We investigate the effect of modeling tropospheric delay gradients on the station position estimates using the Japanese nationwide GPS array. The time series shows spatially coherent temporal fluctuations due to the variability of water vapor distribution. This makes it difficult to identify small crustal deformation signals in GPS time series. Precision and accuracy of station positions are known to be improved by modeling the tropospheric delay gradient, and past studies suggest that delay gradient estimates agree well with the collocated water vapor radiometer measurements. Here we pick up two intervals as long as 2 weeks in 1996 summer, when remarkable tropospheric delay gradients are expected, and investigate various influences of tropospheric delay gradients on station position estimates. First, we study spatial patterns of station position deviations caused by azimuthal asymmetry of water vapor distributions from two solutions, i.e., with and without the tropospheric delay gradient model. Second, we compare them with the estimated delay gradients. Site coordinate deviations in the solution without the delay gradient model are negatively correlated with the estimated delay gradients, but such systematic deviations disappear by introducing the delay gradient model. We found that the improvement in position accuracy is significant not only horizontally but also vertically over both of the two time intervals.

We used various analysis strategies to evaluate zenith tropospheric delays (ZTDs) retrieved from the Japanese nationwide Global Positioning System (GPS) array in the summer of 1996. We compared results from the network solution obtained by daily routine data analyses and two‐point positioning analyses with and without a tropospheric delay gradient model. We investigated two 14‐day periods in summer 1996, when water vapor distributions were highly azimuthally asymmetric. ZTD differences up to 4 mm, dependent on clusters used in the network analysis, were found between the network and the point analyses. We also found that differences in the estimated ZTD between the two‐point positioning analyses were correlated with the north components of the estimated tropospheric delay gradients. This is consistent with our simulation studies based on the north‐south asymmetry in the satellite coverage. The estimated tropospheric delay gradient vectors averaged over the periods pointed southward, which matches with the general meteorological condition in summer over the Japanese Islands. The temporal and spatial variations in the gradient estimates matched well with the moisture field determined by ZTD, in particular, during the passage of a weather front. Thus, tropospheric delay gradients obtained by GPS are expected to contain real horizontally anisotropic distribution of water vapor.

The atmospheric, hydrologic, and ocean current contributions to the annual wobble and length‐of‐day change are investigated, based upon 10 years of monthly values simulated by a coupled ocean‐atmosphere general circulation model (COAGCM). Also used are time series of geodetic excitation inferred from the observed wobble in the “SPACE97” data and, for reference, atmospheric angular momentum (AAM) functions from operational objective analysis data of Japan Meteorological Agency (JMA) for the period 1988–1997. The simulated annual variation in equatorial AAM function, which includes pressure and wind contributions, agrees well with that from the JMA operational analysis for 90°E component, but disagrees for 0° component due to differences in both pressure and wind contributions over ocean hemisphere. In the seasonal cycle the simulated hydrologic and ocean current excitation functions plus the AAM function from the JMA operational analysis tend to be in better agreement with the geodetic excitation function than those from the COAGCM output themselves. Because of the small magnitude of the simulated ocean current effects, the simulated hydrological cycle is found to be an important contributor to the nonatmospheric annual variations in polar motion. This confirms the importance of hydrology as a source of the annual wobble.

The excitation of the Chandler wobble (CW) by atmospheric wind and pressure variations is studied for the period 1983–1998. The data used here are the atmospheric angular momentum (AAM) functions computed from the operational objective analysis data of the Japan Meteorological Agency, compared with observed Earth wobble data. In the vicinity of the Chandler frequency (0.847 cycle per year), the AAM function has a broad spectral peak with somewhat excessive power and about 30° phase lead to that inferred from the observed wobble, with a coherence exceeding 95% confidence threshold. In addition, the temporal variation in the CW excitation calculated from the shifted 8‐year data of the AAM function is found to have nearly the same pattern as that inferred from observation, in which the wind and inverted barometer (IB) pressure contributions complement each other. They reveal that the atmospheric wind and IB pressure variations by themselves maintain a major part of the observed CW during the analysis period.

The atmospheric contributions to seasonal variations in length of day (LOD) and wobble are discussed for the period 1988–1997. The data sets used here are SPACE97 and EOP97C04 for LOD and wobble, respectively. Two atmospheric angular momentum (AAM) functions are calculated from the operational objective analysis data of the Japan Meteorological Agency (JMA) and the reanalysis data of the National Centers for Environmental Prediction (NCEP)/National Center for Atmospheric Research (NCAR). Both the axial AAM functions agree well in annual variation and roughly in semiannual variation, with the function inferred from the observed LOD. The two equatorial AAM functions show the different wind contributions to the function inferred from the observed wobble, in both annual and semiannual variations, which are attributable in part to the different vertical wind integration methods between NCEP/NCAR and JMA. The equatorial tropospheric wind contributions, calculated by the same method, are found to show discrepancies arising from the different tropospheric regional winds associated with the Asian monsoon.

