At periods longer than a few years, extending to many tens of years, the so-called 'decadal variations,' the sources of excitation for both LOD and PM are more enigmatic. The difficulty is that at these periods other effects may be important, including visco-elastic behavior such as post-glacial rebound, and exchange of angular momentum with the fluid core. In particular, it has become common to invoke the core as the major cause of decadal LOD changes. Although some climatic forcing of long period LOD has been recognized, (Salstein and Rosen, 1986; Eubanks, 1993), it is uncertain at what time scale air and water become less important than the core. Unfortunately, the role of the core remains largely unquantified because it is too remote to be easily observed. A further difficulty in assessing the air/water role at long periods is that the torques required to cause decadal LOD variations are utterly insignificant when compared with those applied by the atmosphere at shorter time scales (Hide and Dickey, 1991). This means that atmospheric/oceanic torques of geodetic significance are of second-order importance in general circulation studies. Quantification of momentum budgets among Earth, air, and water reservoirs is thus lacking at long periods. The requirements for progress in this field coincide completely with the central problems of global climate change.
Long period PM is conveniently divided into a linear drift, probably a post-glacial rebound effect (Wu and Peltier, 1984) plus irregular motions with periods of years to decades. The postglacial rebound effect provides some constraint on ice loading and rheological models of the earth, (Peltier and Jiang, 1994) but surface geodetic measurements are likely to be perhaps more effective in constraining the time and space distributions of recent glacial ice loads (Mitrovica et al, 1994). Neglecting the core, it is likely that the irregular decadal PM superimposed on the drift is forced by long term variations in water mass distribution, although the details remain obscure. Decadal PM is clearly polarized along the same longitude that would result from a global rise or fall in sealevel (Chao and O'Connor, 1988). However, the implied sealevel variations, ten or so centimeters over decades, are larger than those inferred from coastal tide gauges (Eubanks, 1993; Wilson, 1993). Storage in terrestrial water reservoirs is a likely contributor, but only surface storage has been accessible to observation. However, the potential contribution of surface reservoirs, alone, is surprisingly large (Chao, 1988). World-wide, subsurface (aquifer) water storage exceeds that in surface reservoirs and may be more important (Kuehne and Wilson, 1991). Another aspect of terrestrial water storage is the balance of water stored in glacial ice. At the present time, earth rotation variations (both PM and LOD), coupled with SLR-determined gravity field changes, provide better constraints on ice balance than do field observations by glaciologists (Trupin, 1993). This condition will probably persist until the development of remote sensing methods for ice sheet monitoring (Schutz and Zwally, 1993).