Because of limited high-quality satellite and in situ observations, less attention has been given to the trends in Arctic sea ice thickness and therefore sea ice volume than to the trends in sea ice extent. This study evaluates the spatial and temporal variability in Arctic sea ice thickness using the Pan-Arctic Ice Ocean Modeling and Assimilation System (PIOMAS). Additionally, the Community Earth System Model Large Ensemble Project (LENS) is used to quantify the forced response and internal variability in the model. A dipole spatial pattern of sea ice thickness variability is shown in both PIOMAS and LENS with opposite signs of polarity between the East Siberian Sea and near the Fram Strait. As future sea ice thins, this dipole structure of variability is reduced, and the largest interannual variability is found only along the northern Greenland coastline. Under a high-emissions scenario (RCP8.5) projection, average September sea ice thickness falls below 0.5 m by the end of the twenty-first century. However, a regional analysis shows internal variability contributes to an uncertainty of 10 to 20 years for the timing of the first September sea ice thickness less than 0.5 m in the marginal seas.
The effect of future Arctic amplification (AA) on the extratropical atmospheric circulation remains unclear in modeling studies. Using a collection of coordinated atmospheric and coupled global climate model perturbation experiments, we find an emergent relationship between the high‐latitude 1,000–500 hPa thickness response and an enhancement of the Siberian High in winter. This wave number‐1‐like sea level pressure anomaly pattern is linked to an equatorward shift of the eddy‐driven jet and a dynamical cooling response in eastern Asia. Additional simulations, where AA is imposed directly into the model domain by nudging, demonstrate how the sea ice forcing is insufficient by itself to capture the vertical extent of the warming and by extension the amplitude of the response in the Siberian High. This study demonstrates the importance of the vertical extent of the tropospheric warming over the polar cap in revealing the “warm Arctic, cold Siberia” anomaly pattern in future projections.
14 June 2021 in Journal of Advances in Modeling Earth Systems
It remains difficult to disentangle the relative influences of aerosols and greenhouse gases on regional surface temperature trends in the context of global climate change. To address this issue, we use a new collection of initial‐condition large ensembles from the Community Earth System Model version 1 that are prescribed with different combinations of industrial aerosol and greenhouse gas forcing. To compare the climate response to these external forcings, we adopt an artificial neural network (ANN) architecture from previous work that predicts the year by training on maps of near‐surface temperature. We then utilize layer‐wise relevance propagation (LRP) to visualize the regional temperature signals that are important for the ANN's prediction in each climate model experiment. To mask noise when extracting only the most robust climate patterns from LRP, we introduce a simple uncertainty metric that can be adopted to other explainable artificial intelligence (AI) problems. We find that the North Atlantic, Southern Ocean, and Southeast Asia are key regions of importance for the neural network to make its prediction, especially prior to the early‐21st century. Notably, we also find that the ANN predictions based on maps of observations correlate higher to the actual year after training on the large ensemble experiment with industrial aerosols held fixed to 1920 levels. This work illustrates the sensitivity of regional temperature signals to changes in aerosol forcing in historical simulations. By using explainable AI methods, we have the opportunity to improve our understanding of (non)linear combinations of anthropogenic forcings in state‐of‐the‐art global climate models.
1 January 2020 in Bulletin of the American Meteorological Society
Evaluating historical simulations from global climate models (GCMs) remains an important exercise for better understanding future projections of climate change and variability in rapidly warming regions, such as the Arctic. As an alternative approach for comparing climate models and observations, we set up a machine learning classification task using a shallow artificial neural network (ANN). Specifically, we train an ANN on maps of annual mean near‐surface temperature in the Arctic from a multi‐model large ensemble archive in order to classify which GCM produced each temperature map. After training our ANN on data from the large ensembles, we input annual mean maps of Arctic temperature from observational reanalysis and sort the prediction output according to increasing values of the ANN's confidence for each GCM class. To attempt to understand how the ANN is classifying each temperature map with a GCM, we leverage a feature attribution method from explainable artificial intelligence. By comparing composites from the attribution method for every GCM classification, we find that the ANN is learning regional temperature patterns in the Arctic that are unique to each GCM relative to the multi‐model mean ensemble. In agreement with recent studies, we show that ANNs can be useful tools for extracting regional climate signals in GCMs and observations.
Recent modeling studies have shown an important role for stratosphere‐troposphere coupling in the large‐scale atmospheric response to Arctic sea ice loss. Evidence is growing that the Quasi‐biennial Oscillation (QBO) can contribute to or even mitigate teleconnections from surface forcing. Here, the influence of QBO phase on the atmospheric response to projected Arctic sea ice loss is examined using an atmospheric general circulation model with a well‐resolved stratosphere and a QBO prescribed from observations. The role of the QBO is determined by compositing seasons with easterly phase (QBO‐E) separately from seasons with westerly phase (QBO‐W). In response to the sea ice forcing in early winter, the polar vortex during QBO‐E weakens with strong stratosphere‐troposphere wave‐1 coupling and a negative Northern Annular Mode‐type response. At the surface, this results in more severe Siberian cold spells. For QBO‐W, the polar vortex strengthens in response to the sea ice forcing.