Global&hydrological&cycle&response& …kuang/LarissaBack.pdfGlobal&hydrological&cycle&response&...
Transcript of Global&hydrological&cycle&response& …kuang/LarissaBack.pdfGlobal&hydrological&cycle&response&...
Global hydrological cycle response to rapid and slow global warming
Larissa Back, Kuniaki Inoue, Karen Russ, Zhengyu Liu
*Simulations implemented by Feng He and Jiaxu Zhang
• Anthropogenic global warming causes “robust” changes in the global hydrological cycle
• “Robust” changes (in this context) means changes are directly related to global mean temperature change*
• QuesFon: Are hydrological cycle changes in the paleoclimate (last 22ky) similarly “robust?” – Answer: Only sort of. We’ll examine why this is.
Outline:
*e.g. Held & Soden 2006
“Robust” changes in the global hydrological cycle due to anthropogenic
global warming
Atmospheric radiative cooling constraints (surface energy
balance)
Held & Soden, 2006
“Robust” rainfall changes: wet get weTer and the dry get drier
S
Fig. 6b From Held & Soden (2006) Latitude
SRES A1B scenario
Zhengyu Liu (UW-Madison), B. Otto-Blienser (NCAR)
Realistic Forcing
Insolation GHGs Meltwater Icesheet
Model: NCAR-CCSM3 – fully coupled ocean-atmosphere GCM
Liu et al., 2009, Science
Greenland Temperature Obs (grey), Model (red)
Unexpected TRACE paleoclimate results: q (global water vapor) versus surface T
Surface Temperature (C)
“surface-CC scaling:” 6.7%
shows 4.2 % q increase per unit T change.
At apparent odds with: Consistent with
Boos 2012 results
Unexpected TRACE paleoclimate results: q (global water vapor) versus surface T
Surface Temperature (C)
“surface-CC scaling:” 6.7%
shows 4.2 % q increase per unit T change.
At apparent odds with:
Slope increases during anthropogenic warming era to 7.2 %
Possible explanaFons for 4.2% water vapor increase per unit K warming
• Rela%ve humidity is not constant over climate-‐change %me scales?
• Clausius-‐Clapeyron rela%onship non-‐lineari%es?
• Tropical upper troposphere warms more than surface at warmer temperatures?
• Rapid CO2-‐induced warming affects global water vapor differently than slow CO2-‐induced warming?
56%
57%
8 C
12 C Total water vapor in atmosphere
Total water vapor atmosphere would have if saturated
Blue shows:
Column relative humidity changes small
Rela%ve humidity is not constant over climate-‐change %me scales?
Try changing CO2 rapidly by running branch simulaFons where CO2 doubles instantaneously
Time (ka BP)
CO2 doubling
CO2 doubling
CO2 doubling
Anthro- pogenic
Try changing CO2 rapidly by running branch simulaFons where CO2 doubles instantaneously
CO2 doubling
CO2 doubling
CO2 doubling
Anthro- pogenic
‘s = “Rapid” warming
Time (ka BP)
Rapid CO2-‐induced warming affects global water vapor differently than slower warming
Surface Temperature (C)
“surface-CC scaling:” 6.7%
Rapid warming d(lnq)/dT greater than slow warming d(lnq)/dT
Slow CO2 increases symmetric warming
Rapid CO2 increases N.H. warms more
dominates response
response.
“Slow”
“Rapid”
Latitude
Figure shows amount of surface warming by latitude normalized for a global surface T increase of 1K
S. Ocean thermal inertia
Slow CO2 increases symmetric warming
Rapid CO2 increases N.H. warms more
dominates response
response.
“Slow”
“Rapid”
Latitude
Figure shows amount of surface warming by latitude normalized for a global surface T increase of 1K
Slow CO2 increases symmetric warming
Rapid CO2 increases N.H. warms more
dominates response
response.
“Slow”
“Rapid”
Latitude
More tropical warming per global T increase
Figure shows amount of warming by latitude normalized for a global T increase of 1K Figure shows amount of surface warming by latitude normalized for a global surface T increase of 1K
Consider a hypotheFcal 3-‐box model of warming paTerns
• Assume most water vapor is in tropics due to warmest temperatures there – Water vapor increases exponenFally with T
• Assume warming paTern in 3 equal-‐area boxes:
• “Slow” has less global d(lnq)/dT than “rapid” case
S. H. N. H. Tropics
Rapid
Slow
0
1 0 1
1 1
Real 3-‐box model numbers, where each box has equal area, support
interpretaFon • For 1 degree global temperature change:
S. H. N. H. Tropics
Rapid
Slow
0.75
1.4 0.42 1.2
0.71 1.5
Why does global water vapor increase more (% per K global T change) in response to rapid (anthropogenic-‐like) CO2 change?
• Most water vapor is in the tropics -‐-‐> amount of tropical warming strongly influences global water vapor change
• Tropics warm more (per unit global T change) if warming is concentrated in one hemisphere
• Therefore, one-‐hemisphere warming -‐-‐> larger global water vapor increase (%) per K global warming
€
q = q0eαδT
Global water vapor
Function of initial water vapor & delta T
Local mixing ratio increases at
Clausius-Clapeyron
AlternaFve predicFon of global water vapor increases:
€
δT = δT +δT ' = Mean surface T change + perturbation T change
Correction due to inhomogeneous temperature increases
Slow
Rapid
4.3 %
6.2 % 6.6 %
6.7%
-0.3 %
-2.6 %
AlternaFve theoreFcal predicFon of water vapor increases matches
simulated increases
TRACE Paleoclimate results: Global precipitaFon vs. surface T
“Slow” and “Rapid” cases behave similarly
P increases at 1.9% per unit surface T warming
Global Temperature (Celsius)
TRACE paleoclimate warming: Wet [mostly] get weTer & dry [mostly] get drier
Latitude
For a 1 degree global surface T change:
“thermodynamic” scaling
Mean Δ(P-E) and +/- a standard deviation
Rapid CO2 doubling: Wet [mostly] get weTer & dry [mostly] get drier
Latitude
For a 1 degree global surface T change:
“thermodynamic” scaling
Mean Δ(P-E) and +/- range (for CO2 doubling cases, anthropogenic)
“Rapid” & “Slow” T changes lead to somewhat different precipitaFon changes
Latitude
For a 1 degree global surface T change:
Rapid Slow Thermodynamic
scaling (dashed lines) is similar for rapid and slow changes.
Modeled changes significantly different
Hydrological cycle changes in TRACE paleoclimate (last 22ky) compared to rapid (anthropogenic-‐like) CO2-‐induced changes: • Global water vapor increases less (per unit warming) in TRACE paleoclimate – Longer Fmescale of S. Ocean adjustment different spaFal paTern of warming for rapid vs. slow warming
– most q in tropics (Clausius-‐Clapeyron non-‐linearity) global dlnq/dT dependent on tropical warming amount
• Global mean precipitaFon changes comparable
• Zonal precipitaFon paTern changes somewhat different due to circulaFon paTern changes
“Local” d(ln q)/dT by la%tude and height
• Is a range of dq/dT (% change water vapor per unit warming)
“Rapid” warming
“Slow” warming
Latitude
“Local” temperature change paHerns:
• Put Karen’s figure in showing T change here
• Or show my figure from my idealized model…