| Citation: | Youguang Zhang, Qingliu Bao, Mingsen Lin, Shuyan Lang. Wind speed inversion and in-orbit assessment of the imaging altimeter on Tiangong-2 space station[J]. Acta Oceanologica Sinica, 2020, 39(12): 114-120. doi: 10.1007/s13131-020-1687-9
|
Ocean surface wind field, significant wave height, sea surface current, and sea temperature are important marine dynamical parameters. They play important roles in the monitoring of ocean environment, the forecasting of marine disaster, as well as the maintaining of maritime rights and interests. A synchronous observation of the ocean dynamical parameters is of particular importance to the overall analysis of oceanic dynamic process. Satellite remote sensing is one of the most effective techniques for acquiring the global ocean dynamic parameters. Specifically, microwave remote sensing can be used to observe the ocean at all-weather and all-time conditions. There is a variety of microwave remote sensing payloads that are operated by different communities at present, such as radar altimeters, microwave scatterometers, and microwave radiometers. Different payloads are designed for observing different ocean dynamical parameters, among which the radar altimeters are one of the most widely used microwave remote sensing payloads for monitoring sea surface height, winds, and waves.
Imaging altimeter (IALT) is a new generation of radar altimeter system, which combines the advantages of the interferometric radar altimeter and the synthetic aperture radar (SAR) together. It aims at providing the information of oceanic topography at an accuracy of several centimeters, as well as the information of land topography at an accuracy of several meters (Zhang et al., 2007). The swath of IALT is wide and the spatial resolution of IALT is high. The first IALT onboard Tiangong-2 space station was launched successfully from Jiuquan Satellite Launch Center on September 15, 2016. The in-orbit assessment of IALT was done before April 30, 2017.
In this paper, Section 2 has a brief overview of the IALT system onboard Tiangong-2 space station. Section 3 presents the inversion method of sea surface wind speed. A flowchart is illustrated to describe the overall inversion procedure. The retrieved wind speeds are compared with the European Center for Medium Weather Forecasting (ECMWF) reanalysis wind speed, the buoys wind speed, and the in-site ship measurements. The accuracy of retrieved IALT wind speed is analyzed in Section 4. Finally, the conclusions are addressed in Section 5.
IALT is a Ku band interferometric synthetic aperture radar. It was developed by the National Space Science Center, Chinese Academy of Sciences (NSSC). Similar to the traditional real-aperture altimeters, IALT can be used to measure the sea surface height (Bao et al., 2015), the significant wave height (Fan et al., 2014), and sea surface wind speed (Xu et al., 2014). More interestingly, IALT is able to provide the sea wave directional spectrum within a wide swath. The main characteristics of IALT onboard Tiangong-2 space station are summarized in Table 1.
| Orbit parameters | |
| Altitude/km | 378.6 |
| Velocity/km·s–1 | 7.3 |
| System parameters | |
| Polarization | VV |
| Carrier frequency/GHz | 13.58 |
| Peak power/W | 300 |
| Baseline/m | 2.3 |
| Band width/MHz | 40 |
| Incident angle/(°) | 2.5–8 |
| Swath/km | 30 |
| Spatial resolution/m | 20–60 |
DownLoad:
CSV
Figure 1 shows a photograph of the Tiangong-2 space station docked with the Shenzhou-11 spaceship. The photo was taken by the companion visible light camera on October 23, 2016. Red arrows indicate the antennas of IALT.
