Fig. 1.
Schematic showing a Z-like pattern whose diagonal is the zero wind stress curl line based on wind climatology, and the top and bottom lines are the zero wind stress curl lines of the NAO anomaly. Regions of warming and cooling as well as the intergyre gyre spun up by the NAO+ are also indicated (from Marshall et al., 2001).
Fig. 2.
Climatological Ekman pumping velocity (1948–2008). (Top) Climatological Ekman pumping velocity (10−6 m s−1) over the North Atlantic Ocean derived from the NCEP Reanalysis winds at 10 m height. (Bottom) Mean climatological Ekman pumping velocity (10−6 m s−1) averaged over each longitude belt of the basin.
Fig. 3.
Simulated time series of ocean temperature response (°C, 3d panel), vertical MOC (blue) and horizontal gyre (red) anomalies (Sv, 4th panel) forced by synthetic forcing including SAT forcing (°C, 1st panel) and Wind forcing (Sv, 2d panel). Note that daily time series of synthetic forcings are shown by dashed lines and 7 yr running mean time series of these forcings are shown by solid lines. Note also that the MOC anomalies are shown using a reverse scale. A striking similarity is apparent between the Wind forcing (2d panel) and MOC and Gyre responses (4th panel).
Fig. 4.
Power spectra of water temperature anomalies for experiments with synthetic H-based and white noise wind forcing. For comparison purposes, spectra from GCM CCM4 simulations are also shown. Vertical bar at 0.3 mo−1 frequency show 95% confidence interval.
Fig. 5.
Simulated 15-year running mean smoothed time series of ocean temperature T, vertical MOC Ψ m and horizontal gyre Ψ g anomalies forced by synthetic forcing for four combinations of signs of the model parameters f and g. All time series are normalised by their SDs in order to fit the same vertical scale.
Fig. 6.
Simulated CCSM4-based North Atlantic horizontal distributions of (a) Ekman pumping velocity (10−6 m s−1) associated with NAO+, (b) surface air temperature (SAT, °C), (c) sea-surface height (SSH, cm) and (d) potential water temperature averaged within the upper 200 m layer (T0–200 m, °C) and (e) vertical cross-section of the meridional overturning circulation (MOC, Sv). In addition, panel (a) shows the location of cross-section (black line segment southward of Greenland) used for diagnostics of simulated water transports shown in Fig. 7. For comparison, panel (a) shows a Z-like pattern of observation-based wind stress curl from Marshall et al. (2001) also shown in Fig. 1 and used in defining the simple model domain.
Fig. 7.
Simulated 7 yr running mean CCSM4-based time series of (a) SAT (°C) anomalies, (b) anomalies of Ekman pumping velocity (10−6 m s−1), (c) anomalous intergyre gyre streamfunction Ψ g (Sv) and its components: barotropic component based on sea level tilt , steric component and baroclinic component , (d) anomalous gyre Ψ g and MOC Ψ m streamfunctions (Sv) and NAO (normalised by 5 in order to match the scale) and (e) water temperature anomalies (°C) averaged over the upper 200 m layer limited by 60°N from the top and by the simulated zero wind stress curl line from the bottom and NAO (normalised by 10). SAT and Ekman pumping velocity anomalies are averaged over the triangle limited by 60°N latitude, 80°W longitude and zero wind stress curl line. Streamfunctions are computed for the cross-section shown by black line segment in Fig. 6a. Note that the MOC anomalies in panel (d) are shown using a reverse scale.
Table A1. Values of non-dimensional (except Hek) factors used in model equations
Fig. 8.
Correlation coefficient R(ω) derived from eq. A42 as a function of ω for (left) ms=1 and (right) ms=−1.
Fig. 9.
Transfer function derived from eq. A45 for (left) a = 1 and b = 4 which correspond to the case and for (right) a = 2 and b = 1 which correspond to the case.
