Advanced Search
Article Contents

Assimilating Surface Observations in a Four-Dimensional Variational Doppler Radar Data Assimilation System to Improve the Analysis and Forecast of a Squall Line Case


doi: 10.1007/s00376-016-5290-0

  • This paper examines how assimilating surface observations can improve the analysis and forecast ability of a four-dimensional Variational Doppler Radar Analysis System (VDRAS). Observed surface temperature and winds are assimilated together with radar radial velocity and reflectivity into a convection-permitting model using the VDRAS four-dimensional variational (4DVAR) data assimilation system. A squall-line case observed during a field campaign is selected to investigate the performance of the technique. A single observation experiment shows that assimilating surface observations can influence the analyzed fields in both the horizontal and vertical directions. The surface-based cold pool, divergence and gust front of the squall line are all strengthened through the assimilation of the single surface observation. Three experiments——assimilating radar data only, assimilating radar data with surface data blended in a mesoscale background, and assimilating both radar and surface observations with a 4DVAR cost function——are conducted to examine the impact of the surface data assimilation. Independent surface and wind profiler observations are used for verification. The result shows that the analysis and forecast are improved when surface observations are assimilated in addition to radar observations. It is also shown that the additional surface data can help improve the analysis and forecast at low levels. Surface and low-level features of the squall line——including the surface warm inflow, cold pool, gust front, and low-level wind——are much closer to the observations after assimilating the surface data in VDRAS.
  • 加载中
  • Barnes S. L., 1964: A technique for maximizing details in numerical weather map analysis. J. Appl. Meteor., 3, 396- 409.10.1175/1520-0450(1964)0032.0.CO;289473f927b54a735e064c9875d5b5d7ahttp%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F1964japme...3..396bhttp://adsabs.harvard.edu/abs/1964japme...3..396bThis paper summarizes the development of a convergent weighted-averaging interpolation scheme which can be used to obtain any desired amount of detail in the analysis of a set of randomly spaced data. The scheme is based on the supposition that the two-dimensional distribution of an atmospheric variable can be represented by the summation of an infinite number of independent waves, i.e., a Fourier integral representation. The practical limitations of the scheme are that the data distribution be reasonably uniform and that the data be accurate. However, the effect of inaccuracies can be controlled by stopping the convergence scheme before the data errors are greatly amplified. The scheme has been tested in the analysis of 500-mb height data over the United States producing a result with details comparable to those obtainable by careful manual analysis. A test analysis of sea level pressure based on the data obtained at only the upper air network stations produced results with essentially the same features as the analysis produced at the National Meteorological Center. Further tests based on a regional sampling of stations reporting airways data demonstrate the applicability of the scheme to mesoscale wavelengths.
    Bryan G. H., J. C. Knievel, and M. D. Parker, 2006: A multimodel assessment of RKW theory's relevance to squall-line characteristics. Mon. Wea. Rev., 134, 2772- 2792.10.1175/MWR3226.1f0815c18a0870d18d9c95d3b59d4515chttp%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F2006MWRv..134.2772Bhttp://adsabs.harvard.edu/abs/2006MWRv..134.2772BThe authors evaluate whether the structure and intensity of simulated squall lines can be explained by “RKW theory,” which most specifically addresses how density currents evolve in sheared environments. In contrast to earlier studies, this study compares output from four numerical models, rather than from just one. All of the authors’ simulations support the qualitative application of RKW theory, whereby squall-line structure is primarily governed by two effects: the intensity of the squall line’s surface-based cold pool, and the low- to midlevel environmental vertical wind shear. The simulations using newly developed models generally support the theory’s quantitative application, whereby an optimal state for system structure also optimizes system intensity. However, there are significant systematic differences between the newer numerical models and the older model that was originally used to develop RKW theory. Two systematic differences are analyzed in detail, and causes for these differences are proposed.