At low incidence angle, the microwave backscattering mechanism over sea surface is dominated by the quasi-specular reflection from small facets oriented perpendicular to the radar look direction. The magnitude of the scattered signal is proportional to the fraction of appropriately oriented facets on the illuminated portion of the sea surface (Cox and Munk, 1954; Barrick, 1974). The theoretical model of microwave backscatter at small incidence angles has been reviewed and analyzed by Brown (1990) and Apel (1994). The normalized radar cross section σ0 is proportional to the probability density function of the short waves forced by the wind stress. Let
| $$ {\sigma _0}\left(\theta \right){\rm{ = \pi }}{\left| {R\left(0 \right)} \right|^2}\left({{{{\rm{sec}} }^4}\theta } \right){p_s}\left({ - {\rm{tan}} \theta,0} \right), $$ | (1) |
where
| $${\sigma _0} = \frac{{{{\left| {R\left(0 \right)} \right|}^2}}}{{s\left(u \right)}}\left({{{{\rm{sec}} }^4}\theta } \right)\exp \left[ {\frac{{ - {{{\rm{tan}} }^2}\theta }}{{s\left(u \right)}}} \right],$$ | (2) |
where
According to the theory as well as the experiments, the sea surface cross section at low incident angle is not very sensitive to the wind speed, especially at the incidence angle of 10° (Li et al., 2004; Hesany et al., 2004). The quasi-specular reflection dominates the radar backscattering near nadir, thus the NRCS decreases as the wind speed increases. However at medium incidence angles, the Bragg scattering is the dominant factor, thus the NRCS increases as the wind speed increases. The curves of NRCS versus incidence angle for different wind speeds intersects at around 10° (Fig. 2), which means that different wind speeds correspond to the same NRCS at this particular incidence angle. Generally, the derivative of NRCS (versus the incident angle) is used in the wind speed inversion.
Freilich and Vanhoff (2003) compared the model function of the Precipitation Radar (PR) with the least square fitting results at certain selected wind speed (see Fig. 5 in Freilich and Vanhoff, 2003), and found that Eq. (2) does not fit the PR NRCS very well, especially at the low wind speed and high incident angle. Thus, it is necessary to develop an empirical GMF model for the wind speed inversion.
For low-incidence observation radar such as the PR on tropical rainfall measuring mission (TRMM) (Kozu et al., 2001), the ocean surface wind speed is usually retrieved by the maximum likelihood estimation (MLE) based on multi-incidence angles (Bao et al., 2016a, b). The accuracy of the MLE-based wind speed inversion is directly determined by the GMF. In order to improve the wind speed retrieval accuracy, a more accurate empirical GMF is established. The empirical GMF is developed by collocating the PR NRCS with QuikScat (Tsai et al., 2000) wind speed data. Here, we use the QuikScat wind speed and PR NRCS data from January 2007 to December 2008 to establish the empirical GMF model. The temporal window and spatial window for the collocation is 10 min and 25 km, respectively.
The established empirical GMF is a two-dimensional look up table, as shown in Fig. 3. The wind speed interval is 0.2 m/s, and the incidence angle interval is 0.8° (determined by the incidence resolution of PR). Figure 3 illustrates the NRCS as a function of wind speed and incident angle, for wind speed lower than 20 m/s and incident angle smaller than 13°. The comparison of theoretical and empirical GMF models at selected wind speeds is shown in Fig. 4. In the wind speed inversion process of IALT, the empirical GMF is used as a reference for quality control in order to remove the abnormal data, such as land and rain contamination (Lin et al., 2014, 2015).
For the IALT data, the antenna pattern is calibrated using the NRCS of Amazon rain forest in different incident angles. The calibration map of Amazon rain forest at low incident angle is calculated by averaging the TRMM PR data during 2003–2010. The spatial resolution and the incidence resolution of the calibration map are 5 km×5 km and 0.8°, respectively. As an example, Fig. 5 shows the calibration map of Amazon rain forest at θ=7.2°. The absolute NRCS of IALT are acquired after the external calibration.
By statistical calculation, we acquire the sea surface NRCS for different incidence angles, and then compare them with the empirical GMF as shown in Section 3.2. The comparison results for the wind speed of 10 m/s are illustrated in Fig. 6. It is clear that the sea surface NRCS measured by IALT is slightly different from the empirical GMF, which difference is typically smaller than 0.3 dB.
Due to the inaccurate absolute calibration of the antenna gain pattern, the IALT NRCS is slightly different from the GMF, and in turn, both the theoretical and the empirical GMF models are not suitable for IALT wind speed inversion. Thus, we use the neural network algorithm to retrieve the sea surface wind speed for IALT. The wind speed inversion flowchart is shown in Fig. 7.
