Investigation of diurnal cycle simulation at SGP in the GFDL single-column model and AM3

 

Authors

Yanluan Lin — Tsinghua University
Patrick Taylor — NASA - Langley Research Center

Category

Modeling

Description

The diurnal cycle is a principal mode of climate system variability spanning processes important for the determination of climate state, including convective and radiative processes. Representation of the diurnal cycle time scale in climate models is a long-standing weakness of climate simulation, especially over land. The cause for poor diurnal cycle representation is uncertain but is likely related to convective, surface turbulent flux, and boundary-layer parameterizations and possible biased correlations between these parameterizations and large-scale meteorology. The inclusion of a more inhibitive convective trigger is a popular speculated solution. This work investigates the GFDL Single-Column Model (SCM) diurnal cycle simulation fidelity using cloud, precipitation, thermodynamic profile, and surface radiative flux observations at the ARM SGP site in combination with NASA CERES Terra top-of-atmosphere (TOA) flux observations. ARM SGP 4D variational analysis was used to force three years of GFDL SCM simulation, corresponding to 1999 through 2001. In this work, mean diurnal cycle biases exhibited by the model are investigated in all seasons and grouped by large-scale forcing. Particular attention is paid to the SGP summer season where convection is most common and the diurnal cycle is most prominent. At SGP, summer season convection generally takes two forms that exhibit different diurnal cycles: (1) locally triggered convection and (2) convective systems propagating from the Rocky Mountains. The second convective regime will not be present in the SCM simulations. To investigate the second regime statistically, a full 20-year GFDL AM3 simulation is used.