24
Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike Ahlgrimm, Richard Forbes, Irina Sandu, Peter Bechtold

Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

  • Upload
    others

  • View
    15

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 1

ECMWF 20128

ECMWF WG meeting

Evaluating the ECMWF model's clouds and radiation with ARM observations

Maike Ahlgrimm, Richard Forbes, Irina Sandu, Peter Bechtold

Page 2: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 2

ECMWF 20128

ECMWF WG meeting

Radiation and precipitation are two big reasons why we care about clouds in models

• Evaluation products for radiation (especially TOA fluxes) and precipitation are readily available and pretty well established

• Invariably, the model will fall short in some area

Page 3: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 3

ECMWF 20128

ECMWF WG meeting

Challenge: Link model errors to specific aspect of model that needs to be improved

• Under what conditions does error occur?

• Can the error be linked to a particular parameterization or aspect of model?

• Compensating errors - need to identify, then address jointly

• “Right result for the right reason”

Page 4: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 4

ECMWF 20128

ECMWF WG meeting

Ground-based observations well suited to establish link with parameterized process

• Provides vertically resolved cloud macrophysical and microphysical properties in conjunction with radiativeobservations

• Model parameterization based on (incomplete) understanding of processes, few idealized LES cases

Page 5: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 5

ECMWF 20128

ECMWF WG meeting

Example: Identify bias in TOA net SW radiation

• Cloud forcing underestimated in Sc regions, southern ocean, North American continent (ARM SGP site)

• Cloud forcing underestimated in trades

Page 6: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 6

ECMWF 20128

• ECMWF WG meeting

Can same biases be found in ground based observations?

• About 50Wm-2 SW bias at noon

• Which clouds/ situations/ conditions contribute to the radiation bias?

Yes!

Page 7: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 7

ECMWF 20128

A priori guess: fair weather cumulus clouds?

ECMWF WG meeting

Cloud forcing spot-on but cloud fraction low-> identified compensating errors, but not the cause of SW bias

Composite of 146 days with fair weather cumulus

Page 8: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 8

ECMWF 20128

ECMWF WG meeting

Can we identify a cloud type that systematically contributes to the SW bias?

Instead of starting with a priori guess of cloud type, be guided by SW bias.

• Classify cloud layers based on cloud base and thickness

• Sort sample pairs (consisting of one hourly sample each from obs and model) into categories based on cloud type combinations

• Rank cloud type combinations by how much they contribute to the SW bias (using cumulative SW bias of each combination as measure)

Page 9: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 9

ECMWF 20128

ECMWF WG meeting

Use radiation bias to identify regimes of interest. Subset: observed and modeled low clouds

lack of cloud occurrence/fraction

clouds not reflectiveenough

broken cloudstoo reflective

Too much SWreaching surface

Not enough SWreaching surface

Page 10: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 10

ECMWF 20128

ECMWF WG meeting

Do conclusions apply at other locations?

Joint PDFs of modeled and observed total cloud cover from Graciosa

Model rarely has fractionsbetween 50-90%

Surface irradiance downward longwaveSample number

Even for correctly forecast cloud fraction, <80% CF cloudstoo optically thick, >80%CF too optically thin

Page 11: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 11

ECMWF 20128

ECMWF WG meeting

Create link to model’s parameterization

• Which model routines contribute to the generation of the clouds?

• Is the scheme intended to deal with regime producing cloud? (Triggering)

• If the intended scheme is active, is it producing the clouds as observed? (No, or we wouldn’t have a problem!)

• Can we find measurements to constrain parameterized processes in parameterization?

Page 12: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 12

ECMWF 20128

Overview of BL/low cloud parameterizations (EDMF scheme)

ECMWF WG meeting

Surface buoyancy

Stable BL Convective BL

Test parcel ascent

Dry convective BL Moist convective BL

stratocumulus Shallow convection

negative positive

No LCL found LCL found

Stable lower troposphere Stability criterion not met

(independent)

Page 13: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 13

ECMWF 20128

SCM stratocumulus study on triggering: Which parameterization is active?

pre-SAC meeting 2012

“dry” BL, no cloud base found,Shallow convection active

Low LWP

“borderline” stratocumulus case

LWPRHCloud FractionCloud breaks up

Page 14: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 14

ECMWF 20128

Parameterization trigger: SCM experiment

pre-SAC meeting 2012

Lower entrainment in test parcel

Parcel rises higher, finds cloud base

Stratocumulus parameterization active

LWPRHCloud Fraction

higher cloud fraction and LWP

Page 15: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 15

ECMWF 20128

Control New parcel

Stable dry stratocumulus decoupled

Impact of trigger experimentation on BL type

Test parcel reaches cloud base more frequently, stratocumulus and decoupled BL more common.

Page 16: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 16

ECMWF 20128

Impact of trigger experimentation on TOA SW radiation

pre-SAC meeting 2012

(Apologies, older CERES version)

Improved TOA radiation for stratocumulus!

Page 17: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 17

ECMWF 20128

ECMWF WG meeting

Cloud microphysics: water path and radiative properties

High LWP too frequent

Low LWP too frequentModel overestimates Reff

ARM SGP

Page 18: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 18

ECMWF 20128

ECMWF WG meeting

Summary• Example of a strategy to link model error directly to

parameterization

• Stratify observations by meteorology/regime/type that is relevant to model error and parameterization/scheme

• Identify compensating errors

• Address all aspects at the same time, else lack of compensation leads to worse results

• Ground-based obs from multiple instruments may provide statistics of quantities (or their distributions) parameterized in GCM based on few LES cases – over long time period and many “real life” conditions

Page 19: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 19

ECMWF 20128

Other observational products potentially useful to constrain model parameterizations

ECMWF WG meeting

Page 20: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 20

ECMWF 20128

Doppler Radar – mass flux, higher order moments

ECMWF WG meeting

BL depth normalized profiles of hourly averaged (a) reflectivity, (b) vertical velocity, (c) fraction, and (d) mass flux for all, core, and vertically coherent updraft samples. Ghate et al. 2011

Example of MMCR recorded Doppler spectral moments. (top) Reflectivity, (middle) Doppler velocity, and (bottom) spectrum width as observed on 25 March 2005. Also shown are the determined cloud boundaries

Page 21: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 21

ECMWF 20128

Vertical motion in subcloud layer: VV statistics, plume dimensions

ECMWF WG meeting

Example of time–height mapping of (a) MMCR reflectivity factor during a cumulus-topped event on 22 Jul 2006. Red dots indicate the cloud bases measured from a ceilometer. Black lines indicate the objectively defined hourly ILH. (b) MMCR Doppler velocity for the period 1200–1400 LST. (c) MMCR reflectivity for the period 1200–1400 LST.

Reflectivity factor

Chandra et al. 2010

Page 22: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 22

ECMWF 20128

Drizzle retrievals

ECMWF WG meeting

O’Connor et al. 2005

Page 23: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 23

ECMWF 20128

High-resolution water vapour retrievals

ECMWF WG meeting

http://www.arm.gov/news/facility/post/11211

Question whether variability in time translates into variability in space (Johannes).

Page 24: Evaluating the ECMWF model's clouds and radiation with ARM ... · Slide 1 ECMWF 2012 8 ECMWF WG meeting Evaluating the ECMWF model's clouds and radiation with ARM observations Maike

Slide 24

ECMWF 20128

References

Ghate VP, BA Albrecht, and P Kollias. 2010. "Vertical velocity structure of nonprecipitating continental boundary

layer stratocumulus clouds." Journal of Geophysical Research – Atmospheres, 115,

doi:10.1029/2009JD013091.

Ghate VP, M Miller, and L DiPretore. 2011. "Vertical velocity structure of marine boundary layer trade wind

cumulus clouds." Journal of Geophysical Research – Atmospheres, 116, D16206, doi:10.1029/2010JD015344.

Chandra, Arunchandra S., Pavlos Kollias, Scott E. Giangrande, Stephen A. Klein, 2010: Long-Term Observations of

the Convective Boundary Layer Using Insect Radar Returns at the SGP ARM Climate Research Facility. J.

Climate, 23, 5699–5714.

O’Connor, Ewan J., Robin J. Hogan, Anthony J. Illingworth, 2005: Retrieving Stratocumulus Drizzle Parameters

Using Doppler Radar and Lidar. J. Appl. Meteor., 44, 14–27.

ECMWF WG meeting