    Chang S.-F., Y.-C. Liou, J.-Z. Sun, and S.-L. Tai, 2015: The implementation of the ice-phase microphysical process into a four-dimensional Variational Doppler Radar Analysis System (VDRAS) and its impact on parameter retrieval and quantitative precipitation nowcasting. J. Atmos. Sci., 73, 1015- 1038.10.1175/JAS-D-15-0184.1fe8c8c456885c856235224125f36c8fehttp%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F2016JAtS...73.1015Chttp://adsabs.harvard.edu/abs/2016JAtS...73.1015CNot Available
    Crook N. A., J. Z. Sun, 2004: Analysis and forecasting of the low-level wind during the Sydney 2000 forecast demonstration project. Wea.Forecasting, 19, 151- 167.10.1175/1520-0434(2004)0192.0.CO;2ad3c3c0b2f9b5cf9c790f0034e50f916http%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F2004WtFor..19..151Chttp://adsabs.harvard.edu/abs/2004WtFor..19..151CDuring the Sydney 2000 Forecast Demonstration Project (FDP) a four-dimensional variational assimilation (4DVAR) scheme was run to analyze the low-level wind field with high spatial and temporal resolution. The 4DVAR scheme finds an optimal fit to the data and a background field under the constraints of a dry boundary layer model. During the FDP, the system assimilated data from two Doppler radars, a surface mesonet, and a boundary layer profiler, and provided low-level analyses every 10 min. After the FDP, a number of experiments have been performed to test the ability of the system to provide short-term forecasts (0-60 min) of the low-level wind and convergence. Herein, the performance of the system during the FDP and the forecast experiment's performed after the FDP are described. Two strong gust front cases and one sea-breeze case that occurred during the FDP are also examined. It is found that for the strong gust front cases, the numerical forecasts improve over persistence in the 1-h time frame, whereas for the slower-moving sea-breeze case, it is difficult to improve over a persistence forecast.
    Dawson D. T., M. Xue, 2006: Numerical forecasts of the 15-16 June 2002 Southern Plains mesoscale convective system: impact of mesoscale data and cloud analysis. Mon. Wea. Rev., 134, 1607- 1629.10.1175/MWR3141.18b3cd1a2cfbb550c197595fc2ac59569http%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F2006MWRv..134.1607Dhttp://adsabs.harvard.edu/abs/2006MWRv..134.1607DHigh-resolution explicit forecasts using the Advanced Regional Prediction System (ARPS) of the 15-16 June 2002 mesoscale convective system (MCS) that occurred over the U.S. central and southern plains during the International HO Project (IHOP_2002) field experiment period are performed. The forecasts are designed to investigate the impact of mesoscale and convective-scale data on the initialization and prediction of an organized convective system. Specifically, the forecasts test the impact of special mesoscale surface and upper-air data collected by, but not necessarily specific to, IHOP_2002 and of level-II data from multiple Weather Surveillance Radar-1988 Doppler radars. The effectiveness of using 30-min assimilation cycles with the use of a complex cloud-analysis procedure and high-temporal-resolution surface data is also examined. The analyses and forecasts employ doubly nested grids, with resolutions of 9 and 3 km. Emphasis is placed on the solutions of the 3-km grid. In all forecasts, a strong, well-defined bow-shaped MCS is produced with structure and behavior similar to those of the observed system. Verification of these forecasts through both regular and phase-shifted equitable threat scores of the instantaneous composite reflectivity fields indicate that the use of the complex cloud analysis has the greatest positive impact on the prediction of the MCS, primarily by removing the otherwise needed -渟pinup- time of convection in the model. The impact of additional data networks is smaller and is reflected mainly in reducing the spinup time of the MCS too. The use of intermittent assimilation cycles appears to be quite beneficial when the assimilation window covers a time period when the MCS is present. Difficulties with verifying weather systems with high spatial and temporal intermittency are also discussed, and the use of both regular and spatially shifted equitable threat scores is found to be very beneficial in assessing the quality of the forecasts.
    Dong J. L., M. Xue, and K. Droegemeier, 2011: The analysis and impact of simulated high-resolution surface observations in addition to radar data for convective storms with an ensemble Kalman filter. Meteorol. Atmos. Phys., 112, 41- 61.10.1007/s00703-011-0130-385d25d3d8a60adedad062c2b442e350ehttp%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F2011map...112...41dhttp://adsabs.harvard.edu/abs/2011map...112...41dObserving system simulation experiments are performed using an ensemble Kalman filter to investigate the impact of surface observations in addition to radar data on convective storm analysis and forecasting. A multi-scale procedure is used in which different covariance localization radii are used for radar and surface observations. When the radar is far enough away from the main storm so that the low level data coverage is poor, a clear positive impact of surface observations is achieved when the network spacing is 20聽km or smaller. The impact of surface data increases quasi-linearly with decreasing surface network spacing until the spacing is close to the grid interval of the truth simulation. The impact of surface data is sustained or even amplified during subsequent forecasts when their impact on the analysis is significant. When microphysics-related model error is introduced, the impact of surface data is reduced but still evidently positive, and the impact also increases with network density. Through dynamic flow-dependent background error covariance, the surface observations not only correct near-surface errors, but also errors at the mid- and upper levels. State variables different from observed are also positively impacted by the observations in the analysis.
