In-Orbit Calibration Uncertainty of Microwave Radiation Imager Onboard Fengyun-3C

风云三号C星微波成像仪在轨定标不确定度分析

+ Author Affiliations + Find other works by these authors
  • Corresponding author: Wanting MENG, wanting_meng@163.com
  • Funds:

    Supported by the National Key Research and Development Program of China (2018YFB0504900 and 2018YFB0504902), National Natural Science Foundation of China (41805024 and 42005105), and Joint Open Research Fund of the State key Laboratory of Hydroscience and Engineering and Tsinghua–Ningxia Yinchuan Joint Institute of Internet of Waters on Digital Water Governance (sklhse-2021-Iow08)

  • doi: 10.1007/s13351-021-0220-1

PDF

  • This study evaluates the in-orbit calibration uncertainty (CU) for the microwave radiation imager (MWRI) onboard China polar-orbiting meteorological satellite Fengyun-3C (FY-3C). The uncertainty analysis of MWRI provides a direct link to the calibration system of the sensor and quantifies the calibration confidence based on the prelaunch and postlaunch measurements. The unique design of the sensor makes the uncertainty in the calibration of the sensor highly correlated to the uncertainty in the brightness temperature (TB) measured at the hot view, while the cold view has negligible impacts on the calibration confidence. Lack of knowledge on the emission of the hot-load reflector hampers the calibration quality of MWRI significantly in the descending passes of the orbits when the hot-load reflector is heated nonuniformly by the solar illumination. Radiance contamination, which originates from the satellite and in-orbit environments, could enter the primary reflector via the hot view and further impinge on the CU, especially at the 10.65-GHz channels where the main-beam width is much broader than that of higher-frequency channels. The monthly-mean CU is lower than 2 K at all channels, depending on the observed earth scenes and in-orbit environments, and the month-to-month variation of CU is also noticed for all channels. Due to the uncertainty in the emissive hot-load reflector, CU in the descending passes is generally larger than that of the ascending orbits. Moreover, up to 1-K CU difference between the ocean and land scenes is found for the 10.65-GHz channels, while at 89 GHz, the CU difference between the land and ocean scenes is less than 0.1 K.
    本研究定量评估了中国极轨气象卫星风云三号C星微波成像仪(MWRI)的在轨定标不确定度。基于发射前后观测数据,建立与传感器定标系统的直接关联,量化MWRI不确定度标定。MWRI的独特设计使定标不确定性与两点定标法中“热点”亮温不确定性密切相关,而“冷点”对不确定度的影响可忽略。热反射面由于太阳光照导致表面温度不均匀,缺乏准确的热反射面自发射,使仪器的在轨定标质量随季节和纬度发生变化。卫星平台和在轨环境的辐射污染影响了仪器的定标不确定度,尤其是主波束较宽的10.65 GH。微波成像仪所有通道的月平均定标不确定度均低于2K,并与观察场景和星上环境相关。10.65GHz的海洋和陆地场景之间的定标不确定度差异最大为1K,而89GHz的陆地和海洋场景之间的定标不确定度差异小于0.1K。
  • 加载中
  • Fig. 1.  Schematic diagram of the MWRI calibration system (Xie et al., 2019). The hot-view brightness temperature (TB) is composed of the (a) hot-load emission, (b) emissive hot-load reflector, and (c) backlobe intrusion from the earth scenes.

    Fig. 2.  Uncertainty components at the hot view as a function of latitude for the 10.65 V channel along the descending pass of the orbit on 1 January 2019.

    Fig. 3.  As in Fig. 2, but for the ascending pass.

    Fig. 4.  Calibration uncertainties of MWRI in the descending tracks on 1 Januray 2019.

    Fig. 5.  As in Fig. 4, but for the ascending tracks.

    Fig. 6.  MWRI one-month mean calibration uncertainty (CU) in January 2019 over land and ocean for the ascending (ASC) and descending (DES) phases of the orbits.

    Fig. 7.  MWIR month-to-month CU variation over the land and ocean during the period of August 2018 and July 2019 for the ASC and DES tracks.

    Table 1.  Nominal values and perturbations of the calibration parameters at the hot view of FY-3C MWRI

    Channel (GHz)Nominal value Perturbation
    ηTηHεHThot* (K)TET* (K)ΔηTΔηHΔεHΔThot* (K)ΔTET* (K)
    10.65 V 0.96520.97340.0434328.0188.00.00410.0050.002617.34.57
    10.65 H 0.96930.97340.0875131.60.00270.0050.00687.26
    18.7 V 0.96970.98610.0503202.90.00890.0020.00336.40
    18.7 H 0.96760.98610.0707151.90.00450.0020.006710.8
    23.8 V 0.99090.99020.0389215.10.00440.0020.00408.13
    23.8 H 0.99170.99020.0541173.10.00360.0020.007914.0
    36.5 V 0.98510.99400.0437214.60.00200.0020.00287.80
    36.5 H 0.99050.99400.0615172.30.00200.0020.006313.9
    89 V0.9980.99260.0342240.80.00200.0020.00338.43
    89 H0.9980.99260.0406214.50.00200.0020.007416.2
    *The value shown in Table 1 is averaged over one descending track of 1 Januray 2019.
    Download: Download as CSV
  • [1]

