• Aragão, L. E. O. C., , Malhi Y. , , Roman-Cuesta R. M. , , Saatchi S. , , Anderson L. O. , , and Shimabukuro Y. E. , 2007: Spatial patterns and fire response of recent Amazonian droughts. Geophys. Res. Lett., 34, L07701, doi:10.1029/2006GL028946.

    • Search Google Scholar
    • Export Citation
  • Bitew, M. M., , and Gebremichael M. , 2011: Evaluation of satellite rainfall products through hydrologic simulation in a fully distributed hydrologic model. Water Resour. Res., 47, W06526, doi:10.1029/2010WR009917.

    • Search Google Scholar
    • Export Citation
  • Changjiang Water Resources Commission, 1997: Hydrology Research for Three Gorges Project (in Chinese). Hubei Science and Technology Press, 313 pp.

    • Search Google Scholar
    • Export Citation
  • Changjiang Water Resources Commission, 1999: Atlas of the Changjiang River Basin (in Chinese). SinoMaps Press, 286 pp.

  • Cong, Z. T., , Yang D. W. , , Gao B. , , Yang H. B. , , and Hu H. P. , 2009: Hydrological trend analysis in the Yellow River basin using a distributed hydrological model. Water Resour. Res., 45, W00A13, doi:10.1029/2008WR006852.

    • Search Google Scholar
    • Export Citation
  • Ding, Y. H., , and Zhang J. Y. , 2009: Rainstorms and Floods in China (in Chinese). China Meteorological Press, 290 pp.

  • Ebert, E. E., , Janowiak J. E. , , and Kidd C. , 2007: Comparison of near-real-time precipitation estimates from satellite observations and numerical models. Bull. Amer. Meteor. Soc., 88, 4764, doi:10.1175/BAMS-88-1-47.

    • Search Google Scholar
    • Export Citation
  • Food and Agricultural Organization, 2003: Digital Soil Map of the World and Derived Soil Properties. Land Water Digital Media Series, Rev. 1, Food and Agriculture Organization, CD-ROM.

    • Search Google Scholar
    • Export Citation
  • Gao, B., , Yang D. , , Zhao T. , , and Yang H. , 2012: Changes in the eco-flow metrics of the Upper Yangtze River from 1961 to 2008. J. Hydrol., 448–449, 3038, doi:10.1016/j.jhydrol.2012.03.045.

    • Search Google Scholar
    • Export Citation
  • Gao, G., , Chen D. , , Xu C.-Y. , , and Simelton E. , 2007: Trend of estimated actual evapotranspiration over China during1960–2002. J. Geophys. Res., 112, D11120, doi:10.1029/2006JD008010.

    • Search Google Scholar
    • Export Citation
  • Gao, Y., , and Liu M. , 2013: Evaluation of high-resolution satellite precipitation products using rain gauge observations over the Tibetan Plateau. Hydrol. Earth Syst. Sci., 17, 837849, doi:10.5194/hess-17-837-2013.

    • Search Google Scholar
    • Export Citation
  • Gebregiorgis, A. S., , Tian Y. , , Peters-Lidard C. D. , , and Hossain F. , 2012: Tracing hydrologic model simulation error as a function of satellite rainfall estimation bias components and land use and land cover conditions. Water Resour. Res., 48, W11509, doi:10.1029/2011WR011643.

    • Search Google Scholar
    • Export Citation
  • Gebremichael, M., , Krajewski W. F. , , Morrissey M. , , Langerud D. , , Huffman G. J. , , and Adler R. , 2003: Error uncertainty analysis of GPCP monthly rainfall products: A data-based simulation study. J. Appl. Meteor., 42, 18371848, doi:10.1175/1520-0450(2003)042<1837:EUAOGM>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Gosset, M., , Viarre J. , , Quantin G. , , and Alcoba M. , 2013: Evaluation of several rainfall products used for hydrological applications over West Africa using two high-resolution gauge networks. Quart. J. Roy. Meteor. Soc., 139, 923940, doi:10.1002/qj.2130.

    • Search Google Scholar
    • Export Citation
  • Hong, Y., , Hsu K. L. , , Gao X. , , and Sorooshian S. , 2004: Precipitation Estimation from Remotely Sensed Imagery using Artificial Neural Network Cloud Classification System (PERSIANN-CCS). J. Appl. Meteor. Climatol., 43, 18341853, doi:10.1175/JAM2173.1.

    • Search Google Scholar
    • Export Citation
  • Hong, Y., , Adler R. F. , , and Huffman G. J. , 2006: Evaluation of the potential of NASA multi-satellite precipitation analysis in global landslide hazard assessment. Geophys. Res. Lett., 33, L22402, doi:10.1029/2006GL028010.

    • Search Google Scholar
    • Export Citation
  • Hu, Q., , Yang D. , , Li Z. , , Mishra A. , , Wang Y. , , and Yang H. , 2014: Multi-scale evaluation of six high-resolution satellite monthly rainfall estimates over a humid region in China with dense rain gauges. Int. J. Remote Sens., 35, 12721294, doi:10.1080/01431161.2013.876118.

    • Search Google Scholar
    • Export Citation
  • Huffman, G. J., and Coauthors, 2007: The TRMM Multi-satellite Precipitation Analysis (TMPA): Quasi-Global, multi-year, combined-sensor precipitation estimates at fine scales. J. Hydrometeor., 8, 3855, doi:10.1175/JHM560.1.

    • Search Google Scholar
    • Export Citation
  • Jiang, S., , Ren L.-L. , , Hong Y. , , Yong B. , , Yang X. , , Yuan F. , , and Ma M. , 2012: Comprehensive evaluation of multi-satellite precipitation products with a dense rain gauge network and optimally merging their simulated hydrological flows using the Bayesian model averaging method. J. Hydrol., 452-453, 213225, doi:10.1016/j.jhydrol.2012.05.055.

    • Search Google Scholar
    • Export Citation
  • Joyce, R. J., , Janowiak J. E. , , Arkin P. A. , , and Xie P. , 2004: CMORPH: A method that produces global precipitation estimates from passive microwave and infrared data at high spatial and temporal resolution. J. Hydrometeor., 5, 487503, doi:10.1175/1525-7541(2004)005<0487:CAMTPG>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Kubota, T., and Coauthors, 2007: Global precipitation map using satellite-borne microwave radiometers by the GSMaP Project: Production and validation. IEEE Trans. Geosci. Remote Sens., 45, 22592275, doi:10.1109/TGRS.2007.895337.

    • Search Google Scholar
    • Export Citation
  • Li, L., and Coauthors, 2009: Evaluation of the real-time TRMM-based multi-satellite precipitation analysis for an operational Flood Prediction System in Nzoia basin, Lake Victoria, Africa. Nat. Hazards, 50, 109123, doi:10.1007/s11069-008-9324-5.

    • Search Google Scholar
    • Export Citation
  • Li, M., , Yang D. , , Chen J. , , and Hubbard S. S. , 2012: Calibration of a distributed flood forecasting model with input uncertainty using a Bayesian framework. Water Resour. Res., 48, W08510, doi:10.1029/2010WR010062.

    • Search Google Scholar
    • Export Citation
  • Li, X. H., , Zhang Q. , , and Xu C. Y. , 2012: Suitability of the TRMM satellite rainfalls in driving a distributed hydrological model for water balance computations in Xinjiang catchment, Poyang Lake basin. J. Hydrol., 426–427, 2838, doi:10.1016/j.jhydrol.2012.01.013.

    • Search Google Scholar
    • Export Citation
  • Li, Z., , Yang D. , , and Hong Y. , 2013: Multi-scale evaluation of high-resolution multi-sensor blended global precipitation products over the Yangtze River. J. Hydrol., 500, 157169, doi:10.1016/j.jhydrol.2013.07.023.

    • Search Google Scholar
    • Export Citation
  • Maidment, D. R., Ed., 1993: Handbook of Hydrology. McGraw-Hill, 1424 pp.

  • McCollum, J. R., , Krajewski W. F. , , Ferraro R. R. , , and Ba M. B. , 2002: Evaluation of biases of satellite rainfall estimation algorithms over the continental United States. J. Appl. Meteor., 41, 10651080, doi:10.1175/1520-0450(2002)041<1065:EOBOSR>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Nash, J. E., , and Sutcliffe J. V. , 1970: River flow forecasting through conceptual models part I—A discussion of principles. J. Hydrol., 10, 282290, doi:10.1016/0022-1694(70)90255-6.

    • Search Google Scholar
    • Export Citation
  • New, M., , Hulme M. , , and Jones P. , 2000: Representing twentieth-century space–time climate variability. Part II: Development of a 1961–96 monthly grids of terrestrial surface climate. J. Climate, 13, 22172238, doi:10.1175/1520-0442(2000)013<2217:RTCSTC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Nijssen, B., , and Lettenmaier D. P. , 2004: Effect of precipitation sampling error on simulated hydrological fluxes and states: Anticipating the Global Precipitation Measurement satellites. J. Geophys. Res., 109, D02103, doi:10.1029/2003JD003497.

    • Search Google Scholar
    • Export Citation
  • Pan, M., , Li H. , , and Wood E. , 2010: Assessing the skill of satellite-based precipitation estimates in hydrologic applications. Water Resour. Res., 46, W09535, doi:10.1029/2009WR008290.

    • Search Google Scholar
    • Export Citation
  • Ryo, M., , Valeriano O. C. S. , , Kanae S. , , and Ngoc T. D. , 2014: Temporal downscaling of daily gauged precipitation by application of a satellite product for flood simulation in a poorly gauged basin and its evaluation with multiple regression analysis. J. Hydrometeor., 15, 563580, doi:10.1175/JHM-D-13-052.1.

    • Search Google Scholar
    • Export Citation
  • Shen, Y., , Xiong A. , , Wang Y. , , and Xie P. , 2010: Performance of high-resolution satellite precipitation products over China. J. Geophys. Res., 115, D02114, doi:10.1029/2009JD012097.

    • Search Google Scholar
    • Export Citation
  • Sorooshian, S., , Hsu K. L. , , Gao X. , , Gupta H. , , Imam B. , , and Braithwaite D. , 2000: Evaluation of PERSIANN system satellite-based estimates of tropical rainfall. Bull. Amer. Meteor. Soc., 81, 20352046, doi:10.1175/1520-0477(2000)081<2035:EOPSSE>2.3.CO;2.

