Sensitivity of zonal-mean circulation to air-sea roughness in climate models

Transcription

1 Sensitivity of zonal-mean circulation to air-sea roughness in climate models Inna Polichtchouk & Ted Shepherd Royal Meteorological Society National Meeting

2 MOTIVATION Question: How sensitive is the large-scale circulation to changes in the surface drag parameters over the ocean? Momentum transfer between the atmosphere and the ocean through turbulence in the boundary layer that must be parameterized. Many uncertainties exist in parameters underlying turbulence parameterizations: e.g., do we know how airsea coupling depends on near-surface wind speed?

3 MOTIVATION: Drag in idealised models Chen et al (2007) showed sensitivity of eddy-driven jet latitude to surface friction in idealized dry GCM: Decrease in Rayleigh drag à jets move poleward. Question: Does the sensitivity to surface friction carry over to more complex GCMs with more realistic boundary layer and moisture? Drag reduces Chen et al (2007, JAS)

4 MOTIVATION: Drag in climate models Inadequate representation of drag over the ocean in climate models, as surface stress biases are not explained by bias in any other momentum transfer term. Question: How does the variation of drag at low and intermediate wind speed affect large-scale circulation? Simpson et al (2014, JAS) F IG. 1. (a) Zonal mean DJF climatologies of t. (b) as (a) but for the zonal me

5 METHOD Hierarchy of model approach to understand impact of ocean drag on large-scale circulation. Use CAM3 climate model mostly in aquaplanet setup (no topography or seasonal cycle, zonally symmetric SST) and vary air-sea momentum roughness length Z 0m. {,E,H} = {C d v,c e q, C h } C d = C d (Z d (Z 0m 0m ), m ) C e = C e (Z e (Z 0m, Z 0m,Z 0e ) 0e, m, e ) C h = C h (Z h (Z 0m, Z 0m,Z 0h ) 0h, m, e ) = surface stress E = evaporation H = sensible heat C {d,e,h} = transfer coe cients {m,e,h} = integrated flux profiles Z {0m,0h,0e} = roughness lengths

6 METHOD Hierarchy of model approach to understand impact of ocean drag on large-scale circulation. Use CAM3 climate model mostly in aquaplanet setup (no topography or seasonal cycle, zonally symmetric SST) and vary air-sea momentum roughness length Z 0m. {,E,H} = {C d v,c e q, C h } C d = C d (Z d (Z 0m 0m ), m ) C e = C e (Z e (Z 0m, Z 0m,Z 0e ) Change Z 0e, m, e ) 0m C h = C h (Z h (Z 0m, Z 0m,Z 0h ) 0h, m, e ) = surface stress E = evaporation H = sensible heat C {d,e,h} = transfer coe cients {m,e,h} = integrated flux profiles Z {0m,0h,0e} = roughness lengths

7 METHOD Two experiments: 1) Original Z 0m ; 2) Reduced Z 0m Other model details: - Eulerian pseudospectral core - T85L26 resolution (top at 3 hpa) NOMINAL REDUCED Neutral drag profiles in CAM3

8 RESULTS: Zonal-mean circulation response Polichtchouk & Shepherd (2016,QJRMS) Response to reduced surface roughness ENSO-like: i) A poleward shift of the mid-latitude westerlies extending to the surface. ii) A weak poleward shift of the subtropical descent region. iii) A weakening of the HC and a poleward shift of the ITCZ. iv) A poleward shift of the tropical surface easterlies.

9 RESULTS: Tropics vs. extratropics Question: Is the response mediated from the tropics or the extratropics? tropics extratropics

10 METHOD Hierarchy of model approach to understand impact of ocean drag on large-scale circulation. Use CAM3 climate model mostly in aquaplanet setup (no topography or seasonal cycle, zonally symmetric SST) and vary air-sea momentum roughness length Z 0m. {,E,H} = {C d v,c e q, C h } C d = C d (Z d (Z 0m 0m ), m ) C e = C e (Z e (Z 0m, Z 0m,Z 0e ) Change Z 0e, m, e ) 0m C h = C h (Z h (Z 0m, Z 0m,Z 0h ) 0h, m, e ) = surface stress E = evaporation H = sensible heat C {d,e,h} = transfer coe cients {m,e,h} = integrated flux profiles Z {0m,0h,0e} = roughness lengths

11 RESULTS: Heat vs. momentum fluxes Question: Is the response mediated thermodynamically or dynamically? H & E only surface stress only

12 RESULTS: AMIP-type and slab ocean setup Question: Does the sensitivity to reduced Z 0m carry over to setups with seasonal cycle and full complexity of surfaceatmosphere interaction? AMIP slab ocean

13 RESULTS: Basic state sensitivity Question: Is the response sensitive to the basic state? Basic states with strong Hadley circulation and subtropical jet less sensitive.

14 SUMMARY Circulation response to the reduced air-sea roughness ENSO-like: i. A poleward shift of the mid-latitude jet extending to the surface. ii. A weak poleward shift of the subtropical descent region. iii. A weakening of the Hadley circulation generally accompanied by the poleward shift in the ITCZ. Response mediated thermodynamically and from the tropics.

15 Understanding the circulation response Question: Why does the circulation change in response to reduced Z 0m? Ensemble of switch-on simulations reveal: 1. Initial reduction in equatorial zonal surface stress leads to decrease in BL meridional winds through Ekman balance à BL wind magnitude decreases à E & H decrease à cooling of the tropics. 2. Cooler tropics ß à reduced meridional temperature gradient à reduced subtropical jet speed and baroclinic eddy generation. 3. Reduced eddy generation and poleward shift in the critical latitude lead to poleward mid-latitude jet shift.

16 MOTIVATION Southern hemisphere wind speed biases and stratospheric circulation improved by updated air-sea momentum drag parametrization in GEOS-5 comprehensive climate model. Grafinkel et al (2011, GRL)