Scott Springer Gary Lagerloef Earth and Space Research Seattle WA esr

1. Satellite-observed changes in the upper ocean heat Budget of the Northeast Pacific During 1993-2004 • Scott Springer • Gary Lagerloef • Earth and Space Research • Seattle WA • http://www.esr.org

2. Pacific Decadal Oscillation Climate Index http://jisao.washington.edu/pdo/ T/P/J Altimeter SSH T/P/J Altimeter SSH

3. Lag=0 R2=0.8 Note: monthly climatology removed Cummins, Lagerloef, and Mitchum (2005) High Spatial-Temporal Correlation between SSH and SST ±0.15 m ±1.5° C SSH CCA Mode 1 SST SSH index is smoother than SST index

4. The principal component of SSH is smoother than the one based on SST Altimeter-based PDO Index in NE Pacific http://www.esr.org/pdo_index.html

5. Mean Ekman pumping of anomalous vertical temperature gradient Figure 2.4 : (Left) Northeast Pacific satellite SSH difference between multi-year averages after minus before 1998, resembling the principal PDO pattern. The north-south CTD transect is shown near the southern perimeter of the SSH anomaly pattern. (Right) Yearly CTD transects before and after the 1998 transition, showing extensive changes to the upper ocean heat content and thermocline structure (courtesy J.Polovina and M.Seki, NOAA/NMFS, Honolulu).

6. We answer these question using simple models and satellite observed quantities. • Sea surface height Aviso gridded altimetry • Sea surface temperatures Reynolds gridded SST • Winds Ocean vector winds from QuikScat • Sea surface velocities from OSCAR • NCEP-2 heat fluxes Questions and Approach • What processes explain the coherent variations between SSH and SST in the northeast Pacific at interannual time and space scales? • 1) What processes control SSH (and pycnocline depth)? • 2) What are the processes that control SST?

7. SSH model • Sea surface height anomalies are due to displacements of pycnocline • Damping to climatology by unresolved ocean dynamics (mixing, etc.) assumed spatially uniform • Integrate and form EOFs of both sides • Best fit for λ-1=18 months and gΔρ/ρ = .02 m2/s. =0

8. SSH Observations Mode 1 SSH Model Mode 1 38% 38% Temporal correlation R=0.92 Spatial correlation R=0.65 damping time = 18 months Comparison of modeled and observed SSH • z

9. Upper ocean heat budget Subscript I denotes anomaly variables regressed against SSH PDO index. Follows approach of Chhak et al. (2009) • Horizontal velocities have Ekman and geostrophic components: • u= ugeo + uekman; v = vgeo+vekman • wE is Ekman pumping velocity • H is wintertime mixed layer depth

10. Obtained by regressing OSCAR velocity components against SSH PC 1 PDO surface velocity anomaly components Contours of SSH Contours of Ekman pumping

11. Magnitude of advection depends on the relative orientation of velocities and temperature gradients PDO surface velocity anomaly components

12. Anomalous advection of mean SST gradients Ekman Geostrophic = 1 degree/year

13. Horizontal advection terms and SST EOF pattern SST EOF 1 (PDO pattern) Horizontal advection Geostrophic + Ekman = 1 degree/year

14. Summary • Variations in SST result primarily from anomalous advection of the mean temperature gradient by both geostrophic and Ekman currents, which are well represented by satellite altimetry and winds. • Damped integral model driven by Ekman pumping accounts for interannual SSH variability during 1993-2003. • The major role of Ekman transport and its divergence accounts for SST and SSH co-variability.

15. Thank you