    Hayden C. M., R. J. Purser, 1995: Recursive filter objective analysis of meteorological fields: Applications to NESDIS operational processing. J. Appl. Meteor., 34, 3- 15.10.1175/1520-0450-34.1.3c532b080c89955f7de70c47d65f2f24fhttp%3A%2F%2Fdocumentacion.ideam.gov.co%2Fcgi-bin%2Fkoha%2Fopac-detail.pl%3Fbiblionumber%3D32686%26shelfbrowse_itemnumber%3D34083http://adsabs.harvard.edu/abs/1995JApMe..34....3Hof these applications are given.
    Hou T. J., F. Y. Kong, X. L. Chen, and H. C. Lei, 2013: Impact of 3DVAR data assimilation on the prediction of heavy rainfall over Southern China. Advances in Meteorology,Article ID 129642, http://dx.doi.org/10.1155/2013/129642.10.1155/2013/1296421b09ad07aa6c09fc0e91db5c88814162http%3A%2F%2Fdownloads.hindawi.com%2Fjournals%2Famete%2F2013%2F129642.xmlhttp://downloads.hindawi.com/journals/amete/2013/129642.xmlThis study examines the impact of three-dimensional variational data assimilation (3DVAR) on the prediction of two heavy rainfall events over Southern China by using a real-time storm-scale forecasting system. Initialized from the European Centre for Medium-Range Weather Forecasts (ECMWF) high-resolution data, the forecasting system is characterized by combining the Advanced Research Weather Research and Forecasting (WRF-ARW) model and the Advanced Regional Prediction System (ARPS) 3DVAR package. Observations from Doppler radars, surface Automatic Weather Station (AWS) network, and radiosondes are used in the experiments to evaluate the impact of data assimilation on short-term quantitative precipitation forecast (QPF) skill. Results suggest that extrasurface AWS data assimilation has slight but general positive impact on rainfall location forecasts. Surface AWS data also improve model results of near-surface variables. Radiosonde data assimilation improves the QPF skill by improving rainfall position accuracy and reducing rainfall overprediction. Compared with radar data, the overall impact of additional surface and radiosonde data is smaller and is reflected primarily in reducing rainfall overestimation. The assimilation of all radar, surface, and radiosonde data has a more positive impact on the forecast skill than the assimilation of either type of data only for the two rainfall events. 1. Introduction Convective storms accompanied with heavy precipitation, hail, and damaging wind occur frequently in summer season in Southern China. To reduce damage from such severe weather, more accurate short-term forecast from convective-scale numerical weather prediction (NWP) models incorporated with robust data assimilation systems have been paid more attention [1-3]. In recent years, several studies have demonstrated that the Advanced Regional Prediction System (ARPS) three-dimensional variational (3DVAR) system is capable of analyzing different data types, by using multiple analysis passes [4-7]. Based on the Advanced Research Weather Research and Forecasting (WRF-ARW) model and the ARPS 3DVAR/Cloud Analysis module, a real-time hourly updated storm-scale forecasting system has been developed collaboratively by the Center for Analysis and Prediction of Storms (CAPS) in the University of Oklahoma, Shenzhen Meteorological Bureau (SZMB) of China and the Shenzhen Institute of Advanced Technology (SIAT), Chinese Academy of Sciences. The forecasting system, called Hourly Assimilation and Prediction System, or HAPS, has been in daily real-time forecast runs since
    Hsu S. A., E. A. Meindl, and D. B. Gilhousen, 1994: Determining the power-law wind-profile exponent under near-neutral stability conditions at sea. J. Appl. Meteor., 33, 757- 765.10.1175/1520-0450(1994)0332.0.CO;2e6afdf3c66e8d33a774d22aec27f18a1http%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F1994JApMe..33..757Hhttp://adsabs.harvard.edu/abs/1994JApMe..33..757HOn the basis of 30 samples from near-simultaneous overwater measurements by pairs of anemometers located at different heights in the Gulf of Mexico and off the Chesapeake Bay, Virginia, the mean and standard deviation for the exponent of the power-law wind profile over the ocean under near-neutral atmospheric stability conditions were determined to be 0.11 - 0.03. Because this mean value is obtained from both deep and shallow water environments, it is recommended for use at sea to adjust the wind speed measurements at different heights to the standard height of 10 m above the mean sea surface. An example to apply this P value to estimate the momentum flux or wind stress is provided.
