Evaluation of Total Precipitable Water over East Asia from FY-3A/VIRR Infrared Radiances

Transcription

1 ATMOSPHERIC AND OCEANIC SCIENCE LETTERS, 2010, VOL. 3, NO. 2, Evaluation of Total Precipitable Water over East Asia from FY-3A/VIRR Infrared Radiances ZHENG Jing 1,2,3,4, SHI Chun-Xiang 2,4, LU Qi-Feng 2,4, and XIE Zheng-Hui 1 1 Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing , China 2 National Satellite Meteorological Center, China Meteorological Administration, Beijing , China 3 Graduate University of Chinese Academy of Sciences, Beijing , China 4 Key Laboratory of Radiometric Calibration and Validation for Environmental Satellites, China Meteorological Administration, Beijing , China Received 21 December 2009; revised 4 February 2010; accepted 4 February 2010; published 16 March 2010 Abstract Satellite retrieval of atmospheric water vapor is intended to further understand the role played by the energy and water cycle to determine the Earth s weather and climate. The algorithm for operational retrieval of total precipitable water (TPW) from the visible and infrared radiometer (VIRR) onboard Fengyun 3A (FY-3A) employs a split window technique for clear sky radiances over land and oceans during both day and night. The retrieved TPW is compared with that from the moderate resolution imaging spectroradiometer (MODIS) onboard the Terra satellite and data from radiosonde observations (RAOB). During the study period, comparisons show that the FY-3A TPW is in general agreement with the gradients and distributions from the Terra TPW. Their zonal mean difference over East Asia is smaller in the daytime than at night, and the main difference occurs in the complex terrain at mid latitude near 30 N. Compared with RAOB, the zonal FY-3A and the Terra TPW have a moist bias at low latitudes and a dry bias at mid and high latitudes; in addition, the FY-3A TPW performs slightly better in zonal mean biases and the diurnal cycle. The temporal variation of the FY-3A and the Terra TPW generally fits the RAOB TPW with the FY-3A more accurate at night while Terra TPW more accurate during the daytime. Comparisons of correlations, root mean square differences and standard deviations indicate that the FY-3A TPW series is more consistent with the RAOB TPW at selected stations. As a result, the FY-3A TPW has some advantages over East Asia in both spatial and temporal dimensions. Keywords: FY-3A/VIRR, total precipitable water, split window technique, evaluation. Citation: Zheng, J., C.-X. Shi, Q.-F. Lu, et al., 2010: Evaluation of total precipitable water over East Asia from FY-3A/VIRR infrared radiances, Atmos. Oceanic Sci. Lett., 3, Introduction Water vapor, which can exist in all three phases, is the most important constituent of the atmosphere for all weather and climate processes (Ruprecht, 1996). The total precipitable water (TPW) is the total amount of water vapor contained in a vertical column of unit crosssectional Corresponding author: XIE Zheng-Hui, zxie@lasg.iap.ac.cn area extending from the surface to the top of the atmosphere. As one of the critical hydro-meteorological parameters (See Fig. 1, from Zeng, 1999), it represents the water vapor content of the current atmosphere. High resolution measurements of the TPW are hardly available because of the limited availability of sounding data as well as the high variability of the TPW in space and time. To fill this gap, remote sensing with satellite-borne instruments has been used to derive the TPW. The significance of TPW retrieval is threefold. First, it is useful in anticipating the distribution of precipitation patterns and defining the precipitation efficiency (e.g., Robinson and Lutz, 1978; Ojo, 2005). Second, applying the TPW as an additional source of moisture information in data assimilation has a significant impact on numerical weather prediction and climate studies (e.g., Filiberti et al., 1994; Rakesh et al., 2009). Third, TPW retrievals from different satellites can be merged with the TPW from radiosonde observations, GPS stations and other observational data to obtain the global TPW distribution under all weather conditions. These TPW products also contain valuable information on moisture transport in the atmosphere and the hydrological cycle on the Earth (e.g., Wittmeyer and Haar, 1994; Gao et al., 2004). Driven by the importance of the TPW from satellite observations, a number of algorithms have been developed in the near-infrared (e.g., Kaufman and Gao, 1992; Kleidman et al., 2000), infrared (e.g., Aoki and Inoue, 1982; Li et al., 2000) and microwave (e.g., Alishouse