Jon Shonk – Research

Model Drifts in the Tropics

Now that climate models are being used for seasonal and decadal forecasting, it is more important than ever that we deal with the systematic biases in the models that can damage their ability to produce a reliable forecast. By analysing how different model fields drift into error in both fully operational coupled hindcasts and atmosphere-only hindcasts, I have been working to understand the chain of causality of errors that ultimately should point towards the root cause of the error. I have focussed so far on the tendency of the Intertropical Convergence Zone in the western Pacific to drift northwards, and am now investigating the development of errors in the Atlantic. For more information, see paper 7 under "Publications".

Horizontal and Vertical Cloud Structure

Clouds are an important part of the radiation budget of the Earth. However, they remain a great source of uncertainty, posing many challenges to climate science on account of their immensely complex structure and the complicated process that govern their creation, evolution and dissipation. One problem is the physical representation of cloud structure. Traditionally, models use homogeneous clouds (in other words, lacking in any horizontal structure within a model gridbox layer) aligned vertically using maximum-random overlap. In this project, I developed improved representations of both horizontal and vertical cloud structure and implemented them both in an offline version of the Met Office's Edwards–Slingo radiation code and a version within their Unified Model. For more information, see papers 1, 2, 3 and 4 under "Publications".

Three-Dimensional Radiative Transfer

A second problem is the lack of radiation passing through cloud sides in a radiation calculation. In a standard calculation, radiation is only allowed to travel vertically within regions of cloud and clear sky. This work focused on developing a method of accounting for three-dimensional radiative transfer, which was achieved by modifying the Two-Stream equations at a fundamental level. Extra terms were included that account for the lateral transport of radiation between clear sky and cloud in a model gridbox layer. See paper 5 for more details on the method. This 3D Two-Stream scheme is currently being implemented in the Edwards–Slingo radiation code.

Retrievals of Mixed-Phase Cloud

To verify the performance of our cloud representations in climate models, we need high-quality cloud observations. In this short project, I worked with a cloud retrieval algorithm that infers cloud optical depth from surface measurements of radiance at two wavelengths. The algorithm makes the assumption that all retrieved optical depth is due to the presence of liquid cloud. I investigated the effect of this assumption on the accuracy of the retrieval.

Things to do now



[an error occurred while processing this directive]