    Kalnay E., H. Li, T. Miyoshi, S.-C. Yang, and J. Ballabrera-Poy, 2007: 4-D-Var or ensemble Kalman filter? Tellus A, 59, 758- 773.10.1111/j.1600-0870.2007.00261.xeeaeb031663375c71249d7ee9fdd5858http%3A%2F%2Fonlinelibrary.wiley.com%2Fdoi%2F10.1111%2Fj.1600-0870.2007.00261.x%2Fcitedbyhttp://onlinelibrary.wiley.com/doi/10.1111/j.1600-0870.2007.00261.x/citedbyABSTRACT We consider the relative advantages of two advanced data assimilation systems, 4-D-Var and ensemble Kalman filter (EnKF), currently in use or under consideration for operational implementation. With the Lorenz model, we explore the impact of tuning assimilation parameters such as the assimilation window length and background error covariance in 4-D-Var, variance inflation in EnKF, and the effect of model errors and reduced observation coverage. For short assimilation windows EnKF gives more accurate analyses. Both systems reach similar levels of accuracy if long windows are used for 4-D-Var. For infrequent observations, when ensemble perturbations grow non-linearly and become non-Gaussian, 4-D-Var attains lower errors than EnKF. If the model is imperfect, the 4-D-Var with long windows requires weak constraint. Similar results are obtained with a quasi-geostrophic channel model. EnKF experiments made with the primitive equations SPEEDY model provide comparisons with 3-D-Var and guidance on model error and -榦bservation localization-. Results obtained using operational models and both simulated and real observations indicate that currently EnKF is becoming competitive with 4-D-Var, and that the experience acquired with each of these methods can be used to improve the other. A table summarizes the pros and cons of the two methods.
    Klazura G. E., D. A. Imy, 1993: A description of the initial set of analysis products available from the NEXRAD WSR-88D system. Bull. Amer. Meteor. Soc., 74, 1293- 1311.7e0b54ac7f512a03266f58420213036ehttp%3A%2F%2Fadsabs.harvard.edu%2Fcgi-bin%2Fnph-data_query%3Fbibcode%3D1993BAMS...74.1293K%26db_key%3DPHY%26link_type%3DABSTRACT%26high%3D05623http://xueshu.baidu.com/s?wd=paperuri%3A%28c084bb17cf169a3d543988969a187570%29&filter=sc_long_sign&tn=SE_xueshusource_2kduw22v&sc_vurl=http%3A%2F%2Fadsabs.harvard.edu%2Fcgi-bin%2Fnph-data_query%3Fbibcode%3D1993BAMS...74.1293K%26db_key%3DPHY%26link_type%3DABSTRACT%26high%3D05623&ie=utf-8&sc_us=1416873342511334516
    Lilly D. K., 1990: Numerical prediction of thunderstorms as its time come? Quart. J. Roy. Meteor. Soc., 116, 779- 798.
