Extending the Dynamical Core Test Case Hierarchy: Moist Baroclinic Waves with Topography

Christiane Jablonowski & Owen Hughes

University of Michigan, Ann Arbor, USA�

CESM Atmospheric Model Working Group (AMWG) Winter Meeting

February 9^th, 2021

Overarching Questions

• How well is topographic forcing simulated in dynamical cores?
• What is the impact of moisture on the topographically-triggered waves?
• How does the shape and peak height of the topography impact the flow field?
• Does the impact of the topography differ in different dynamical cores?
• What can we learn about the choice of the (topography-following) vertical coordinate and the physics-dynamics coupling strategy?

​

1

Answer some of these questions with the help of a GCM model hierarchy

A comprehensive simulation-to-science infrastructure that tackles the needs of next-generation, high-resolution, data intensive climate modeling activities.

CESM Simpler-Models Hierarchy

Isolated Dynamics: Deterministic dry dynamical core tests

Isolated Physics: Single Column Modeling

Deterministic moist �dynamical core tests

Dry dynamical core (climate)

Models with simplified physics (climate)

Radiative Convective Equilibrium (RCE) Models

Full-physics Aqua Planet Models

Atmosphere models with prescribed ocean/ice data (AMIP, CAPT)

Coupled Earth System Models

https://www.cesm.ucar.edu/models/simpler-models/

DCMIP

Almost always in �simpler model hierarchy:�no topography

Idealized World

Real �World

Deterministic dry/moist dynamical core tests with idealized mountains

this research project

Design of the Test Case: �Inspired by Atmospheric Rivers

Vertically-integrated�water vapor transport (IVT)�Jan/28/2021

Precipitable water (IWV)�Jan/28/2021

• Land-falling atmospheric river in California on Jan/28/2021
• (Tropical) moisture gets squeezed out by mountain range upon �landfall of baroclinic wave, long & narrow moisture band,�presence of low-level jet

​

Dynamical Cores and Configuration

Models

• CESM 2.1.3 / CESM 2.2:
• Spectral Element SEne60L30 (≈ 50 km)
• Finite Volume FV05L30 (0.47˚ x 0.63˚ grid, ≈ 50 km x 65 km)
• Finite Volume Cubed Sphere FV3C192L30 (≈ 50 km), new in CESM 2.2
• Standalone repository:
• Model for Prediction Across Scales MPAS (60 km L30)�

Configuration

• FKESSLER compset: Kessler warm-rain physics (precipitation only) in CESM
• Analytic moist baroclinic wave initial condition (DCMIP-2016, dry test described in �Ullrich et al., 2014), added topography, initial zonal wind perturbation removed

Initial Conditions

• Inspired by Staniforth and White (ASL, 2011) & DCMIP-2016 (Ullrich, Melvin, Jablonowski, Staniforth (QJ, 2014))

• Well-balanced moist initial �conditions (baroclinic wave)
• Ridge mountains, 2 km peaks

u

q

T

p_s

z_s

Height (km)

Latitude

Characteristics of the Test Case

Snapshots of the CESM2.2 SE ne60L30 (50 km)�dynamical core with ∆t_phys = 900 s, rsplit = 3, �nsplit = 2, qsplit = 1, ftype = 2 (hybrid)

• Well-balanced moist initial �conditions (baroclinic wave),�analytically prescribed
• 10-day simulation reveals �flow pattern
• Mountains serve as initial �perturbations and provide�continuous forcing

Application Examples: Moist versus Dry

• Moisture processes (warm-rain Kessler physics) intensify the evolution of the baroclinic wave
• At day 5, the minimum sea level pressures are 941 hPa (moist) and 966 hPa (dry)

Application Examples: Physics-Dynamics Coupling

CESM 2.1.3 �SEne60L30 (50 km):�∆t_phys = 900 s, rsplit = 3, �nsplit = 2, qsplit = 1

• Test case reveals�impact of SE’s�physics-dynamics �coupling strategy�(ftype)
• hybrid: sudden �adjustments of tracers�like specific humidity,�dribbled otherwise

colors saturate to highlight small scales

• Numerical noise in SE: Consequence of the long physics time step with subcyled dynamics (here with ∆t_phys = 900 s, rsplit = 3, nsplit = 2, qsplit = 1, ftype=2)
• Using the same short physics and dynamics time step of ∆t_phys = ∆t_dyn = 150 s �eliminates the numerical noise in SE
• Likely: increasing the strength of the horizontal diffusion / divergence damping will �also eliminate the noise (small-scale gravity wave oscillations)

​

colors saturate

Application Examples: Physics-Dynamics Coupling

CESM 2.2 SEne60L30

Application Examples: Dycore Intercomparisons

Resolutions: ≈ 50 km L30

CESM 2.1.3

CESM 2.1.3

CESM 2.2

MPAS (standalone)

Time series: Minimum sea level pressure

• Overall: sea level pressure patterns are similar
• But: considerable differences in the intensification (by day 5)

Application Examples: Dycore Intercomparisons

• Mountain test case reveals differences in the rain response in the dynamical cores
• Comparisons between leading rain band (no mountain interference) and middle rain band (hitting the mountain) are insightful
• Evolution of frontal zones with sharp (vertically integrated) precipitable water signatures that have similarities to flows in atmospheric rivers�

Application Examples: Dycore Intercomparisons

• Height-longitude cross sections of the cloud liquid water mixing ratio at 45N (at day 5)�

Application Examples: Dycore Intercomparisons

• Height-longitude cross sections of the vertical pressure velocity �at 45N (at day 5)�

Application Examples: Dycore Intercomparisons

• Local spectral method in SE shows signatures of spectral ringing�(numerical noise)
• Noise not obvious in other dycores

colors saturate to highlight small scales

Application Examples: Dycore Intercomparisons

• Initially: all dycores have �signatures of global �high-speed gravity waves�triggered by slight �initial imbalance
• In SE: global gravity waves �are persistent (little �damping), have high�amplitudes and are still �present by day 5-10
• FV3 shows signs of the �cubed-sphere grid, grids �not obvious in SE and MPAS

colors saturate to highlight small scales

Summary & Future Work

• Test case with focus on topography: Additional element in the simpler-model hierarchy
• Helps answer many fundamental dynamics questions:
• moist versus dry dynamics
• impact of mountain shape, size and peak heights on clouds, rain and flow field
• Topographic gravity wave studies
• Sheds light on numerical designs of dynamical cores and their physics interplay
• Physics-dynamics coupling
• Hydrostatic versus nonhydrostatic designs
• Diffusion
• Simulation of clouds and rain (placement, rain amount, shape of rain bands, etc.)
• Two publications in development: (1) Characteristics of the test case, �(2) fundamental dynamical behavior

​

16