et al., 1990; Ferraro et al., 1996) spectra. In China, the available data from Fengyun 3A (FY-3A), which was launched successfully on May 27, 2008, provide a new source for TPW retrieval. Measurements of the visible and infrared radiometer (VIRR) onboard FY-3A are used to retrieve the clear sky TPW by a physical split window technique based on the work of Jedlovec (1987). Meanwhile, bias evaluation is particularly important because it will imply shortcomings in the observing system (Fetzer et al., 2006), identify the uncertainties that may exist during the retrieving process, and improve the application of the algorithm. To preliminarily assess the FY-3A VIRR infrared TPW (hereafter FY-3A TPW), in section 3, we compare the clear sky TPW retrieved from the IR bands of the moderate resolution imaging spectroradiometer (MODIS) onboard the U.S. National Aeronautics and Space Administration (NASA) Earth Observing

2 94 ATMOSPHERIC AND OCEANIC SCIENCE LETTERS System (EOS) Terra satellite (hereafter Terra TPW) and the TPW calculated from the radiosonde observations (hereafter, RAOB TPW) in both spatial and temporal dimensions. The remainder of this work is as follows: section 2 describes the TPW data and algorithms from the FY-3A VIRR and Terra MODIS observing systems as well as the RAOB observations. The analysis of their comparisons are presented in section 3, which is focused on the East Asian region (EAR) from N and E (shown in Fig. 1). Finally, the major conclusions and additional discussions are contained in section Algorithms and data sets TPW from VIRR onboard FY-3A spacecraft The VIRR onboard FY-3A scans a swath width of 2800 km and provides images in 10 spectral bands between 0.44 and 12.5 μm with its spatial resolution of 1.1 km at nadir. It is widely used in monitoring cloud, vegetation, sediment, snow, ice, land surface temperature, sea surface temperature, and atmospheric water vapor content (Dong et al., 2009; Zhang et al., 2009). Measurements from VIRR channel 4 ( μm) and channel 5 ( μm) are used to retrieve the clear sky TPW by using a physical split window technique, which implements a perturbation solution of the radiative transfer equation to obtain the total precipitable water. The basic algorithm was first developed by Jedlovec (1987) and subsequently evaluated by Guillory et al. (1993) and Suggs et al. (1998). This retrieval algorithm requires at least two longwave infrared window channel observations to simultaneously solve for perturbations or departures of the TPW and the surface skin temperature from guess values of these quantities. The additional input guess values required by this algorithm include the estimated profiles of the temperature, precipitable water and channel transmittance. A perturbed profile of the precipitable water and the transmittance in those two retrieval channels are also needed. From these profiles, the coefficients are calculated and used to solve the radiative transfer equation for the perturbation TPW. A detailed derivation can be found in Guillory et al. (1993), Suggs et al. (1998) and Haines et al. (2004). This technique has been widely used to retrieve the TPW from Geostationary Operational Environmental Satellites (GOES) data (Haines, et al., 2004) and the Fengyun-2 series (Shi and Xie, 2005; Shi, 2005) among others. The FY-3A TPW can be retrieved twice a day in clear sky pixels during both the day and night. All of the data can be downloaded from Figure 1 Spatial distribution of the radiosonde observation (RAOB) stations during the study period of 51 days from November 11 to December 31, There are around 162 stations at (a) 0000 UTC and 153 stations at (b) 1200 UTC for each day, respectively. The red dot denotes that there are 24 and 16 stations with more than 25 days of data for clear sky conditions, which matched with the FY-3A and Terra for both day and night, respectively. The blue cross mark denotes the other stations. The study area is inside of the dashed rectangle and includes EAR from 10 N to 60 N and 70 E to 140 E. VOL. 3 TPW from MODIS onboard Terra spacecraft MODIS is a scanning spectroradiometer onboard the NASA EOS Terra and Aqua platforms with 36 visible (VIS), near-infrared (NIR), and infrared (IR) spectral bands between and μm (King et al., 1992). The Terra TPW product is more synchronized and comparable with the FY-3A TPW because the local overpass time is similar (1030 LST for Terra and 1005 LST for FY-3A at descending node). In contrast with the FY-3A TPW, the MODIS atmospheric moisture retrieval algorithm is a statistical synthetic regression with the