    • Search Google Scholar
    • Export Citation
  • Sorooshian, S., and Coauthors, 2011: Advanced concepts on remote sensing of precipitation at multiple scales. Bull. Amer. Meteor. Soc., 92, 13531357, doi:10.1175/2011BAMS3158.1.

    • Search Google Scholar
    • Export Citation
  • Su, F., , Hong Y. , , and Lettenmaier D. P. , 2008: Evaluation of TRMM Multisatellite Precipitation Analysis (TMPA) and its utility in hydrologic prediction in the La Plata basin. J. Hydrometeor., 9, 622640, doi:10.1175/2007JHM944.1.

    • Search Google Scholar
    • Export Citation
  • Tian, Y., , and Peters-Lidard C. D. , 2007: Systematic anomalies over inland water bodies in satellite-based precipitation estimates. Geophys. Res. Lett., 34, L14403, doi:10.1029/2007GL030787.

    • Search Google Scholar
    • Export Citation
  • Tian, Y., , Peters-Lidard C. D. , , Choudhury B. J. , , and Garcia M. , 2007: Multitemporal analysis of TRMM-based satellite precipitation products for land data assimilation applications. J. Hydrometeor., 8, 11651183, doi:10.1175/2007JHM859.1.

    • Search Google Scholar
    • Export Citation
  • Tian, Y., and Coauthors, 2009: Component analysis of errors in satellite-based precipitation estimates. J. Geophys. Res., 114, D24101, doi:10.1029/2009JD011949.

    • Search Google Scholar
    • Export Citation
  • Tong, K., , Su F. , , Yang D. , , and Hao Z. , 2014: Evaluation of satellite precipitation retrievals and their potential utilities in hydrologic modeling over the Tibetan Plateau. J. Hydrol., 519, 423437, doi:10.1016/j.jhydrol.2014.07.044.

    • Search Google Scholar
    • Export Citation
  • Turk, F. J., , and Miller S. D. , 2005: Toward improving estimates of remotely sensed precipitation with MODIS/AMSR-E blended data techniques. IEEE Trans. Geosci. Remote Sens., 43, 10591069, doi:10.1109/TGRS.2004.841627.

    • Search Google Scholar
    • Export Citation
  • Villarini, G., , Mandapaka P. V. , , Krajewski W. F. , , and Moore R. J. , 2008: Rainfall and sampling uncertainties: A rain gauge perspective. J. Geophys. Res., 113, D11102, doi:10.1029/2007JD009214.

    • Search Google Scholar
    • Export Citation
  • Villarini, G., , Krajewski W. F. , , and Smith J. A. , 2009: New paradigm for statistical validation of satellite precipitation estimates: Application to a large sample of the TMPA 0.25° 3-hourly estimates over Oklahoma. J. Geophys. Res., 114, D12106, doi:10.1029/2008JD011475.

    • Search Google Scholar
    • Export Citation
  • Wang, X. F., , and Zhang S. M. , 2013: Analysis on characteristics and forecasting of “12.7” Flood in upper Yangtze River basin (in Chinese). Yangtze River, 2013 (19), 14.

    • Search Google Scholar
    • Export Citation
  • Wood, E. F., and Coauthors, 2011: Hyperresolution global land surface modeling: Meeting a grand challenge for monitoring Earth’s terrestrial water. Water Resour. Res., 47, W05301, doi:10.1029/2010WR010090.

    • Search Google Scholar
    • Export Citation
  • Wu, H., , Adler R. F. , , Tian Y. , , Huffman G. J. , , Li H. , , and Wang J. , 2014: Real-time global flood estimation using satellite-based precipitation and a coupled land surface and routing model. Water Resour. Res., 50, 26932717, doi:10.1002/2013WR014710.

    • Search Google Scholar
    • Export Citation
  • Xu, J., , Yang D. , , Yi Y. , , Lei Z. , , Chen J. , , and Yang W. , 2008: Spatial and temporal variation of runoff in the Yangtze River basin during the past 40 years. Quat. Int., 186, 3242, doi:10.1016/j.quaint.2007.10.014.

    • Search Google Scholar
    • Export Citation
  • Yang, D., , and Musiake K. , 2003: A continental scale hydrological model using the distributed approach and its application to Asia. Hydrol. Processes, 17, 28552869, doi:10.1002/hyp.1438.

    • Search Google Scholar
    • Export Citation
  • Yilmaz, K. K., , Hogue T. S. , , Hsu K. , , Sorooshian S. , , Gupta H. V. , , and Wagener T. , 2005: Intercomparison of rain gauge, radar, and satellite-based precipitation estimates with emphasis on hydrologic forecasting. J. Hydrometeor., 6, 497517, doi:10.1175/JHM431.1.

    • Search Google Scholar
    • Export Citation
  • Yong, B., , Hong Y. , , Ren L.-L. , , Gourley J. J. , , Huffman G. J. , , Chen X. , , Wang W. , , and Khan S. I. , 2012: Assessment of evolving TRMM-based multisatellite real-time precipitation estimation methods and their impacts on hydrologic prediction in a high latitude basin. J. Geophys. Res., 117, D09108, doi:10.1029/2011JD017069.

    • Search Google Scholar
    • Export Citation
  • Yong, B., , Liu D. , , Gourley J. J. , , Tian Y. , , Huffman G. J. , , Ren L.-L. , , and Hong Y. , 2015: Global view of real-time TRMM Multisatellite Precipitation Analysis: Implication to its successor Global Precipitation Measurement mission. Bull. Amer. Meteor. Soc., doi:10.1175/BAMS-D-14-00017.1, in press.

    • Search Google Scholar
    • Export Citation
  • Zhou, T., , Yu R. , , Chen H. , , Dai A. , , and Pan Y. , 2008: Summer precipitation frequency, intensity, and diurnal cycle over China: A comparison of satellite data with rain gauge observations. J. Climate, 21, 39974010, doi:10.1175/2008JCLI2028.1.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    (a) Topography and CMA gauges and (b) streamflow gauges with hydrological subregions in the Yangtze River basin.

  • View in gallery

    Schematic of the GBHM over the Yangtze River basin.

  • View in gallery

    Spatial maps of (a) annual precipitation (mm) and of bias (mm day−1) for annual mean precipitation between the satellite precipitation products of (b) 3B42 V7 and gauge, (c) 3B42 RT and gauge, and (d) CMORPH and gauge estimates during 2003–12.

  • View in gallery

    Spatial maps of multiyear (2003–12) averaged error components for the warm season (AMJ).

  • View in gallery

    As in Fig. 4, but for the cold season (OND).

  • View in gallery

    Scatterplots of daily basin-averaged rainfall estimated by satellite vs interpolated rainfall by gauges over eight hydrological subregions in the Yangtze River basin.

  • View in gallery

    Comparison of the simulated water balance over eight subregions in the Yangtze River (RB_P, RB_E, and RB_R represent relative bias of precipitation, evapotranspiration, and runoff, respectively).

  • View in gallery

    Comparison between monthly observed discharge and simulation results forced by different precipitation inputs over major tributaries in the Yangtze River.

  • View in gallery

    Comparison between monthly observed discharge and simulation results forced by different precipitation inputs along the mainstream of the Yangtze River.

  • View in gallery

    Flow duration curves of the observed and simulated daily discharge at representative gauges in the Yangtze River.

  • View in gallery

    Evaluation statistics for flood events at different streamflow gauges: (a) RB, (b) CC, and (c) probability of detection.

  • View in gallery

    Observed vs simulated daily streamflow during the July 2012 extreme flood event in the upper Yangtze River.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 106 106 34
PDF Downloads 56 56 24

Multiscale Hydrologic Applications of the Latest Satellite Precipitation Products in the Yangtze River Basin using a Distributed Hydrologic Model

View More View Less
  • 1 State Key Laboratory of Hydroscience and Engineering, Department of Hydraulic Engineering, Tsinghua University, Beijing, China
  • 2 School of Water Resources and Environment, China University of Geosciences, Beijing, China
  • 3 State Key Laboratory of Hydroscience and Engineering, Department of Hydraulic Engineering, Tsinghua University, Beijing, China
  • 4 State Key Laboratory of Hydroscience and Engineering, Department of Hydraulic Engineering, Tsinghua University, Beijing, China, and Department of Civil Engineering and Environmental Science, University of Oklahoma, Norman, Oklahoma
  • 5 Three Gorges Water Cascade Dispatch and Communication Center, Yichang, China
© Get Permissions
Full access

Abstract

The present study aims to evaluate three global satellite precipitation products [TMPA 3B42, version 7 (3B42 V7); TMPA 3B42 real time (3B42 RT); and Climate Prediction Center morphing technique (CMORPH)] during 2003–12 for multiscale hydrologic applications—including annual water budgeting, monthly and daily streamflow simulation, and extreme flood modeling—via a distributed hydrological model in the Yangtze River basin. The comparison shows that the 3B42 V7 data generally have a better performance in annual water budgeting and monthly streamflow simulation, but this superiority is not guaranteed for daily simulation, especially for flood monitoring. It is also found that, for annual water budgeting, the positive (negative) bias of the 3B42 RT (CMORPH) estimate is mainly propagated into the simulated runoff, and simulated evapotranspiration tends to be more sensitive to negative bias. Regarding streamflow simulation, both near-real-time products show a region-dependent bias: 3B42 RT tends to overestimate streamflow in the upper Yangtze River, and, in contrast, CMORPH shows serious underestimation in those downstream subbasins while it is able to effectively monitor streamflow into the Three Gorges Reservoir. Using 394 selected flood events, the results indicate that 3B42 RT and CMORPH have competitive performances for near-real-time flood monitoring in the upper Yangtze, but for those downstream subbasins, 3B42 RT seems to perform better than CMORPH. Furthermore, the inability of all satellite products to capture some key features of the July 2012 extreme floods reveals the deficiencies associated with them, which will limit their hydrologic utility in local flood monitoring.