Firstly, the NRCS data of IALT are obtained by preprocessing the data with the real aperture mode. The sea surface NRCS are averaged in both along-track and cross-track directions. The spatial window for the along-track average is 25 km, and the incidence angle window for the cross-track average is 0.25°. Secondly, the NRCS data are evaluated by quality control. The empirical GMF model introduced in Section 3.2 is used as a reference to define the quality of the IALT NRCS data. Thirdly, a neural network is applied to train the model parameters. Part of the quality-controlled IALT NRCS data are collocated with the ECMWF reanalysis wind speed data, and then are used as the training samples. The model parameters obtained by neural network algorithm are consequently used in the IALT wind speed inversion. Finally, the inversion accuracies are analyzed by the ECMWF, National Data Buoy Center (NDBC) buoy, and in-situ measurements. The inversion results of IALT at Eastern Pacific Ocean and the adjacent sea areas of China (in December 2016) are shown in Fig. 8.
The ECMWF reanalysis wind speed data (ERA-Interim) are used in the assessment. Its spatial resolution and temporal resolution are 0.25°×0.25° and 6 h respectively. The collocated ERA-Interim and IALT data are acquired by spatially and temporally interpolating the ECMWF reanalysis data to the IALT acquisitions. First, the wind cells of IALT are partitioned to a size of 25 km×25 km. Then the central position and the UTC time of IALT wind cells are derived. The ECMWF reanalysis data is actually interpolated spatially and temporally to the above geolocation and observing time.
Practically, the IALT NRCS data and the collocated ECMWF reanalysis wind speeds on the first half of December 2016 are used as the training samples. The wind speed inversion model of IALT is trained by the neural network algorithm. There are 2631 samples used in the training. We use a two-layer feed-forward network with sigmoid hidden neurons (25) and linear output neurons. The network is trained with Levenberg-Marquardt backpropagation algorithm. The schematic diagram of the two-layer feed-forward network is shown in Fig. 9.
Using the wind speed inversion model trained by neural network algorithm, the IALT NRCS data from second half of December 2016 to January 2017 are retrieved. The retrieved wind speeds are compared with the ECMWF reanalysis wind speeds, and the wind speed inversion accuracy is analyzed. Figure 10 shows the scatter plot of the retrieved wind speed versus ECMWF winds. Black line indicates the correspondence along the diagonal. There are 5848 wind cells in total. Comparing with the ECMWF reanalysis wind speeds, the IALT retrieved wind speeds have a negative bias of 0.21 m/s, a root mean square (RMS) of 1.85 m/s, and a correlation coefficient of 0.78. Basically, the IALT wind speed accuracy meets its specification, which requires that the RMS error of wind speed should be smaller than 2 m/s.
The buoys of NDBC meteorological/ocean program belong to the NDBC. They mostly locate near the east and the west coast of United State. The NDBC buoys have different types. The anemometers are installed on different heights depending on the sizes of the platforms. For the sake of comparison, the wind speeds measured by NDBC buoy are also converted to 10 m equivalent neutral stability wind speed. In this paper, we use three buoys (51000, 51004, 51101) that locate over the open sea (near Hawaii) to validate the retrieved wind speeds. The positions of these NDBC buoys are illustrated in Fig. 11.
Moreover, the buoy data with high temporal resolution (10 min) are used. Consequently, the IALT retrieved wind speed can be collocated with the buoy data using a temporal window of ±5 min. To acquire as many buoy collocations as possible, the spatial window of IALT-buoy collocations is set to 50 km. Finally, the IALT data from December 2016 to March 2017 are used, and 15 matchup samples with the buoys are obtained, as shown in Table 2. The rightmost column of Table 2 indicates that all of the wind speed differences are within ±2 m/s. The retrieved wind speeds have a mean bias of 0.144 m/s and a RMS of 1.32 m/s.
