succeeded in retrieving the refractivity profiles over the ... Although GPS-LEO occultation data have the ... receiver memory into the hard disk of the computer. A.
Down looking GPS occultation measurement on the top of Mt. Fuji Yuichi AOYAMA(1), Yoshinori SHOJI(2), Ashraf MOUSA(1), Toshitaka TSUDA(1), and Hajime NAKAMURA(3) (1) (2)
Radio Science Center for Space and Atmosphere, Kyoto University, Uji, Japan Meteorological Research Institute, Japan Meteorological Agency, Tsukuba, Japan (3)
Numerical Prediction Division, Japan Meteorological Agency, Tokyo, Japan
Abstract Down looking (DL) GPS occultation experiment was performed in collaboration with NASA/JPL on the top of Mt. Fuji from July 10 to September 25, 2001, in order to obtain the water vapor profile near the Earth's surface. The GPS receiver, Turbo Rogue SNR-8000, and chock ring antenna were installed in the Mt. Fuji weather station located at an altitude of about 3.8 km. On average, the numbers of daily occultation that included the negative elevation angle were six events. We succeeded in retrieving the refractivity profiles over the southern area of Mt. Fuji from these DL measurement data by applying Abel inversion. These profiles were consistent with the radio sonde observation near Mt. Fuji. 1. Introduction Application of radio occultation technique using Global Positioning System (GPS) transmitters in high orbiters and receivers onboard low Earth orbiter (LEO) has been providing accurate atmospheric refractivity profiles with high vertical resolution [Melbourne et al., 1994; Ware et al., 1996; Kursinski et al., 1996, 1997; Rocken et al., 1997; Feng and Herman, 1999; Wickert et al., 2001]. The basic idea behind the radio occultation is to measure how radio waves are propagated into the atmosphere. The ray path of the radio wave between LEO and occulting GPS satellite traversing the atmosphere is bent by atmospheric refractivity gradients. The bending angle of the ray path is derived from a change in the phase delay (Doppler shift) of the received GPS signal. Assuming spherical symmetry, this bending information can be inverted with an Abel inversion to produce a vertical profile of refractive index. The vertical profile of atmospheric parameters, i.e., pressure, temperature, water vapor, and electron density, can be estimated from the refractive index profile. The atmospheric parameters from GPS-LEO occultation measurements have been widely used in the atmospheric studies [e.g., Tsuda et al., 2000; Hocke and Tsuda, 2001]. Although GPS-LEO occultation data have the advantage of being global (One occultation antenna connected with receiver in LEO provides about 150 globally distributed occultation events per day), the horizontal resolution in any region is relatively sparse without a large number of LEOs [e.g. Kursiniski et al, 1997]. By contrast, the occultation measurement with the receiver located inside the Earth’s atmosphere, such as
on a mountaintop or an airplane, can provide data over specific areas of interest for the purpose of a meso-scale weather prediction. This measurement, which is known as down looking (DL) occultation, has been proposed by Zuffada et al. [1999]. A mountain-based or airborne receiver would track any GPS satellite while it sets behind the Earth’s limb. These occultation data routinely produce the refractivity profiles below the height of the receiver in the specific area, with a diffraction-limited, vertical resolution of 150-250m. Although fundamentally DL measurement is very similar to the LEO measurement, it was originally thought that Abel transform could not be implemented by the limits of integration when the receiver is inside the atmosphere [Zuffada et al., 1999]. In fact, it is possible to use an Abel inversion for the DL case. The measurement geometry is similar to the one considered by Bruton and Kattawar [1997] when inverting solar occultation data. This paper introduces DL experiment that is conducted on the top of Mt. Fuji, Japan starting from year 2001 summer. Preliminary results of the refractivity profile derived by using Abel inversion are shown. 2. Experimental setup The DL occultation experiment aims at the use of GPS signals received at the top of Mt. Fuji to retrieve the water vapor distribution in the lower troposphere. The instruments used in this experiment are a Turbo Rouge SNR 8000 GPS AOA receiver and a Dorne- Margolin chokering antenna as well as a Linux OS laptop computer for controlling observation and archiving observed data. This system is provided by NASA/JPL. The observation is controlled by a shell script which is coded by NASA/JPL and named as GNEX. The GNEX checks azimuth and elevation angles of GPS satellites every 30 seconds, and triggers the high sampling measurement with rate of 50 Hz when there is occulting GPS satellite in the view of the antenna. After the signal from occulting GPS satellite can’t be detected by the receiver, GNEX saves the measurement data stocked in receiver memory into the hard disk of the computer. A mountain-based or airborne receiver would track any GPS satellites while they set behind the Earth’s limb. Therefore measurement data can be collected at both negative and positive elevation angles relative to the local horizon at the receiver site.
