FIGURE 8.2 (a) Monthly Ceven (even zonal harmonics of the Earth's gravity field) series spanning 1980-1992 recovered from LAGEOS I satellite laser ranging data (the mean value has been removed). The error bars represent a I sigma uncertainty.

(b) Monthly Ceven calculated from NMC (1984-1992) atmospheric surface pressure data using the same linear combination of the spherical harmonic coefficients to which the observations are sensitive. Solid line: NIB model; dashed line: IB model.

(c) Monthly Ceven calculated from an equilibrium seasonal ocean tide model (solid line) and the surface water (dashed line) results of Chao and O'Connor (1988). Figure from Dong et al. (1996).

tudes. This is a reflection of the chaotic nature of weather. Some anomalies are quasi-persistent, existing for weeks or months. Surface pressure anomalies in the tropics are small (a few mbar)—exceptions include hurricanes, which cause pressure anomalies of up to 100 mbar for a short time.

There are many atmospheric phenomena that cause variations in pressure—Table 8.1 lists those that will likely be sensed by a gravity satellite. The list is not exhaustive and does not include phenomena that are unlikely to significantly influence gravity. For purposes of modeling, each phenomenon is presented in terms of a spatially-isolated mass anomaly, with its characteristic diameter on the Earth's surface, magnitude (in mbars of surface pressure), and typical lifetime and speed. These specifications were used to generate the atmospheric curves in Figure 1.2.

The geographic structure of, and  dynamical mechanisms responsible for much of the atmospheric contribution to time-dependent gravitational anomalies have not yet been fully explored. A possible source for these effects on seasonal time scales, for example, might be the Indian monsoon system, which involves large mass transfers between the Asian continent and