Atmospheric Modeling I: Physics in the Community Atmosphere Model (CAM)

Size: px

Start display at page:

Download "Atmospheric Modeling I: Physics in the Community Atmosphere Model (CAM)"

Dwight Murphy
5 years ago
Views:

1 Atmospheric Modeling I: Physics in the Community Atmosphere Model (CAM) CESM Tutorial Rich Neale, NCAR August 8, 2016 NCAR is sponsored by the National Science Foundation

$types), cloud frackon and microphysics RadiaKon Boundary layers, surface fluxes and gravity waves Process interac2ons Model complexity, sensi2vity and climate feedbacks$

2 Outline Physical processes in an atmosphere GCM DisKnguishing GCMs from other models (scales) Concept of Parameteriza2on Physics representakons (CESM) Clouds (different types), cloud frackon and microphysics RadiaKon Boundary layers, surface fluxes and gravity waves Process interac2ons Model complexity, sensi2vity and climate feedbacks

3 Community Atmosphere Model (CAM) Atmosphere Model Working Group (AMWG) BioGeo- Chemistry Polar Land Ice Chemistry- Climate Software Engineering Societal Dimensions Atmosphere Paleo- Climate Whole Atmosphere Land Ocean CESM Climate Variability and Change

4 1 week Scales of Atmospheric Processes Determines the formulakon of the model Resolved Scales Cloud Drops Microphysics 1 sec Climate Models Forecast models Future Climate Models Cloud/Mesoscale/Turbulence Models 1 m km

5 HydrostaKc PrimiKve EquaKons Where do we put the physics? Horizontal scales >> verkcal scales VerKcal accelerakon << gravity F V F T +transport F QV, F QL, F QI

6 What is a ParameterizaKon? Usually based on Basic physics (conservakon laws of thermodynamics) Empirical formulakons from observakons In many cases: no explicit formulakon based on first principles is possible at the level of detail desired. Why? Non-lineariKes & interackons at sub-grid scale Ocen coupled with observakonal uncertainty Insufficient informakon in the grid-scale parameters Vertical eddy transport of χ Unresolved sub-grid Diffusivity Resolved grid-scale

7 Community Atmosphere Model Representing the key atmospheric processes in CAM5 Park SWCF ZM Cloud Macrophysics and Microphysics Detr ainm Dynamics Deep Convection ent Aerosols MG DSD Radiation SW LW A LWCF RH CAPE M1 M2 M3 Shallow Convection TKE CIN UW + Abs/Emis/Ref UW Wind P - + Gravity Wave Drag Wind Planetary Boundary Layer Emission Albedo σt4s Deposition Spectral Element Surface Fluxes MO stabilityty FUV, FLH, FSH Turbulent Mountain Stress

8 Clouds

$Clouds MulKple Categories Stra2form (large-scale) clouds Responds to large-scale saturakon frackon, RH (parameterized) Coupled to presence of condensate (microphysics, adveckon) Shallow convec2on$

9 Clouds MulKple Categories Stra2form (large-scale) clouds Responds to large-scale saturakon frackon, RH (parameterized) Coupled to presence of condensate (microphysics, adveckon) Shallow convec2on clouds Symmetric turbulence in lower troposphere Non precipitakng (mostly) Responds to surface forcing Deep convec2on clouds Asymmetric turbulence PenetraKng conveckon (surface -> tropopause) PrecipitaKng Responds to surface forcing and condikonal instability

10 StraKform Clouds (macrophysics) Sub-Grid Humidity and Clouds ü Liquid clouds form when RH = 100% (q=q sat ) ü But if there is variakon in RH in space, some clouds will form before mean RH = 100% ü RHcrit determines cloud frackon > 0; Value is lower over land due to higher humidity variance Fraction 1.0 Clear (RH < Rhcrit) 0. RHcrit Mean RH 50% 100% Relative Humidity Cloudy (RH = 100%) Assumed Cumulative Distribution function of Humidity in a grid box with sub-grid variation

