About

Kyle Armour is a professor and the director of the Program on Climate Change at the University of Washington. He holds a joint professor position in the School of Oceanography and the Department of Atmospheric and Climate Science. His research focuses on understanding the dynamics of Earth’s climate through the analysis of observations and numerical simulations with both idealized and comprehensive climate models. His work includes studying recent and future sea ice changes, polar oceanography, and global climate change. Recently, he has worked on topics such as the assessment of the reversibility of Arctic sea ice loss, the time-dependence of atmospheric feedbacks, and regional climate predictability.

Research topics

• Environmental science
• Oceanography
• Geology
• Climatology
• Physics
• Computer Science
• Engineering
• Political Science
• Geography
• Meteorology
• Atmospheric sciences
• Remote sensing
• Library science
• Geophysics
• Astrobiology
• Economics
• Econometrics

Selected publications

• Basin-wide sea-surface observations reveal post-2000 emergence of AMOC weakening

  2026-04-21

  articleOpen access

  Despite its importance for climate, the Atlantic Meridional Overturning Circulation (AMOC) is not well constrained before in-situ RAPID array observations began in 2004. Extending the record further back in time requires proxies, the most widely used being a subpolar North Atlantic sea-surface temperature (SST) index, which implies substantial AMOC decline over the 20th century. In contrast, climate model simulations generally show a stable or strengthening AMOC prior to a decline only over recent decades. Here, we move beyond the subpolar SST index by training deep-learning models on a diverse ensemble of coupled climate model simulations, reconstructing AMOC from observed SST records since the late 19th century and sea-surface height (SSH) records since 1993. When applied to observations, the deep-learning reconstruction shows no statistically significant AMOC decline over the 20th century, consistent with compensating effects of greenhouse-gas and aerosol forcing. However, the reconstruction reveals sustained post-2000 AMOC weakening, consistent with in-situ observations and climate model simulations, to a state that is now weaker than at any point over the 20th century. These findings provide an observational constraint on historical AMOC variability that is consistent with climate model simulations, and place the post-2000 weakening in a longer historical context.

  PublisherDOI

• The Strength of Coupling to the Southern Ocean Modulates Tropical Eastern Pacific Variability and Forced Response

  Journal of Climate · 2026-02-04

  articleSenior author

  Abstract Despite rising global-mean temperatures, large parts of the Southern Ocean and tropical eastern Pacific Ocean have cooled during the satellite era. These regions may be linked by teleconnections, with Southern Ocean cooling contributing to tropical eastern Pacific cooling. We demonstrate that, on average, state-of-the-art Earth system models (ESMs) underestimate the magnitude of interaction between the Southern Ocean and tropical eastern Pacific Ocean. The strength of the teleconnection is shown to be mediated by the magnitude of the positive cloud–sea surface temperature (SST) feedback in the subtropical eastern Pacific Ocean and the strength of the wind–evaporation–SST (WES) feedback. We link excessive precipitation in the tropical Pacific south of the equator to the strength of the Southern Ocean–eastern Pacific teleconnection. This model bias, known as the double intertropical convergence zone (ITCZ), is shown to be related to erroneous convection south of the equator, weakened cross-equatorial trade winds, and unfavorable meteorological conditions for marine boundary layer subtropical clouds. We postulate there is a two-way interaction, in which a double-ITCZ occurs with weaker cloud–SST and WES feedbacks, which in turn impact local SSTs and amplify the double-ITCZ. Models with a stronger Southern Ocean to tropical Pacific teleconnection tend to exhibit more multidecadal variability in the Walker circulation, ITCZ, and west–east equatorial SST gradient, as well as greater delayed warming in the tropical eastern Pacific Ocean resulting from delayed Southern Ocean warming under greenhouse gas forcing. These results provide insight into why ESMs struggle to replicate observed tropical Pacific temperature trend patterns and point to ITCZ location as a key target for improvement in future model development. Significance Statement The key advancement of this study is to demonstrate that, on average, state-of-the-art Earth system models underestimate the magnitude of interaction between the Southern Ocean and tropical east Pacific. As a result, historical cooling in the Southern Ocean may explain a larger fraction of observed east Pacific cooling than previously appreciated. Initial evidence suggests unrealistic precipitation simulated by models in the southeast equatorial Pacific may result in a “blocking” of high latitude influence due to its impact on the magnitude of the cloud–SST feedback and response of easterly trade winds. These results improve our understanding of the processes controlling the Southern Ocean–eastern Pacific teleconnection and provide a guide for future model development and climate trend attribution.

