Developing an Experimental Predictive Framework for Climate Regime Shifts and Their Impacts Within a 2-20 Year Outlook Window
Dr. Michael Alexander | NOAA/ERSL/PSD
The aim of this work is to enhance the ability to understand, assess and predict decadal-scale climate variations and related regime shifts, with a longer-term goal of applying this framework to gauging related climate impacts upon natural resource management.
Predicting decadal climatic events and their impacts remains a challenge. Numerous uncertainties in climate systems exist that have yet to be adequately assessed. The short observational record limits understanding of hydrological and ecological system responses to both amplitude and sequencing of climate events over decades. National Center for Atmospheric Research’s (NCAR) large ensembles of coupled climate models (most recently, the Community Earth System Model, version 1 [CESM1]) allow for better estimation of unpredictable climate noise relative to the climate change signal. A key element of the work will be to demonstrate how such a large number of lengthy climate simulations can be used to quantify predictability of climate regime shifts and uncertainty in regional climate assessments, leading to an actionable 2-20-year outlook.
However, the current generation of coupled climate models is not entirely able to simulate the pattern or time scale of observed large scale decadal variations in the Pacific basin, such as that of the recent regime shift of the Pacific decadal oscillation (PDO) occurring in 1998. This likely limits their use for quantifying related atmospheric and land surface predictability on a 2-20-year time frame, even in the CESM1 large ensemble. On the other hand, an empirical dynamical model called a linear inverse model (LIM) of low-frequency global ocean dynamics can provide better simulation of decadal Pacific variations, both in terms of its statistics of decadal variability, including the pattern, amplitude and frequency of regime shifts, and in terms of initialized decadal forecast skill. However, these models are limited in providing the necessary level of detail for atmospheric and land climate impacts, and can only extrapolate anthropogenic effects. Researchers believe that a combination of the two techniques will provide the best estimate of future potential for climate regime shifts and the related atmospheric response envelope of decadal variability. Therefore, the plan is to develop a hybrid model in which researchers link a Linear Inverse Model (LIM) to an Atmospheric General Circulation Model (AGCM), thereby combining the LIM’s better simulation of oceanic decadal variability with the AGCM’s detailed land and atmosphere simulations and emissions scenario-based projections of anthropogenic climate change. Researchers will first develop the technique by using the CESM1 as a testbed for this approach.
The development of a usable predictive framework to quantify uncertainty in regional assessments on the 2-20-year time frame. This will lead to the ability to insert the full range of low-frequency climate responses into different natural resource models, resulting in probability distributions of ecological impacts on decadal timescales in the presence of climate change. This work will advance the ability to effectively translate climate model predictions and uncertainties into the scientific basis for resource management decisions on decadal scales.