option for a subsequent nonlinear physical retrieval based on the regularization method (Li et al., 2000). The retrieval procedure involves linearization of the radiative transfer model and inversion of the radiance measurements. The retrievals are performed by using MODIS clear sky radiance over the land and ocean during both the day and night. In the regression procedure, the primary predictors are 11 infrared spectral band brightness temperatures with wavelengths between 4.5 μm and 14.2 μm. Determination of the Terra TPW is performed by integrating 101 level retrieved mixing ratio profiles through the atmospheric column (King et al., 2003; Seemann et al., 2003). This can be obtained from its atmospheric level 2 products (MOD05_L2) at the 5 km pixel resolution (ftp://ladsftp.nascom.nasa.gov/alldata/5/ MOD05_L2/2009) and then interpolated into grid boxes to match the daily FY-3A TPW data grid resolution over EAR during both the day and night. Only the

3 NO. 2 ZHENG ET AL.: EVALUATIONOF TPW OVER EAST ASIA FROM FY-3A/VIRR 95 grids that share the same clear sky cloud mask with the FY-3A will be used in the following analysis. 2.3 Calculated TPW from radiosonde observations The available radiosonde observations (RAOB TPW) are used to validate both the FY-3A and the Terra TPW. The distribution of RAOB stations is shown in Fig. 1 including around 162 and 153 stations that observe at 0000 UTC and 1200 UTC for each day, respectively (data from The RAOB TPW is calculated by integrating the water vapor from the surface to the top of the atmosphere: 1 0 W = x( p)dp ρ g, (1) ps where the subscript s stands for the surface; x(p) is the mixing ratio calculated from the observed temperature and the dew-point at pressure level p; g is the acceleration of gravity; ρ is the density of water; and W is the RAOB TPW in millimeters. 3 Comparisons and results To evaluate the spatial and temporal characteristics, the FY-3A TPW is compared with the Terra and RAOB TPW in the study period from November 11 to December 31, Since moisture transport plays a critical role in the energy and water cycle of the East Asian monsoon system (Ding and Chan, 2005), we have focused the TPW evaluation over EAR in the following analysis. A similar evaluation can also be done over any other interested region because the FY-3A TPW is a global dataset. 3.1 Spatial distribution Figure 2 shows the spatial distributions of the mean FY-3A and Terra TPW during the study period. The valid number of clear skies for both satellites reveals their cloud mask differences during both the day and night (Figs. 2a and 2b). The frequent clear skies (> 30 days) mainly occur in the south of the Tibetan Plateau and north of both the Indian peninsula and Indo-China peninsula, where the correlation will be more reliable. The FY-3A TPW distribution shows general agreement with the Terra for moist air in the south and dry air in the north (Figs. 2c f), but the FY-3A TPW is apparently moister than Terra in areas with frequently clear skies, where the terrain is complex and steep (Figs. 2g and 2h). Their correlation is positive over most of EAR (Figs. 2i and 2j), and their root mean square (RMS) differences are acceptable (< 10 mm) over most of the area, except in the complex terrain areas (Figs. 2k and 2l). This may indicate that the FY-3A TPW has more variation over mountainous areas. 3.2 Zonal mean The zonal mean comparisons for every 0.05 zonal belt (figure omitted) show that the FY-3A TPW is about 3.8 mm more than Terra in the daytime and 6.3 mm at night. This difference is apparent near the latitude of 30 N, where complex terrain lies. In the south of 35 N, their zonal correlations during both the day and night are similar and significant (> of 0.05 significance level). In the north of 35 N, their zonal correlation at night is significant, while in the daytime, it is smaller and insignificant. Moreover, the variable TPW s range between the day and night is small for FY-3A, but for Terra, the zonal TPW at night is apparently drier than that during the daytime at low and mid latitudes. This indicates that the diurnal cycle of the FY-3A TPW is probably weaker than Terra, especially in moist areas at low and mid latitudes. To better understand both of the TPW retrievals, the available RAOB TPW is averaged in every 5 zonal belt over EAR. Both of the TPW retrievals follow a pattern of observation with a high TPW at low latitudes and a gradual change to a low TPW at high latitudes (Fig. 3a); however, they have a moist bias at low latitudes and a dry bias at mid and high latitudes. The FY-3A fits the RAOB better (< 5 mm bias for each zone), except in the N where its moist bias is slightly larger. The dry bias of Terra occurs as far south as the 25 N latitude, while the FY-3A s dry bias is observable around 35 N (Fig. 3b). For the diurnal cycle of the TPW (Fig. 3c), the RAOB shows a weak cycle at low and high latitudes but a strong variability in the mid latitude of N. The FY-3A is generally closer to the RAOB; however, at the mid latitude of N, the RAOB is wet at night, while both retrievals are dry at night (Terra is slightly drier). 