Corresponding author address: Dawen Yang, Room 312, New Hydraulic Engineering Building, Department of Hydraulic Engineering, Tsinghua University, Beijing 100084, China. E-mail: yangdw@tsinghua.edu.cn

Abstract

The present study aims to evaluate three global satellite precipitation products [TMPA 3B42, version 7 (3B42 V7); TMPA 3B42 real time (3B42 RT); and Climate Prediction Center morphing technique (CMORPH)] during 2003–12 for multiscale hydrologic applications—including annual water budgeting, monthly and daily streamflow simulation, and extreme flood modeling—via a distributed hydrological model in the Yangtze River basin. The comparison shows that the 3B42 V7 data generally have a better performance in annual water budgeting and monthly streamflow simulation, but this superiority is not guaranteed for daily simulation, especially for flood monitoring. It is also found that, for annual water budgeting, the positive (negative) bias of the 3B42 RT (CMORPH) estimate is mainly propagated into the simulated runoff, and simulated evapotranspiration tends to be more sensitive to negative bias. Regarding streamflow simulation, both near-real-time products show a region-dependent bias: 3B42 RT tends to overestimate streamflow in the upper Yangtze River, and, in contrast, CMORPH shows serious underestimation in those downstream subbasins while it is able to effectively monitor streamflow into the Three Gorges Reservoir. Using 394 selected flood events, the results indicate that 3B42 RT and CMORPH have competitive performances for near-real-time flood monitoring in the upper Yangtze, but for those downstream subbasins, 3B42 RT seems to perform better than CMORPH. Furthermore, the inability of all satellite products to capture some key features of the July 2012 extreme floods reveals the deficiencies associated with them, which will limit their hydrologic utility in local flood monitoring.

Corresponding author address: Dawen Yang, Room 312, New Hydraulic Engineering Building, Department of Hydraulic Engineering, Tsinghua University, Beijing 100084, China. E-mail: yangdw@tsinghua.edu.cn

1. Introduction

The current era of satellite remote sensing has provided unprecedented opportunities for the monitoring and prediction of Earth’s terrestrial water conditions (Wood et al. 2011). As a part of the effort to establish the Earth Observing System, a growing number of satellite-based, quasi-global precipitation datasets have been developed and routinely released during the past decade (Ebert et al. 2007). These multisensor blended estimators, including Precipitation Estimation from Remotely Sensed Imagery using Artificial Neural Networks (PERSIANN; Sorooshian et al. 2000), Climate Prediction Center morphing technique (CMORPH; Joyce et al. 2004), PERSIANN–Cloud Classification System (PERSIANN-CCS; Hong et al. 2004), NRL-Blend (Turk and Miller 2005), the TRMM Multisatellite Precipitation Analysis (TMPA) (Huffman et al. 2007), and Global Satellite Mapping of Precipitation (GSMaP; Kubota et al. 2007), are designed to provide high spatial (≤0.25°) and temporal (≤3 h) resolution forcing datasets to facilitate regional and global investigations into the weather, climate, and hydrology, especially for those natural hazards trigged by precipitation extremes, such as floods, landslides, and droughts (Hong et al. 2006; Aragão et al. 2007; Wu et al. 2014).

Despite the continuing great efforts to develop high-resolution satellite precipitation products, the remotely sensed datasets are always subject to significant error sources, arising from indirect measurements, retrieval algorithms, and sampling uncertainties (e.g., McCollum et al. 2002; Nijssen and Lettemaier 2004; Villarini et al. 2009). Furthermore, the error characteristics vary in different climatic regions, storm regimes, seasons, and surface conditions, suggesting that region-specific evaluation of the quality and applicability of these satellite precipitation products is necessary (Sorooshian et al. 2011). In general, the evaluation can be implemented through two strategies: 1) statistics-based direct quantification and 2) modeling-based indirect inference. The former one carries out the straightforward comparison of the satellite estimates with the ground rainfall data (e.g., gauge or weather radar) according to a set of meaningful diagnostic statistics, which has been the subject of numerous validation studies (e.g., Tian et al. 2007; Ebert et al. 2007; Gosset et al. 2013). In contrast, the rationale for the latter approach is application oriented with a focus on the applicability of satellite products to hydrological modeling. The model, as the surrogate of natural watersheds, can diminish (or amplify) and propagate input errors into other simulated fluxes or states, reflecting useful information to the hydrologists who use them to drive the simulation of terrestrial water cycle (Pan et al. 2010).

With this application-oriented view, many studies have investigated the performance of satellite precipitation products by hydrological modeling in specific watersheds around the world (e.g., Su et al. 2008; Li et al. 2009; Bitew and Gebremichael 2011; Gebregiorgis et al. 2012; Yong et al. 2012), indicating that there is an increasing potential to use these products in hydrological modeling with continuing upgrades of retrieval algorithms (Su et al. 2008; Yong et al. 2012). In the future, it is expected that the recently launched Global Precipitation Measurement (GPM) mission will further improve flood monitoring in medium-to-large river basins substantially. However, these products are rarely used in operational hydrologic applications (Bitew and Gebremichael 2011), perhaps partly because of the suspicion of their accuracy for operational use by the hydrologic prediction community and partly because of the lack of long-term systematic assessment with well-calibrated modeling systems to demonstrate their applicability.

In China, rainstorms are the major cause of riverine floods, which is the most frequent and significant type of natural disaster in summer and autumn, ranging from local to regional scales (Ding and Zhang 2009). Compared with conventional rain gauges, satellites provide an alternative way to monitor large-scale rainfall dynamics in time. To our knowledge, many previous studies in China have been conducted to analyze the performance of satellite precipitation data in a statistics-based manner (Zhou et al. 2008; Shen et al. 2010; Gao and Liu 2013; Hu et al. 2014), while modeling-based investigations are only available in very few medium-sized watersheds in China (X. H. Li et al. 2012; Jiang et al. 2012; Yong et al. 2012; Tong et al. 2014). Obviously, without comprehensive hydrological evaluation work emphasizing the application purpose over diverse regions, especially for the large-scale basins where complicated hydrometeorological regimes and land surface conditions exist, the efforts to enhance disaster predictability using satellite data will not bear results.

The present work focuses on the modeling-based evaluation of satellite precipitation data in the Yangtze River, the largest river of China, to bridge the gap between the remote sensing technologies development community and the hydrologic prediction community. We extend our previous 4-yr statistical analysis (Li et al. 2013) into a decade-long (2003–12) evaluation via a physically based distributed modeling framework of the Yangtze, aimed at the following essential goals. First, a physically based distributed model is established over the whole Yangtze River, to serve as the fundamental tool for quantitative prediction as well as the critical platform for integrating increasing amounts of remote sensing data. Second, using popular satellite precipitation products and traditional gauge-based estimates, we compare the simulated hydrological fluxes and states with different forcing data across the Yangtze River basin and implement a long-term systematic evaluation to discuss the potential and challenges for hydrologic applications of the latest satellite precipitation datasets, including annual water balance analysis, streamflow simulation, and extreme flood modeling. We believe this decadal investigation will provide valuable information to the end users who are interested in water resources management, reservoir operation, and natural hazards early warnings (Gosset et al. 2013). Finally, for a better understanding of the error propagation when using a satellite product as an input, we also discuss this issue via water balance analysis, examining the relationship between the bias of rainfall estimation and the bias of other simulated components of water cycle (e.g., evapotranspiration and runoff), as the water balance analysis is claimed to be another important indicator for testing the validity of precipitation data (X. H. Li et al. 2012).

The remaining part of this paper is organized as follows. Section 2 describes the study area and the datasets applied in this study. Section 3 briefly describes the distributed hydrological model, including the setup, calibration and validation of the model, and modeling-based evaluation method. The results are presented and discussed in section 4. Finally, section 5 summarizes major conclusions of this study.

2. Study area and data

a. Yangtze River basin

The Yangtze River basin with a drainage area of 1.8 million km2 is the largest river basin in China. It originates from the Tibetan Plateau and flows eastward for more than 6300 km before draining into the East China Sea (Fig. 1a). The Yangtze River spans nearly one-fifth of mainland China, it experiences diverse landforms and complicated hydroclimatic conditions affected by both East and South Asian monsoon activities (Changjiang Water Resources Commission 1997). As a consequence, this basin suffers frequent floods during the warm season from April to September.

Fig. 1.
Fig. 1.

(a) Topography and CMA gauges and (b) streamflow gauges with hydrological subregions in the Yangtze River basin.

Citation: Journal of Hydrometeorology 16, 1; 10.1175/JHM-D-14-0105.1

To evaluate regional performance of satellite precipitation products, we divide the whole basin into eight hydrological subregions (Fig. 1b), according to its drainage system (Changjiang Water Resources Commission 1999). From upstream to downstream, these subregions are called Jinsha River, Min and Tuo Rivers, Jialing River, Wu River and Three Gorges region, Han River, Dongting Lake river system, Poyang Lake river system, and the middle and lower Yangtze mainstream (see Table 1).

Table 1.

Streamflow gauges and hydrological regions in the Yangtze River basin. In the second column, the full gauge names are: PS, Pingshan; GC, Gaochang; BB, Beibei; WL, Wulong; HJG, Huangjiagang; TY, Taoyuan; XT, Xiangtan; WZ, Waizhou; CT, Cuntan; YC, Yichang; HK, Hankou; and DT, Datong. In the fifth column, the full hydrological subregion names are: I, Jinsha River; II, Min and Tuo Rivers; III, Jialing River; IV, Wu River and Three Gorges region; V, Han River; VI, Dongting Lake river system; VII, Poyang Lake river system; and VIII, middle and lower Yangtze mainstream.

Table 1.

b. Ground gauge data

The daily meteorological observations from 1961 to 2012 are obtained from 141 China Meteorological Administration (CMA) gauges (circles in Fig. 1a), including daily precipitation; wind speed; maximum, minimum, and mean air temperature; relative humidity; and hours of sunshine.