    Bates, J. J., J. L. Privette, E. J. Kearns, et al., 2016: Sustained production of multidecadal climate records: Lessons from the NOAA climate data record program. Bull. Amer. Meteor. Soc., 97, 1573–1581. doi: 10.1175/BAMS-D-15-00015.1.
    [2]

    Burgdorf, M., I. Hans, J. Mittaz, et al., 2019: FIDUCEO Project: (Microwave): Report on the MW FCDR: Uncertainty. Accessed on 10 August 2021, available online at http://cedadocs.ceda.ac.uk/id/eprint/1402.
    [3]

    Carminati, F., B. Candy, W. Bell, et al., 2018: Assessment and assimilation of FY-3 humidity sounders and imager in the UK Met Office global model. Adv. Atmos. Sci., 35, 942–954. doi: 10.1007/s00376-018-7266-8.
    [4]

    Carminati, F., N. Atkinson, B. Candy, et al., 2020: Insights into the microwave instruments onboard the Fengyun 3D satellite: Data quality and assimilation in the Met Office NWP system. Adv. Atmos. Sci. . doi: 10.1007/s00376-020-0010-1.
    [5]

    Fox, N., 2010: A guide to expression of uncertainty of measurements. In A Quality Assurance Framework for Earth Observation (QA4EO). Ed., GEO, Geneva, Switzerland, 9 pp. Accessed on 10 August 2021, available online at http://www.qa4eo.org/docs/QA4EO-QAEO-GEN-DQK-006_v4.0.pdf.
    [6]

    Gu, S. Y., Y. Guo, Z. Z. Wang, et al., 2012: Calibration analyses for sounding channels of MWHS onboard FY-3A. IEEE Trans. Geosci. Remote Sens., 50, 4885–4891. doi: 10.1109/TGRS.2012.2214391.
    [7]

    Hewison, T. J., 2013: An evaluation of the uncertainty of the GSICS SEVIRI-IASI intercalibration products. IEEE Trans. Geosci. Remote Sens., 51, 1171–1181. doi: 10.1109/TGRS.2012.2236330.
    [8]

    Hollmann, R., C. J. Merchant, R. Saunders, et al., 2013: The ESA climate change initiative: Satellite data records for essential climate variables. Bull. Amer. Meteor. Soc., 94, 1541–1552. doi: 10.1175/BAMS-D-11-00254.1.
    [9]

    Kang, C. S., T. J. Zhao, J. C. Shi, et al., 2021: Global soil moisture retrievals from the Chinese FY-3D Microwave Radiation Imager. IEEE Trans. Geosci. Remote Sens., 59, 4018–4032. doi: 10.1109/TGRS.2020.3019408.
    [10]

    Srianand, R., P. Petitjean, and C. Ledoux, 2000: The cosmic microwave background radiation temperature at a redshift of 2.34. Nature, 408, 931–935. doi: 10.1038/35050020.
    [11]

    Tang, F., and X. L. Zou, 2017: Liquid water path retrieval using the lowest frequency channels of Fengyun-3C Microwave Radiation Imager (MWRI). J. Meteor. Res., 31, 1109–1122. doi: 10.1007/s13351-017-7012-7.
    [12]

    Tobin, D., H. Revercomb, R. Knuteson, et al., 2013: Suomi-NPP CrIS radiometric calibration uncertainty. J. Geophys. Res. Atmos., 118, 10,589–10,600. doi: 10.1002/jgrd.50809.
    [13]

    Wang, Y., Y. F. Fu, X. Fang, et al., 2014: Estimating ice water path in tropical cyclones with multispectral microwave data from the FY-3B satellite. IEEE Trans. Geosci. Remote Sens., 52, 5548–5557. doi: 10.1109/TGRS.2013.2290320.
    [14]

    Weng, F. Z., and H. Yang, 2016: Validation of ATMS calibration accuracy using Suomi NPP pitch maneuver observations. Remote Sens., 8, 332. doi: 10.3390/rs8040332.
    [15]

    Wu, B. H., Y. Wang, C.-Z. Zou, et al., 2020: A fundamental climate data record derived from AMSR-E, MWRI, and AMSR2. IEEE Trans. Geosci. Remote Sens., 58, 5450–5461. doi: 10.1109/TGRS.2020.2966055.
    [16]

    Xie, X. X., S. L. Wu, H. X. Xu, et al., 2019: Ascending–descending bias correction of microwave radiation imager on board FengYun-3C. IEEE Trans. Geosci. Remote Sens., 57, 3126–3134. doi: 10.1109/TGRS.2018.2881094.
    [17]

    Xie, X. X., K. S. Dong, W. M. Yu, et al., 2020: In-orbit calibration of FengYun-3C microwave radiation imager: Nonlinearity correction. IEEE Trans. Geosci. Remote Sens. . doi: 10.1109/TGRS.2020.3034302.
    [18]