11 Shallow and Deep ConvecKon ExploiKng conservakon properkes Common properkes Parameterize consequences of verkcal displacements of air parcels Unsaturated: Parcels follow a dry adiabat (conserve dry sta2c energy) Saturated: Parcels follow a moist adiabat (conserve moist sta2c energy) Shallow (10s-100s m) Parcels remain stable (buoyancy<0) Shallow cooling mainly Some latent heakng and precipitakon Generally a source of water vapor Small cloud radius large entrainment Deep (100s m-10s km) Parcels become unstable (buoyancy>0) Deep heakng Latent heakng and precipitakon Generally a sink of water vapor Large cloud radius small entrainment

(CAPE) CAM4 and CAM5 CAPE>CAPE trigger Timescale=1 hour ConvecKve inhibikon and

12 Shallow and Deep ConvecKon Closure: How much and when? Shallow Local condikonal instability CAM4 Deep ConvecKve Available PotenKal Energy (CAPE) CAM4 and CAM5 CAPE>CAPE trigger Timescale=1 hour ConvecKve inhibikon and turbulent kinekc energy (TKE) CAM5 Shallow and deep convection and stratiform cloud fractions combined for radiation

Cloud Microphysics Condensed phase water processes ProperKes of condensed species (=liquid, ice) size distribukons, shapes DistribuKon/transformaKon of condensed species PrecipitaKon, phase

13 Cloud Microphysics Condensed phase water processes ProperKes of condensed species (=liquid, ice) size distribukons, shapes DistribuKon/transformaKon of condensed species PrecipitaKon, phase conversion, sedimentakon Important for other processes: Aerosol scavenging RadiaKon In CAM = stra2form cloud microphysics ConvecKve microphysics simplified FormulaKons currently being implemented into conveckon

14 CAM5 Microphysics

15 RadiaKon The Earth s Energy Budget Trenberth & Fasullo, 2008 Thanks to: Bill Collins, Berkeley & LBL Gas SW AbsorpKon (Wm-2) CO2 1 O2 2 O3 14 H2O 43

16 Goals of GCM RadiaKon Codes Accurately represent the input and output of energy in the climate system and how it moves around Solar Energy Thermal Emission Gases Condensed species: Clouds & Aerosols

17 (TOA) 1000nm = 1µm Input at TOA, Radiation at surface From: Sunlight, Wikipedia

18 IR absorpkon 1000nm = 1µm

19 k-distribukon Band Models Sort Line-by-line calculations Very expensive/slow, accurate k-distribution band model, sort absorption coefficients by magnitude Cheaper/fast, less accurate

20 Planetary Boundary Layer (PBL) Regime dependent representakons Vital for near-surface environment (humidity, temperature, chemistry) Exploit thermodynamic conserva2on (liquid virtual potenkal temperature θ vl ) Conserved for rapidly well mixed PBL Not conserved for stable PBL CriKcal determinant is the presence of turbulence Richardson number <<1, flow becomes turbulent CAM4: Gradient Ri # + non-local transport (Holtslag and Boville, 1993) CAM5: TKE-based Moist turbulence (Park and Bretherton, 2009)

Gravity Waves and Mountain Stresses Sub-grid scale dynamical forcings Gravity Wave Drag Determines flow effect of upward propagakng (sub-grid scale) gravity waves that break and dump momentum

21 Gravity Waves and Mountain Stresses Sub-grid scale dynamical forcings Gravity Wave Drag Determines flow effect of upward propagakng (sub-grid scale) gravity waves that break and dump momentum Generated by surface orography (mountains) and deep conveckon Important for closing off jet cores in the upper troposphere (strat/mesosphere) Turbulent mountain stress Local near-surface stress on flow Roughness length < scales < grid-scale Impacts mid/high-laktude flow (CAM5) More difficult to parameterize than thermodynamic impacts (conservakon?)

