Camille Li

Camille Li

Leader: Camille Li
Co-leader: Martin King


To apply mechanism-based approach to evaluate performance of CMIP6 models in simulating the observed tropical Pacific variability (and its teleconnections) and dynamics of heat and carbon in the ocean.


Despite increasing confidence in ESMs performance through reproducing many important aspects of the observed long climate-relevant time scale, the complex nature of the specific climatic events involves multiple key interactions between different Earth system components. For instance, the accurate projection of impacts (frequency and amplitude) related to tropical Pacific variability depends on correctly representing: the tropical variability itself (e.g., ENSO), the background climate, and adjustments due to wave-mean flow feedbacks. Evaluating how these features are represented in the current generation of models, and how large the model spread is compared to changes over the historical period, can help constrain which simulations to use for future projections. Hence, a more thorough assessment focusing on these key processes and feedback remain a great challenge for the climate modelling community.

Similarly, there is a lot of uncertainty as to why the ocean heat and carbon uptakes are spatially and temporally non-linear. Long-term anthropogenic climate change stems from the anthropogenic carbon emissions, which lead to increasing the Earth’s radiative budget. The ocean plays a critical role in buffering and redistributing both the excess heat and carbon in the atmosphere through the ocean circulation (Hansen et al., 1985). The latter is tightly coupled to climate modes such as the Atlantic Multidecadal Oscillation (AMO), and Southern Ocean Annual Mode (SAM), and Pacific Decadal Oscillation (PDO), among others. Here, we will evaluate the relationships between ocean heat/carbon fluxes to high latitude ventilation rate, which is modulated by climate modes mentioned above (e.g. Meijers, 2014). Comparing the consistency of these relationships across models as well as to observations is crucial to determine the reliability of long-term projections of global mean temperature, sea ice, sea level high, as well as regional climate change.

Description of work

Task 2.1. Characterize the representation of ENSO teleconnection in CMIP6. Evaluate the simulated ENSO and apply the numerical algorithm developed in Task 1.2 to determine the primary mode of tropical Pacific variability over the historical period (1850-2010) in all models, and evaluate them against the reanalysis data. Apply similar approach to atmospheric circulation fields and examine fidelity of models in reproducing the observed ENSO teleconnections depending on their ability to simulate the mean background flow (e.g., position and variability of jet stream) and temporal evolution of teleconnections from autumn to winter.

Task 2.2. ENSO impact on European climate. Building from Task 2.1, evaluate simulated ENSO impacts on North Atlantic-European climate by assessing the ability of models to represent two hypothesized mechanisms that seem to be activated during “Central Pacific” El Niño events in particular: the stratospheric pathway, and changes in synoptic processes that control the anchor point of storm tracks.

Task 2.3. Dynamics of ocean heat and carbon budget. Determine the primary physical mechanisms responsible for the temporal variations in the global oceanic heat and carbon uptakes under the multi-centennial equilibrium preindustrial simulations. Repeat for all models and for the transient preindustrial period. Elucidate sources of inconsistencies among the models. Applying the non-linear correlation approach in machine learning on CMIP6 models to discover previously unknown relationships or climate interactions affecting the regional air-sea heat and carbon fluxes.