    Xie, X. X., W. T. Meng, K. S. Dong, et al., 2021: In-orbit calibration of FengYun-3C microwave radiation imager: Characterization of backlobe intrusion for the hot-load reflector. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. . doi: 10.1109/JSTARS.2021.3075969.
    [19]

    Yang, H., F. Z. Weng, L. Q. Lyu, et al., 2011: The FengYun-3 Microwave Radiation Imager on-orbit verification. IEEE Trans. Geosci. Remote Sens., 49, 4552–4560. doi: 10.1109/TGRS.2011.2148200.
    [20]

    Yang, Z. D., P. Zhang, S. Y. Gu, et al., 2019: Capability of Fengyun-3D satellite in earth system observation. J. Meteor. Res., 33, 1113–1130. doi: 10.1007/s13351-019-9063-4.
    [21]

    Zhang, Y., Y. Wang, G. S. Liu, et al., 2019: Satellite-based assessment of various cloud microphysics schemes in simulating typhoon hydrometeors. Adv. Meteor., 3168478. doi: 10.1155/2019/3168478.
  • 加载中
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

In-Orbit Calibration Uncertainty of Microwave Radiation Imager Onboard Fengyun-3C

    Corresponding author: Wanting MENG, wanting_meng@163.com
  • 1. School of Atmospheric Sciences, Sun Yat-Sen University, Zhuhai 519082
  • 2. Key Laboratory of Tropical Atmosphere–Ocean System, Ministry of Education, Zhuhai 519082
  • 3. Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai), Zhuhai 519082
  • 4. Shanghai Spaceflight Institute of TT&C and Telecommunication, Shanghai 201109
Funds: Supported by the National Key Research and Development Program of China (2018YFB0504900 and 2018YFB0504902), National Natural Science Foundation of China (41805024 and 42005105), and Joint Open Research Fund of the State key Laboratory of Hydroscience and Engineering and Tsinghua–Ningxia Yinchuan Joint Institute of Internet of Waters on Digital Water Governance (sklhse-2021-Iow08)

Abstract: This study evaluates the in-orbit calibration uncertainty (CU) for the microwave radiation imager (MWRI) onboard China polar-orbiting meteorological satellite Fengyun-3C (FY-3C). The uncertainty analysis of MWRI provides a direct link to the calibration system of the sensor and quantifies the calibration confidence based on the prelaunch and postlaunch measurements. The unique design of the sensor makes the uncertainty in the calibration of the sensor highly correlated to the uncertainty in the brightness temperature (TB) measured at the hot view, while the cold view has negligible impacts on the calibration confidence. Lack of knowledge on the emission of the hot-load reflector hampers the calibration quality of MWRI significantly in the descending passes of the orbits when the hot-load reflector is heated nonuniformly by the solar illumination. Radiance contamination, which originates from the satellite and in-orbit environments, could enter the primary reflector via the hot view and further impinge on the CU, especially at the 10.65-GHz channels where the main-beam width is much broader than that of higher-frequency channels. The monthly-mean CU is lower than 2 K at all channels, depending on the observed earth scenes and in-orbit environments, and the month-to-month variation of CU is also noticed for all channels. Due to the uncertainty in the emissive hot-load reflector, CU in the descending passes is generally larger than that of the ascending orbits. Moreover, up to 1-K CU difference between the ocean and land scenes is found for the 10.65-GHz channels, while at 89 GHz, the CU difference between the land and ocean scenes is less than 0.1 K.

风云三号C星微波成像仪在轨定标不确定度分析

本研究定量评估了中国极轨气象卫星风云三号C星微波成像仪(MWRI)的在轨定标不确定度。基于发射前后观测数据,建立与传感器定标系统的直接关联,量化MWRI不确定度标定。MWRI的独特设计使定标不确定性与两点定标法中“热点”亮温不确定性密切相关,而“冷点”对不确定度的影响可忽略。热反射面由于太阳光照导致表面温度不均匀,缺乏准确的热反射面自发射,使仪器的在轨定标质量随季节和纬度发生变化。卫星平台和在轨环境的辐射污染影响了仪器的定标不确定度,尤其是主波束较宽的10.65 GH。微波成像仪所有通道的月平均定标不确定度均低于2K,并与观察场景和星上环境相关。10.65GHz的海洋和陆地场景之间的定标不确定度差异最大为1K,而89GHz的陆地和海洋场景之间的定标不确定度差异小于0.1K。
1.   Introduction
  • Microwave radiation imager (MWRI) onboard China second generation polar-orbiting meteorological satellite Fengyun-3 (FY-3) has been supporting Earth sciences and severe weather monitoring (Wang et al., 2014; Tang and Zou, 2017; Zhang et al., 2019; Kang et al., 2021). Operating at the frequencies of 10.65, 18.7, 23.8, 36.5, and 89 GHz, each of which has dual polarization, the sensor has provided continuous global observations for more than one decade and currently the consecutive MWRI aboard FY-3D launched in November 2017 is operational in orbit (Yang et al., 2019).

