New Insights Into Atmospheric Predictability From Global High-resolution Non-hydrostatic Simulations And The New-generation Geostationary Satellite Observations

Abstract:

Global convection-permitting models can produce realistic simulations of the Earth’s atmosphere and are often expected to transform the weather prediction enterprise. However, this potential is dependent on the intrinsic predictability of the atmosphere, which was explored using the currently operational Finite-Volume-Cubed-Sphere (FV3) with Global Forecast System (GFS) physics (fvGFS) modeling system and high-spatiotemporal brightness temperature (BT) observations from the geostationary weather satellite, GOES-16. The control simulation was produced on a quasi-uniform 3-km mesh, which allowed the explicit representation of convection. The FV3 kinetic energy (KE) spectra compare well with previous observations and possess a well-resolved mesoscale (<500 km) structure. The divergent component of KE was found to be dominant in the mesoscale at both the troposphere and stratosphere, indicating that it is possible for deep convection to play an important role in mesoscale predictability. The dynamical process behind the thermal energy distribution was first explored through BT and its spectra from FV3 simulations and GOES-16. BT spectra possess a canonical structure of -5/3 slope in the mesoscale and synoptic scale. At the mesoscale, the BT spectrum is likely associated with convection-related systems including thunderstorms and short gravity waves. The spectrum in the synoptic scale is likely related to the convectively coupled equatorial waves in the tropics and baroclinic systems in the mid-latitudes. To investigate the role of convection behind the mesoscale predictability, maps of BT are separated into convective and stratiform regions. The results indicate deep active convection in convective regions rather than stratiform regions is an important source of thermal energy in the mesoscale. The ability of a model to produce the realistic deep convection regions, therefore impact the energy spectra, has a significant impact on the predictability of the model.