The daily discharge records (1961–2011) are collected from the Hydrological Year Book published by the Hydrological Bureau of the Ministry of Water Resources the People’s Republic of China, and discharge data from 2012 is separately obtained from China Three Gorges Corporation. In this study, 12 streamflow gauges located in the major tributaries and along the Yangtze mainstream are selected (squares indexed with numbers in Fig. 1b; also see Table 1).

c. Satellite precipitation data

According to previous statistical evaluation work in the same basin by Li et al. (2013), we select three sets of satellite products as the forcing data for distributed modeling over the Yangtze River: the latest released version 7 of TMPA data (the post-real-time product and near-real-time product are hereinafter referred to as 3B42 V7 and 3B42 RT, respectively) and CMORPH. All these products are generated by combining information from both passive microwave (PMW) and infrared (IR) observations with high spatial (0.25°) and temporal (3 h) resolutions. The three sets of precipitation estimators are all available during the study period of 2003–12.

d. Other geographical information

The basin-scale geographical information is extracted from a series of global datasets. The digital elevation model (DEM) data are acquired from the USGS Hydrological Data and Maps based on Shuttle Elevation Derivatives at Multiple Scales (HydroSHEDS) (http://hydrosheds.cr.usgs.gov/index.php) at 100-m resolution. Land use/land cover (LULC) data are obtained from the Environmental and Ecological Science Data Center of West China (http://westdc.westgis.ac.cn/), consisting of three images from different periods (1980s, 1990s, and 2000s) at 100-m resolution. The soil type is obtained from the Digital Soil Map of the World (Food and Agricultural Organization 2003) at 10-km resolution, while the corresponding soil properties (e.g., the porosity and the saturated hydraulic conductivity) and other soil water parameters can be estimated from IGBP–Data Information Systems (DIS) Global Soil Database (www.daac.ornl.gov). The vegetation dynamics are described by monthly leaf area index (LAI), derived from the 8-km Global Inventory Modeling and Mapping Studies (GIMMS) normalized difference vegetation index (NDVI) products, which are also archived from the Environmental and Ecological Science Data Center of West China.

3. Modeling framework

In the present study, a geomorphology-based hydrological model (GBHM; Yang and Musiake 2003; Cong et al. 2009; M. Li et al. 2012; Ryo et al. 2014) is applied to the whole Yangtze River basin to simulate the hydrological processes with various precipitation inputs. This distributed modeling framework takes advantage of the geomorphologic similarities to reduce the spatial-structure complexity within a grid and to characterize the catchment topography by hillslope–stream formulation. In brief, GBHM includes the following components: a gridded discretization scheme, a subgrid parameterization scheme, a hillslope-based hydrological modeling module, and a kinematic wave flow routing module.

a. Model setup

Hillslope is the computational unit for hydrological simulation in GBHM (Fig. 2), and the hillslope-based hydrological modeling includes the following processes: snowmelt, canopy interception, evapotranspiration, infiltration, surface and subsurface flow, and the water exchange between groundwater and the river channel (Yang and Musiake 2003). A detailed description of GBHM physical representations can be found in several recent papers (Cong et al. 2009; Ryo et al. 2014). All parameters used in the GBHM of the Yangtze River are summarized in Table 2.

Fig. 2.
Fig. 2.

Schematic of the GBHM over the Yangtze River basin.

Citation: Journal of Hydrometeorology 16, 1; 10.1175/JHM-D-14-0105.1

Table 2.

Summary of parameters used in the Yangtze GBHM.

Table 2.

The runoff generated at the hillslope is the lateral inflow into the river channel (Fig. 2) within the flow interval. Flow routing by the kinematic wave approach is then applied to each flow interval sequentially to get the discharge at the outlet of the river basin.

Considering the computational capacity and the available geographical datasets, we construct GBHM of the Yangtze River basin using the grid size of 10 km. The whole Yangtze River is then divided into 137 subbasins organized by the Pfafstetter system (Yang and Musiake 2003), and all of the grids within a subbasin are classified into a series of flow intervals. Within a flow interval, the grids are grouped into different categories of hillslopes by the subgrid parameterization procedure. The hillslopes are identified according to land cover types and topographical characteristics, which are calculated from the geographical information data with finer resolutions (e.g., 100-m DEM and 100-m LULC).

The 10-km meteorological forcing fields are interpolated from 141 CMA gauges: the daily precipitation, wind speed, relative humidity, and sunshine hours are interpolated using an angular distance weighting method (New et al. 2000), whereas the maximum, minimum, and mean temperature are interpolated via an elevation-corrected angular direction weighting method (Yang and Musiake 2003).

It is known that initial condition plays a critical role in hydrologic prediction, and thus, we implement a 10-yr warm-up simulation to determine it. A predefined initial field is given at the beginning of a warm-up period, and then at the end of the 10-yr warm-up the simulated soil moisture and groundwater table will approximate the real condition, which is recorded to preset the initial condition for future simulations.

b. Model calibration and validation

As shown in Table 2, there are three GBHM parameters in total (snowmelt factor, groundwater hydraulic conductivity, and specific yield for unconfined aquifer) that need to be calibrated manually. Considering the available data and recent human impacts, we use daily discharge data during the period of 1961–65 to calibrate the model and then use data from 1966 to 2002 for a long-term validation. It is noted that simulations during both calibration and validation periods are forced by the CMA gauge-interpolated forcing data.

To quantitatively evaluate the model performance, the Nash–Sutcliffe coefficient of efficiency (NSCE; Nash and Sutcliffe 1970) and the relative error (RE) are calculated:
e1
and
e2
where Qobs,i is the observed discharge at the ith day, Qsim,i is the simulated discharge at the ith day, and is the mean value of the observed discharge series.

Table 3 shows the long-term performance of GBHM during the period of 1961–2002. It is found that, during calibration and validation periods, the NSCE value is greater than 0.70 while RE is constrained within a reasonable range from −4.0% to 12.0% at most of the streamflow gauges. The result also indicates there is a significant decline of the performance at Huangjiagang during the validation period. This is caused by the impoundment of the Danjiangkou Reservoir since 1968 in the upper Han River, and therefore, this gauge is excluded from the following analysis.

Table 3.

The performance of GBHM during the calibration and validation periods.

Table 3.

c. Modeling-based evaluation

As shown in the results of model calibration and validation, the Yangtze River GBHM can provide a reasonable modeling framework to discuss the pros and cons of satellites precipitation data for prediction applications during the past decade (2003–12).

The three sets of satellite precipitation products (3B42 V7, 3B42 RT, and CMORPH) are integrated from their original 3-h scale into the daily scale first to match the local time, and then daily precipitation fields are projected and resampled into the 10-km modeling grid system.

With exactly the same initial condition (given by the long-term modeling of 1961–2002), GBHM is then driven by the gauge-interpolated fields, 3B42 V7, 3B42 RT, and CMORPH, respectively. Based on the 10-yr inputs and simulated results, including precipitation, runoff, actual evapotranspiration, and discharge in the river channels, the utilities of the latest satellite precipitation estimators to hydrologic applications are further presented and discussed in the following text.

4. Results and discussion

In this section, for a better understanding of the modeling results with different input datasets, statistical comparison of various precipitation estimates is first revisited, then a comprehensive multiscale evaluation based on hydrological modeling is discussed, focusing on three elements from the perspective of hydrologic applications: 1) water balance analysis for water resources assessment, 2) streamflow simulation for river management, and 3) near-real-time flood forecasting for natural disasters warning and mitigation.

a. Comparison of satellite precipitation with the gauge estimates

Figure 3 shows the spatial maps of gauge-observed annual precipitation (mm) and mean bias of annual precipitation (mm day−1) between different satellite estimates and gauge estimates during 2003–12. In general, precipitation increases from the west to the east of the basin, ranging from 348 mm (over the Tibetan Plateau) to 1700 mm (over the Dongting Lake river system and Poyang Lake river system), although there is an obvious enhancement of precipitation over the Sichuan basin. Looking into the spatial distribution of bias, it is clear that 3B42 V7 shows the closest agreement with the gauge estimate, indicating the critical role of monthly gauge correction algorithm for bias removal. Comparatively, the other two near-real-time estimators without gauge correction (3B42 RT and CMORPH) have an evident local bias (Figs. 3c,d).

Fig. 3.
Fig. 3.

Spatial maps of (a) annual precipitation (mm) and of bias (mm day−1) for annual mean precipitation between the satellite precipitation products of (b) 3B42 V7 and gauge, (c) 3B42 RT and gauge, and (d) CMORPH and gauge estimates during 2003–12.

Citation: Journal of Hydrometeorology 16, 1; 10.1175/JHM-D-14-0105.1

In the upper Yangtze River, 3B42 RT overestimates (by >1 mm day−1) precipitation while CMORPH presents a mixed pattern with both negative (from −0.2 to −1 mm day−1 in the western part) and positive (from 0.5 to 2 mm day−1 in the eastern part) bias. This bias pattern is consistent with previous studies (Shen et al. 2010; Li et al. 2013), suggesting there are great uncertainties associated with satellite-based precipitation retrievals over the Tibetan Plateau. As 3B42 RT is developed by PMW-calibrated IR technique, perhaps the positive biases can be attributed to the dominant usage of IR images without sufficient calibration by PMW data, considering most PMW overpasses have been screened out because of the snow covers and complex terrains in the Tibetan Plateau (Yong et al. 2015). In contrast, as CMORPH utilizes a different technique that propagates PMW estimates by IR-derived advection vectors, its bias pattern has been contaminated by uncertainties associated with both IR and PMW measurements, and it thus shows more complicated error features over there. When focusing on the middle and lower Yangtze River, CMORPH seems to underestimate precipitation consistently (by from −0.2 to −2 mm day−1), but 3B42 RT shows a similar mixed error pattern as 3B42 V7.

Following the error component decomposition scheme proposed by Tian et al. (2009), we also decompose the total bias E into three independent parts (i.e., hit bias H, false precipitation F, and missed precipitation M, related by E = H + FM) at the seasonal scale to understand the error sources and their evolution. Following our previous study (Li et al. 2013), in the Yangtze River, we define the spring as January–March (JFM), the summer as April–June (AMJ), the autumn as July–September (JAS), and the winter as October–December (OND). This seasonal division scheme is consistent with the flood season (AMJ and JAS) in the Yangtze and is thus useful for studying the hydrological cycle of this region. Because of the limited space, we only take AMJ and OND to represent the warm season and cold season, respectively (as in the same season, the error pattern tends to be similar during different months; see Li et al. 2013). Figures 4 and 5 present the spatial maps of multiyear (2003–12) averaged error components for the warm season (AMJ) and the cold season (OND), respectively.

Fig. 4.
Fig. 4.

Spatial maps of multiyear (2003–12) averaged error components for the warm season (AMJ).

Citation: Journal of Hydrometeorology 16, 1; 10.1175/JHM-D-14-0105.1

Fig. 5.
Fig. 5.

As in Fig. 4, but for the cold season (OND).

