IPCC Fourth Assessment Report: Climate Change 2007
Climate Change 2007: Working Group I: The Physical Science Basis

3.5.6 Winds, Waves and Surface Fluxes

Changes in atmospheric circulation imply associated changes in winds, wind waves and surface fluxes. Surface wind and meteorological observations from Voluntary Observing Ships (VOS) became systematic around 150 years ago and are assembled in ICOADS (Worley et al., 2005). Apparent significant trends in scalar wind should be considered with caution as VOS wind observations are influenced by time-dependent biases (Gulev et al., 2007), resulting from the rising proportion of anemometer measurements, increasing anemometer heights, changes in definitions of Beaufort wind estimates (Cardone et al., 1990), growing ship size, inappropriate evaluation of the true wind speed from the relative wind (Gulev and Hasse, 1999) and time-dependent sampling biases (Sterl, 2001; Gulev et al., 2007). Consideration of time series of local surface pressure gradients (Ward and Hoskins, 1996) does not support the existence of any significant globally averaged trends in marine wind speeds, but reveals regional patterns of upward trends in the tropical North Atlantic and extratropical North Pacific and downward trends in the equatorial Atlantic, tropical South Atlantic and subtropical North Pacific (see also Sections 3.5.1 and 3.5.3).

Visual VOS observations of wind waves for more than a century, often measured as significant wave height (SWH, the highest one-third of wave (sea and swell) heights), have been less affected than marine winds by changes in observational practice, although they may suffer from time-dependent sampling uncertainty, which was somewhat higher at the beginning of the record. Local wind speed directly affects only the wind-sea component of SWH, while the swell component is largely influenced by the frequency and intensity of remote storms. Linear trends in the annual mean SWH from ship data (Gulev and Grigorieva, 2004) for 1900 to 2002 were significantly positive almost everywhere in the North Pacific, with a maximum upward trend of 8 to 10 cm per decade (up to 0.5% yr–1). These are supported by buoy records for 1978 to 1999 (Allan and Komar, 2000; Gower, 2002) for annual and winter (October to March) mean SWH and confirmed by the long-term estimates of storminess derived from the tide gauge residuals (Bromirski et al., 2003) and hindcast data (Graham and Diaz, 2001), although Tuller (2004) found primarily negative trends in wind off the west coast of Canada. In the Atlantic, centennial time series (Gulev and Grigorieva, 2004) show weak but statistically significant negative trends along the North Atlantic storm track, with a decrease of 5.2 cm per decade (0.25% yr–1) in the western Atlantic storm formation region. Regional model hindcasts (e.g., Vikebo et al., 2003; Weisse et al., 2005) show growing SWH in the northern North Atlantic over the last 118 years.

Box 3.3: Stratospheric-Tropospheric Relations and Downward Propagation

The troposphere influences the stratosphere mainly through planetary-scale waves that propagate upward during the extended winter season when stratospheric winds are westerly. The stratosphere responds to this forcing from below to produce long-lived changes to the strength of the polar vortices. In turn, these fluctuations in the strength of the stratospheric polar vortices are observed to couple downward to surface climate (Baldwin and Dunkerton, 1999, 2001; Kodera et al., 2000; Limpasuvan et al., 2004; Thompson et al., 2005). This relationship occurs in the zonal wind and can be seen clearly in annular modes, which explain a large fraction of the intra-seasonal and interannual variability in the troposphere (Thompson and Wallace, 2000) and most of the variability in the stratosphere (Baldwin and Dunkerton, 1999). Annular modes appear to arise naturally as a result of internal interactions within the troposphere and stratosphere (Limpasuvan and Hartmann, 2000; Lorenz and Hartmann, 2001, 2003).

The relationship between NAM anomalies in the stratosphere and troposphere can be seen in Box 3.3, Figure 1, in which the NAM index at 10 hPa is used to define events when the stratospheric polar vortex was extremely weak (stratospheric warmings). On average, weak vortex conditions in the stratosphere tend to descend to the troposphere and are followed by negative NAM anomalies at the surface for more than two months. Anomalously strong vortex conditions propagate downwards in a similar way.

Long-lived annular mode anomalies in the lowermost stratosphere appear to lengthen the time scale of the surface NAM. The tropospheric annular mode time scale is longest during winter in the NH, but longest during late spring (November–December) in the SH (Baldwin et al., 2003). In both hemispheres, the time scale of the tropospheric annular modes is longest when the variance of the annular modes is greatest in the lower stratosphere.

