Precipitation response to climate change over tropical oceans: - PowerPoint PPT Presentation

Precipitation response to climate change over tropical oceans: Importance of changes in surface convergence driven by near-surface temperature gradients Margaret L. Duffy (MIT EAPS) Paul A. O’Gorman (MIT EAPS) Larissa E. Back (UW Madison AOS) Funding by NSF 1 photo from NASA GOES satellite

Model (dis)agreement on mean precipitation response to climate change The response of tropical precipitation to climate change will be impactful. However, GCMs do not agree on the response △ P = P RCP 8.5 - P historical P RCP 8.5 2080 - 2099 P historical 1980 - 1999 Precipitation response to climate change in two GCMs 2

Three proposed mechanisms for precipitation response over tropical oceans 1. Wet-get-wetter : Historical pattern of precipitation enhanced by warming (Chou and Neelin 2004, Held and Soden 2006) 2. Warmer-get-wetter : Precipitation response corresponds to SST response, relative to tropical mean (Xie et al. 2010) 3. Surface convergence important for convection (and thus precipitation) and SST gradients are important for setting this convergence via pressure gradients (Lindzen and Nigam 1987, Back and Bretherton 2009) 3

Three proposed mechanisms for precipitation response over tropical oceans 1. Wet-get-wetter : Historical pattern of precipitation enhanced by warming (Chou and Neelin 2004, Held and Soden 2006) 2. Warmer-get-wetter : Precipitation response corresponds to SST response, relative to tropical mean (Xie et al. 2010) 3. Surface convergence important for convection (and thus precipitation) and SST gradients are important for Mechanism for observed setting this convergence via pressure gradients (Lindzen precipitation and Nigam 1987, Back and Bretherton 2009) 4

What is the relative contribution of the different proposed mechanisms to the tropical precipitation response to climate change over oceans? Approach: 1. Use a simple model of precipitation (adapted from the “two-mode model” of Back and Bretherton 2009) to estimate the response of precipitation to climate change 2. Compare annual-mean precipitation response between the two-mode model and GCMs from CMIP5 3. Evaluate the relative importance of terms corresponding to proposed mechanisms of precipitation response 5

The two-mode model uses dry static energy budget to relate monthly-mean precipitation to monthly-mean SST and surface convergence • Key approximation: Two modes of vertical motion ( ω ) variability, one shallow and one deep • Shallow mode amplitude is related to surface convergence via mass continuity • Deep mode amplitude is empirically related to surface convergence and relative SST 6

Expression for monthly-mean precipitation from two-mode model 𝑀𝑄 = 𝐼 𝜓 𝑁 (),( 𝑏 ( 𝑇𝐷 + 𝑁 (),/ 𝑏 / 𝑇𝑇𝑈 1)2 + 𝑐 / 𝑇𝐷 + 𝑑 / − 𝑆 7 χ LP Precipitation (W m -2 ) SC Surface convergence SST rel Anomaly from tropical mean SST M se,s Shallow dry effective stability M se,d Deep dry effective stability a s , a d , b d , c d Coefficients R 0 Clear sky radiation H [ χ ] Heaviside function 7

Precipitation from two-mode model compares favorably to observations Mean precipitation (2000- 2008) from GPCP observations (top) and from the two-mode model using coefficients from reanalysis and inputs from observations (bottom) 8

Two-mode model accurately reproduces GCM precipitation response Precipitation response to climate change in different GCMs (top) and from two-mode model using coefficients and inputs from the GCM (bottom) 9

Evaluate relative contribution of each mechanism Wet-get-wetter LP = H [ χ ] { M se,s a s SC + M se,d ( a d SC + b d SST rel + c d ) − R 0 } Relatively warmer-get-wetter Surface convergence important Wet-get-wetter: M se,s and M se,d change Relatively warmer-get-wetter: SST rel changes Surface convergence important: SC changes All other coefficients remain fixed 10

GFDL-CM3 decomposition of precipitation response Precipitation response to climate change in GFDL-CM3 from GCM (top) and from two-mode model (bottom) Contribution of each proposed mechanism to precipitation response to climate change 11

MPI-ESM-MR decomposition of precipitation response Precipitation response to climate change in MPI-ESM-MR from GCM (top) and from two-mode model (bottom) Contribution of each proposed mechanism to precipitation response to climate change 12

What is the relative contribution of different proposed mechanisms to the tropical precipitation response to climate change over oceans? • The pattern of precipitation response differs between the GCMs studied, but the relative contributions are robust • The surface convergence contribution is surprisingly strong and the “warmer-get-wetter” contribution is surprisingly weak 13

The surface convergence response is key to the precipitation response; to what extent is it set by changes in SST gradients? 14

Relationship between boundary layer convergence and surface temperature gradients • Boundary layer winds are related to boundary layer pressure gradients, and thus temperature gradients by the hydrostatic relationship (Lindzen and Nigam 1987) • We use the mixed layer model from Back and Bretherton 2009 to derive approximations for SC 15

Surface convergence approximations from mixed layer model • Proportional to the Laplacian of boundary layer virtual temperature r 2 T v Z p s g ✏ i � r 2 T BL � = � SC dz p s Surface pressure R ⇢ 0 ( ✏ 2 i + f 2 ) T 2 g acceleration due to gravity BL v parameter proportional to mixing and friction ✏ i R gas constant reference density • ρ 0 Proportional to the Laplacian of SST f coriolis parameter T v virtual tempearature Z vertical integral over boundary layer BL p s ✏ i Φ 850 ¯ Φ 850 geopotential at 850 hPa SC ( r 2 SST ) = � i + f 2 ) SST 2 r 2 SST R ⇢ 0 ( ✏ 2 16

Changes in surface convergence are related to changes in the Laplacian of low- level temperatures Response of surface convergence to climate change in the ensemble mean of 8 GCMs (top), approximated as proportional to the Laplacian of boundary layer virtual temperature (middle), and approximated as proportional to the Laplacian of SST (bottom) 17

Precipitation response is captured by two-mode model when surface convergence is approximated as proportional to the Laplacian of low-level temperature Only spatially- varying inputs are Response of precipitation to low-level T v and climate change (a) in ensemble SST mean of 8 GCMs, (b) approximated using two-mode model, (c) using two-mode model with SC( ▽ 2 T BL ), and (d) using two-mode model with Only spatially- SC( ▽ 2 SST). varying input is SST 18

Conclusions • All three proposed mechanisms contribute to precipitation response to climate change over tropical oceans • Surprisingly, surface convergence contribution is stronger than warmer-get-wetter contribution • Surface convergence contribution is approximated well by the Laplacian of boundary layer temperature and, to a lesser extent, by the Laplacian of SST • Implication: pattern of SST change affects precipitation via both warmer-get-wetter and its effect on changes in surface convergence Margaret L. Duffy, MIT EAPS, mlduffy@mit.edu 19