Deeply learning global temperature patterns

Another very high quality blogpost from John Kennedy with his usual mix of insight and wit.

This one struck me as especially interesting as I’m also starting to investigate deep learning for regional climate and surface mass balance models. Lots of bear traps for the unwary clearly, but also genuine promise.

Read on and of course, follow!

A preprint has appeared on infilling of global temperature data using “deep learning”. On their tests, it performs better than the Kadow et al. method. That’s quite interesting and new methods for filling the gaps in HadCRUT are always great to see. What’s more exciting, potentially, is that they used the same method to infill […]

Deeply learning global temperature patterns