    As one of the essential sensors in microwave radiometer constellation, the calibration accuracy of MWRI is extremely important for the demanding weather and climate applications. The four MWRIs deployed on FY-3A/B/C/D have the same design and architecture and all are calibrated by the algorithm introduced by Yang et al. (2011). Strong efforts have been made to mitigate the calibration anomalies originating from MWRI hot-load reflector, via the correction of its in-orbit calibration parameters and intercalibration with other microwave radiometers (Carminati et al., 2018; Xie et al., 2019; Wu et al., 2020). To ensure a consistent long-term observation of FY-3C MWRI, special handling has been applied to the nonlinearity system of the instrument which deviated from its optimal status at the time of power-ON/OFF (Xie et al., 2020). The successful removal of the calibration bias further promotes the future use of MWRI observations in numerical weather predictions and climate reanalysis systems (Carminati et al., 2020).

    Uncertainty in the calibration accuracy reflects the lack of exact knowledge of the calibration system during observations. Understanding the calibration uncertainty (CU) of space-borne sensors is required when it comes to accurately characterize the global radiation data records for weather and climate research (Hollmann et al., 2013; Bates et al., 2016). To assess the quantitative statement and evaluate the confidence of the data quality, parameters at each calibration step have to be reviewed during the calibration process of the sensor (Fox, 2010; Hewison, 2013; Tobin et al., 2013). Similar techniques were applied to the correction of the Advanced Technology Microwave Sounder (ATMS) aboard the Suomi National Polar-orbiting Partnership (SNPP) Satellite by analyzing the error budget of the instrument (Weng and Yang, 2016). More recently, the fidelity and uncertainty in climate data records from Earth observations (FIDUCEO) project conducted a comprehensive overview on the uncertainty analysis of Special Sensor Microwave-Humidity (SSM/T2), Advanced Microwave Sounding Unit-B (AMSU-B), and Microwave Humidity Sounder (MHS), leading to metrological traceability for the fundamental climate data records (FCDRs; Burgdorf et al., 2019).

    This paper intends to analyze the calibration uncertainty of MWRI and assessits the in-orbit calibration confidence, focusing on the calibration traceability of the sensor. With the same design and configuration, differences among the four MWRIs aboard FY-3A/B/C/D are supposed to be minimal and FY-3C MWRI is taken as an example sensor in this study. Followed by a brief overview of the calibration algorithm, we first introduce the CU analysis model of MWRI and then quantify the impacts of the calibration parameters at the hot and cold views in Section 2. The in-orbit CU of MWRI and its temporal variability are discussed in detail in Section 3. Finally, Section 4 concludes this study.

2.   Uncertainty estimation model of MWRI
  • MWRI aboard FY-3C is calibrated with the conventional two-point calibration method by viewing the hot-load and cold sky alternately (Fig. 1). The detailed calibration algorithm has been documented in Yang et al. (2011). The received brightness temperature (TB) of MWRI is composed of the linear and nonlinear parts and expressed as follows,

    Figure 1.  Schematic diagram of the MWRI calibration system (Xie et al., 2019). The hot-view brightness temperature (TB) is composed of the (a) hot-load emission, (b) emissive hot-load reflector, and (c) backlobe intrusion from the earth scenes.

    $$\hspace{15pt} {\rm{TB}}_{\rm{E}}={\rm{TB}}_{\rm{C}}+\frac{{\mathrm{TB}}_{\mathrm{W}}-\mathrm{T}{\mathrm{B}}_{\mathrm{C}}}{V_{\mathrm{W}}-V_{\mathrm{C}}}(V_{\mathrm{E}}-V_{\mathrm{C}})+\mathrm{T}{\mathrm{B}}_{\mathrm{N}\mathrm{L}}, $$ (1)

    where the subscripts E, C, and W denote the earth, cold, and hot views, respectively; V is the MWRI-observed voltage; and TBNL is the nonlinear TB depending on the operation status of the receiver system (Yang et al., 2011; Gu et al., 2012),

    $$\hspace{15pt} \mathrm{T}{\mathrm{B}}_{\mathrm{N}\mathrm{L}}=\mathrm{\mu }\frac{(V_{\mathrm{E}}-V_{\mathrm{C}})(V_{\mathrm{E}}-V_{\mathrm{W}})}{(V_{\mathrm{W}}-V_{\mathrm{C}}{)}^{2}}(\mathrm{T}{\mathrm{B}}_{\mathrm{W}}-\mathrm{T}{\mathrm{B}}_{\mathrm{C}}{)}^{2}, $$ (2)

    where μ is the nonlinear parameter.