    Lima M. A., J. W. Wilson, 2008: Convective storm initiation in a moist tropical environment. Mon. Wea. Rev., 136, 1847- 1864.10.1002/jcb.2402802099744e14538d4a4970606c699bff1d903http%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F2008MWRv..136.1847Lhttp://adsabs.harvard.edu/abs/2008MWRv..136.1847LRadar and satellite data from the Tropical Rainfall Measuring Mission–Large-Scale Biosphere–Atmosphere (TRMM–LBA) project have been examined to determine causes for convective storm initiation in the southwest Amazon region. The locations and times of storm initiation were based on the National Center for Atmospheric Research (NCAR) S-band dual-polarization Doppler radar (S-Pol). Both the radar and the () visible data were used to identify cold pools produced by convective precipitation. These data along with high-resolution topographic data were used to determine possible convective storm triggering mechanisms. The terrain elevation varied from 100 to 600 m. Tropical forests cover the area with numerous clear-cut areas used for cattle grazing and farming. This paper presents the results from 5 February 1999. A total of 315 storms were initiated within 130 km of the S-Pol radar. This day was classified as a weak monsoon regime where convection developed in response to the diurnal cycle of solar heating. Scattered shallow cumulus during the morning developed into deep convection by early afternoon. Storm initiation began about 1100 LST and peaked around 1500–1600 LST. The causes of storm initiation were classified into four categories. The most common initiation mechanism was caused by forced lifting by a gust front (GF; 36%). Forcing by terrain (>300 m) without any other triggering mechanism accounted for 21% of the initiations and colliding GFs accounted for 16%. For the remaining 27% a triggering mechanism was not identified. Examination of all days during TRMM–LBA showed that this one detailed study day was representative of many days. A conceptual model of storm initiation and evolution is presented. The results of this study should have implications for other locations when synoptic-scale forcing mechanisms are at a minimum. These results should also have implications for very short-period forecasting techniques in any location where terrain, GFs, and colliding boundaries influence storm evolution.
    Marquis J., Y. Richardson, P. Markowski, D. Dowell, J. Wurman, K. Kosiba, P. Robinson, and G. Romine, 2014: An investigation of the Goshen county, Wyoming, tornadic supercell of 5 June 2009 using EnKF assimilation of mobile mesonet and radar observations collected during VORTEX2. Part I: Experiment design and verification of the EnKF analyses. Mon. Wea. Rev., 142, 530- 554.10.1175/MWR-D-13-00007.16dac103a-c968-4418-9953-c542ac388ff6de406dc87a0bd96677cc874d73779150http%3A%2F%2Fwww.researchgate.net%2Fpublication%2F273920253_An_Investigation_of_the_Goshen_County_Wyoming_Tornadic_Supercell_of_5_June_2009_Using_EnKF_Assimilation_of_Mobile_Mesonet_and_Radar_Observations_Collected_during_VORTEX2._Part_I_Experiment_Design_and_Verification_of_the_EnKF_Analysesrefpaperuri:(c5c7c6b90513f26f104a9bc099de8243)http://www.researchgate.net/publication/273920253_An_Investigation_of_the_Goshen_County_Wyoming_Tornadic_Supercell_of_5_June_2009_Using_EnKF_Assimilation_of_Mobile_Mesonet_and_Radar_Observations_Collected_during_VORTEX2._Part_I_Experiment_Design_and_Verification_of_the_EnKF_AnalysesAbstract High-resolution Doppler radar velocities and in situ surface observations collected in a tornadic supercell on 5 June 2009 during the second Verification of the Origins of Rotation in Tornadoes Experiment (VORTEX2) are assimilated into a simulated convective storm using an ensemble Kalman filter (EnKF). A series of EnKF experiments using a 1-km horizontal model grid spacing demonstrates the sensitivity of the cold pool and kinematic structure of the storm to the assimilation of these observations and to different model microphysics parameterizations. An experiment is performed using a finer grid spacing (500 m) and the most optimal data assimilation and model configurations from the sensitivity tests to produce a realistically evolving storm. Analyses from this experiment are verified against dual-Doppler and in situ observations and are evaluated for their potential to confidently evaluate mesocyclone-scale processes in the storm using trajectory analysis and calculations of Lagrangian vorticity budgets. In Part II of this study, these analyses will be further evaluated to learn the roles that mesocyclone-scale processes play in tornado formation, maintenance, and decay. The coldness of the simulated low-level outflow is generally insensitive to the choice of certain microphysical parameterizations, likely owing to the vast quantity of kinematic and in situ thermodynamic observations assimilated. The three-dimensional EnKF wind fields and parcel trajectories resemble those retrieved from dual-Doppler observations within the storm, suggesting that realistic four-dimensional mesocyclone-scale processes are captured. However, potential errors are found in trajectories and Lagrangian three-dimensional vorticity budget calculations performed within the mesocyclone that may be due to the coarse (2 min) temporal resolution of the analyses. Therefore, caution must be exercised when interpreting trajectories in this area of the storm.