| Buoy No. | UTC time | IALT wind speed/m·s–1 | Buoy wind speed/m·s–1 | Wind speed difference/m·s–1 |
| 51000 | 2016/12/04 12:26:35.4 | 11.15 | 9.8 | 1.35 |
| 2016/12/04 12:26:38.9 | 10.89 | 9.8 | 1.09 | |
| 2016/12/04 12:26:42.3 | 11.78 | 9.8 | 1.98 | |
| 2016/12/04 12:26:45.8 | 11.39 | 9.8 | 1.59 | |
| 2017/03/19 11:36:46.5 | 3.63 | 3.8 | –0.17 | |
| 2017/03/19 11:36:49.9 | 4.57 | 3.8 | 0.77 | |
| 2017/03/19 11:36:53.4 | 3.60 | 3.8 | –0.2 | |
| 51004 | 2016/12/05 11:23:49.4 | 7.46 | 6.8 | 0.66 |
| 2016/12/05 11:23:52.9 | 5.36 | 6.8 | –1.44 | |
| 2016/12/05 11:23:56.3 | 6.15 | 6.8 | –0.65 | |
| 2016/12/16 06:26:15.1 | 4.58 | 3.3 | 1.28 | |
| 2016/12/16 06:26:18.5 | 4.58 | 3.3 | 1.28 | |
| 51101 | 2017/03/15 14:03:43.9 | 3.94 | 5.8 | –1.86 |
| 2017/03/15 14:03:47.4 | 4.03 | 5.8 | –1.77 | |
| 2017/03/15 14:03:50.8 | 4.05 | 5.8 | –1.75 |
DownLoad:
CSV
In March 2017, the National Satellite Ocean Application Service (NSOAS) carried out several in-situ measurements on the south of Hainan Island in order to verify the observation accuracy of IALT on Tiangong-2 space station. The in-situ wind speeds were measured on March 16, 19, and 22 respectively. The ship positions, the actual wind speed, and the retrieved wind speed are shown in Fig. 12. The comparison with ship measurements are presented in Table 3.
| Station No. | North latitude/(°) | East longitude/(°) | Beijing time | IALT wind speed/m·s–1 | In-situ wind speed/m·s–1 | Wind speed difference/m·s–1 |
| TB01 | 18.1011 | 109.2353 | 2017/3/16 12:15 | 6.9 | 8.3 | –1.4 |
| TB02 | 18.1026 | 110.0554 | 2017/3/19 10:31 | 3.3 | 2.8 | 0.5 |
| TB03 | 19.0894 | 110.6672 | 2017/3/22 09:17 | 6.2 | 3.3 | 2.9 |
DownLoad:
CSV
Note that the positions of in-situ measurements are nearshore, which is primarily designed to validate the sea surface height measurements of IALT. From Table 3 we know that the wind speed difference on March 22 is about 2.9 m/s, and the wind speed differences on March 16 and 19 are smaller than ±2 m/s. The main objective of TB03 measurements is to validate the ocean color retrieved from the Multi-angle wide band imager on Tiangong-2, such that TB03 is about 10 km offshore, which is much closer to coast than TB01 and TB02. This may be the main reason that the estimation error at TB03 is quite larger than that at TB01 and TB02.
IALT is a new generation of radar altimeter system, which operates at Ku band microwave frequency. It can be used to derive the sea surface height, the significant wave height, the sea surface wind speed, and the sea wave directional spectrum simultaneously. Comparing with the traditional radar altimeters such as Jason-2 and HY-2A altimeter, its swath is wider and its spatial resolution is higher.
In this paper, we briefly introduce the IALT radar system on board Tiangong-2 space station. Then a neural network algorithm is used to train the inversion model, and consequently is applied to the wind speed inversion of IALT. In order to assess the accuracy of IALT wind speeds, we compare the retrieved wind speed with the ECMWF, buoy, and in-situ measurements. The RMS error of the retrieved wind speed is about 1.85 m/s, and the mean bias is smaller than 0.21 m/s, implying that the accuracy of IALT wind speed satisfies its specification.
