Angular momentum has been shown to be essentially in balance among the components of the Earth system, namely the mantle, the oceans/hydrosphere, the atmosphere, and the core. Whereas internal transfers of momentum between the core and the rest of the system are likely to be restricted to time scales of several years and longer, those involving the atmosphere, ocean, and mantle occur primarily on shorter time scales. Understanding the causes for the changes in the angular momentum of the various components requires a knowledge of the mechanisms that effect its transfer across the component boundaries. Fortunately, these transfers can now be estimated using observations and modern modeling approaches. For example, an atmospheric general circulation model is used to study the details and mechanisms related to the transfer of momentum on the seasonal scale during a simulation of a major El Niño event [Ponte et al., 1994]. Here the pressure and friction torques combined to help produce a strong maximum in atmospheric angular momentum, typical of an El NiNiño/Southern Oscillation peak [e.g., Rosen et al., 1984].
The two basic mechanisms for torque-induced transfer of angular momentum are friction and pressure. In the first case, friction creates momentum flux between the atmosphere or ocean and the solid Earth or between the two fluid components. The stress of the winds over the ocean or land depends on several factors, including the wind speed itself, the stability of the lower boundary layer, and the roughness of the underlying surface, which depends upon surface characteristics. In the case of pressure torques, angular momentum is transferred by horizontal differences in pressure across topographic features. Examples of the pressure torque are those resulting from differential pressure across mountain chains (mountain torques), and differential ocean water mass on opposite sides of continents. A higher pressure on the western side of a mountain, or continent, than on its eastern side, will transfer (positive) momentum to the solid Earth, reducing that of the atmosphere or ocean.
We have helped compute pressure (or mountain) and friction torques applied by the atmosphere across its lower boundary by using results from a global atmospheric model within the analysis/forecast system of the U.S. National Meteorological Center. Studying the torques allows one to understand the angular momentum transfer during a particular period in detail, as we did in the case of the special campaign known as SEARCH'92 to measure Earth rotation and related geophysical interactions. At high frequencies, the mountain torque tends to vary more than the friction torque, reflecting the fluctuations of synoptic weather patterns. During one weekly portion of the SEARCH'92 campaign, we were able to isolate the mechanism responsible for a rapid acceleration of the atmosphere: the passage of a high pressure center from the western to the eastern side of the South American Andes in July/August 1992 [Salstein and Rosen, 1994a]. Because the presence of topography shapes much of Earth's climate, identifying such interactions as these is critical to understanding the dynamical basis of climate change. At the same time, the axial rotation of the Earth during the period was observed very clearly by space-based geodetic systems to decelerate, consistent with the nature of the torque being applied. During the northern hemisphere winter of 1994, another intensive campaign captured a similar event of atmospheric angular momentum transfer by mountain torques, although this time acting across the Rocky Mountains of North America [Salstein, 1994b].
When time series of the mountain and friction torques on the atmosphere are combined, their sum is typically quite close to the series of the time derivative of atmospheric angular momentum, derived separately from analyzed wind and pressures fields (Figure 4). Nevertheless, some discrepancy does exist, which may be related to the use of model-derived values for the friction torque or other approximations. An alternate approach to estimating the friction torque uses wind values at the surface of the ocean from satellite-based measurements.
We first employed such remotely sensed values to study angular momentum/torque issues [Ponte and Rosen, 1993] with output from the Special Sensor Microwave Imager (SSM/I), a passive radiometer aboard a U.S. Defense Meteorological Satellite Program spacecraft that measures the emissivity of the ocean leading to estimates of wind speed. Using winds from a combined analysis of SSM/I and a set of operational weather analyses, we showed the importance of ocean/atmosphere torques, especially in the 1-3 month range of variability. Now that the scatterometer aboard the ERS-1 satellite has produced stress vectors for several years, based on reductions of the scatter patterns from an active microwave system, we have a new opportunity to assess frictional torques between the atmosphere and ocean [Salstein et al., 1994]. Although the torques produced by the ERS-1 scatterometer (Figure 5) are very similar in pattern to those produced by the NMC model, when they are substituted for the NMC-based values over the ocean they appear at times to provide better quantitative assessments. For example, in mid-1993, when particularly strong wintertime variations occur over the southern oceans, the torques produced with the ERS-1 scatterometer winds are an improvement over those from the numerical model in the context of the global axial angular momentum balance. We look forward to the use of ERS-2 scatterometer, EOS NSCAT and SeaWinds/Adeos II observations to further our studies of stress torques.
Regarding angular momentum transfers to and by the oceans, stress over the sea surface created by low-level winds can transfer angular momentum to the water below which, according to Ponte [1995b], can then be transferred laterally by quick barotropic adjustment processes. Via this adjustment, at external edges of the ocean basin, a continental torque is created by the subsequent difference in sea level, transferring this angular momentum quickly to the solid Earth. More quantitatively, Ponte and Rosen  were able to use output from the eddy-resolving model of Semtner and Chervin  to assess the nature of the torque mechanisms. The pressure torque calculated implied a rapid transfer of momentum between ocean and solid Earth. Regionally, results show that seasonal signals in torques are largest in the subtropics, for example. We see great potential of the ocean models and data assimilation approaches for these sorts of angular momentum and torque calculations.