In the post
I presented reasons why the application of multi-decadal regional climate predictions for use in policy decisions regarding threats to key societal and environmental resources is scientifically flawed. These projections are providing a level of scientific confidence to the impact assessment and policy communities that is unwarranted.
This post summarizes several of the key points with respect to the climate scenarios to use:
1. The assessment of risk needs to start from “what if” past historical and paleo-history climate conditions reoccurred (e.g. the 1930s or 1950s drought in the central and southern Great Plains of the USA; the mega-drought of the 16th century in the western USA – see Meko et al. 2007), but with the current societal conditions.
2. These historical conditions, recent paleo-history and worst case sequence of historical events can then be used to estimate threats in response to plausible changes in society in the coming decades.
3. Realistic possible changes in the climatology of a region (e.g., increases by 1C in the summer average maximum temperatures; a decrease of 10% in growing season precipitation, etc.) can be superimposed on the historical climatological data. These can be made by asking the question as to what changes in the climatology would have to occur before a significant risk threshold is passed? This approach is an application of the bottom-up, resource-based (contextual vulnerability) approach that we present in our paper
Pielke Sr., R.A., R. Wilby, D. Niyogi, F. Hossain, K. Dairuku, J. Adegoke, G. Kallos, T. Seastedt, and K. Suding, 2011: Dealing with complexity and extreme events using a bottom-up, resource-based vulnerability perspective. AGU Monograph on Complexity and Extreme Events in Geosciences, in press.
The multi-decadal climate models are intended to provide information as to what are the “plausible changes”. However, in my view there is no need (and much added expense) to use the models to create changed climatology when these models have never demonstrated skill in the prediction for the past decades of changes in regional climatology. Thus, the large expense being applied, as illustrated in my post on the Mote et al 2011 EOS article, is not justified.
The argument has been made that the models are physically based so that even if they could not be verified, they still are “consistent with the physics“. I disagree whether they actually are consistent with the physics, but, in any case, the historical and recent-paleo-data is obviously consistent with the real world, as these conditions actually occurred.
It is this climatological data that should be the focus of impact assessments in terms of the input of possible weather conditions in the future. If one insists, the multi-decadal climate model predictions can be inserted as another set of scenarios to see where they fit within the climatology obtained from the historical, recent paleo, and worst case sequence of climate events. However, these climate predictions should not be the starting point for the risks from climate in the coming decades.