Note also that the retrieved wind speeds are lower than the ECMWF reanalysis wind speeds under high wind speed condition (Fig. 10), especially for wind speed larger than 15 m/s. While at low wind speeds, the retrieved wind speeds are a little overestimated. The reasons could be: first, the small amount of training samples limits the wind speed range in neural network algorithm. Second, the quality control is imperfect at present. The future work will focus on the improvement of the wind speed inversion method, as well as the selection of the training samples.
| [1] |
Apel J R. 1994. An improved model of the ocean surface wave vector spectrum and its effects on radar backscatter. Journal of Geophysical Research: Oceans, 99(C8): 16269–16291. doi: 10.1029/94JC00846
|
| [2] |
Bao Lifeng, Gao Peng, Peng Hailong, et al. 2015. First accuracy assessment of the HY-2A altimeter sea surface height observations: cross-calibration results. Advances in Space Research, 55(1): 90–105
|
| [3] |
Bao Qingliu, Zhang Youguang, An Wentao, et al. 2016a. Sea surface wind speed inversion using low incident NRCS. In: Proceedings of 2016 IEEE International Geoscience and Remote Sensing Symposium (IGARSS). Beijing: IEEE, 4619–4622
|
| [4] |
Bao Qingliu, Zhang Youguang, Lang Shuyan, et al. 2016b. Sea surface wind speed inversion using the low incident NRCS measured by TRMM precipitation radar. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 9(11): 5262–5271. doi: 10.1109/JSTARS.2016.2581215
|
| [5] |
Barrick D. 1974. Wind dependence of quasi-specular microwave sea scatter. IEEE Transactions on Antennas and Propagation, 22(1): 135–136. doi: 10.1109/TAP.1974.1140736
|
| [6] |
Brown G S. 1990. Quasi-Specular scattering from the air-sea interface. In: Geernaert G L, Plant W L, eds. Surface Waves and Fluxes. Dordrecht: Springer, 1–39
|
| [7] |
Cox C, Munk W. 1954. Statistics of the sea surface derived from sun glitter. Journal of Marine Research, 13(2): 198–227
|
| [8] |
Fan Chenqing, Zhang Jie, Meng Junmin, et al. 2014. Significant wave height operational inversion algorithm of HY-2A altimeter. Haiyang Xuebao (in Chinese), 36(3): 121–126
|
| [9] |
Freilich M H, Vanhoff B A. 2003. The relationship between winds, surface roughness, and radar backscatter at low incidence angles from TRMM precipitation radar measurements. Journal of Atmospheric and Oceanic Technology, 20(4): 549–562
|
| [10] |
Hesany V, Plant W J, Keller W C. 2004. The normalized radar cross section of the sea at 10° incidence. IEEE Transactions on Geoscience and Remote Sensing, 38(1): 64–72
|
| [11] |
Holliday D, St-Cyr G, Woods N E. 1986. A radar ocean imaging model for small to moderate incidence angles. International Journal of Remote Sensing, 7(12): 1809–1834. doi: 10.1080/01431168608948971
|
| [12] |
Kozu T, Kawanishi T, Kuroiwa H, et al. 2001. Development of precipitation radar onboard the Tropical Rainfall Measuring Mission (TRMM) satellite. IEEE Transactions on Geoscience and Remote Sensing, 39(1): 102–116. doi: 10.1109/36.898669