Citation: Journal of Hydrometeorology 16, 1; 10.1175/JHM-D-14-0105.1

During the warm season (Fig. 4), the TMPA series estimators (3B42 V7 and 3B42 RT) share considerable similarities in their spatial distributions of the total bias and its components over most parts of the Yangtze River, except the upper reaches. Over the Jinsha River, 3B42 RT has remarkable overestimation (>100 mm), which is mainly attributed to positive hit bias combined with several false alarms. However, the mixed error pattern in the middle and lower Yangtze is dominated by both positive hit bias (from 60 to 200 mm) and negative missed precipitation (from −60 to −200 mm) for 3B42 RT. Because of the complex precipitation regime itself over the Yangtze River shaped by monsoon climate as well as complex terrains, we suspect the missed events may be caused by the inability to catch warm rainfall or short-lived convective storms, and the positive hit biases perhaps can be related to the overestimation by PMW land algorithm for the convectively active regimes (Tian et al. 2009). In addition, the discrepancies between the hit bias (false precipitation) maps of 3B42 V7 and 3B42 RT demonstrate the effect of gauge-based adjustment, but unfortunately, the correction procedure cannot recover undetected events to further reduce the amounts of missed precipitation. In contrast, CMORPH presents a different error pattern from the TMPA series: its total biases are dominated by underestimation (from −30 to −200 mm), with only a part of the upper reaches along the eastern edge of the Tibetan Plateau covered by overestimations. Considering the complex terrains in that region (Fig. 1a), we speculate this is mostly linked to the anomalously high PMW rainfall contaminated by snow cover in the surrounding high mountains.

For the cold season (Fig. 5), the total bias of all three products is still dominated by both hit bias and missed precipitation. The most obvious common feature shared by them is that the most serious biases are concentrated largely over the lower Yangtze River, as the precipitation bands have “shifted back” to this area because of the movement of the East Asian monsoon during the winter. Over the upper and middle Yangtze, the precipitation amount during this season is very limited, but still, there is a long-lasting overestimation for 3B42 RT and CMORPH, respectively. The former is caused by hit bias, which is similar to the situation in the warm season, while the latter, unlike the warm season, is evidently caused by false alarms. Qualitatively, we also note there are pixels with obviously high false alarm rate (the false precipitation amount over these pixels is more than twice the values of their surrounding pixels) over the lower Yangtze for all products (see box 1 and box 2 in Fig. 5; as estimated precipitation is much lower than TMPA series, CMORPH does not show these pixels when its map is scaled into the same color bar). It is found that these pixels correspond to the locations of Poyang Lake and Dongting Lake, and this false rainfall-like signal over the inland water bodies is apparently caused by the deficiencies of PMW-based retrievals for emissivity characterization (Tian and Peters-Lidard 2007). Since this effect only exits at a relatively small spatial scale (~1000 km2), it is believed to have little impact on basin-scale hydrologic modeling.

At the daily scale, we further compare basin-scale rainfall interpolated by gauges against that estimated based on satellites in Fig. 6. Several statistical metrics are calculated, including relative bias (RB), root-mean-square error (RMSE), and Pearson’s correlation coefficient (CC). Generally, all three satellite estimators show a strong correlation with gauge estimation results, interpreted by the high values of CC (0.62–0.86). According to the relative bias, 3B42 V7 gives a fairly good approximation of gauge estimates, with a slight overestimation (RB lies within the range of 0%–10%) over most subregions. In summary, estimated biases are smaller in Jinsha River, Min and Tuo Rivers, Wu River, and Three Gorges region, and Han River, compared to other subregions over the lower Yangtze River. As the mixed error pattern exits at these regions (Figs. 35), the positive and negative biases will cancel each other out and make basin-scale estimates unbiased or less biased.

Fig. 6.
Fig. 6.

Scatterplots of daily basin-averaged rainfall estimated by satellite vs interpolated rainfall by gauges over eight hydrological subregions in the Yangtze River basin.

Citation: Journal of Hydrometeorology 16, 1; 10.1175/JHM-D-14-0105.1

Without gauge adjustment, 3B42 RT is found to have overestimated daily basin-scale rainfall over the whole Yangtze Rivers, especially in the upper Yangtze (RB is 75.67%, 36.96%, and 18.28% for Jinsha River, Min and Tuo Rivers, and Jialing River, respectively). At the same time, another unadjusted near-real-time product, CMORPH, is found to work as the complementary dataset for 3B42 RT, slightly underestimating basin-scale rainfall over the upper Yangtze and seriously underestimating in the lower reaches (RB is −31.44%, −38.79%, −35.13%, and −31.41% for Wu River and Three Gorges region, Dongting Lake region, Poyang Lake region, and the middle and lower Yangtze mainstream, respectively). Combined with the error features characterized at annual and seasonal scales (Figs. 35), the daily result also suggests that there is a consistent overestimation (underestimation) across multiple time scales for 3B42 RT (CMORPH).

Additionally, it is demonstrated that 3B42 V7 does not always show its superiority over other products at the daily scale, in particular CMORPH, in terms of RMSE and CC (e.g., Figs. 6b–d). This implies, although the monthly satellite–gauge (SG) combination algorithm tends to make 3B42 V7 statistically closer to monthly gauge observation (to minimize the bias), there is no guarantee for the improvement of daily precipitation estimates within a month (Li et al. 2013; Yong et al. 2015).

b. Analysis of annual water balance simulation

Figure 7 compares averaged simulation results of annual water balance components (basin-scale precipitation, evapotranspiration, and runoff) over the eight hydrological subregions in the Yangtze River during the study period (2003–12).

Fig. 7.
Fig. 7.

Comparison of the simulated water balance over eight subregions in the Yangtze River (RB_P, RB_E, and RB_R represent relative bias of precipitation, evapotranspiration, and runoff, respectively).

Citation: Journal of Hydrometeorology 16, 1; 10.1175/JHM-D-14-0105.1

First of all, this result clearly provides a general idea of annual water budgeting of the Yangtze River: precipitation and evapotranspiration increase from the upper reaches to the lower reaches, with an amount of 700–1600 and 400–800 mm, respectively. However, runoff, ranging from 300 to 800 mm, shows a more complicated pattern as it is controlled by the variations of runoff coefficient. The gauge-based modeling results are also consistent with many previous studies for regional water budget (Gao et al. 2007; Xu et al. 2008).

To quantitatively characterize the error for water balance simulation and understand how the input error propagates into other simulated hydrologic fluxes and states, we calculate the relative bias (%) by dividing the bias of satellite-driven simulation by the gauge-driven modeling result (for each water balance component). As expected from the comparisons discussed above, the 3B42 V7 result shows the closest agreement with the gauge-based simulation. Overall, this dataset can provide very reasonable simulated results of evapotranspiration (with a relative bias from −1.8% to 4.9%) as well as runoff (with a relative bias from −1.9% to 6.8%). It is demonstrated that 3B42 V7 should be the most appropriate dataset for long-term regional water balance studies.

As both 3B42 RT and CMORPH show region-dependent bias for rainfall estimation, the error will thus be propagated into other simulated variables correspondingly. In summary, 3B42 RT tends to substantially overestimate runoff in those upstream basins (e.g., Jinsha River and Min and Tuo Rivers; Figs. 7a,b) while CMORH seriously underestimates it in the downstream subregions (e.g., Wu River and Three Gorges Region, Han River, Dongting Lake and Poyang Lake region, and middle and lower mainstream; Figs. 7d–h). Furthermore, by comparing the relative bias of the estimated precipitation, simulated evapotranspiration, and simulated runoff among various products, we can provide implications on error propagation from the input forcing into other simulated hydrologic fluxes and states.

Comparing modeling results from 3B42 RT and gauges, we find that basin-scale actual evapotranspiration presents no substantial change (relative bias varies between −1.1% and 4.4%) with increased precipitation over all regions, reflecting the energy-control nature without water stress for evapotranspiration over the whole Yangtze during the warm season. As the total precipitation bias is dominated by errors in summer (Li et al. 2013), the positive bias in the 3B42 RT estimate has thus been mainly propagated into simulated runoff bias. Looking into the relative biases of precipitation and runoff, we also find that the relative bias has been enhanced through hydrologic modeling because of the nonlinear nature of watershed rainfall–runoff processes. Roughly, from rainfall to runoff, the bias has been multiplied by a factor of 2. At the same time, comparison between the results of gauge and CMORPH indicates that reduced inputs will cause decreases in simulated runoff and in evapotranspiration simultaneously, suggesting evapotranspiration tends to be more sensitive to negative bias in precipitation over the Yangtze. Similarly, the negative relative bias in precipitation has also been enhanced by a factor around 1.5 for most subregions.

c. Evaluation of streamflow simulation

Figure 8 presents the simulated monthly streamflow against observations over major tributaries of the Yangtze, and the performance of both monthly and daily simulations is also quantified by NSCE and RE in Table 4. It is evident that 3B42 V7, the gauge-adjusted rainfall data, works fairly well to capture the general streamflow dynamics of all tributary basins with a medium size (from 8 × 104 to 46 × 104 km2). Comparing with gauge-driven results in Table 4, we can find that there is a substantial decline of model performance when we use 3B42 V7 as the forcing data for daily simulation in the downstream tributaries (e.g., Yuan River and Gan River), though it shows good agreement with gauge-based results at the monthly scale. This can be explained by Fig. 6, which shows more scattered and less correlated daily rainfall estimates by 3B42 V7 in the downstream subbasins (Figs. 6f–h), indicating the challenge for satellite data to adequately capture short-lived heavy rainstorms there. As for another two sets of near-real-time product, 3B42 RT and CMORPH, their streamflow simulation results also present remarkable region-dependent errors, which have close correspondence to the local bias contained in precipitation fields (Figs. 35). The 3B42 RT seriously overestimated streamflow in the upstream subbasins (e.g., Jinsha River and Min and Tuo Rivers) during the warm season, which can be traced back to large positive total bias in precipitation (Fig. 4); as a result of the significantly underestimated precipitation, CMORPH cannot reproduce flood events in the summer for the subbasins over the middle and lower Yangtze (e.g., Wu River and Three Gorges region, Yuan River, and Gan River). In other words, this suggests 3B42 RT can get better streamflow modeling results in those midstream and downstream subbasins while CMORPH should be applied to the upstream subbasins to get reasonable simulations. However, considering the mixed error pattern in the Yangtze River discussed above (for CMORPH, mainly in Jinsha River and Min and Tuo Rivers, and for 3B42 RT, mainly in the middle and lower Yangtze; see Figs. 3, 4), special caution should therefore be taken when 3B42 RT and CMORPH are applied to streamflow modeling at catchments with smaller scales compared to subbasins discussed in this study, since the local positive bias and negative bias perhaps cannot cancel each other out. For instance, in the very upstream part of the Yangtze River, CMORPH still shows a dominant negative bias (Fig. 3d), and as a result, the simulated streamflow is largely underestimated over the source region of the Yangtze (Tong et al. 2014).

