burkett et al. 2007). Therefore, to suggest decision makers all scenarios equally will be bring some confusion or debates (Groves, Lempert 2007 p. 74).

A forecasting model based on known probabilities is possible only if most factors are predetermineds (Figure x). A scenario approach can handle more uncertainty factors, but if a decision maker has to handle too many uncertainties and does not have many predetermineds, the approach will not be useful, either. Hence, it is extremely important not only to consider multiple methods, but also use appropriate methods to assess complex decision options for their purposes. ADx will help to select multiple appropriate methods or approaches to a particular decision making situation to adapt climate change. This issue is revisited in the section the later section [1].