|
| [13] |
Li Li, Im E, Connor L N, et al. 2004. Retrieving ocean surface wind speed from the TRMM Precipitation Radar measurements. IEEE Transactions on Geoscience and Remote Sensing, 42(6): 1271–1282. doi: 10.1109/TGRS.2004.828924
|
| [14] |
Lin Wenming, Portabella M, Stoffelen A, et al. 2014. Rain identification in ASCAT winds using singularity analysis. IEEE Geoscience and Remote Sensing Letters, 11(9): 1519–1523. doi: 10.1109/LGRS.2014.2298095
|
| [15] |
Lin Wenming, Portabella M, Stoffelen A, et al. 2015. ASCAT wind quality control near rain. IEEE Transactions on Geoscience and Remote Sensing, 53(8): 4165–4177. doi: 10.1109/TGRS.2015.2392372
|
| [16] |
Tsai W T, Spencer M, Wu C, et al. 2000. SeaWinds on QuikSCAT: sensor description and mission overview. In: Proceedings of IEEE 2000 International Geoscience and Remote Sensing Symposium, Taking the Pulse of the Planet: The Role of Remote Sensing in Managing the Environment. Honolulu: IEEE, 1021–1023
|
| [17] |
Valenzuela G R. 1978. Theories for the interaction of electromagnetic and oceanic waves: A review. Boundary-Layer Meteorology, 13(1): 61–85
|
| [18] |
Xu Yuan, Yang Jingsong, Zheng Gang, et al. 2014. Calibration and verification of sea surface wind speed from satellite altimeters. Haiyang Xuebao (in Chinese), 36(7): 125–132
|
| [19] |
Zhang Yunhua, Zhang Xiangkun, Meng Xin, et al. 2007. An interferometric imaging altimeter applied for both ocean and land observation. In: Proceedings of 2007 IEEE International Geoscience and Remote Sensing Symposium. Barcelona: IEEE, 3821–3824
|
| 1. | Xiaofan Hong, Zuozhi Chen, Jun Zhang, et al. Construction and analysis of a coral reef trophic network for Qilianyu Islands, Xisha Islands. Acta Oceanologica Sinica, 2022, 41(12): 58. doi:10.1007/s13131-022-2047-8 | |
| 2. | Guo Li, Yunhua Zhang, Xiao Dong. Approaches for Joint Retrieval of Wind Speed and Significant Wave Height and Further Improvement for Tiangong-2 Interferometric Imaging Radar Altimeter. Remote Sensing, 2022, 14(8): 1930. doi:10.3390/rs14081930 | |
| 3. | Guo Li, Yunhua Zhang, Xiao Dong. New Ku-Band Geophysical Model Function and Wind Speed Retrieval Algorithm Developed for Tiangong-2 Interferometric Imaging Radar Altimeter. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2022, 15: 1981. doi:10.1109/JSTARS.2022.3150390 |
Figures(12) / Tables(3)
Supported by:
Beijing Renhe Information Technology Co. Ltd
Youguang Zhang, Qingliu Bao, Mingsen Lin, Shuyan Lang. Wind speed inversion and in-orbit assessment of the imaging altimeter on Tiangong-2 space station[J]. Acta Oceanologica Sinica, 2020, 39(12): 114-120. doi: 10.1007/s13131-020-1687-9
| Orbit parameters | |
| Altitude/km | 378.6 |
| Velocity/km·s–1 | 7.3 |
| System parameters | |
| Polarization | VV |
| Carrier frequency/GHz | 13.58 |