2-08-1
Figure 1. Distribution map of DL experiment sites. The arrows indicate the direction of the view of antenna. At first, the observing system was tested at the meteorological station of Japan Meteorological Agency (JMA) on the top of Mt. Tsukuba located 20km northeast of Tsukuba city, Japan. The antenna was installed in the roof of the meteorological station (140.10E, 36.23N, and 922.6m in the terrestrial reference frame, ITRF 97). The antenna was mounted in the direction of south-west (Fig. 1) and tilted 60 degrees from the zenith to allow for negative elevation angle observation. The test experiment was carried out from August, 2000 to June, 2001, continuously. This result was satisfactory, and hence the system was operated on the top of Mt. Fuji. The observing system was installed at JMA Mt. Fuji weather station on the top of Mt. Fuji (138.7E, 35.4N, 3810.9m in ITRF 97) from July 10 up to September 25, 2001. The antenna was mounted on the wooden wall of Mt. Fuji weather station which is tilted 65 degrees. The antenna is directed to the southeast direction as shown in Figure 1. From the antenna site, we can overlook the Suruga bay in 30km south of Mt. Fuji and the horizon isn't hidden behind the landform in the southward view. On the other hand, the receiver and the computer are placed inside the hut behind the wall. The antenna and receiver were connected by cable with length of 30m. A surge cutter is also inserted between them because thunder occurs frequently near Mt. Fuji. In addition, the antenna cable took off from receiver while thunder occurs, that is, observation is also interrupted. The measurement data archived in the computer are downloaded by off line. Therefore we climb Mt. Fuji every two weeks for downloading data and checking the observing system. To verify the DL occultation measurement, the radiosonde measurement was carried out in Shizuoka University from July 23 to July 28, and in Honkawane from July 30 to August 3. Shizuoka University is located in 50km southwest of Mt. Fuji, and Honkawane is in 100km southwest of Mt. Fuji. We can also obtain the radiosonde data measured operationally in Hamamatsu located in 120km southwest of Mt. Fuji. 3. Observation Data The observation sampling rate is set to 50 Hz when occultation starts. The observation starts at 3 degrees elevation angle of a setting satellite and last up to loss of look. We have no data on 1 August due to thunder, and
from 29 August to 8 September due to mal-functioning of the receiver. During the observation period, about 9 occultation events were observed daily on average. However, only 6 events have negative elevation angle. Details of the observation statistics are given in Figure 2. This figure shows that the range of minimum observed elevation angle is always higher than -2 degree. This range is very much higher than the expected minimum elevation angle of about -4 degree as indicated by simulation results [Mousa and Tsuda, 2002]. This is considered to be mainly caused by the very low signal to noise ratio (SNR) of the received signal and by the narrow range of signal detection of the receiver. The pressure, temperature, and humidity observed at Mt. Fuji weather station are used for calculating accurate refractivity at receiver site.