    Morrison H., G. Thompson, and V. Tatarskii, 2009: Impact of cloud microphysics on the development of trailing stratiform precipitation in a simulated squall line: Comparison of one- and two-moment schemes. Mon. Wea. Rev., 137, 991- 1007.10.1175/2008MWR2556.1d294baef7ea15176f30d98d77aab0486http%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F2009mwrv..137..991mhttp://adsabs.harvard.edu/abs/2009mwrv..137..991mNot Available
    Parker M. D., 2010: Relationship between system slope and updraft intensity in squall lines. Mon. Wea. Rev., 138, 3572- 3578.10.1175/2010MWR3441.191804b3dbc8bbcbc8e816ed9f3f5551chttp%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F2010MWRv..138.3572Phttp://adsabs.harvard.edu/abs/2010MWRv..138.3572PIn recent years there has been debate about whether squall lines have an 'optimal state.' It has been repeatedly demonstrated that the slope of a squall line's convective region is related to the comparative magnitudes of the squall line's cold pool and the base-state vertical wind shear. The present work addresses a related assertion, that squall-line intensity ought to be maximized for an upright updraft zone. A simple demonstration shows that upright systems realize more of their buoyancy because their attendant downward-directed perturbation pressure gradient accelerations are weaker.
    Peterson E. W., J. P. Hennessey Jr., 1978: On the use of power laws for estimates of wind power potential. J. Appl. Meteor., 17, 390- 394.b4cc986e83eef7a3e93312a229d8d112http%3A%2F%2Fwww.nrcresearchpress.com%2Fservlet%2Flinkout%3Fsuffix%3Drefg17%2Fref17%26dbid%3D16%26doi%3D10.1139%252Fx2012-038%26key%3D10.1175%252F1520-0450%281978%290172.0.CO%253B2http://xueshu.baidu.com/s?wd=paperuri%3A%28956b3f3ced921aacadeebde5dc198be1%29&filter=sc_long_sign&tn=SE_xueshusource_2kduw22v&sc_vurl=http%3A%2F%2Fwww.nrcresearchpress.com%2Fservlet%2Flinkout%3Fsuffix%3Drefg17%2Fref17%26dbid%3D16%26doi%3D10.1139%252Fx2012-038%26key%3D10.1175%252F1520-0450%281978%290172.0.CO%253B2&ie=utf-8&sc_us=13409902568254460047
    Pleim J. E., 2007: A combined local and nonlocal closure model for the atmospheric boundary layer. Part II: Application and evaluation in a mesoscale meteorological model. Journal of Applied Meteorology and Climatology, 46, 1396- 1409.10.1175/JAM2534.1f9278fcfd83c541dd043d594ea6d3ce4http%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F2007JApMC..46.1396Phttp://adsabs.harvard.edu/abs/2007JApMC..46.1396PA new combined local and nonlocal closure atmospheric boundary layer model called the Asymmetric Convective Model, version 2, (ACM2) was described and tested in one-dimensional form and was compared with large-eddy simulations and field data in Part I. Herein, the incorporation of the ACM2 into the fifth-generation Pennsylvania State University-揘CAR Mesoscale Model (MM5) is described. Model simulations using the MM5 with the ACM2 are made for the summer of 2004 and evaluated through comparison with surface meteorological measurements, rawinsonde profile measurements, and observed planetary boundary layer (PBL) heights derived from radar wind profilers. Overall model performance is as good as or better than similar MM5 evaluation studies. The MM5 simulations with the ACM2 compare particularly well to PBL heights derived from radar wind profilers during the afternoon hours. The ACM2 is designed to simulate the vertical mixing of any modeled quantity realistically for both meteorological models and air quality models. The next step, to be described in a subsequent article, is to incorporate the ACM2 into the Community Multiscale Air Quality (CMAQ) model for testing and evaluation.
    Pu Z. X., H. L. Zhang, and J. Anderson, 2013: Ensemble Kalman filter assimilation of near-surface observations over complex terrain: Comparison with 3DVAR for short-range forecasts. Tellus A, 65, 19620.10.3402/tellusa.v65i0.196202e0cf404d7cedde57a0eeda7d9d9adedhttp%3A%2F%2Fwww.oalib.com%2Fpaper%2F2230977http://www.oalib.com/paper/2230977Surface observations are the main conventional observations for weather forecasts. However, in modern numerical weather prediction, the use of surface observations, especially those data over complex terrain, remains a unique challenge. There are fundamental difficulties in assimilating surface observations with three-dimensional variational data assimilation (3DVAR). In this study, a series of observing system simulation experiments is performed with the ensemble Kalman filter (EnKF), an advanced data assimilation method to compare its ability to assimilate surface observations with 3DVAR. Using the advanced research version of the Weather Research and Forecasting (WRF) model, results from the assimilation of observations at a single observation station demonstrate that the EnKF can overcome some fundamental limitations that 3DVAR has in assimilating surface observations over complex terrain. Specifically, through its flow-dependent background error term, the EnKF produces more realistic analysis increments over complex terrain in general. More comprehensive comparisons are conducted in a short-range weather forecast using a synoptic case with two severe weather systems: a frontal system over complex terrain in the western US and a low-level jet system over the Great Plains. The EnKF is better than 3DVAR for the analysis and forecast of the low-level jet system over flat terrain. However, over complex terrain, the EnKF clearly performs better than 3DVAR, because it is more capable of handling surface data in the presence of terrain misrepresentation. In addition, results also suggest that caution is needed when dealing with errors due to model terrain representation. Data rejection may cause degraded forecasts because data are sparse over complex terrain. Owing to the use of limited ensemble sizes, the EnKF analysis is sensitive to the choice of horizontal and vertical localisation scales.
    Putnam B. J., M. Xue, Y. Jung, N. Snook, and G. F. Zhang, 2014: The analysis and prediction of microphysical states and polarimetric radar variables in a mesoscale convective system using double-moment microphysics, multinetwork radar data, and the ensemble Kalman filter. Mon. Wea. Rev., 142, 141- 162.10.1175/MWR-D-13-00042.1b3e39f14-feaa-42fe-8e97-8d8d1399d243042843a5ae6f1ade44911ebda729c9c3http%3A%2F%2Fwww.researchgate.net%2Fpublication%2F257943233_The_Analysis_and_Prediction_of_Microphysical_States_and_Polarimetric_Variables_of_a_Mesoscale_Convective_System_Using_Double-Moment_Microphysics_Multi-Network_Radar_Data_and_the_Ensemble_Kalman_Filter%3Fev%3Dauth_pubrefpaperuri:(94d6f164e4aabd191845d757a7dbbe2e)http://www.researchgate.net/publication/257943233_The_Analysis_and_Prediction_of_Microphysical_States_and_Polarimetric_Variables_of_a_Mesoscale_Convective_System_Using_Double-Moment_Microphysics_Multi-Network_Radar_Data_and_the_Ensemble_Kalman_Filter?ev=auth_pubAbstract Doppler radar data are assimilated with an ensemble Kalman Filter (EnKF) in combination with a double-moment (DM) microphysics scheme in order to improve the analysis and forecast of microphysical states and precipitation structures within a mesoscale convective system (MCS) that passed over western Oklahoma on 8-9 May 2007. Reflectivity and radial velocity data from five operational Weather Surveillance Radar-1988 Doppler (WSR-88D) S-band radars as well as four experimental Collaborative and Adaptive Sensing of the Atmosphere (CASA) X-band radars are assimilated over a 1-h period using either single-moment (SM) or DM microphysics schemes within the forecast ensemble. Three-hour deterministic forecasts are initialized from the final ensemble mean analyses using a SM or DM scheme, respectively. Polarimetric radar variables are simulated from the analyses and compared with polarimetric WSR-88D observations for verification. EnKF assimilation of radar data using a multimoment microphysics scheme for an MCS case has not previously been documented in the literature. The use of DM microphysics during data assimilation improves simulated polarimetric variables through differentiation of particle size distributions (PSDs) within the stratiform and convective regions. The DM forecast initiated from the DM analysis shows significant qualitative improvement over the assimilation and forecast using SM microphysics in terms of the location and structure of the MCS precipitation. Quantitative precipitation forecasting skills are also improved in the DM forecast. Better handling of the PSDs by the DM scheme is believed to be responsible for the improved prediction of the surface cold pool, a stronger leading convective line, and improved areal extent of stratiform precipitation.