3.3 Temporal pattern Based on the available total clear sky data from 2531 cases during the day and 2693 cases at night, comparisons of both satellite TPW retrievals and the RAOB in scatter plots (Figure omitted) show that the FY-3A and Terra agree with the RAOB; the correlation coefficients are 0.73 and 0.71 during the day and 0.63 and 0.58 at night, respectively. Thus, the FY-3A is slightly better. On average, there is about 3.5 mm of moist bias for the FY-3A during the day and less bias (< 1 mm) overestimated at night; there is about 4.1 mm of dry-bias for Terra at night, while less bias (< 1 mm) underestimated during the day. The standard deviation (STD) between the FY-3A and RAOB TPW is 0.72 mm and 1.29 mm during the day and night, respectively, in comparison with 1.01 and 0.69 between the Terra and RAOB TPW. Therefore, their variation amplitudes are within the level of 5% 10% deviation from the RAOB TPW. In all, compared with the RAOB, the FY-3A TPW has less bias with more variable amplitude at night than the Terra; while the Terra shows less bias during the day with a higher variable amplitude than the FY-3A. To synthetically study how closely a pattern of time series matches observations, Taylor diagrams are used to illustrate three statistics characteristics (i.e., the correlation coefficient, the RMS difference and the standard deviation) at selected stations that have more than 25 days of clear sky over EAR (see Fig. 1) (Taylor, 2001). As shown in Fig. 4, for most stations, both the TPW data agree well with the observations (RAOB TPW) and have a correlation coefficient higher than 0.38 (i.e., 5% sig-

4 Figure 2 Spatial distributions of the mean total precipitable water (TPW, units: mm) that was retrieved from the FY-3A VIRR and the Terra MODIS during the study period as well as their statistics characteristics during the day (on the frist and third columns) and night (on the second and the fourth columns). 96 ATMOSPHERIC AND OCEANIC SCIENCE LETTERS VOL. 3

5 NO. 2 ZHENG ET AL.: EVALUATIONOF TPW OVER EAST ASIA FROM FY-3A/VIRR 97 Figure 3 Comparisons of the zonal TPW means between the RAOB and both the FY-3A and Terra for every 5 zonal belt over EAR. nificance level), which indicates that the time series of both TPW retrievals fit with the observations quite well. Meanwhile, the amplitude variation of the FY-3A TPW is quite reasonable because the fraction of its STD to that of the RAOB TPW is in the range of at most stations; for comparison, the amplitude variation of the Terra TPW at some stations is too small (a normalized STD of under 0.5) during both the day and night. Therefore, the temporal pattern of the FY-3A TPW series is more consistent with the RAOB than the Terra TPW. Nevertheless, the RMS differences of both the FY-3A and Terra are comparable with the reference RAOB field (i.e., REF point in Fig. 4), which indicates that many challenges still remain to improve the potential ability for both TPW products. 4 Conclusions and discussions This paper examines the FY-3A TPW that was retrieved by using a split window technique under the clear sky condition. The evaluation focuses on the spatial distribution, zonal mean and temporal pattern over East Asia from November 11 to December 31 in Generally, the FY-3A TPW is in agreement with the gradients and distributions from the Terra TPW. Their zonal mean shows that the difference is smaller during the day than at night, while the main difference occurs at mid-latitude near 30 N. Comparisons of the zonal mean TPW between the RAOB and both FY-3A and Terra show that both retrievals seem to have a moist bias at low latitudes and a dry bias at mid and high latitudes, while the FY-3A TPW seems to perform slightly better than Terra in the zonal mean biases and diurnal cycle. The analysis of correlations, root mean square differences and standard deviations between the RAOB and both the FY-3A and Terra indicate that the FY-3A TPW series is more consistent with the RAOB TPW at selected stations with respect to its variation and amplitude.