Fig. 8.
Fig. 8.

Comparison between monthly observed discharge and simulation results forced by different precipitation inputs over major tributaries in the Yangtze River.

Citation: Journal of Hydrometeorology 16, 1; 10.1175/JHM-D-14-0105.1

Table 4.

Performance of streamflow simulation by different precipitation datasets.

Table 4.

In addition, Fig. 9 compares the simulated monthly streamflow with observations along the mainstream. This result, combined with those summary statistics (Table 4), shows again that the performance of 3B42 V7 is in exceptionally good agreement with the observation from the upper to the lower Yangtze mainstream. As the drainage area becomes larger for these mainstream stations (from 87 × 104 to 170 × 104 km2), the simulation results indicate that modeling performance will be better with larger watershed size. As the general pattern of 3B42 RT and CMORPH estimated rainfall is dominated by positive and negative bias, respectively (Fig. 3), 3B42 RT therefore tends to generally overestimate streamflow (with RB from 80.7% to 144.3%) along the main Yangtze while CMORPH underestimates it overall (with RB from −23.1% to −45.7%). It is also noted that, at Cuntan station in the upper Yangtze, CMORPH is able to get reasonable results with moderate underestimation for streamflow modeling (NSCE value is 0.64 and RB is around −20% for daily results). As Cuntan is the inflow gauge to the Three Gorges Reservoir, CMORPH thus offers an alternative way to effectively monitor the streamflow poured into the reservoir in a near-real-time manner, which can potentially provide useful information for the regulation of Three Gorges Dam.

Fig. 9.
Fig. 9.

Comparison between monthly observed discharge and simulation results forced by different precipitation inputs along the mainstream of the Yangtze River.

Citation: Journal of Hydrometeorology 16, 1; 10.1175/JHM-D-14-0105.1

To get a better understanding of the full spectrum of daily streamflow regime, we further compare the simulated streamflow against observation at several gauges in the form of flow duration curves (Fig. 10). Evidently, this figure provides illustration on how the simulated flow regime changes with different river flow levels among various input datasets.

Fig. 10.
Fig. 10.

Flow duration curves of the observed and simulated daily discharge at representative gauges in the Yangtze River.

Citation: Journal of Hydrometeorology 16, 1; 10.1175/JHM-D-14-0105.1

Similarly, the results confirm the region-dependent bias for 3B42 RT and CMORPH again. However, additional information can be found: the discrepancies of CMORPH-driven simulation over the upstream basins (Figs. 10a,b) mainly come from high-flow conditions (when the exceedance probability falls below 40%), but it works similarly as a gauge for simulation of low-flow conditions. In contrast, 3B42 RT tends to seriously overestimate it at the full range of the streamflow regime in the upper Yangtze River. For the downstream basins, the regime reproduced by 3B42 RT becomes very similar to that simulated by both 3B42 V7 and gauge (Figs. 10c,d), which can be inferred from their similar spatial error pattern over this region (Figs. 35), but the significant negative biases attached to CMORPH estimations will distort the flow regime seriously at full range. Moreover, we also noted that, for the stations over Dongting River region and Poyang Lake region (taking Waizhou as an example here; Fig. 10d), all satellite estimates as well as the gauge estimates cannot reproduce the observed flow regime accurately for the medium- to low-flow conditions.

d. Applicability for near-real-time flood forecasting

Finally, we extend our investigations into the applicability of satellite rainfall products to flood forecasting, especially for near-real-time monitoring of floods in the Yangtze River. As suggested by previous global evaluation research (Wu et al. 2014), we also apply the percentile-based (95th percentile value of the flow duration curves) method to the observed streamflow data and eventually select 394 typical flood events in total for this study.

Figure 11 summarizes the diagnostic statistics calculated by comparing the observed flood events and corresponding simulations. Here we flag the flood events as detected only when the absolute relative bias of simulated discharge is no more than 20%, and the ratio of the detected events to total flood events is then defined as the probability of detection. Obviously, 3B42 V7 outperforms the other two satellite products for flood forecasting in terms of all three statistics, especially because it has remarkably reduced and stabilized the bias to yield almost exactly the same performance as the gauge (Fig. 11a). In addition, we find that the probability of detection for both gauge and 3B42 V7 drops from ~0.7 to ~0.3 as the drainage area decreases (Fig. 11a), indicating that the flood detection capabilities for both gauge and 3B42 V7 seem scale dependent, as they have a better performance over large watersheds.

Fig. 11.
Fig. 11.

Evaluation statistics for flood events at different streamflow gauges: (a) RB, (b) CC, and (c) probability of detection.

Citation: Journal of Hydrometeorology 16, 1; 10.1175/JHM-D-14-0105.1

Considering the potential to further utilize satellite estimates for near-real-time flood monitoring, we also compare the performance of 3B42 RT and CMORPH during the selected flood events. Clearly, we can find distinct features for the two data (Fig. 11a): from the upper to the lower Yangtze mainstream, the negative bias in CMORPH’s simulation will accumulate (from −40% to −50%) while the positive bias in 3B42 RT’s simulation will diminish slightly (from 70% to 60%); over the tributaries, CMORPH’s simulations also present increased biases from the west to the east while the bias of 3B42 RT’s simulation drops down and stays around 10%. According to the other two statistics, it is also found that the performance of 3B42 RT approximates that of 3B42 V7 over downstream tributaries. In summary, event-based analysis shows that 3B42 RT and CMORPH should have competitive performances for near-real-time flood forecasting in the upper Yangtze River, but over the downstream tributaries, 3B42 RT definitely performs better than CMORPH. It is emphasized that 3B42 RT has higher probability of detection for flood events, but it generates more false alarm events simultaneously because of overestimation. This performance is declared to be better than that of CMORPH for near-real-time flood monitoring, since false alarm events have greater values than “undetected” events in the context of disaster early warnings and risk management.

In July 2012, the Three Gorges Dam was hit by a record flood (with the largest inflow of 71 200 m3 s−1), which was reported as the biggest one since the establishment of the dam in 2003, with even a higher local river flow than the devastating Yangtze floods in 1954 and 1998 (Wang and Zhang 2013). As a case study of extreme floods in the Yangtze River, we analyze the performance of satellite precipitation products when they are applied into the modeling of this extreme event. Figure 12 shows the results driven by different inputs at Gaochang (Min River) and Beibei (Jialing River), the two flooding source areas of this event, as well as the result at Cuntan, which is the inflow gauge to the Three Gorges Reservoir.

Fig. 12.
Fig. 12.

Observed vs simulated daily streamflow during the July 2012 extreme flood event in the upper Yangtze River.

Citation: Journal of Hydrometeorology 16, 1; 10.1175/JHM-D-14-0105.1

In general, all three sets of remote sensing products can reproduce the key pattern of this extreme flood in two tributaries (Gaochang and Beibei); however, gauge-adjusted 3B42 V7 obviously no longer shows its superiority over the other two near-real-time satellite products during this event. This result suggests that the monthly gauge-based correction scheme does not promise a better result for floods simulation and thus calls for a new strategy to calibrate satellite data in a near-real-time manner to improve their performance for flood forecasting. Meanwhile, it is apparent that all satellite estimates have missed the 15–19 July floods in Min River, indicating the inherent deficiencies associated with all satellite data. Our investigation on the precipitation map shows that this is caused by a localized rainstorm on 15–16 July that was not been captured by satellite products (figure not shown). Therefore, it seems that, because of the indirect way of taking measurements and the sampling issue, there is still a great challenge for the application of current satellite-based estimates into local flood monitoring.

e. Discussions on gauge data, model calibration, and human activities

In the present study, the gridded “ground truth” of rainfall is obtained by interpolation of gauge data. Therefore, uncertainties may arise because of 1) inadequate sampling by gauge measurement and 2) interpolation method applied to infer the spatial fields of rainfall. The inherent scale mismatching associated with gauge-based validation can make the comparison and evaluation potentially problematic (Gebremichael et al. 2003); however, we believe this problem is unlikely to cause substantial errors in the investigations discussed above, since we mainly focus on the basin-scale hydrologic simulation. It is also found that large temporal integration scales (e.g., monthly and daily) will depress the spatial sampling error and then get a reasonable approximation to areal true rainfall (Villarini et al. 2008). On the other hand, we would like to highlight the rationale of “modeling-based evaluation,” which emphasizes using gauge-driven simulations as the benchmark to investigate the applicability of satellite precipitation products for regional hydrological modeling. In practice, these CMA gauges are the most widely used dataset to force hydrological modeling in China; therefore, in this study, we only provide a “relative judgment” on the skills of newly available multisensor merged rainfall estimates rather than state the “absolute advantages” of satellite estimates.

As the basic tool for evaluation, GBHM is calibrated by historical gauge-observed rainfall and streamflow data, and then all the model parameters are fixed during both validation period (1966–2002) and evaluation period (2003–12). However, many previous studies (e.g., Yilmaz et al. 2005; Yong et al. 2012) also suggested a satellite-based recalibration to improve the performance of hydrological modeling when the satellite product is applied as input. Nevertheless, we did not adopt this strategy because of the following reasons. First, this study aims to comparatively evaluate hydrological simulations forced by various rainfall inputs. Hence, the same model with fixed parameters is the prerequisite for a meaningful comparison. Second, long-term gauge data records in history promise a wide range of hydrologic conditions to ensure a model’s validity (Su et al. 2008). Finally, it is also reported that the improved performance of streamflow modeling by recalibration will work at the expense of distorting other simulated fluxes, such as evapotranspiration (Bitew and Gebremichael 2011).

Another critical issue that needs to be discussed is the impact of the reservoir projects in the Yangtze River basin and their implications to our evaluation work. For the major tributaries of the Yangtze River basin, we can only identify obvious alternation of the natural river flow in the Han River due to the impoundment of the Danjiangkou Reservoir since 1968; thus, this subregion has been excluded from our analysis of streamflow simulation. For the Yangtze mainstream, with limited literature (Gao et al. 2012), we knew that only the Three Gorges Dam impacts the lower mainstream’s flow regime from October through the following February. The greatest effect was found at the Yichang gauge (very close to the Three Gorges Dam), and the effect became less at the Hankou and Datong gauges, where more water flows into the mainstream from the middle and downstream tributaries. In summary, the overall impacts of the reservoir projects should be limited, especially over those major tributaries and the upper mainstream. However, as human activities exerted increasing impacts on natural hydrologic processes in the Yangtze River, we would like to leave it as an open question to be addressed quantitatively in future studies.

