Earth’s energy budget in the presence of internal climate variability *Yu Kosaka (RCAST, University of Tokyo) Shang-Ping Xie (Scripps Institution of Oceanography, UCSD) Yuko M Okumura (University of Texas at Austin) WCRP Grand Challenge Workshop: Earth’s Climate Sensitivities Schloss Ringberg, Germany Mar 23, 2015
13
Embed
Earth’s energy budget in the presence of internal climate ... › fileadmin › atmosphaere › ...Earth’s energy budget in the presence of internal climate variability *Yu Kosaka
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Earth’s energy budget in the presence of internal climate variability
*Yu Kosaka (RCAST, University of Tokyo)
Shang-Ping Xie (Scripps Institution of Oceanography, UCSD)
Yuko M Okumura (University of Texas at Austin)
WCRP Grand Challenge Workshop: Earth’s Climate Sensitivities Schloss Ringberg, Germany
Mar 23, 2015
The global-warming hiatus and internal variability
• GMST has been flat for the past ~17 yrs • Slowdown of radiative forcing increase
o Solar cycle, volcano, anthropogenic aerosol, stratospheric water vapor, …
• Internal climate variability o Tropical Pacific decadal variability
GMST anom (ref: 1970-1999)
Radiative forcing only (historical + RCP4.5)
HadCRUT4
Radiative forcing +
Tropical Pacific variability
The global-warming hiatus and internal variability
• GMST has been flat for the past ~17 yrs • Slowdown of radiative forcing increase
o Solar cycle, volcano, anthropogenic aerosol, stratospheric water vapor, …
• Internal climate variability o Tropical Pacific decadal variability
Radiative forcing only (historical + RCP4.5)
HadCRUT4
Radiative forcing +
Tropical Pacific variability
Mostly explains the recent hiatus!
GMST anom (ref: 1970-1999)
• If N = F – λΔT N increase would have accelerated during hiatus
• Observed OHC increase does not show such acceleration
Is the radiative forcing slowdown the key? … Need an energy equation considering internal variability!
• High correlation for each of shortwave and longwave radiation o Longwave radiation peaks at lag 0 → negative feedback o Shortwave radiation leads GMST → atmospheric stochastic forcing
• Due to a large cancellation, the correlation for net radiation is not high
(only ~ –0.4 at peak)
→ NI is not a major constraint on ΔTI
ΔTI (MME)
NI (MME)
Lag correlation with ΔTI Lag correlation with ΔTI
NI (each model)
net shortwave
–OLR
inco
min
g ou
tgoi
ng
Time scale dependence: interannual variability
• Interannual variability: 2-10 yr band-pass filtered • Again θ ~ 45º - 90º • Peak correlation (~ –0.65) is higher than in decadal variability
• Peak NI /ΔTI exceeds the (forced) feedback parameter λ (For internal variability, λI has a strong time-scale dependence)
→ Care needed in estimating λ from short observations
ΔTI (MME)
NI (MME) NI (MME)
–λ
Lag correlation with ΔTI Lag regression against ΔTI [W/m2/ºC]
NI (each model)
NI (each model)
MME mean±1σ
inco
min
g ou
tgoi
ng
Dynamical ocean vs slab ocean
• Large interannual variability in CCSM4 ← ENSO
• Comparable power of GMST variability for timescales > 30 yrs
o Ocean dynamics increases GMST variance by 5% for >30 yrs
→ Ocean dynamics (and therefore subsurface ocean heat uptake) is secondary for decadal hiatus and accelerated GW (≳15 yrs)
Power spectrum of ΔTI • CCSM4 vs • CAM4 with a slab-ocean model
(CAM4-SOM)
• A common AGCM (CAM4) • For SOM, “Q-flux” and mixed-layer
depth are based on CCSM4 climatology
• 500-year piControl runs
CCSM4
CAM4-SOM
Structure of decadal GMST variability
• Internal decadal GMST variability is structured to an IPO-like pattern with warming in the polar regions
• Tropical Pacific decadal variability arises from thermodynamic coupling (Clement et al. 2011, Okumura 2013)
→ Applicable to internal GMST variability
• GMST variability arises from atmospheric stochastic forcing and thermodynamic air-sea feedback
SAT (shading) & SLP (contours) anom regressed against ΔTI (10-yr low-pass filtered)
[ºC]
CCSM4 CAM4-SOM
Decadal GMST and TOA radiation variability
• Decadal variability (10-yr low-pass filtered)
• CAM4-SOM:
→ NI = 0 at lag 0, θ = 90º No ocean dynamics (and vertical heat redistribution), but still the correlation is low
• CCSM4: Dynamical ocean redistributes heat vertically, dynamically inducing ΔTI and perturbing net TOA radiation to damp ΔTI
→ NI < 0 at lag 0, θ = 45º - 90º (from a coherence analysis)
c dΔTIdt
≈ NI
ΔTI
NI
net shortwave
–OLR
ΔTI
NI
CCSM4 CAM4-SOM
c: global-mean heat capacity of the ocean mixed layer
Cor
rela
tion
GMST-TOA radiation variability in the two models
• NI power decreases with time scale, but ΔTI power increases → λI decreases as time scale gets longer
• NI is much weaker in CAM4-SOM despite comparable GMST variability → Again NI is not a major constraint on GMST
Power spectrum of ΔTI
CCSM4
CAM4-SOM CAM4-SOM
CCSM4
Power spectrum of NI
Energy budget in the presence of internal variability
• N is smaller than pure forced case (F – λΔTF) during hiatus • Global OHC increase slows down during hiatus (statistically) • Consistent with the OHC observations (despite the loose GMST-TOA rad relation)
hiatus
ΔTF
ΔT
ΔTI
F – λΔTF
hiatus t t
F – λΔTF + NI
NI
hiatus hiatus
OHC anom [×1022 J] dOHC/dt [W/m2]
Based on top 700m OHC
Assuming θ = 90º for simplicity
Conclusions
• GMST-TOA radiation relationship is distinct b/w forced change and internal variability (time lag b/w GMST and TOA radiation)
• During hiatus, net incoming energy decreases instead of accelerates (the traditional energy equation predicts the latter)
Observed OHC tendency is consistent with the internal variability hypothesis of the current hiatus
• Internal GMST variability is only loosely correlated with net TOA radiation
Net TOA radiation is not a major constraint on internal GMST variability
• Net TOA radiation perturbation change associated with internal GMST change strongly depends on time scale
Care needs to be taken in estimating (forced) feedback parameter from short observations