  PublisherDOI

• The Double-ITCZ Bias Weakens the Southern Ocean-Tropical Pacific Teleconnection

  2026-05-19

  article
  PublisherDOI

• Paleoclimate pattern effects help constrain climate sensitivity and 21st-century warming

  Proceedings of the National Academy of Sciences · 2026-01-22

  articleOpen access

  Paleoclimates provide examples of past climate change that inform estimates of modern warming from greenhouse-gas emissions, known as Earth’s climate sensitivity. However, differences between past and present climate change must be accounted for when inferring climate sensitivity from paleoclimate evidence. The closest paleoclimate analog to near-term warming from greenhouse-gas emissions is the Pliocene (5.3 to 2.6 Ma), a warm epoch with atmospheric CO 2 concentrations similar to today. Recent reconstructions indicate the Pliocene was 1 °C warmer than previously thought, implying higher climate sensitivity, which is also supported by recent reconstructions showing more cooling with reduced CO 2 at the Last Glacial Maximum (LGM; 19 to 23 thousand years ago). However, large-scale patterns of paleoclimate temperature change differ strongly from modern projections. Climate feedbacks and sensitivity depend on temperature patterns, and such “pattern effects” must be accounted for when using paleoclimates to constrain modern climate sensitivity. Here we combine data-assimilation reconstructions with atmospheric general circulation models to show Earth’s climate is more sensitive to Pliocene forcing than modern CO 2 forcing. Pliocene ice sheets, topography, and vegetation alter patterns of ocean warming and excite destabilizing cloud feedbacks, and LGM feedbacks are similarly amplified by the North American ice sheets. Accounting for paleoclimate pattern effects produces a best estimate (median) for modern climate sensitivity of 2.8 °C and 66% CI of 2.4 to 3.4 °C (90% CI: 2.1 to 4.0 °C), substantially reducing uncertainty in projections of 21st-century warming.

  PublisherOA PDFDOI

• Paleoclimate Pattern Effects in the Pliocene and Last Glacial Maximum Help Constrain Modern Climate Sensitivity

  2026-03-14

  articleOpen access

  Paleoclimates provide examples of past climate change that inform estimates of modern warming from greenhouse-gas emissions, known as Earth's climate sensitivity. However, differences between past and present climate change must be accounted for when inferring climate sensitivity from paleoclimate evidence. The closest paleoclimate analog to near-term warming from greenhouse-gas emissions is the Pliocene (5.3-2.6 Ma), a warm epoch with atmospheric CO2 concentrations similar to today. Recent reconstructions indicate the Pliocene was 1°C warmer than previously thought, implying higher climate sensitivity, which is also supported by recent reconstructions showing more cooling with reduced CO2 at the Last Glacial Maximum (LGM; 19-23 thousand years ago).However, large-scale patterns of paleoclimate temperature change differ strongly from modern projections under CO2 forcing. Climate feedbacks and sensitivity depend on temperature patterns, and such "pattern effects" must be accounted for when using paleoclimates to constrain modern climate sensitivity. Here we combine data-assimilation reconstructions with atmospheric general circulation models to show Earth's climate is more sensitive to Pliocene and LGM forcing than modern CO2 forcing. Pliocene ice sheets, topography, and vegetation alter patterns of ocean warming and excite destabilizing cloud feedbacks, and LGM feedbacks are similarly amplified by massive ice sheets. Accounting for paleoclimate pattern effects produces a best estimate (median) for modern climate sensitivity of 2.8°C and 66% confidence interval of 2.4-3.4°C (90% CI: 2.1-4.0°C), substantially revising climate sensitivity's upper bound and projections of 21st-century warming.

  PublisherDOI

• Impacts of Mean State Ocean Heat Transport on Climate and Its Response to CO ₂ Forcing

  Geophysical Research Letters · 2026-03-16

  articleOpen access

  Abstract Simulations of the slab ocean configuration of the coupled Energy Exascale Earth System Model (E3SM) were used to isolate the role of poleward ocean heat transport (OHT) in shaping the climate and its response to CO 2 forcing. Imposed changes to mean‐state OHT produce compensating changes in atmospheric heat transport (AHT) that are mediated by changes in surface evaporation. A reduction of maximum OHT by 0.56 PW (32%) reduces the global mean surface air temperature by 3.6°C. However, this cooler mean state exhibits 1.2°C more warming under CO 2 quadrupling, with the largest differences occurring at high latitudes. The amplified warming arises from stronger surface albedo and lapse rate feedbacks in polar regions and a shortwave cloud feedback in the southern midlatitudes. These results highlight the critical role of mean‐state OHT in modulating mean‐state climate, the partitioning between the OHT and AHT, and climate sensitivity.

  PublisherDOI

• Making sense of ZEC and TCRE: A conceptual model for the coupled climate response to carbon emissions. 

  2026-03-14

  articleOpen access

  Effective climate policy requires quantifying the temperature response to CO2 emissions. The current policy framework centers around Remaining Carbon Budgets, and depends heavily on there being a linear Transient Climate Response to Cumulative Emissions (TCRE) and a low Zero Emission Commitment (ZEC). The linearity of TCRE and the smallness of ZEC are based on emergent behaviors of a small number of Earth System Models (ESMs) and lack both conceptual understanding and uncertainty quantification. Here we present an analytically tractable conceptual model for the coupled interaction of the thermal component of the climate system with the carbon cycles. Unlike previous decompositions our model is built by assembling dynamical energy balance and carbon flux models. Thus, we obtain closed-form approximations for TCRE and ZEC in terms of well-established conceptual parameters such as the radiative feedback, ocean heat uptake efficiency, the average timescale ocean carbon uptake, the Q10 temperature sensitivity of respiration, etc. We derive conditions for both long-term (millennial-scale) low ZEC, as well as conditions for transient (centennial-scale) low ZEC, along with conditions for the near-linearity of TCRE. We find that there is no intrinsic physical reason for a low ZEC or a linear TCRE, and they arise from fortuitous compensations between unrelated parameters. We also show the system has the potential for significant centennial-scale transient amplification, arising from non-normal system dynamics.In addition to providing conceptual insight, the model allows us to easily explore the limits of the traditional assumptions surrounding TCRE and ZEC. For example, we show that a pattern effect derived from models with observed Sea Surface Temperature patterns (AMIP), can lead to a much larger ZEC than that derived from coupled ESMs.