    The CU analysis model of MWRI is built upon the confidence in the calibration process where the uncertainty of each calibration parameter propagates to CU associated with the observed scenes and in-orbit environments. This work follows the guidance provided by Fox, 2010. For each calibration parameter, the standard uncertainty is referred as $\Delta {X_i}$ and the corresponding sensitivity $ \dfrac{\partial {\rm T{B}_{E}}}{\partial {X}_{i}} $ of the earth-view TB to the perturbation of each calibration parameter can be derived according to Eqs. (1, 2). Thus, ui, the uncertainty of the estimated earth-view TBE resulted from the calibration parameter can be given,

    $$\hspace{1pt} {u}_{i}\left({X}_{i}\right)=\frac{\partial {\mathrm{T}\mathrm{B}}_{\mathrm{E}}}{\partial {X}_{i}} \Delta {X}_{i},\;{X}_{i}={\mathrm{T}\mathrm{B}}_{\mathrm{W}},{\mathrm{T}\mathrm{B}}_{\mathrm{C}},V_{\mathrm{W}},V_{\mathrm{C}},V_{\mathrm{E}},\; \mathrm{or}\; \mu, \!\!\! $$ (3)

    with $ {X}_{i} $ equals to $ {\mathrm{T}\mathrm{B}}_{\mathrm{W}},{\mathrm{T}\mathrm{B}}_{\mathrm{C}},V_{\mathrm{W}},V_{\mathrm{C}},V_{\mathrm{E}},\;\mathrm{o}\mathrm{r}\;\mu $ the calibration parameters defined in Eqs. (1, 2).

    The uncertainty ui at each calibration parameter contributes to the CU of MWRI-observed TBE via,

    $$\hspace{1pt} {\mathrm{\Delta }\mathrm{T}\mathrm{B}}_{\rm E}=\sqrt{\sum _{i}{\left({u}_{i}\right)}^{2}}. $$ (4)

    For MWRI, the propagation path of microwave radiance is individual for the hot/cold/Earth views in the calibration system and no interference has been detected between each two channels. The correlation between the calibration parameters is thus proved to be negligibly small.

  • TB at the cold view is composed of the cosmic background radiation reflected by the cold-sky mirror and the emission of the cold-sky mirror. With the cold mirror emissivity (εC), cosmic background (TCS), and measured temperature of the cold mirror (TCM), TBC is given by,

    $$\hspace{1pt} \mathrm{T}{\mathrm{B}}_{\mathrm{C}}=(1-{\varepsilon }_{{\rm{C}}}){T}_{\mathrm{C}\mathrm{S}}+{\varepsilon }_{{\rm{C}}}{T}_{\mathrm{C}\mathrm{M}}, $$ (5)

    and the TB uncertainty at the cold view is derived as follows,

    $$\hspace{1pt} {\mathrm{\Delta }\mathrm{T}\mathrm{B}}_{\mathrm{C}}=\sqrt{\sum _{j}{\left(\frac{\partial {\mathrm{T}\mathrm{B}}_{\mathrm{C}}}{\partial {X}_{j}}\Delta {X}_{j}\right)}^{2}}, \; {X}_{j}={\varepsilon }_{{\rm{C}}},{T}_{\mathrm{C}\mathrm{S}},\;\mathrm{a}\mathrm{n}\mathrm{d}\;{T}_{\mathrm{C}\mathrm{M}}, \!\!\! $$ (6)

    where the partial derivative term $\dfrac{{\partial {\rm{T}}{{\rm{B}}_{\rm{C}}}}}{{\partial {{X}_{j}}}}$ is the sensitivity of the TB at the cold view to the calibration parameter ${{X}_{j}}$, and Δ denotes the perturbation of each parameter.

    The cold mirror was made from pure aluminum with high conductivity on the order of 107 Siemens m−1, which was measured on-ground in the prelaunch phase. Thus, the emissivity of the cold mirror is rather low as expected, varying from 0.000357 at the 10.65-GHz channel to 0.001031 at the 89-GHz channel (Xie et al., 2019). Meanwhile, the cosmic background has been proved to be with high precision at the level of 0.01 K (Srianand et al., 2000). Taking advantage of the fact that the emissivity of the cold mirror and the cosmic background are well-defined in the previous analysis and measurements, the uncertainty in the cold-view TB is thus trivial and it is not the predominant factor of the CU.

  • The design of MWRI is different from other microwave imagers in orbit, with a unique hot-load reflector deployed in the calibration system. According to the architecture of MWRI, three components (i.e., the emission from the hot load, emission, and backlobe intrusion of the hot-load reflector), contribute to the TB at the hot view (Fig. 1). Therefore, TBH the TB at the hot view is composed of three parts,

    $$\hspace{5pt} \begin{aligned}[b] \mathrm{T}{\mathrm{B}}_{\mathrm{H}}=& {\eta }_{\rm T}(1-{\varepsilon }_{\mathrm{H}})[{\eta }_{\mathrm{H}}\varepsilon {T}_{\mathrm{H}}+(1-{\eta }_{\mathrm{H}}\left){T}_{\mathrm{E}\mathrm{C}}\right]\\ & +{\eta }_{\rm T}{\varepsilon }_{\mathrm{H}}{T}_{\mathrm{h}\mathrm{o}\mathrm{t}}+(1-{\eta }_{\rm T}){T}_{\mathrm{E}\mathrm{T}}, \end{aligned}$$ (7)

    where (1−ηT) is the backlobe spillover factor of the hot-load reflector, ηH is the forward efficiency of the hot-load reflector describing the effective factor of the hot-view radiance received by the main reflector, εH is the emissivity of the hot-load reflector, ε is the emissivity of the hot target, Thot is the temperature of the hot-load reflector, TH is the temperature of the hot load, TEC is from the cosmic background radiation entering the antenna, and TET is the backlobe radiation of the hot reflector.