We investigated the accuracy and features of precipitable water vapor (PWV) obtained from the Japanese Global Positioning System (GPS) network. We compared PWVs estimated from the Geographical Survey Institute (GSI) routine geodetic analysis using the Bernese software with those derived from the nearest 10 radiosonde stations over the Japanese islands. The comparisons for a half year showed that the agreement of the GPS‐derived PWV was 3.7 mm in terms of rms difference, with a standard deviation of 2.6 mm and a mean bias of −2.7 mm, which is less precise than those obtained in central North America. This may reflect the high temporal and spatial variability of water vapor over the Japanese islands, causing differences between PWVs if GPS and radiosonde launch stations are not colocated. We also found systematic biases in the GPS‐derived PWV. The absolute value of the mean bias at 1200 UT was systematically larger than that at 0000 UT. Their difference reached as much as 2 mm. Moreover, the mean biases tended to increase as PWV increased. We also confirmed that there were similar systematic biases in the PWV obtained from our GPS analysis using the GIPSY software with high temporal resolution. This fact indicates that the systematic biases depend neither on analysis software nor on temporal resolution of PWV estimation. We analyzed GPS data collected at the Tsukuba station using GIPSY for 2 years and showed a possibility that the systematic biases also found at the Tsukuba station were attributed to ocean tidal loading and seasonal variation of the mapping function both of which were not taken into account in the GPS analysis with the routine analysis using Bernese and our analysis using GIPSY. These results suggest the importance of implementing them in the analysis for accurate estimates of absolute value of PWV from the Japanese GPS network.

Behavior of precipitable water vapor (PWV) routinely retrieved from the nationwide array of the Global Positioning System (GPS) established by Geographical Survey Institute (GSI) of Japan are compared with the Japan area objective analysis data for numerical weather prediction (NWP) of the Japan Meteorological Agency (JMA). The array used here has a spatial resolution of about 50 km for monitoring crustal deformation. The 3‐hourly zenith tropospheric delay (ZTD) data obtained in GSI's routine analysis system are converted into PWV data by using the 12‐hourly NWP data. While a front accompanying with heavy rainfall moved eastward across the Japanese Islands from 0900 local standard time (LST) September 1 to 0900 LST September 3, 1996, the GPS array successfully detected the temporal anomalies of GPS PWV moving along with the front, in which the internal errors are estimated to be less than 3 mm. The results reveal that GSI's GPS array can work as an all weather giant array sensor of PWV over the Japanese Islands. It is found, however, that the GPS shows systematically fewer PWV than NWP data in mountainous areas reaching about 10 mm. This bias results mainly from the fact that most GPS sites located at bottom of valley in mountainous areas. After removing the topographical effects, there still remain significant differences amounting to 2–4 mm associated with errors of GPS observations and/or NWP objective analyses data.

Axial and equatorial atmospheric angular momentum (AAM) functions for the rotational dynamics of the Earth are calculated monthly from ensemble mean data of three independent 40-year simulations during 1955-1994 by the global model of the Japan Meteorological Agency (JMA) forced by observed near-global sea surface temperature (SST) conditions. The model results are compared with those from the reanalysis data of the National Centers for Environmental Prediction (NCEP) and the operational objective analysis data of JMA and with the functions inferred from the observed length of day (LOD) and polar motion. The annual term of the simulated axial wind AAM function (dimensionless relative angular momentum of atmosphere due to zonal wind) during 1984-1994 agrees well with those from the two analysis data sets and roughly with the inferred function from LOD, while the semi-annual term is considerably over-estimated, suggesting an incompleteness in the simulated subtropical zonal winds. The annual term of the simulated equatorial pressure AAM function (dimensionless atmospheric inertia products due to atmospheric mass redistribution) is considerably over-estimated with respect to those from the two analysis data sets, presumably due to the large simulated redistribution of atmospheric mass between the Eurasian continent and the North Pacific Ocean. For interannual Variations during 1955-1994, only the axial wind AAM function is reasonably simulated and shows good agreement with that from NCEP data as well as the Southern Oscillation Index. The above results lead to an understanding that the SST-forced AGCM simulates reasonably the atmospheric axial modes exciting LOD change but not the equatorial (non-axial) modes exciting the polar motion.