6 98 ATMOSPHERIC AND OCEANIC SCIENCE LETTERS VOL. 3 Figure 4 Taylor diagrams for the TPW at the selected stations that have more than 25 days of clear sky over the East Asia Region (a) during the day and (b) at night, respectively. There are 24 available stations in the daytime and 16 available stations at night (see Fig. 1). REF is RAOB TPW. The radial distance from the origin is proportional to STD of the retrieved (FY-3A and Terra) TPW to that of the RAOB TPW. The centered RMS difference between the retrieved TPW and the RAOB TPW is proportional to their distance apart (in the same units as the standard deviation). The correlation between the retrieved TPW and the RAOB TPW is given by the cosine of the azimuth. The correlation coefficient lines at , , , and represent the significance level of α = 0.1, 0.05, 0.01, and 0.001, respectively. In addition, there are some factors that may introduce uncertainties when comparing the FY-3A and Terra TPW, such as differences in their cloud mask and algorithms, the impact of the first-guess value (e.g., Knabb and Fuelberg, 1997), the surface emissivity (e.g., Seemann et al., 2007), etc. Uncertainty may also exist in the RAOB TPW because it is about two hours earlier than the FY-3A and Terra s local over-passing time. More sensitivity studies are required for further validation. In conclusion, the preliminary results of the evaluation show that the FY-3A TPW has some advantages over East Asia in both the spatial and temporal dimensions. Continuous TPW monitoring from the FY-3 series will help to advance our understanding of the role that is played by the energy and water cycle processes to determine the Earth s weather and climate. Acknowledgements. This research is supported by the National High Technology Research and Development Program of China (Grant No. 2007AA12Z144), the Professional Projects (Grant Nos. GYHY and GYHY ), and the China Meteorological Administration New Technology Promotion Project (Grant No. CMATG2008Z04). The authors thank Dr. HU Xiu-Qing at the National Satellite Meteorological Center (NSMC/CMA), Dr. TIAN Xiang-Jun at the Institute of Atmospheric Physics (IAP/CAS), Prof. Q. S. HU at the School of Natural Resources and the Department of Geosciences from the University of Nebraska-Lincoln, and also two anonymous reviewers for their helpful comments and suggestions for this work. References Alishouse, J. C., S. A. Snyder, J. Vongsathorn, et al., 1990: Determination of oceanic total precipitable water from the SSM/I, IEEE Trans. Geosci. Remote Sens., 28, Aoki, T., and T. Inoue, 1982: Estimation of precipitable water from the IR channel of the geostationary satellite, Remote Sens. Environ., 12, Ding, Y., and J. C. L. Chan, 2005: The East Asian summer monsoon: An overview, Meteor. Atmos. Phys., 89, Dong, C., J. Yang, W. Zhang, et al., 2009: An overview of a new chinese weather satellite FY-3A, Bull. Amer. Meteor. Soc., 90(10), Ferraro, R. R., N. C. Grody, F. Z. Weng, et al., 1996: An eight year ( ) time series of rainfall, clouds, water vapor, snow cover, and sea ice derived from SSM/I measurements, Bull. Amer. Meteor. Soc., 77, Fetzer, E. J., B. H. Lambrigtsen, A. Eldering, et al., 2006: Biases in total precipitable water vapor climatologies from Atmospheric Infrared Sounder and Advanced Microwave Scanning Radiometer, J. Geophys. Res., 111, D09S16, doi: /2005jd Filiberti, M. A., L. Eymard, and U. Urban, 1994: Assimilation of satellite precipitable water in a meteorological forecast model, Mon. Wea. Rev., 122, Gao, B.-C., P. K. Chan, and R.