A New Generation of Climate Models
<p>As we embark on the third year of <a href="https://m2lines.github.io/" rel="noopener ugc nofollow" target="_blank">M²LInES</a>, we want to share our progress and what comes next.</p>
<p><strong>M²LInES’ mission is to improve coupled climate models by reimagining physics model development through innovative use of data and AI.</strong> We aim to accelerate the pace of climate model development by learning physics from data with scientific machine learning, and ultimately enhance climate model fidelity and reliability for future projections.</p>
<p> As we continue to develop and generalize AI-enhanced models of ocean, sea-ice, and atmospheric processes from data, we can now begin to assess their impact on the large-scale climate in a suite of global model configurations.</p>
<p>Climate models are known to have stubborn biases <em>(model error relative to observations)</em> due to incorrect representations of unresolved physics. We can now demonstrate in <strong>GFDL </strong><a href="https://www.gfdl.noaa.gov/om4-0/" rel="noopener ugc nofollow" target="_blank"><strong>OM4</strong></a>, <strong>Global Ocean and Sea Ice Model at 1/4 degree horizontal resolution:</strong></p>
<ul>
<li> A reduction in upper ocean temperature biases in the summer, through the enhancement of a physics-based parameterization with a data-driven vertical diffusivity profile across many forcing regimes (<a href="https://arxiv.org/abs/2306.09045" rel="noopener ugc nofollow" target="_blank">Sane et al., 2023</a>);</li>
</ul>
<p><a href="https://medium.com/@lz1955/a-new-generation-of-climate-models-aefd851d47bd"><strong>Learn More</strong></a></p>