Since February 1992, measurements of atmospheric CO2 at 200 m above the ground have been taken from a meteorological tower (lat 36.1°N, long 140.1°E, 25 m a.s.l.) in Tsukuba, central Japan. We have examined seasonal/long-term variations in atmospheric CO2, based on atmospheric 85Kr, a noble gas with well-known source and sink terms, monitored continuously at weekly intervals in Tsukuba since May 1995. During the annual cycle, the daily mean values of atmospheric CO2 in the afternoons (1300-1600 JST) showed maximum in late April, except in 2002, and minimum in late August or early September. The seasonal variations in atmospheric 85Kr, which is low in summer and high in winter, were caused by the seasonal variations in the large-scale atmospheric circulation, and latitudinal distribution of atmospheric 85Kr. Compared with seasonal variations in atmospheric 85Kr, the draw-down (DD) and buildup (BU) for atmospheric CO2 occurred about 1 week and 3 weeks later, respectively. The peak-to-trough amplitude of de-trended seasonal variations in atmospheric CO2 concentrations ranged from 11 to 14 ppm. A decrease in the seasonal CO2 minimum occurred with a decrease in the 85Kr concentration, showing the effect of air mass transport from different origins. Unlike the minimum, the maximum CO2 did not vary as expected from the air mass transport. The daily mean values of atmospheric CO2 increased at an average rate of 2.0 ppm yr−1, with a range from −0.5 to 4 ppm yr−1. While both CO2 and 85Kr are emitted into the atmosphere, due to energy consumption, the growth rate of atmospheric CO2 showed a pattern different from that of atmospheric 85Kr (0.03 Bq m−3 yr−1).
In this study, numerical simulations of ‘Yamase’ clouds were conducted using a nonhydrostatic model (NHM). Yamase clouds are typical maritime boundary-layer clouds that frequently appear over the sea off the east coast of the Sanriku District (northeast Honshu Island, Japan) during the summer season. The simulation period was from July to August 1993, when cool easterly winds (termed Yamase winds) prevailed. The NHM, with a horizontal grid of 40 km (NHM40) was integrated, nested within global objective analysis data, for one month by adopting sponge boundaries widely to lateral grids and upper layers, in order to reproduce a synoptic situation. The synoptic weather pattern is precisely reproduced in the long term integration. Then, in order to investigate the formation and evolution processes and structures of Yamase low-level clouds, multi-nesting simulations were conducted using the results of the one-month integration. The characteristic features of simulated Yamase clouds in terms of horizontal distribution, cloud shape and structures, and cloud water content, as well as the dominant mechanisms of cloud formation and evolution, strongly depend on the horizontal resolution of NHMs. The NHM with the high resolution of 100 m (NHM01), simulates convective structures similar to those observed from satellite images. Furthermore, the simulated liquid water path is close to the values observed in other Yamase cases. These results indicate that NHM01 reproduces the characteristic properties of the low-level clouds. In contrast, in simulations using the NHM with the low-resolution of 10 km (NHM10), the mixed layer is too moist and its height too low compared with observed data, and the liquid water path is overestimated. These errors are brought from the fact that the interaction between cloud radiation and formation at night time is too strong compared with the case of NHM01. Vertical heating profiles related to radiation, condensation, and evaporation in NHM10 are different from those generated by NHM01. This study suggests that improved parameterizations, such as a partial condensation scheme, and an adequate scheme of the buoyant production, should be introduced in low-resolution models such as NHM10.