22 Surface Exchange Surface fluxes (bulk formulakons) Stresses Specific Heat Latent heat (evaporation) Resistances r ax based on Monin-Obhukov similarity theory

23 ParameterizaKon InteracKons Direct and Indirect Process CommunicaKon Cloud Processes & RadiaKon Feedbacks Boundary Layer / Cumulus & Dynamics PrecipitaKon & Scavenging Chemical (gas phase) consktuents Aerosols (condensed phase consktuents) Microphysics and Aerosols Physics and surface components (ice, land, ocean) Resolved scales and unresolved scales

24 Clouds in GCMs State of the Art from CMIP3 Outgoing Long-wave Radiation (Annual, ) CMIP3 Models CMIP3 Models (~20 models)

25 Clouds in GCMs State of the Art from CMIP3 Total Cloud Fraction (Annual, ) CMIP3 Models (~20 models)

26 Structural Clouds in GCMs State of the Art from CMIP3 Liquid Water Path (Annual, ) CMIP3 Models (~20 models)

27 Future Clouds in GCMs State of the Art from CMIP3 response to climate change Total Cloud Fraction Change (Annual, SRES A1B: minus ) CMIP3 Models (~20 models)

28 Climate Sensitivity What happens to clouds when we double CO2? GFDL Model +4.2K NCAR Model +1.8K Change in low cloud amount (%) Significant range in low-cloud sensitivity (low and high end of models) Cloud regimens are largely oceanic stratocumulus (difficult to model) Implied temperatures change is due to (higher/lower) solar radiation reaching the ground because of clouds feedbacks.

29 Community Atmosphere Model Representing the key atmospheric processes in CAM5 Park SWCF ZM Cloud Macrophysics and Microphysics Detr ainm Dynamics Deep Convection ent Aerosols MG DSD Radiation SW LW A LWCF RH CAPE M1 M2 M3 Shallow Convection TKE CIN UW + Abs/Emis/Ref UW Wind P - + Gravity Wave Drag Wind Planetary Boundary Layer Emission Albedo σt4s Deposition Spectral Element Surface Fluxes MO stabilityty FUV, FLH, FSH Turbulent Mountain Stress

Community Atmosphere Model Representing the key atmospheric processes in CAM6 CLUBB SWCF ZM MG2 Microphysics Dynamics Deep Convection Detrainment M1 M2 M3 M4 LS DSD CAPE Aerosols Cu

30 Community Atmosphere Model Representing the key atmospheric processes in CAM6 CLUBB SWCF ZM MG2 Microphysics Dynamics Deep Convection Detrainment M1 M2 M3 M4 LS DSD CAPE Aerosols Cu Spectral Element Radiation SW LW Abs/Emis/Ref Gravity Wave Drag LWCF Wind Wind Emission Albedo σt4s Deposition P PBL SCu Surface Fluxes Turbulent Mountain Stress MO stabilityty FUV FLH, FSH

31 Zhang McFarlane (ZM) PBL o Unifies moist and dry turbulence (except deep conveckon) o Unifies microphysics o High order closures (1 third order, 9 second order) o Use two Gaussians to described the sub-grid mulkvariate PDF: P=P(w,q t,θ L )

32 Model physics: The future 1. How to operate in varying grid scale environments 2. Advanced representakon of processes. New and more complex processes Cloud super-parameterization High Resolution, Regional grid refinement and scale-aware physics

33 Summary GCMs physics=unresolved processes=parameteriza2on ParameterizaKon (CESM) = approxima2ng reality Starts from and maintains physical constraints Tries to represent effects of smaller sub-grid scales Fundamental constraints, mass & energy conserva2on Clouds are fiendishly hard: lots of scales, lots of phase changes, lots of variability Clouds are coupled to radia2on (also hard) = biggest uncertainkes (in future climate); largest dependencies CESM physics increasingly complex and comprehensive Future parameterizakons aim to be process scale-aware and model grid-scale independent

34 Questions?

Rich Neale, Peter Caldwell, Christiane Jablonowski and Cecile Hannay

Rich Neale, Peter Caldwell, Christiane Jablonowski and Cecile Hannay and many, many others! AMP/CGD National Center for Atmospheric Research Boulder, Colorado 1 Rich Neale 2010 Julio Bacmeister 2 AIM: