The Interaction between Climate Forcing and Feedbacks From the global scale to the process level

Transcription

1 The Interaction between Climate Forcing and Feedbacks From the global scale to the process level A. Gettelman (NCAR), L. Lin (U. Lanzhou), B. Medeiros, J. Olson (NCAR)

2 The interaction of Forcing & Feedbacks DT s Known l T = DR/DT s If know DR (forcing) then l T (=1/climate sensitivity) can be estimated Processes: Aerosols may play a large role in forcing. Role in Feedbacks uncertain.

3 IPCC, 2013, SPM.5 Climate Forcing

4 Forcing uncertainty Dashed: Uncertainty in Ocean Heat content Kiehl 2007: TAR Models F DT 2x = g = 1/l T Total forcing is different between models DT obs = F/g Different F and g can yield same DT obs DT obs is known. Better constraints on F could constrain g

5 Climate Models agree for the 20 th Century Models Multi-Model Mean Observations IPCC 2013, Ch9 (Hartmann et al 2013), Fig 9.8

6 Current Climate Models Models that reproduce 20 th Century Forster et al 2013, Figure 7

7 Process Level Forcing - Feedback Interaction Aerosols are affected by climate state E.g.: Sea salt emissions depend on wind speed So, climate change causes Aerosol Feedbacks Sea salt changes affect cloud forcing = Feedback Aerosol Feedbacks contribute to spread in diagnosed cloud feedbacks! Aerosols may directly alter cloud feedback Gettelman, A., L. Lin, B. Medeiros, and J. Olson. Climate Feedback Variance and the Interaction of Aerosol Forcing and Feedbacks. Journal of Climate 29, (2016): doi: /jcli-d

8 Models/Methods Test with 3 ensembles of CESM1-CAM5 (coupled) 30 members with high forcing (RCP 8.5) Large Ensemble = LE 15 members with medium forcing (RCP4.5) Medium Ensemble =ME 15 members with high forcing (RCP8.5,2000 aerosols) Fixed aerosol Ensemble = FixA Feedbacks Kernel adjusted cloud feedback (Soden et al., 2008) Difference 20 years of control with (LE,ME) Difference with (FixA) Years do not affect results Number of ensembles affects variance (spread)

9 Aerosol Feedback Change in AOD S. Ocean DAOD: not forced Tpcs & NH: forced LE: larger climate changes Sea salt and dust emissions depend on climate Aerosol lifetimes depend on climate (humidity, clouds) Affects CRE, but not a forcing (a feedback) Can this be factored out? AOD Change (unitless) control = PI (1850)

10 First: Correct for aerosol changes Aerosol Kernel ( R/ AOD) to remove effects of changing aerosols on clouds: otherwise diagnosed as a feedback, when actually a forcing Diagnose kernel with fixed SST simulations in CAM5 Annual SW All Sky Aerosol Kernel ( R/ AOD) Solid Line: Threshold of AOD > 0.01

11 Aerosol Kernel All-sky SW (indirect) dominates (-200Wm -2 /AOD) Clear sky (direct) is -20 to -40 Wm -2 /AOD Note: Scales differ!

12 Cloud Feedback Total cloud feedbacks Includes differences in Aerosol Now Apply Aerosol Kernel Note that variability across the ensemble is SMALL. Applies to other feedbacks too.

13 Cloud Feedback Adjust Feedbacks for Aerosol Feedbacks more positive (aerosol cooling from +DAOD misdiagnosed as feedback) NH Mid-latitude feedback differences smaller Tropical & Subtropical differences remain

14 Aerosol Effects on Clouds Adjust Feedbacks for Aerosol Tropical Changes: FixA LE or ME Difference in feedbacks due to different cloud state (different LWP, drop #) = 20% of cloud feedbacks S. Ocean difference is an aerosol feedback (LE larger AOD change, ~20%) Total difference between LE and FixA cloud feedbacks is 50% of cloud feedback!

15 Conclusions Feedback internal variability is small Uncertainty in forcing limits our ability to understand feedbacks Aerosol Feedbacks aliased into cloud feedbacks May need to remove aerosol feedback Different aerosol loading gives different cloud feedback Combined effect is 50% of cloud feedback in CESM ~0.5K of climate sensitivity in CESM Need to factor in aerosols when comparing models! Aerosol Kernel is an effective way to remove effect Kernel is probably model specific ( R x / AOD) Is there a physical (not social) relationship behind higher sensitivity = higher historical forcing? Do aerosols change how clouds are described?

16 Thanks! Questions, Comments or Feedback? Google Gettelman Papers Gettelman, A., L. Lin, B. Medeiros, and J. Olson. Climate Feedback Variance and the Interaction of Aerosol Forcing and Feedbacks. Journal of Climate 29, (2016): doi: /jcli-d

17 Water Vapor Feedback Variance Gettelman et al 2016, J. Climate

18 Albedo Feedback Variance Gettelman et al 2016, J. Climate

19 Cloud Feedback Variance Gettelman et al 2016, J. Climate