Downward coupling to the surface depends on having large circulation anomalies in the lowermost stratosphere. In such cases, the stratosphere can be used as a statistical predictor of the monthly mean surface NAM on time scales of up to two months (Baldwin et al., 2003; Scaife et al., 2005). Similarly, SH trends in temperature and geopotential height, associated with the ozone hole, appear to couple downward to affect high-latitude surface climate (Thompson and Solomon, 2002; Gillett and Thompson, 2003). As the stratospheric circulation changes with ozone depletion or increasing greenhouse gases, those changes will likely be reflected in changes to surface climate. Thompson and Solomon (2005) showed that the spring strengthening and cooling of the SH polar stratospheric vortex preceded similarly signed trends in the SH tropospheric circulation by one month in the interval 1973 to 2003. They argued that similar downward coupling is not evident in the NH geopotential trends computed using monthly radiosonde data. An explanation for this difference may be that the stratospheric signal is stronger in the SH, mainly due to ozone depletion, giving a more robust downward coupling.

The dynamical mechanisms by which the stratosphere influences the troposphere are not well understood, but the relatively large surface signal implies that the stratospheric signal is amplified. The processes likely involve planetary waves (Song and Robinson, 2004) and synoptic-scale waves (Wittman et al., 2004), which interact with stratospheric zonal wind anomalies near the tropopause. The altered waves would be expected to affect tropospheric circulation and induce surface pressure changes corresponding to the annular modes (Wittman et al., 2004).

Box 3.3, Figure 1

Box 3.3, Figure 1. Composites of time-height development of the NAM index for 18 weak vortex events. The events are selected by the dates on which the 10 hPa annular mode index crossed –3.0. Day 0 is the start of the weak vortex event. The indices are non-dimensional; the contour interval for the colour shading is 0.25, and 0.5 for the white lines. Values between –0.25 and 0.25 are not shaded. Yellow and red shading indicates negative NAM indices and blue shading indicates positive indices. The thin horizontal lines indicate the approximate boundary between the troposphere and the stratosphere. Modified from Baldwin and Dunkerton (2001).

Linear trends in SWH for the period 1950 to 2002 (Figure 3.25) are statistically significant and positive over most of the mid-latitudinal North Atlantic and North Pacific, as well as in the western subtropical South Atlantic, the eastern equatorial Indian Ocean and the East China and South China Seas. The largest upward trends (14 cm per decade) occur in the northwest Atlantic and the northeast Pacific. Statistically significant negative trends are observed in the western Pacific tropics, the Tasman Sea and the south Indian Ocean (–11 cm per decade).


Figure 3.25. Estimates of linear trends in significant wave height (cm per decade) for regions along the major ship routes of the global ocean for 1950 to 2002. Trends are shown only for locations where they are significant at the 5% level. Adapted from Gulev and Grigorieva (2004).

Hindcasts of waves with global and basin-scale models by Wang and Swail (2001, 2002) and Sterl and Caires (2005), based on NRA and ERA-40 winds, respectively, show an increasing mean SWH as well as intensification of SWH extremes during the last 40 years, with the 99% extreme of the winter SWH increasing in the northeast Atlantic by a maximum of 0.4 m per decade. Wave height hindcasts driven with NRA surface winds suggest that worsening wave conditions in the northeastern North Atlantic during the latter half of the 20th century were connected to a northward displacement in the storm track, with decreasing wave heights in the southern North Atlantic (Lozano and Swail, 2002). Increases of SWH in the North Atlantic mid-latitudes are further supported by a 14-year (1988–2002) time series of the merged TOPography EXperiment (TOPEX)/Poseidon and European Remote Sensing (ERS-1/2) satellite altimeter data (Woolf et al., 2002).

Since the TAR, research into surface fluxes has continued to be directed at improving the accuracy of the mean air-sea exchange fields (particularly of heat) with less work on long-term trends. Significant uncertainties remain in global fields of the net heat exchange, stemming from problems in obtaining accurate estimates of the different heat flux components. Estimates of surface flux variability from reanalyses are strongly influenced by inhomogeneous data assimilation input, especially in the Southern Ocean, and Sterl (2004) reported that variability of the surface latent heat flux in the Southern Ocean became much more reliable after 1979, when observations increased. Recent evaluations of heat flux estimates from reanalyses and in situ observations indicate some improvements but there are still global biases of several tens of watts per square metre in unconstrained products based on VOS observations (Grist and Josey, 2003). Estimates of the implied ocean heat transport from the NRA, indirect residual techniques and some coupled models are in reasonable agreement with hydrographic observations (Trenberth and Caron, 2001; Grist and Josey, 2003). However, the hydrographic observations also contain significant uncertainties (see Chapter 5) due to both interannual variability and assumptions made in the computation of the heat transport, and these must be recognised when using them to evaluate the various flux products. For the North Atlantic, there are indications of positive trends in the net heat flux from the ocean of 10 W m–2 per decade in the western subpolar gyre and coherent negative changes in the eastern subtropical gyre, closely correlated with the NAO variability in the interval 1948 to 2002 (Marshall et al., 2001; Visbeck et al., 2003; Gulev et al., 2007).