-R. Li, 2004: A global water vapor data set obtained by merging the SSMI and MODIS data, Geophys. Res. Lett., 31, L18103, doi: /1004GL Guillory, A. R., G. J. Jedlovec, and H. E. Fuelberg, 1993: A technique for deriving column-integrated water content using VAS split window data, J. Appl. Meteor., 32, Haines, S. L., G. J. Jedlovec, and R. L. Suggs, 2004: The GOES product generation system, NASA Marshall Space Flight Center Tech. Memo., NASA/TM , Huntsville, Alabama, 64pp. Jedlovec, G. J., 1987: Determination of Atmospheric Moisture Structure from High-Resolution MAMS Radiance Data, PhD. thesis, University of Wisconsin-Madison, Madison, 187pp. Kaufman, Y. J., and B.-C. Gao, 1992: Remote sensing of water vapor in the near IR from EOS/MODIS, IEEE Trans. Geosci. Remote Sens., 30, King, M. D., Y. J. Kaufman, W. P. Menzel, et al., 1992: Remote sensing of cloud, aerosol, and water vapor properties from the Moderate Resolution Imaging Spectrometer (MODIS), IEEE Trans. Geosci. Remote Sens., 30, King, M. D., W. P. Menzel, Y. J. Kaufman, et al., 2003: Cloud and aerosol properties, precipitable water, and profiles of temperature and water vapor from MODIS, IEEE Trans. Geosci. Remote Sens., 41, Kleidman, R. G., Y. J. Kaufman, B. Gao, et al., 2000: Remote sensing of total precipitable water vapor in the near-ir over ocean

7 NO. 2 ZHENG ET AL.: EVALUATIONOF TPW OVER EAST ASIA FROM FY-3A/VIRR 99 glint, Geophys. Res. Lett., 27(17), Knabb, R. D., and H. E. Fuelberg, 1997: A comparison of the first-guess dependence of precipitable water estimates from three techniques using GOES data, J. Appl. Meteor., 36, Li, J., W. Wolf, W. P. Menzel, et al., 2000: Global soundings of the atmosphere from ATOVS measurements: The algorithm and validation, J. Appl. Meteor., 39, Ojo, O., 2005: The distribution of mean monthly precipitable water vapor and annual precipitation efficiency in Nigeria, Theor. Appl. Climatol., 18(3 4), Rakesh, V., R. Singh, and P. K. Pal, et al., 2009: Impact of satellite-observed winds and total precipitable water on WRF short-range forecasts over the Indian region during the 2006 summer monsoon, Wea. Forecasting, 24, Robinson, P. J., and J. T. Lutz, 1978: Precipitation efficiency of cyclonic storms, Ann. Assoc. Amer. Geogr., 68(1), Ruprecht, E., 1996: Atmospheric water vapour and cloud water: An overview, Adv. Space Res., 18(7), Seemann, S. W., E. E. Borbas, R. O. Knuteson, et al., 2007: Development of a global infrared land surface emissivity database for application to clear sky sounding retrievals from multi-spectral satellite radiance measurements, J. Appl. Meteor. Climatol., 47, Seemann, S. W., J. Li, W. P. Menzel, et al., 2003: Operational retrieval of atmospheric temperature, moisture, and ozone from MODIS infrared radiances, J. Appl. Meteor., 42, Shi, C.-X., 2005: The Operational Products of Fengyun-2C. Chapter 11: Clear-Sky Total Precipitable Water from FY-2C, National Satellite Meteorological Center, Beijing, Shi, C.-X., and Z.-H. Xie, 2005: Operational method of total precipitable water retrieved from satellite multi-channels infrared data. J. Infrared Millim. Waves (in Chinese), 24(4), Suggs, R. J., G. J. Jedlovec, and A. R. Guillory, 1998: Retrieval of geophysical parameters evaluation of a split window retrieval technique, J. Appl. Meteor., 37, Taylor, K., 2001: Summarizing multiple aspects of model performance in a single diagram, J. Geophys. Res., 106(D7), Wittmeyer, I. L., and T. H. vonder Haar, 1994: Analysis of the global ISCCP TOVS water vapor climatology, J. Climate, 6, Zhang, P., J. Yang, C. Dong, et al., 2009: General introduction on payloads, ground segment and data application of Fengyun 3A, Front. Earth Sci. China, 3(3), Zeng, N., 1999: Seasonal cycle and interannual variability in the Amazon hydrologic cycle, J. Geophys. Res., 104, D8,