5. Conclusions

With decade-long (2003–12) observation datasets, this thorough evaluation aims at assessing the multiscale hydrologic utilities for three sets of the most popular high-resolution multisensor blended global precipitation products (3B42 V7, 3B42 RT, and CMORPH) via a physically based distributed hydrological modeling framework over the Yangtze River, the largest watershed in China. To accomplish this application-oriented evaluation work, we first establish and validate a 10-km distributed GBHM; then statistically compare different precipitation estimates at multiple temporal scales; and finally examine their utilities in terms of various hydrologic applications, including annual water balance simulation, streamflow modeling, and near-real-time flood monitoring. The major conclusions of this modeling-based evaluation work can be summarized as follows.

  1. For comparisons of precipitation input datasets, in summary, 3B42 V7 shows the closest agreement with gauge estimates in terms of the bias, and comparatively, the other two near-real-time estimators present evident local bias. In the upper Yangtze, 3B42 RT seriously overestimates precipitation while CMORPH has a mixed error pattern, all indicating there are great uncertainties for satellite-based precipitation retrievals over the Tibetan Plateau; for the middle and lower Yangtze, 3B42 RT shows a similar mixed error pattern as 3B42 V7, but CMORPH tends to underestimate precipitation substantially. By the decomposition scheme, it is found that the total bias for all satellite estimators is dominated by hit bias and missed precipitation during both warm and cold seasons. Additionally, daily comparison implies that 3B42 V7 does not always show superiority over other products at daily scale, suggesting the monthly SG combination algorithm provides no guarantee for improvement of daily precipitation estimates.
  2. For annual water balance simulation, as the most appropriate dataset for regional water budgeting study, 3B42 V7 works fairly well to get results comparable to gauge-driven simulations (with relative bias from −1.8% to 4.9% for evapotranspiration and from −1.9% to 6.8% for runoff). Comparing the results of 3B42 RT and CMORPH, it is also found that the bias in precipitation estimates has been mainly propagated into simulated runoff, and simulated evapotranspiration tends to be more sensitive to negative bias.
  3. For streamflow modeling, it is found that the 3B42 V7–driven simulation shows fair agreement with observations in those upstream subbasins at both monthly and daily scales, but its performance declines obviously for daily modeling over the downstream basins, reflecting the challenges for satellite estimators to adequately capture heavy rainstorms over the lower Yangtze. The results also suggest that 3B42 RT tends to get better modeling results in the midstream and downstream subbasins while CMORPH can be applied to the upstream subbasins. However, as the mixed error pattern exists, special caution should be taken when we apply 3B42 RT and CMORPH to modeling at catchments with smaller scale compared to subbasins discussed in this study, since local positive bias and negative bias perhaps cannot cancel each other out.
  4. For near-real-time flood monitoring, with the 394 selected flood events during the study period, we find that 3B42 RT and CMORPH should have competitive performances for near-real-time flood monitoring in the upper Yangtze River, but in the downstream tributaries, 3B42 RT performs better than CMORPH. During the extreme flood event in July 2012, the inability of all products to reproduce key features suggests there are inherent deficiencies associated with current satellite rainfall products when they are applied to the monitoring and warning of local floods.
With the four aspects discussed above, we believe the present study will promote better utilization of satellite precipitation products in various hydrologic applications over the Yangtze River. Clearly, satellite precipitation products provide valuable information to regional water resources assessment, river management, and natural hazards warning, and this paper also initially illustrates a demo of a physically based distributed modeling and forecasting framework over the Yangtze River. Future efforts will be made to complement comprehensive evaluation for flood prediction in small- to medium-sized basins, as well as to develop multiscale multisource merging techniques to effectively combine ground observations and remote sensing estimates. Furthermore, as the Global Precipitation Mission (GPM) Core Observatory was successfully launched in February 2014, the modeling framework presented here can be readily employed to benchmark the upcoming GPM-era satellite precipitation data into a regional operational hydrological prediction system over the Yangtze River.

Acknowledgments

This research was supported by the National Natural Science Funds for Distinguished Young Scholars (Project 51025931) and the National Natural Science Foundation of China (Project 51190092). The first author also acknowledges the HyDROS Lab (http://hydro.ou.edu) at the National Weather Center, Norman, OK, for their support during his visiting. The authors wish to thank the two reviewers and Dr. Christa D. Peters-Lidard for their insightful comments.

REFERENCES

  • Aragão, L. E. O. C., , Malhi Y. , , Roman-Cuesta R. M. , , Saatchi S. , , Anderson L. O. , , and Shimabukuro Y. E. , 2007: Spatial patterns and fire response of recent Amazonian droughts. Geophys. Res. Lett., 34, L07701, doi:10.1029/2006GL028946.

    • Search Google Scholar
    • Export Citation
  • Bitew, M. M., , and Gebremichael M. , 2011: Evaluation of satellite rainfall products through hydrologic simulation in a fully distributed hydrologic model. Water Resour. Res., 47, W06526, doi:10.1029/2010WR009917.

    • Search Google Scholar
    • Export Citation
  • Changjiang Water Resources Commission, 1997: Hydrology Research for Three Gorges Project (in Chinese). Hubei Science and Technology Press, 313 pp.

    • Search Google Scholar
    • Export Citation
  • Changjiang Water Resources Commission, 1999: Atlas of the Changjiang River Basin (in Chinese). SinoMaps Press, 286 pp.

  • Cong, Z. T., , Yang D. W. , , Gao B. , , Yang H. B. , , and Hu H. P. , 2009: Hydrological trend analysis in the Yellow River basin using a distributed hydrological model. Water Resour. Res., 45, W00A13, doi:10.1029/2008WR006852.

    • Search Google Scholar
    • Export Citation
  • Ding, Y. H., , and Zhang J. Y. , 2009: Rainstorms and Floods in China (in Chinese). China Meteorological Press, 290 pp.

  • Ebert, E. E., , Janowiak J. E. , , and Kidd C. , 2007: Comparison of near-real-time precipitation estimates from satellite observations and numerical models. Bull. Amer. Meteor. Soc., 88, 4764, doi:10.1175/BAMS-88-1-47.

    • Search Google Scholar
    • Export Citation
  • Food and Agricultural Organization, 2003: Digital Soil Map of the World and Derived Soil Properties. Land Water Digital Media Series, Rev. 1, Food and Agriculture Organization, CD-ROM.

    • Search Google Scholar
    • Export Citation
  • Gao, B., , Yang D. , , Zhao T. , , and Yang H. , 2012: Changes in the eco-flow metrics of the Upper Yangtze River from 1961 to 2008. J. Hydrol., 448–449, 3038, doi:10.1016/j.jhydrol.2012.03.045.

    • Search Google Scholar
    • Export Citation
  • Gao, G., , Chen D. , , Xu C.-Y. , , and Simelton E. , 2007: Trend of estimated actual evapotranspiration over China during1960–2002. J. Geophys. Res., 112, D11120, doi:10.1029/2006JD008010.

    • Search Google Scholar
    • Export Citation
  • Gao, Y., , and Liu M. , 2013: Evaluation of high-resolution satellite precipitation products using rain gauge observations over the Tibetan Plateau. Hydrol. Earth Syst. Sci., 17, 837849, doi:10.5194/hess-17-837-2013.

    • Search Google Scholar
    • Export Citation
  • Gebregiorgis, A. S., , Tian Y. , , Peters-Lidard C. D. , , and Hossain F. , 2012: Tracing hydrologic model simulation error as a function of satellite rainfall estimation bias components and land use and land cover conditions. Water Resour. Res., 48, W11509, doi:10.1029/2011WR011643.

    • Search Google Scholar
    • Export Citation
  • Gebremichael, M., , Krajewski W. F. , , Morrissey M. , , Langerud D. , , Huffman G. J. , , and Adler R. , 2003: Error uncertainty analysis of GPCP monthly rainfall products: A data-based simulation study. J. Appl. Meteor., 42, 18371848, doi:10.1175/1520-0450(2003)042<1837:EUAOGM>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Gosset, M., , Viarre J. , , Quantin G. , , and Alcoba M. , 2013: Evaluation of several rainfall products used for hydrological applications over West Africa using two high-resolution gauge networks. Quart. J. Roy. Meteor. Soc., 139, 923940, doi:10.1002/qj.2130.

    • Search Google Scholar
    • Export Citation
  • Hong, Y., , Hsu K. L. , , Gao X. , , and Sorooshian S. , 2004: Precipitation Estimation from Remotely Sensed Imagery using Artificial Neural Network Cloud Classification System (PERSIANN-CCS). J. Appl. Meteor. Climatol., 43, 18341853, doi:10.1175/JAM2173.1.

    • Search Google Scholar
    • Export Citation
  • Hong, Y., , Adler R. F. , , and Huffman G. J. , 2006: Evaluation of the potential of NASA multi-satellite precipitation analysis in global landslide hazard assessment. Geophys. Res. Lett., 33, L22402, doi:10.1029/2006GL028010.

    • Search Google Scholar
    • Export Citation
  • Hu, Q., , Yang D. , , Li Z. , , Mishra A. , , Wang Y. , , and Yang H. , 2014: Multi-scale evaluation of six high-resolution satellite monthly rainfall estimates over a humid region in China with dense rain gauges. Int. J. Remote Sens., 35, 12721294, doi:10.1080/01431161.2013.876118.

    • Search Google Scholar
    • Export Citation
  • Huffman, G. J., and Coauthors, 2007: The TRMM Multi-satellite Precipitation Analysis (TMPA): Quasi-Global, multi-year, combined-sensor precipitation estimates at fine scales. J. Hydrometeor., 8, 3855, doi:10.1175/JHM560.1.

    • Search Google Scholar
    • Export Citation
  • Jiang, S., , Ren L.-L. , , Hong Y. , , Yong B. , , Yang X. , , Yuan F. , , and Ma M. , 2012: Comprehensive evaluation of multi-satellite precipitation products with a dense rain gauge network and optimally merging their simulated hydrological flows using the Bayesian model averaging method. J. Hydrol., 452-453, 213225, doi:10.1016/j.jhydrol.2012.05.055.