| Peak power/W | 300 |
| Baseline/m | 2.3 |
| Band width/MHz | 40 |
| Incident angle/(°) | 2.5–8 |
| Swath/km | 30 |
| Spatial resolution/m | 20–60 |
DownLoad:
CSV
| Buoy No. | UTC time | IALT wind speed/m·s–1 | Buoy wind speed/m·s–1 | Wind speed difference/m·s–1 |
| 51000 | 2016/12/04 12:26:35.4 | 11.15 | 9.8 | 1.35 |
| 2016/12/04 12:26:38.9 | 10.89 | 9.8 | 1.09 | |
| 2016/12/04 12:26:42.3 | 11.78 | 9.8 | 1.98 | |
| 2016/12/04 12:26:45.8 | 11.39 | 9.8 | 1.59 | |
| 2017/03/19 11:36:46.5 | 3.63 | 3.8 | –0.17 | |
| 2017/03/19 11:36:49.9 | 4.57 | 3.8 | 0.77 | |
| 2017/03/19 11:36:53.4 | 3.60 | 3.8 | –0.2 | |
| 51004 | 2016/12/05 11:23:49.4 | 7.46 | 6.8 | 0.66 |
| 2016/12/05 11:23:52.9 | 5.36 | 6.8 | –1.44 | |
| 2016/12/05 11:23:56.3 | 6.15 | 6.8 | –0.65 | |
| 2016/12/16 06:26:15.1 | 4.58 | 3.3 | 1.28 | |
| 2016/12/16 06:26:18.5 | 4.58 | 3.3 | 1.28 | |
| 51101 | 2017/03/15 14:03:43.9 | 3.94 | 5.8 | –1.86 |
| 2017/03/15 14:03:47.4 | 4.03 | 5.8 | –1.77 | |
| 2017/03/15 14:03:50.8 | 4.05 | 5.8 | –1.75 |
DownLoad:
CSV
| Station No. | North latitude/(°) | East longitude/(°) | Beijing time | IALT wind speed/m·s–1 | In-situ wind speed/m·s–1 | Wind speed difference/m·s–1 |
| TB01 | 18.1011 | 109.2353 | 2017/3/16 12:15 | 6.9 | 8.3 | –1.4 |
| TB02 | 18.1026 | 110.0554 | 2017/3/19 10:31 | 3.3 | 2.8 | 0.5 |
| TB03 | 19.0894 | 110.6672 | 2017/3/22 09:17 | 6.2 | 3.3 | 2.9 |
DownLoad:
CSV
| Orbit parameters | |
| Altitude/km | 378.6 |
| Velocity/km·s–1 | 7.3 |
| System parameters | |
| Polarization | VV |
| Carrier frequency/GHz | 13.58 |
| Peak power/W | 300 |
| Baseline/m | 2.3 |
| Band width/MHz | 40 |
| Incident angle/(°) | 2.5–8 |
| Swath/km | 30 |
| Spatial resolution/m | 20–60 |
| Buoy No. | UTC time | IALT wind speed/m·s–1 | Buoy wind speed/m·s–1 | Wind speed difference/m·s–1 |
| 51000 | 2016/12/04 12:26:35.4 | 11.15 | 9.8 | 1.35 |
| 2016/12/04 12:26:38.9 | 10.89 | 9.8 | 1.09 | |
| 2016/12/04 12:26:42.3 | 11.78 | 9.8 | 1.98 | |
| 2016/12/04 12:26:45.8 | 11.39 | 9.8 | 1.59 | |
| 2017/03/19 11:36:46.5 | 3.63 | 3.8 | –0.17 | |
| 2017/03/19 11:36:49.9 | 4.57 | 3.8 | 0.77 | |
| 2017/03/19 11:36:53.4 | 3.60 | 3.8 | –0.2 | |
| 51004 | 2016/12/05 11:23:49.4 | 7.46 | 6.8 | 0.66 |
| 2016/12/05 11:23:52.9 | 5.36 | 6.8 | –1.44 | |
| 2016/12/05 11:23:56.3 | 6.15 | 6.8 | –0.65 | |
| 2016/12/16 06:26:15.1 | 4.58 | 3.3 | 1.28 | |
| 2016/12/16 06:26:18.5 | 4.58 | 3.3 | 1.28 | |
| 51101 | 2017/03/15 14:03:43.9 | 3.94 | 5.8 | –1.86 |
| 2017/03/15 14:03:47.4 | 4.03 | 5.8 | –1.77 | |
| 2017/03/15 14:03:50.8 | 4.05 | 5.8 | –1.75 |
| Station No. | North latitude/(°) | East longitude/(°) | Beijing time | IALT wind speed/m·s–1 | In-situ wind speed/m·s–1 | Wind speed difference/m·s–1 |
| TB01 | 18.1011 | 109.2353 | 2017/3/16 12:15 | 6.9 | 8.3 | –1.4 |
| TB02 | 18.1026 | 110.0554 | 2017/3/19 10:31 | 3.3 | 2.8 | 0.5 |
| TB03 | 19.0894 | 110.6672 | 2017/3/22 09:17 | 6.2 | 3.3 | 2.9 |
DownLoad:
DownLoad:
