Library Header Image
LSE Research Online LSE Library Services

Irreducible uncertainty in near-term climate projections

Hawkins, Ed, Smith, Robin S., Gregory, Jonathan M. and Stainforth, David A. ORCID: 0000-0001-6476-733X (2016) Irreducible uncertainty in near-term climate projections. Climate Dynamics, 46 (11). pp. 3807-3819. ISSN 0930-7575

PDF - Published Version
Available under License Creative Commons Attribution.

Download (2MB) | Preview
Identification Number: 10.1007/s00382-015-2806-8


Model simulations of the next few decades are widely used in assessments of climate change impacts and as guidance for adaptation. Their non-linear nature reveals a level of irreducible uncertainty which it is important to understand and quantify, especially for projections of near-term regional climate. Here we use large idealised initial condition ensembles of the FAMOUS global climate model with a 1 %/ year compound increase in CO2 levels to quantify the range of future temperatures in model-based projections. These simulations explore the role of both atmospheric and oceanic initial conditions and are the largest such ensembles to date. Short-term simulated trends in global temperature are diverse, and cooling periods are more likely to be followed by larger warming rates. The spatial pattern of near-term temperature change varies considerably, but the proportion of the surface showing a warming is more consistent. In addition, ensemble spread in inter-annual temperature declines as the climate warms, especially in the North Atlantic. Over Europe, atmospheric initial condition uncertainty can, for certain ocean initial conditions, lead to 20 year trends in winter and summer in which every location can exhibit either strong cooling or rapid warming. However, the details of the distribution are highly sensitive to the ocean initial condition chosen and particularly the state of the Atlantic meridional overturning circulation. On longer timescales, the warming signal becomes more clear and consistent amongst different initial condition ensembles. An ensemble using a range of different oceanic initial conditions produces a larger spread in temperature trends than ensembles using a single ocean initial condition for all lead times. This highlights the potential benefits from initialising climate predictions from ocean states informed by observations. These results suggest that climate projections need to be performed with many more ensemble members than at present, using a range of ocean initial conditions, if the uncertainty in nearterm regional climate is to be adequately quantified.

Item Type: Article
Official URL:
Additional Information: © 2015 The Authors.
Divisions: Grantham Research Institute
Subjects: G Geography. Anthropology. Recreation > GB Physical geography
G Geography. Anthropology. Recreation > GC Oceanography
G Geography. Anthropology. Recreation > GE Environmental Sciences
Date Deposited: 27 Oct 2015 16:43
Last Modified: 05 Jun 2024 06:54
Projects: ES/K006576/1
Funders: Economic and Social Research Council

Actions (login required)

View Item View Item


Downloads per month over past year

View more statistics