    • Search Google Scholar
    • Export Citation
  • Joyce, R. J., , Janowiak J. E. , , Arkin P. A. , , and Xie P. , 2004: CMORPH: A method that produces global precipitation estimates from passive microwave and infrared data at high spatial and temporal resolution. J. Hydrometeor., 5, 487503, doi:10.1175/1525-7541(2004)005<0487:CAMTPG>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Kubota, T., and Coauthors, 2007: Global precipitation map using satellite-borne microwave radiometers by the GSMaP Project: Production and validation. IEEE Trans. Geosci. Remote Sens., 45, 22592275, doi:10.1109/TGRS.2007.895337.

    • Search Google Scholar
    • Export Citation
  • Li, L., and Coauthors, 2009: Evaluation of the real-time TRMM-based multi-satellite precipitation analysis for an operational Flood Prediction System in Nzoia basin, Lake Victoria, Africa. Nat. Hazards, 50, 109123, doi:10.1007/s11069-008-9324-5.

    • Search Google Scholar
    • Export Citation
  • Li, M., , Yang D. , , Chen J. , , and Hubbard S. S. , 2012: Calibration of a distributed flood forecasting model with input uncertainty using a Bayesian framework. Water Resour. Res., 48, W08510, doi:10.1029/2010WR010062.

    • Search Google Scholar
    • Export Citation
  • Li, X. H., , Zhang Q. , , and Xu C. Y. , 2012: Suitability of the TRMM satellite rainfalls in driving a distributed hydrological model for water balance computations in Xinjiang catchment, Poyang Lake basin. J. Hydrol., 426–427, 2838, doi:10.1016/j.jhydrol.2012.01.013.

    • Search Google Scholar
    • Export Citation
  • Li, Z., , Yang D. , , and Hong Y. , 2013: Multi-scale evaluation of high-resolution multi-sensor blended global precipitation products over the Yangtze River. J. Hydrol., 500, 157169, doi:10.1016/j.jhydrol.2013.07.023.

    • Search Google Scholar
    • Export Citation
  • Maidment, D. R., Ed., 1993: Handbook of Hydrology. McGraw-Hill, 1424 pp.

  • McCollum, J. R., , Krajewski W. F. , , Ferraro R. R. , , and Ba M. B. , 2002: Evaluation of biases of satellite rainfall estimation algorithms over the continental United States. J. Appl. Meteor., 41, 10651080, doi:10.1175/1520-0450(2002)041<1065:EOBOSR>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Nash, J. E., , and Sutcliffe J. V. , 1970: River flow forecasting through conceptual models part I—A discussion of principles. J. Hydrol., 10, 282290, doi:10.1016/0022-1694(70)90255-6.

    • Search Google Scholar
    • Export Citation
  • New, M., , Hulme M. , , and Jones P. , 2000: Representing twentieth-century space–time climate variability. Part II: Development of a 1961–96 monthly grids of terrestrial surface climate. J. Climate, 13, 22172238, doi:10.1175/1520-0442(2000)013<2217:RTCSTC>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Nijssen, B., , and Lettenmaier D. P. , 2004: Effect of precipitation sampling error on simulated hydrological fluxes and states: Anticipating the Global Precipitation Measurement satellites. J. Geophys. Res., 109, D02103, doi:10.1029/2003JD003497.

    • Search Google Scholar
    • Export Citation
  • Pan, M., , Li H. , , and Wood E. , 2010: Assessing the skill of satellite-based precipitation estimates in hydrologic applications. Water Resour. Res., 46, W09535, doi:10.1029/2009WR008290.

    • Search Google Scholar
    • Export Citation
  • Ryo, M., , Valeriano O. C. S. , , Kanae S. , , and Ngoc T. D. , 2014: Temporal downscaling of daily gauged precipitation by application of a satellite product for flood simulation in a poorly gauged basin and its evaluation with multiple regression analysis. J. Hydrometeor., 15, 563580, doi:10.1175/JHM-D-13-052.1.

    • Search Google Scholar
    • Export Citation
  • Shen, Y., , Xiong A. , , Wang Y. , , and Xie P. , 2010: Performance of high-resolution satellite precipitation products over China. J. Geophys. Res., 115, D02114, doi:10.1029/2009JD012097.

    • Search Google Scholar
    • Export Citation
  • Sorooshian, S., , Hsu K. L. , , Gao X. , , Gupta H. , , Imam B. , , and Braithwaite D. , 2000: Evaluation of PERSIANN system satellite-based estimates of tropical rainfall. Bull. Amer. Meteor. Soc., 81, 20352046, doi:10.1175/1520-0477(2000)081<2035:EOPSSE>2.3.CO;2.

    • Search Google Scholar
    • Export Citation
  • Sorooshian, S., and Coauthors, 2011: Advanced concepts on remote sensing of precipitation at multiple scales. Bull. Amer. Meteor. Soc., 92, 13531357, doi:10.1175/2011BAMS3158.1.

    • Search Google Scholar
    • Export Citation
  • Su, F., , Hong Y. , , and Lettenmaier D. P. , 2008: Evaluation of TRMM Multisatellite Precipitation Analysis (TMPA) and its utility in hydrologic prediction in the La Plata basin. J. Hydrometeor., 9, 622640, doi:10.1175/2007JHM944.1.

    • Search Google Scholar
    • Export Citation
  • Tian, Y., , and Peters-Lidard C. D. , 2007: Systematic anomalies over inland water bodies in satellite-based precipitation estimates. Geophys. Res. Lett., 34, L14403, doi:10.1029/2007GL030787.

    • Search Google Scholar
    • Export Citation
  • Tian, Y., , Peters-Lidard C. D. , , Choudhury B. J. , , and Garcia M. , 2007: Multitemporal analysis of TRMM-based satellite precipitation products for land data assimilation applications. J. Hydrometeor., 8, 11651183, doi:10.1175/2007JHM859.1.

    • Search Google Scholar
    • Export Citation
  • Tian, Y., and Coauthors, 2009: Component analysis of errors in satellite-based precipitation estimates. J. Geophys. Res., 114, D24101, doi:10.1029/2009JD011949.

    • Search Google Scholar
    • Export Citation
  • Tong, K., , Su F. , , Yang D. , , and Hao Z. , 2014: Evaluation of satellite precipitation retrievals and their potential utilities in hydrologic modeling over the Tibetan Plateau. J. Hydrol., 519, 423437, doi:10.1016/j.jhydrol.2014.07.044.

    • Search Google Scholar
    • Export Citation
  • Turk, F. J., , and Miller S. D. , 2005: Toward improving estimates of remotely sensed precipitation with MODIS/AMSR-E blended data techniques. IEEE Trans. Geosci. Remote Sens., 43, 10591069, doi:10.1109/TGRS.2004.841627.

    • Search Google Scholar
    • Export Citation
  • Villarini, G., , Mandapaka P. V. , , Krajewski W. F. , , and Moore R. J. , 2008: Rainfall and sampling uncertainties: A rain gauge perspective. J. Geophys. Res., 113, D11102, doi:10.1029/2007JD009214.

    • Search Google Scholar
    • Export Citation
  • Villarini, G., , Krajewski W. F. , , and Smith J. A. , 2009: New paradigm for statistical validation of satellite precipitation estimates: Application to a large sample of the TMPA 0.25° 3-hourly estimates over Oklahoma. J. Geophys. Res., 114, D12106, doi:10.1029/2008JD011475.

    • Search Google Scholar
    • Export Citation
  • Wang, X. F., , and Zhang S. M. , 2013: Analysis on characteristics and forecasting of “12.7” Flood in upper Yangtze River basin (in Chinese). Yangtze River, 2013 (19), 14.

    • Search Google Scholar
    • Export Citation
  • Wood, E. F., and Coauthors, 2011: Hyperresolution global land surface modeling: Meeting a grand challenge for monitoring Earth’s terrestrial water. Water Resour. Res., 47, W05301, doi:10.1029/2010WR010090.

    • Search Google Scholar
    • Export Citation
  • Wu, H., , Adler R. F. , , Tian Y. , , Huffman G. J. , , Li H. , , and Wang J. , 2014: Real-time global flood estimation using satellite-based precipitation and a coupled land surface and routing model. Water Resour. Res., 50, 26932717, doi:10.1002/2013WR014710.

    • Search Google Scholar
    • Export Citation
  • Xu, J., , Yang D. , , Yi Y. , , Lei Z. , , Chen J. , , and Yang W. , 2008: Spatial and temporal variation of runoff in the Yangtze River basin during the past 40 years. Quat. Int., 186, 3242, doi:10.1016/j.quaint.2007.10.014.

    • Search Google Scholar
    • Export Citation
  • Yang, D., , and Musiake K. , 2003: A continental scale hydrological model using the distributed approach and its application to Asia. Hydrol. Processes, 17, 28552869, doi:10.1002/hyp.1438.

    • Search Google Scholar
    • Export Citation
  • Yilmaz, K. K., , Hogue T. S. , , Hsu K. , , Sorooshian S. , , Gupta H. V. , , and Wagener T. , 2005: Intercomparison of rain gauge, radar, and satellite-based precipitation estimates with emphasis on hydrologic forecasting. J. Hydrometeor., 6, 497517, doi:10.1175/JHM431.1.

    • Search Google Scholar
    • Export Citation
  • Yong, B., , Hong Y. , , Ren L.-L. , , Gourley J. J. , , Huffman G. J. , , Chen X. , , Wang W. , , and Khan S. I. , 2012: Assessment of evolving TRMM-based multisatellite real-time precipitation estimation methods and their impacts on hydrologic prediction in a high latitude basin. J. Geophys. Res., 117, D09108, doi:10.1029/2011JD017069.

    • Search Google Scholar
    • Export Citation
  • Yong, B., , Liu D. , , Gourley J. J. , , Tian Y. , , Huffman G. J. , , Ren L.-L. , , and Hong Y. , 2015: Global view of real-time TRMM Multisatellite Precipitation Analysis: Implication to its successor Global Precipitation Measurement mission. Bull. Amer. Meteor. Soc., doi:10.1175/BAMS-D-14-00017.1, in press.

    • Search Google Scholar
    • Export Citation
  • Zhou, T., , Yu R. , , Chen H. , , Dai A. , , and Pan Y. , 2008: Summer precipitation frequency, intensity, and diurnal cycle over China: A comparison of satellite data with rain gauge observations. J. Climate, 21, 39974010, doi:10.1175/2008JCLI2028.1.

    • Search Google Scholar
    • Export Citation
Save