Figure 2. Statistics of DL measurement data obtained on the top of Mt. Fuji. In the upper panel, the number of occultation events observed daily is shown, with those having negative elevations indicated in darker color. The middle panel shows the minimum observed elevation angle of each events. The lower panel displays the vertical gradient of the refractivity estimated from radio sonde measurement data at Hamamatsu. Unit of refractivity gradient is km-1. 4. Data analysis and Results In the present analysis, we use 1 Hz sampling rate data converted from original measurement data with 50Hz sampling rate. The occultation events with negative elevation angle are listed in occultation event table. For the listed data, the double differencing technique was used to isolate the excess phase delay due to the atmosphere at both L1 and L2 on the occulting ray path. We analyze the data with Guam International GPS Service (IGS) reference site at the 1Hz sampling rate and orbit data which are interpolated into 1Hz sampling rate from IGS orbit data with 15 minute sampling rate. To smooth the time series of resulting excess phase delay, we apply a low pass filter with cut off period of 50 seconds. In the case that cycle slip appears, it is corrected using low passed time series. On the other hand, the ionosphere effect is removed (to the first order) using the L3 linear combination. Then the Doppler shift as the time derivative of the L3 carrier time series was
2-08-2
calculated by applying a fourth-order Spline function with 60 seconds sliding window. Knowing the positions and velocities of the GPS from the occultation geometry, the Doppler shift equation is solved to calculate the bending angle as a function of the impact parameter. In the analysis, the rigorous ellipsoidal correction was considered to satisfy the symmetry conditions needed for Abel inversion. Figure 3 shows an example of the Mt. Fuji data analysis starting with the observed phase at L1 and L2 and including the derived Doppler shift and bending angle. However the bending angle in the period from 650 to 800 seconds epoch time disappears in this figure. Since these epoch times correspond near zero elevation, the bending angles are disturbed by slight noise in Doppler shift.
Figure 3. Example of the Mt. Fuji occultation data observed between 19:17 and 19:36 UT, on September 23, 2001 is shown. Phase, SNR(C/N), atmospheric excess phase, Doppler shift as will as bending angle are given. The bending angle measurements are grouped into a set of negative and positive elevation angles, which are separated based on the fact that the transition between them correspond to the maximum impact parameter [e.g., Mousa and Tsuda, 2001]. Finally, the bending angle profile is then used to produce the partial bending profile. Using the partial bending profile, the vertical profile of the refractivity is calculated by numerically integration. The refractivity was compared with that estimated from the nearest radiosonde data. Three examples of this comparison are given in Figure 4. These figures show the data from Mt. Fuji as compared to radiosonde data at Hamamatsu city where is about 300 km away from the tangent point (Fig. 5). One of Figure 4 graphs demonstrates the refractivity profile estimated from occultation event that occurred at about 19UT(4:00 local time) of September, 23, 2001. The Hamamatsu radiosonde data used here are lunched at 0UT and 12UT (9:00 and 21:00 local time, respectively), so our derived refractivity lies between the values of the two lunches. In addition, it is clearly shown that the refractivity retrieved from DL measurement data at 19UT using Abel
inversion agrees with refractivity estimated from the radiosonde observation at 00UT.
Figure 4. Retrieved refractivity profiles compared with the Hamamatsu radio sonde observation.