    Similar to the TB uncertainty at the cold and earth views, the uncertainty of TBH can be derived,

    $$\hspace{5pt} {\mathrm{\Delta }\mathrm{T}\mathrm{B}}_{\mathrm{H}}=\sqrt{\sum _{j}{\left(\frac{\partial {\mathrm{T}\mathrm{B}}_{\mathrm{w}}}{\partial {X}_{j}}\Delta {X}_{j}\right)}^{2}}, $$ (8)

    with the calibration parameters ($ {X}_{j} $) presented at the hot view in Eq. (7).

    Table 1 gives an overview of the nominal values and perturbations of the calibration parameters related to the hot-load reflector. The backlobe spillover and the emissivity of the hot-load reflector have been examined in detail and their uncertainties are given according to the previous work (Xie et al., 2019, 2021).

    Channel (GHz)Nominal value Perturbation
    ηTηHεHThot* (K)TET* (K)ΔηTΔηHΔεHΔThot* (K)ΔTET* (K)
    10.65 V 0.96520.97340.0434328.0188.00.00410.0050.002617.34.57
    10.65 H 0.96930.97340.0875131.60.00270.0050.00687.26
    18.7 V 0.96970.98610.0503202.90.00890.0020.00336.40
    18.7 H 0.96760.98610.0707151.90.00450.0020.006710.8
    23.8 V 0.99090.99020.0389215.10.00440.0020.00408.13
    23.8 H 0.99170.99020.0541173.10.00360.0020.007914.0
    36.5 V 0.98510.99400.0437214.60.00200.0020.00287.80
    36.5 H 0.99050.99400.0615172.30.00200.0020.006313.9
    89 V0.9980.99260.0342240.80.00200.0020.00338.43
    89 H0.9980.99260.0406214.50.00200.0020.007416.2
    *The value shown in Table 1 is averaged over one descending track of 1 Januray 2019.

    Table 1.  Nominal values and perturbations of the calibration parameters at the hot view of FY-3C MWRI

    For the hot-load reflector of FY-3 MWRIs, only two platinum resistance thermometers (PRTs) are embedded symmetrically at the lower back of the hot-load reflector with a distance of 58 cm away from each other and a distance of 31.2 cm away from the center of the hot reflector (Xie et al., 2019). The mean temperature measured by the two PRTs is assumed to be the representative physical temperature of the hot-load reflector, despite of the inhomogeneous thermal condition of the hot-load reflector exposing in the complicated in-orbit environments. The temperatures measured by the two PRTs exhibit significant discrepancies and seasonal- and latitudinal- dependent variations have been noticed in the descending tracks when the instrument endures solar illumination (Fig. 4 in Xie et al., 2019). The extreme temperature bias of the two PRTs could be a good approximation for the temperature perturbation of the hot-load reflector (Table 1). Note that, the orbit drifting of the satellite leads to varying thermal conditions of the hot reflector and further changes the CU contributed by the emission of the hot-load reflector, though the spacecraft was designed to cross the equator in the descending nodes at around 1020 Beijing Time (BT). Long-term evaluation of the MWRI calibration stability is required to investigate the impacts of the orbit drifting on the CU.

    The earth radiation enters the backlobe view of the hot-load reflector and further introduces the CU. However, a “true” backlobe background is still difficult to be acquired either in the on-ground measurements or in-orbit observations. Instead, the monthly-mean Advanced Microwave Scanning Radiometer (AMSR-2) observations are considered as the background radiances viewed by MWRI. The TB standard deviation of AMSR-2 at each pixel is considered as the perturbation of the backlobe background. The errors introduced by the uncertainty in the TB standard deviation are negligible since the value is at least a magnitude less than the standard deviation. With the backlobe spillover factor of the hot-load reflector given in Xie et al. (2021), the backlobe contribution to the CU is thus characterized.

    At the calibration “hot point”, the emission from the satellite platform may intrude into the microwave propagation path and contaminate the radiance received by the primary reflector. However, lack of knowledge on the complex mixture of the satellite temperatures and emissivity makes it even more impossible to accurately characterize the effective radiance of the satellite platform. As a compromise, we empirically assume that the uncertainty of the forward efficiency, which is the portion of radiation reflected by the hot-load reflector into the primary reflector, is a magnitude less than its nominal value derived in the electromagnetic analysis on-ground. Considering a relatively wider antenna beam width at low frequencies, the perturbation in ηH is assumed to be 0.005 at 10.65-GHz channels while the value is 0.002 in magnitude for other channels (Table 1).

    Characterization of the uncertainty in TBH now reduces to the performance of the hot load. The physical temperature of the hot load is measured by three PRTs embedded in the hot load in different depths and locations. The temperatures of the three PRTs are expected to be identical, when the hot-load array is an ideal blackbody with thermal control in orbit. However, nonuniform hot-load temperatures have been detected and the uncertainty is found to be on the order of 0.3 K according to the in-orbit measurements. The emissivity of the hot-load array has been examined in the prelaunch phase, of which the perturbation is less than 0.0009 at all channels of MWRI.