We have analyzed the behavior of Kelvin waves in the upper troposphere and lower stratosphere (UTLS), through an intensive radiosonde campaign conducted in November 2002 at Koto Tabang (0.2°S, 100.32°E), Indonesia. In addition, we employed simultaneous global CHAllenging Mini satellite Payload (CHAMP) Global Positioning System (GPS) radio occultation (RO) measurements, in which we fitted the temperature perturbations (T´F) along the longitude, assuming only zonal wave number 1 and 2 components. Height-time comparison of the Kelvin waves well above the tropopause (16.5 km) revealed good agreement between the two techniques, but disagreement is observed around and below the tropopause, probably due to the effects ofhigher zonal wave number (>2) components. We have derived longitude-time section of the deviation from the T´F (residual, T´R) with GPS RO measurements in order to confirm the existence of higher zonal wave number perturbations around the tropopause. By combining the T´F and T´R, we were able to reproduce well the major features of the radiosonde results even below the tropopause. Clear eastward motion of the convective centers, inferred from the outgoing long-wave radiation (OLR) distribution, was observed from the Indian Ocean, reaching Koto Tabang during 18-19 November, 2002. This coincided with the enhancement of Kelvin-wave-like perturbations around the tropopause, having the zonal wave number 4 in a limited longitude region around Koto Tabang. We also found modulation of the tropopause structure by both the global scale Kelvin waves, and those with higher wave number components. The cold-point tropopause (on average at 16.5 km) jumped to 19 km, which is the height with the minimum temperature phase of the global-scale Kelvin wave. Hence, caution is advised in relating the tropopause variations observed via the radiosonde measurements with large-scale Kelvin waves, as the Kelvin waves with higher zonal wave numbers (>2) could also be responsible for the modification of the tropopause structure.
An algorithm has been developed to retrieve the vertical profiles of cloud microphysical properties from the surface observation based on three radars operated at three microwave bands (X-band, Ka-band, and W-band) and a microwave-radiometer. This algorithm has the advantage that it can be applied to any mixing condition of liquid and ice particles in clouds (i.e., pure-water clouds, pure-ice clouds, and mixed-phase clouds). To precisely treat the scattering processes of the radar waves in clouds, we pay attention to non-sphericity of ice-particles and radar wave attenuation by scattering, which are frequently neglected in the previous studies. The hexagonal ice-particles with various aspect ratios are considered. On the other hand, the observational data used in this algorithm need two assumptions: 1) the observed cloud does not contain precipitating and melting particles, and 2) the water vapor profiles are approximated by successive steady-state representations. The accuracy of the radar calibration required for this algorithm would be within ±0.1 dB, which is about the order of uncertainty of radar measurements. In the retrieval procedure, the mixing condition of water-droplets and ice-particles in the target cloud sub-layer can be identified by utilizing the wavelength-dependence of the scattering and attenuation properties. Once the mixing condition of the cloud sub-layer is determined, the microphysical properties of pure-water and pure-ice clouds are estimated by the algorithm, similar to the widely-used dual-wavelength technique, except for the fact that in the present algorithm, an equivalent aspect ratio of ice-particles can be estimated, in addition to ice-water-content and ice-particle size of pure-ice clouds. For mixed-phase clouds, the microphysical profiles, which satisfy the observed radar and microwave radiometer signals, can be estimated by iteratively changing aspect ratios of ice-particles in each sublayer. The retrieval algorithm is applied to the observational data obtained in the Vertically Pointing Measurements, Tsukuba, 2001 (VPM_TKB01). The retrieved microphysical profiles for the pure-ice clouds show reasonable performance of the present algorithm, although there are no in-situ measurements to validate the estimated values.
Long-term changes in the intensity and frequency of heavy precipitation in Japan were analyzed using quality checked daily precipitation data at 51 stations from 1901 to 2004. The analysis is based on ten categories defined from precipitation intensity and frequency, and some indices of heavy precipitation, such as ≥100 mm days, the annual maximum, and the top 100 cases during the 104 years. The result indicates that heavy precipitation based on these indices has increased during the 104 years. The linear trend of precipitation corresponding to the upper 10% is 2.3% per decade, and that of the number of top 100 cases is 2.6% per decade on the average over the stations. The increase is most pronounced in western Japan and in autumn, while weak, but similar signals are found in other regions and seasons as well. However, no increasing trend is found for less intense precipitation, such as ≥50 mm days and the number of top 1000 cases. Analysis was also made for 5, 11, and 31 day precipitations, and some indices of dry weather. It is found that the frequency of dry weather has increased during the 104 years. The number of days with precipitation less than 1 mm has increased in all the seasons and regions, with a trend of 0.4-0.7% per decade on the average, while the lower 1% ofcases of 31-day precipitation have doubled with a trend of 10% per decade.