Figure 5. Geographical location of the tangent point of ray path. The triangle mark indicates the location of Mt. Fuji, while square mark shows radio sonde launch point in Hamamatsu. 5. Discussion and summary The paper introduces the Mt. Fuji DL and the first refractivity profiles derived from DL occultation measurement. The refractivity retrieved here using Abel inversion shows good agreement with the radiosonde observation. This result is encouraging to continue observation and to try to assimilate the data to numerical weather prediction system. However, the Abel inversion relies on the assumption of local spherical symmetry. This assumption may be poorer than in the LEO case. For our Mt. Fuji data, the tangent point is drifted by about 200 to 400 km as shown in Figure 5. This suggests that we will be assuming local spherical symmetry over a larger horizontal scale. The errors from this assumption could be investigated by simulating measurements within the domain of a high resolution mesoscale model [Healy, 2001]. The observation statistics indicates that we have problem in the tracking at very low elevation angles. This limits our retrieved refractivity profile to about 2 km above the Earth’s surface as shown in Figure 4. This stop problem may be mainly due to very low signal to
2-08-3
noise ratio of the observed signal due to ducting propagation because the change of refractivity gradient is open seen in the height range of 2-3km near Mt. Fuji (Fig. 2). In order to get data to cover the lower part of the atmosphere, we consider that a high gain antenna should be installed on the top of Mt. Fuji, or that DL measurement should be performed within atmospheric boundary layer. In addition, the receiver firmware is going to be upgraded as well. Possibility of using open loop tracking will be investigated also. It is to be noted here that the Abel inversion is used even though the generalized ray tracing algorithm proposed by Zuffada et al [1999] could be also used. Abel inversion is chosen here as it does not require preliminary refractivity information as the case with the generalized ray tracing. Also, due to expected multi-path and other noise in the lower troposphere, the actual bending angle profile has many biases and the generalized ray tracing is not going to converge in that case. Besides, the calculation of partial bending, needed for the Abel inversion, is expected to reduce some noise. In summary, Abel inversion was used to retrieve refractivity profile from Mt. Fuji DL GPS data. The result is very much encouraging as it shows very good agreement with radiosonde observation. Experiment statistics suggest that more careful consideration is needed to be able to track the GPS signals at very low elevation angle. For such a very low elevation angle, Abel inversion is not expected to give good results due to the inversion layers and more advanced inversion algorithms are to be used.
monitoring, JPL Publication 94-18, 147 pp., Jet Propulsion Lab, Pasadena, CA. Mousa and Tsuda (2001): Retrieval of Key Climate Variables Using Occultation Geometry of a Mountain top GPS Receiver, ION GPS 2001 proceedings, pp. 1117-1126. Mousa and Tsuda (2002): Refractivity profile retrieved from Down-looking GPS radio Occultation using Abel Inversion: simulation study, Japan Earth and planetary science Joint Meeting, Tokyo, 27-31. Rocken et al. (1997): Analysis and validation of GPS/MET data in the neutral atmosphere, J. G. R.., 102, pp. 29849-29866. Tsuda et al. (2000): A global morphology of gravity wave activity in the stratosphere revealed by the GPS occultation data (GPS/MET), J. Geophys. Res., 105, 7257-7273. Ware et al. (1996): GPS sounding of the atmosphere from Low Earth Orbit: preliminary results, Bull. Am. Meteorol. Soc., vol. 77, No 1, pp. 19-40. Wickert et al. (2001), Atmospheric sounding by GPS radio occultations: first results from Champ., GRL, 28, 17, pp. 3263-3266. Zuffada et al.(1999): A Novel approach to atmospheric profiling with a mountain-based or airborne GPS receiver, J. Geophys., Res., 104, No. D20, pp. 24435-24447.
Acknowledgments This work has been performed in collaboration with NASA/JPL and Mt. Fuji weather station. References Bruton and Kattawar (1997): Unique temperature profiles for the atmosphere below an observer from sunset images, Appl. Opt 36, 27, pp. 6957-6961. Feng and Herman (1999): Remotely Sensing the Earth’s atmosphere using Global Positioning System (GPS) – the GPS/MET data analysis, J. Atmosphric and Oceanic Technology, 16, pp. 989-1002. Healy (2001): Radio occultation bending angle and impact parameter errors caused by Horizontal gradients in the troposphere: A simulation study. J.G.R., Vol. 106, No. D11, pp. 11875-11889. Hocke and Tsuda (2001): Gravity waves and ionospheric irregularities over tropical convection zones observed by GPS/MET radio occultation, Geophys. Res. Lett., 28, 2815. Kursinski, et al. (1996): Initial results of radio occultation observations of Earth’s atmosphere using the Global Positioning System, Science, 271, pp. 1107-1110. Kursinski, et al. (1997): Observing Earth’s atmosphere with radio occultation measurement using the Global Positioning System, J. Geophys., Res., 102, pp. 23429-23465. Melbourne et al. (1994): The application of spaceborne GPS to atmospheric limb sounding and global change
2-08-4