A hydrologic excitation of ten-yearly polar motion appearing in the combined five-daily polar motion data of International Polar Motion Service (IPMS) and SPACE94 during 1962-1995 is discussed. The discussion is based on time-dependent exponential decay models of land water storage supplied by precipitation, in which the National Oceanic and Atmospheric Administration (NOAA) monthly gridded precipitation anomaly data for 5 degrees x 5 degrees is used. The decay parameter of the land water storage tau(d) is determined when square of the difference between the observed ten-yearly polar motion and calculated one becomes minimum. Amplitude in the calculated polar motion can explain about 36 % of that in observed ten-yearly polar motion when tau(d) takes 3.7 months. The behavior of the calculated polar motion is similar to the observed one though the phase of the former leads that of the latter by about one year. When we do not consider the small phase difference, about 52 % of amplitude in the observed polar motion can be explained when tau(d) takes 5.1 months, suggesting a globally averaged decay time of the land water storage. The seesaw-like changes of precipitation that excite efficiently the ten-yearly polar motion are also confirmed between the North American continent and the western region of the Eurasian continent. These facts suggest a conceptual model of the feedback system on decadal hydrologic cycle centered in the North Atlantic Ocean, in which the system consists of atmospheric variations connected with the North Atlantic Oscillation, seesaw-like sea level changes, and precipitation changes in the North American and Eurasian continents.

1Kyoto University, Uji, Kyoto 611-0011, Japan 2National Astronomical Observatory, Mizusawa, Iwate 023-0861, Japan 3Geographical Survey Institute, Tsukuba, Ibaraki 305-0811, Japan 4Meteorological Research Institute, Tsukuba, Ibaraki 305-0052, Japan 5Nagoya University, Nagoya 464-0814, Japan 6Kochi University, Kochi 780-8520, Japan 7Kyushu University, Fukuoka 812-0053, Japan 8Tsukuba University, Tsukuba, Ibaraki 305-0006, Japan 9Shizuoka University, Shizuoka 422-8017, Japan 10University of Tokyo, Tokyo 113-0032, Japan

Atmospheric effects on gravity observations at Kyoto were estimated by using meteorological data sets at an interval of 12 h during a 4 month period from July to October in 1993. The effects owing to the air mass near the gravity station were evaluated by numerical integrals, and those distant from the station were calculated by using spherical harmonic expansions of meteorological data. The error in the calculated atmospheric effects was of the order of 0.1 μgal at most, except for the error related to the response of the oceans near the station based on the inverted barometric loading mode. About 90% of the atmospheric effects were attributed to local atmospheric variations within 50 km of the station. The remaining effects owing to the air mass outside this zone were of the order of 1 μgal, in which different features were recognized as compared with the effects owing to the regional air mass around Kyoto. The atmospheric effects thus estimated were compared with gravity data obtained by a superconducting gravity meter at Kyoto. The residuals showed gravity changes of a few microgals, a part of which might be caused by sources such as variations of the ground water level around the station.

Atmospheric wind and pressure contributions to non-seasonal variations in the length of day (LOD) determined by the International Radio Interferometry Surveying axe evaluated during the years of 1984-1991 using Japan Meteorological Agency data that include winds up to 10mb. We confirm that intra-annual LOD variations are the result of tropospheric zonal wind changes, which include a pronounced quasi-seven month oscillation in addition to the well-known intra-seasonal oscillation. It is also shown that inter-annual LOD variations on time scales of two to five years are the superposed contributions from both a stratospheric quasi-biennial oscillation and tropospheric low frequency variations of two-to-five year time scales in zonal wind. The residual LOD after removing these atmospheric contributions shows a decadal-like variation, which may be caused by a core - mantle coupling torque.

The seasonal variation of the earth's axial angular momentum budget is discussed with an error on the order of a few percent. The budget is based on the core‐mantle decoupling (CMD) hypothesis, using the length of day (LOD) data observed astronomically by the International Radio Interferometry Surveying and the atmospheric angular momentum data calculated from the forecast/analysis data set for numerical weather prediction of the Japan Meteorological Agency. The atmospheric relative angular momentum changes due to zonal wind account for an additional 23 % contribution over the annual budget. However, this is counterbalanced by the effects of the redistributions of air and water masses. Namely, about 16 % is accounted for by the redistribution of air mass, and remained about 7 % agrees with the contribution from surface water storage estimate on continents by Chao & O'Connor [1988]. These facts demonstrate a confirmation of the CMD hypothesis on time scale of a year. At the semi‐annual period, however, there still is a shortage of about 6 % in atmospheric and hydrospheric contributions to the budget.