In this paper, we describe a new algorithm for analyzing aerosol optical propenies from sun and sky radiance measurements. The Skyrad package (SKYRAD.pack) is one of the well-known software which analyzes the sky radiometer data. The SKYRAD.pack has been improved continuously since the first version appeared in 1996, and the programming procedures have become increasingly more complicated. Therefore, we rewrote the software by applying the statistical optimization method to its inversion algorithm to make it easier to improve, and extend the analytic capabilities in the future. In this method, the normal distribution was assumed as a probability density function of measurement error, and the maximum likelihood method was applied. Our new software can retrieve the columnar volume size distribution, and the complex refractive index as well as the SKYRAD.pack. We tested it using the simulated data from aerosol models, and it was able to retrieve the size distributions more accurately and more stably than the SKYRAD.pack. We also confirmed that the new program could analyze the real observed data more stably than the SKYRAD.pack with little difference between the results of the two programs.
This paper describes land-surface parameters, including the roughness lengths of momentum, heat, and water vapor, and the surface moisture availability derived from 1 year of field measurements at a suburban site in Tokyo, Japan. The main results are as follows. The estimated ratio of roughness length of momentum to heat, kBT−1, had an average value of 7. This value of kBT−1 was larger than those docu mented for vegetated and agricultural surfaces, but less than that reported for a light industrial area. In winter, kBT−1 tended to be smaller, suggesting dependence on the vertical position of the heat source within the canopy. The kBT−1 results also showed a diurnal change, in which higher values occurred in the afternoon. The parameter kBq−1, the ratio of the roughness length of momentum to water vapor, ranged from 40 to 600. The parameter kBq−1 depended on the water vapor deficit, the friction velocity, and the number of elapsed days after precipitation, all of which can change surface water availability. The value of surface moisture availability, β, ranged from 0.02 to 0.3. β gradually decreased with the number of elapsed days after precipitation.
In this study, numerical simulations of the summertime ice formation at Ice Valley in Korea are conducted using a simple 2D model, which is driven by the observed air temperature. The Ice Valley in Korea is a famous summer resort, as the Natural Monument where natural ice forms in spring, and remains till summer along the slope, and the ice disappears in fall to winter. It is interesting to note that the hotter the outside air is, the larger the ice grows in spring. The mysterious behavior of the summertime ice has been partly explained by a series of numerical experiments in terms of the convection theory, which is explained by the seasonally reversing wind-hole circulations. The numerical model in this study is updated from our former versions considering the new finding by the in situ observations. The result of the simulation is compared with the observations at the Ice Valley in Korea, and Nakayama in Japan. According to the result of the numerical simulation, the summertime wind-hole circulation activates the downward flow when the outside air is getting hot. The downward flow transfers the cold accumulated in the talus during the previous winter toward the outlet of the cold wind hole. As a result, the intensified cold advection grows the ice when the outside air is getting hot. The wind hole circulation appears to be about 17 mm/s in April, and the residence time of the air in the talus is estimated as 2.8 hours for this case. It is shown, that the seasonal reversal of the wind hole circulation is the essential mechanism of the summertime ice at the Ice Valley, acting as a natural thermal filter which effectively accumulates only the winter cold in the talus.
The Japan Meteorological Agency (JMA) started the operation of a wind profiler network, the WInd profiler Network and Data Acquisition System (WINDAS), in April 2001. The WINDAS is a network consisting of thirty-one 1.3 GHz-band wind profilers, with dense spatial resolution of 130 km on the average over the main islands of Japan. Operated with high data accuracy, under strict data quality control and high data availability, from reliable system operation. Height coverages of wind measurement are 6-7 km in summer, 3-4 km in winter and 5.3 km on the average through a year. The main purpose of the operation of WINDAS is to provide upper-air wind data to the numerical weather prediction (NWP) of the JMA, panicularly to the hydrostatic (till 2004) and non-hydrostatic (from 2004) mesocale numerical model (MSM). The WINDAS data are assimilated into the MSM using a fu11 forecast-analysis system, with 4-dimensional variational method. From statistical analyses and some case studies, it was conformed that the WINDAS data has contributed to improve accuracy of the MSM for mesoscale weather systems, particularly for heavy rainfall events. Being put on GTS, the wind data are distributed to the world in real-time. Although a problem of data contamination from migrating birds had occurred in the first year of the operation of the WINDAS, it was practically solved by developing a removal algorithm.