October 16, 2013

Chaos in climate change scenarios means chaos in climate change impact estimates?

I finished a first complete draft of the paper "Chaos in climate change impact estimates".

Why the 2011-2030 years averages of temperature anomaly in December, January, February are so different for the same General Circulation model that uses virtually identical emissions trajectories?

(Thanks to Paola Marson for generating these maps!)

What are the implications for the impacts literature?

Is it really possible to use high-resolution climate change scenarios to predict impacts at sub-regional level?

These are the questions that I address in this paper at the cross-road of climate science, economics and the impacts literature.

I copy the abstract below. The draft of the paper on GCM scenarios is ready and available here. A lot of maps and other Supplementary Material is available here.

Global Circulation Models incorporate chaotic dynamics to reflect real-world weather patterns. This implies that extremely small perturbations of the climate system may generate very different weather patterns. Here I show that the SRES climate change scenarios generated by the Coupled Model Intercomparison Project phase 3 (CMIP3) - ubiquitous in the impact literature - display strong chaotic dynamics at regional and sub-regional level, at least until 2065. Chaos is triggered by changes to historic forcing in the year 2000 to reflect different emissions trajectories. This suggests that large uncertainty exists on how to link local climate change and global forcing. Furthermore,  short- and mid-term differences in local climate change across different SRES emission scenarios reflect chaotic dynamics rather than different forcing patterns. I show that the "chaos" in the climate scenarios generates a "chaotic" relationship between exogenous forcing and local economic impacts. "Perturbed exogenous forcing" model ensemble would resolve this uncertainty.