    Figure 2 depicts the uncertainty components at the hot view due to the perturbations of calibration parameters for the 10.65 V channel along one descending pass. The estimated uncertainties of the hot-view TB are mainly on the order of 0.4 K. Exceptions are for the uncertainties due to the temperature and the forward efficiency of the hot-load reflector, where the corresponding values are found to be greater than 1.2 K. While the hot-view TB uncertainty due to the forward efficiency of the hot-load reflector is nearly constant, the uncertainty due to the temperature of the hot-load reflector exhibits notable variation along the descending orbit. Presumably, the hot-load reflector exposing to the sunlight is heated nonuniformly, and severe temperature discrepancies between the two reflector PRTs further lead to significant uncertainties in the hot-view TB. Extreme values of the hot-view TB uncertainty due to the temperature of the hot reflector are detected near the latitudes of 10°S and 50°N.

    Figure 2.  Uncertainty components at the hot view as a function of latitude for the 10.65 V channel along the descending pass of the orbit on 1 January 2019.

    As a contrast, the hot-view TB uncertainty due to the perturbation of the calibration parameters is shown at the 10.65 V channel for the ascending orbit in Fig. 3. Similar to the descending phases, the hot-view TB uncertainty is mainly caused by the forward efficiency of the hot load in the ascending pass. Note that the impacts of the temperature of the hot-load reflector are mitigated as the hot-load reflector is cooling down in the shade and the temperature bias of the two PRTs is alleviated (Fig. 4 in Xie et al., 2019).

    Figure 3.  As in Fig. 2, but for the ascending pass.

    Figure 4.  Calibration uncertainties of MWRI in the descending tracks on 1 Januray 2019.

  • The main uncertainty components of the received voltage are caused by the system noise and digital quantization of MWRI. These values are determined in the on-ground thermal/vacuum (T/V) tests where the liquid nitrogen and room temperature are used as the “cold and hot points” in the calibration process. At a fixed scene temperature varying from ~90 to 300 K at a step of 15 K in the T/V tests, the dispersion of the voltage is negligibly small and thus TB perturbation is indiscernible, i.e., on the order of 0.01 K.

    The nonlinearity is determined by the quadratic curve fitting to the T/V test results and a maximum value occurs near the scene temperature of 200 K in the vacuum chamber. For FY-3C MWRI, the nonlinear TB is required to be less than 2 K (1 K) at the 10.65-GHz (89-GHz) channels. According to the T/V test results, the error between the curve fitting and T/V measurements is a magnitude less than the maximum nonlinear TB. Nonlinear TB uncertainty due to the nonlinear parameter is supposed to be less than 0.5 K for all channels according to the T/V test results.

3.   Results and discussion
  • Following Eqs. (1–4), the CU can be calculated with the nominal values and the perturbations of calibration parameters. The CU of FY-3C MWRI is examined at each scan pixel for all channels, in order to evaluate the calibration quality.

    Figures 4 and 5 shows the CU in the descending and ascending phases of the orbits during one day (1 January 2019), respectively. The CU exhibits latitudinal dependent variation at all channels and a maximum CU up to 3 K is found at the 10.65 H channel in the descending tracks (Fig. 4). Generally, the CU decreases with the increasing channel frequency. Taking benefit of the mitigated impacts from the hot-load reflector, the 89-GHz channels provide relatively confident observations with the CU less than 1 K in the descending orbits. The variation of CU is also limited at 89 GHz with a change on the order of 0.5 K when the spacecraft flies towards the South.

    Figure 5.  As in Fig. 4, but for the ascending tracks.

    The forward efficiency of the hot-load reflector is the driving factor affecting the calibration confidence and introduces near-constant uncertainty during the entire flight of the spacecraft (Figs. 2, 3). However, when viewing Fig. 4, the CU is found to be highly correlated to the Earth scenes, especially at the lower-frequency channels where the CU over land is much stronger than that of the ocean. This is because the constant uncertainty due to the forward efficiency of the hot-load reflector propagates in the calibration process, accompanying by the sensitivity of the observed TB to the perturbations of the forward efficiency. Meanwhile, solar illumination on the emissive hot-load reflector induces severe environment-dependent uncertainties in the observations, as shown in Fig. 2. The uncertainty due to the temperature of the hot-load reflector causes the CU to vary along the latitude. The CU is further weighed by the emissivity of the hot-load reflector, especially at the horizontally-polarized channels where the emission of the hot-load reflector is much stronger than that of the vertical polarization. It is thus found that, at the 36.5 H channel, the CU reaches up to 1.5 K near 10°S where a maximum uncertainty in the hot reflector temperature was also noticed (Xie et al., 2019), coincident with Fig. 2.

