Correlations of atmospheric angular momentum (M) and change in length of day (ΔLOD) generally indicate conservation of momentum between the atmosphere and the crust, yet short-term weekly to monthly discrepancies of as much as 30% of the annual variation in either parameter are unexplained. Given the uncertainties present in each data set, it is possible that significant parts of these discrepancies may arise either from undetermined contributions or from incompleteness in the sampling of the atmospheric data. We examine the contributions of upper stratosphere winds (100-0.4 mbar) and of atmospheric moment of inertia terms (MI) for possible evidence of the discrepancies apparent between M and ΔLOD. For the period December 1978 to December 1979, time series of these parameters are calculated from the U.S. National Meteorological Center meteorological data set. Contributions of MI to M are not negligible. This parameter exhibits monthly scale variations of as much as 1.2¿1025 kg m2 s-1. Assuming that momentum is conserved between the atmosphere and the earth's crust and that the ΔMI variations are fully effective in exciting changes in the earth rotation, these monthly variations in MI would be equivalent to variations of as much as 0.2 ms in the change in the length of day (ΔLOD). This would be an upper limit for the M1 contribution, however, since not all of the variation in M1 may equate to variation in ΔLOD. The momentum contribution from upper stratospheric winds to M and ΔLOD is also nontrival, exhibiting variations which would stimulate ΔLOD changes of the order of 0.2-0.3 ms.

The time series of the high-altitude momentum component is seasonally shifted with respect to that of MI. Combining both terms with either values for M, we find that the main discrepancies between ΔLOD and M remain. Since the upper winds and atmospheric moment of inertia cannot explain the discrepancies, we comment on some other uncertainties which must be clarified to understand the global ΔLOD-M relationship. We conclude that other unknowns, among the components of atmospheric and crustal angular momentum or within the measurement of ΔLOD, are probably responsible for the apparent discrepancies.