Hindcast-Based Estimates of Recent Climate Trends

Abstract: In initialised seasonal and decadal prediction systems, retrospective forecasts (“hindcasts”) are routinely used to assess model skill on seasonal or interannual timescales. The frequent initialisation of the hindcasts reduces the development of biases relative to free-running simulations with historical forcing. Here, we apply a recent version of the Met Office’s coupled decadal prediction model to study multidecadal trends over the satellite era. Forty hindcast members are initialised twice annually since 1980 and each run for between 13-66 months. By studying trends between successive hindcast runs, a distribution is built of plausible alternative histories that are consistent with the observed climate state and observed phases of multidecadal variability, while minimising the impact of mean state biases.

A broad survey of circulation trends in the hindcasts is presented from a global perspective, with a focus on changes in temperature and zonal winds in the troposphere. We show that the initialisation largely brings the hindcast trends closer to those of ERA5, as compared to the equivalent free-running model. Nevertheless, although minimised, some persistent model biases remain in the hindcasts. Initialised hindcasts offer a bridge between observations and free-running climate models, and understanding their differences from each will build our understanding of both the observed and modelled climate.

Biography: I am a final year DPhil student in the Atmospheric, Oceanic and Planetary Physics subdepartment of the University of Oxford, supervised by Prof. Tim Woollings and co-supervised by Dr. Nick Dunstone at the Met Office. My research interests are broadly in trends in the large-scale circulation, from the tropics to the poles.