    Same as the descending track, in the ascending orbit, the CU also decreases when the channel frequency increases, from ~1.5 K at the 10.65-GHz channels to ~0.5 K at the 89-GHz channels (Fig. 5). The CU in the ascending orbits is also dominated by the uncertainty in the forward efficiency of the hot-load reflector. The effects of the temperature of the hot-load reflector on the CU are damped in the ascending orbits, comparing to the descending passes. When the spacecraft begins to enter the ascending pass, MWRI is not illuminated by the sunlight and the hot-load reflector cools down gradually. The temperature discrepancy between the two PRTs of the hot-load reflector is narrowed down and the impacts of the reflector emission are thus alleviated at all channels. Therefore, generally, the CU in the ascending tracks is much smaller than that of the descending passes.

    In the ascending pass, the CU contrast between the land and ocean is noticed at the lower-frequency channels (Fig. 5), while similar phenomena are not found for the 89-GHz channels. Both the narrow antenna beam-width and the reduced emission of the hot reflector at 89 GHz enhance the confidence of the calibration accuracy.

    To further quantify the calibration quality of MWRI, the monthly-mean CU is examined with the one-month observations in January 2019 (Fig. 6). The ocean and land scenes in the ascending and descending tracks are considered separately in this study. For most of the MWRI channels, the monthly-mean CU is around 1 K in magnitude. Only at the 10.65 H channels, the corresponding mean value reaches up to 2 K for the land scenes. The CU over land is greater than that of the ocean at all channels, either in the descending or ascending tracks. The 1-K CU difference between the ocean and land scenes is found for the sensor at the low frequencies. Only at the 89-GHz channels, the CU over ocean is comparable to that of the land, consistent with the results presented in Figs. 4, 5. Furthermore, the lower CU in the ascending passes is attributed to the confidence in the temperature of the hot-load reflector without solar illumination, as discussed above.

    Figure 6.  MWRI one-month mean calibration uncertainty (CU) in January 2019 over land and ocean for the ascending (ASC) and descending (DES) phases of the orbits.

    Figure 7 shows the month-to-month variation of the CU for the MWRI channels during August 2018 and July 2019. The CU difference between the land and ocean scenes occurs, especially for the low-frequency channels in the descending orbits. Negligible CU differences between the land and ocean in the ascending orbits are found for the 89-GHz channels. Both in the descending and ascending orbits, the variation of the monthly-mean CU at all channels is less than 0.5 K in one year.

    Figure 7.  MWIR month-to-month CU variation over the land and ocean during the period of August 2018 and July 2019 for the ASC and DES tracks.

4.   Conclusions
  • This study evaluated the in-orbit CU of MWRI aboard FY-3C. The CU analysis model was established to assess the confidence of the MWRI observations. Variable sources correlated to the calibration uncertainties of MWRI were investigated in the study.

    Uncertainties in the cold-view TB, the measured voltages, and in-orbit nonlinearity were proved to have limited impacts in the confidence of the observation quality. Uncertainties in the TB observed at the hot view, including the emission, the forward efficiency, and the backlobe of the hot-load reflector, were found to contribute significantly to the CU. Among the calibration parameters of the hot-load reflector, lack of knowledge on the exact emission makes the calibration quality vary along the orbits, especially in the descending phases when the sensor exposes to the sunlight. The satellite platform and its complex emissivity complicate the characterization of the forward radiation reflected by the hot-load reflector into the primary reflector, either in the ascending or descending orbits. Accompanied by the sensitivity of the earth-view TB to the perturbations in the hot-view TB, the uncertainty in the hot-view TB deteriorates the calibration accuracy.

    The CU of MWRI is found to be correlated to the observation scenes and in-orbit environments. Special attention should be paid to the observations in the descending orbits when the emission of the hot-load reflector is still not accurately characterized. Without the sunlight, the CU uncertainty is improved by 0.5 K at the 10.65-GHz channels in the ascending tracks. Meanwhile, the data users of MWRI need to be aware that MWRI provides relatively confident observations over the ocean, especially at high-frequency channels. The CU over land is generally greater than that of the ocean at each channel, and up to 1-K CU difference between the land and ocean is found for the 10.65 H channel. For the higher-frequency channels, the CU difference between the warm and cold scenes is mitigated. Temporal variability of the CU is found to be on the order of 0.5 K during August 2018 and July 2019 for all the channels.

    Recommendations could be made for the adjustments of the calibration system and algorithm to produce more consistent and reliable observations. The variability of the CU would be alleviated with more accurate knowledge of the calibration system of the instrument. To reduce the uncertainty from the radiance entering the primary reflector from the hot view, the design of the reflector should be upgraded to improve the main beam efficiency of the reflectors and mitigate the contamination originating from the space environment and satellite, especially at the 10.65-GHz channels. Meanwhile, knowledge of the in-orbit precise thermal conditions is necessary to achieve better performance of the sensor. Comprehensive on-ground measurements are also required to benefit accurate estimates of the calibration system for the in-orbit observations.

    Acknowledgments. The authors would like to thank for the technical supports from Shanghai Spaceflight Institute of TT&C and Telecommunication for MWRI on-ground measurements.

Reference (21)

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return