  PublisherDOI

• Observational constraints imply limited future Atlantic meridional overturning circulation weakening

  Nature Geoscience · 2025-05-29 · 10 citations

  article
  PublisherDOI

• A novel framework to evaluate climate model emulators for global warming projections

  2025-08-07

  preprintOpen access

  Climate model emulators were extensively used in the IPCC’s Sixth Assessment Report due to their computational efficiency and consistency with key climate metrics. The emulators were calibrated to historical observations and used for future climate projections without systematic evaluation of their robustness. Here, we develop a framework to evaluate emulator performance against global climate model (GCM) large ensembles. This is demonstrated by constraining a two-layer energy balance model (EBM) to historical simulations of four GCMs and comparing their 21st century warming projections. The EBM matches projected warming in three of the four GCMs but exhibits substantial spread across ensemble members. It fails to reproduce the time-evolving global climate feedback in GCMs, with compensating biases between feedbacks and ocean heat uptake efficiency allowing seemingly accurate projections for incorrect reasons. Our results underscore the importance of evaluating the accuracy and physical realism of climate model emulators before using them for warming projections.

  PublisherOA PDFDOI

• Mid-Pliocene Climate Forcing, Sea Surface Temperature Patterns, and Implications for Modern-Day Climate Sensitivity

  Journal of Climate · 2025-05-06 · 2 citations

  articleOpen access

  Abstract Characterized by similar-to-today CO 2 levels (∼400 ppm) and surface temperatures approximately 3°–4°C warmer than the preindustrial, the mid-Pliocene warm period (mPWP) has often been used as an analog for modern CO 2 -driven climate change and as a constraint on the equilibrium climate sensitivity (ECS). However, model intercomparison studies suggest that non-CO 2 boundary conditions—such as changes in ice sheets—contribute substantially to the higher global mean temperatures and strongly shape the pattern of sea surface warming during the mPWP. Here, we employ a set of CESM2 simulations to quantify mPWP effective radiative forcings, study the role of ocean circulation changes in shaping the patterns of sea surface temperatures, and calculate radiative feedbacks during the mPWP. We find that the non-CO 2 boundary conditions of the mPWP, enhanced by changes in ocean circulation, contributed to larger high-latitude warming and less-stabilizing feedbacks relative to those induced by CO 2 alone. Accounting for differences in feedbacks between the mPWP and the modern (greenhouse gas–driven) climate provides stronger constraints on the high-end of modern-day ECS. However, a quantification of the forcing of non-CO 2 boundary condition changes combined with the distinct radiative feedbacks that they induce suggests that Earth system sensitivity may be higher than previously estimated. Significance Statement Climate records from past warm intervals in Earth’s geologic history are frequently used to constrain modern-day equilibrium climate sensitivity. Yet, climate warmth during these periods is not solely driven by the radiative effect of CO 2 , but also by environmental and geographic features, such as large-scale ice sheet and vegetation changes, that are not expected to occur in the near future. Using the mid-Pliocene warm period as an example, we find that the surface patterns of warming induced by non-CO 2 paleoenvironmental boundary conditions lead to a more-sensitive climate state than the modern. This implies that near-term warming under CO 2 forcing may be smaller than expected but that on geological time scales, future warming may be significantly larger.

  PublisherOA PDFDOI

Recent grants

• The role of oceans in climate asymmetries

  NSF · $359k · 2019–2023

• OCE-RIG: Identifying the role of ocean circulation in polar climate change

  NSF · $96k · 2015–2018

• Identifying Climate Model Biases in the Pattern of Ocean Warming and their Influence on Regional Climate Change

  NSF · $479k · 2022–2026

• CAREER: Understanding the Time- and State-Dependence of Climate Sensitivity

  NSF · $800k · 2018–2025

Frequent coauthors

• Gerard H. Roe

  University of Washington

  143 shared

• Aaron Donohoe

  University of Washington

  108 shared

• David S. Battisti

  University of Washington

  59 shared

• Nicholas Siler

  Oregon State University

  58 shared

• Ian Eisenman

  University of California, San Diego

  47 shared

• Cecilia M. Bitz

  University of Washington

  43 shared

• Cristian Proistosescu

  University of Illinois Urbana-Champaign

  41 shared

• Yue Dong

  Lamont-Doherty Earth Observatory

  38 shared

Labs

• Kyle Armour LaboratoryPI