    Rotunno R., J. B. Klemp, and M. L. Weisman, 1988: A theory for strong, long-lived squall lines. J. Atmos. Sci., 45, 463- 485.10.1175/1520-0469(1988)0452.0.CO;27340669ce79a156a8c5d08b0f5e87b54http%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F1988JAtS...45..463Rhttp://adsabs.harvard.edu/abs/1988JAtS...45..463RNot Available
    Schenkman A. D., M. Xue, A. Shapiro, K. Brewster, and J. D. Gao, 2011a: The analysis and prediction of the 8-9 May 2007 Oklahoma tornadic mesoscale Convective system by assimilating WSR-88D and CASA radar data using 3DVAR. Mon. Wea. Rev., 139, 224- 246.182a191f6387f27bd0454a0b6ed76247http%3A%2F%2Fadsabs.harvard.edu%2Fcgi-bin%2Fnph-data_query%3Fbibcode%3D2011MWRv..139..224S%26db_key%3DPHY%26link_type%3DABSTRACT%26high%3D12602http://xueshu.baidu.com/s?wd=paperuri%3A%286f6fdfe93c08e7720b569feb3e50daca%29&filter=sc_long_sign&tn=SE_xueshusource_2kduw22v&sc_vurl=http%3A%2F%2Fadsabs.harvard.edu%2Fcgi-bin%2Fnph-data_query%3Fbibcode%3D2011MWRv..139..224S%26db_key%3DPHY%26link_type%3DABSTRACT%26high%3D12602&ie=utf-8&sc_us=15673482078073305551
    Schenkman A. D., M. Xue, A. Shapiro, K. Brewster, and J. D. Gao, 2011b: Impact of CASA radar and Oklahoma mesonet data assimilation on the analysis and prediction of tornadic mesovortices in an MCS. Mon. Wea. Rev., 139, 3422- 3445.10.1175/MWR-D-10-05051.14edf6fb9d2e0f93b488d39944232105fhttp%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F2011MWRv..139.3422Shttp://adsabs.harvard.edu/abs/2011MWRv..139.3422SQualitative comparison with observations shows highly accurate forecasts of mesovortices up to 80 min in advance of their genesis are obtained when the low-level shear in advance of the gust front is effectively analyzed. Accurate analysis of the low-level shear profile relies on assimilating high-resolution low-level wind information. The most accurate analysis (and resulting prediction) is obtained in experiments that assimilate low-level radial velocity data from the CASA radars. Assimilation of 5-min observations from the Oklahoma Mesonet has a substantial positive impact on the analysis and forecast when high-resolution low-level wind observations from CASA are absent; when the low-level CASA wind data are assimilated, the impact of Mesonet data is smaller. Experiments that do not assimilate low-level wind data from CASA radars are unable to accurately resolve the low-level shear profile and gust front structure, precluding accurate prediction of mesovortex development.
    Snook N., M. Xue, and Y. Jung, 2015: Multiscale EnKF assimilation of radar and conventional observations and ensemble forecasting for a tornadic mesoscale convective system. Mon. Wea. Rev., 143, 1035- 1057.10.1175/MWR-D-13-00262.1b70d2a68a3390c8d93eefd7254f44d4bhttp%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F2015MWRv..143.1035Shttp://adsabs.harvard.edu/abs/2015MWRv..143.1035SNot Available
    Sobash R. A., D. J. Stensrud, 2015: Assimilating surface mesonet observations with the EnKF to improve ensemble forecasts of convection initiation on 29 May 2012. Mon. Wea. Rev., 143, 3700- 3725.10.1175/MWR-D-14-00126.16cb97de5f1e1755a2ff21776665bf885http%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F2015MWRv..143.3700Shttp://adsabs.harvard.edu/abs/2015MWRv..143.3700SThe 5-min assimilation of mesonet data improved forecasts of the placement and timing of CI for this particular event due to the ability of mesonet data to capture rapidly evolving mesoscale features and to constrain model biases, particularly surface moisture errors, during the cycling period.
    Sun J. Z., N. A. Crook, 1997: Dynamical and microphysical retrieval from Doppler radar observations using a cloud model and its adjoint. Part I: Model development and simulated data experiments. J. Atmos. Sci., 54, 1642- 1661.10.1175/1520-0469(1997)0542.0.CO;28b6a26dd-0480-497e-94f5-b3ac0e10a8d14d65fad43479cedc983061dbc7fe9168http%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F1997jats...54.1642srefpaperuri:(3442f4439a64c47314ac768eea6c1f42)http://adsabs.harvard.edu/abs/1997jats...54.1642sThe purpose of the research reported in this paper is to develop a variational data analysis system that can be used to assimilate data from one or more Doppler radars. In the first part of this two-part study, the technique used in this analysis system is described and tested using data from a simulated warm rain convective storm. The analysis system applies the 4D variational data assimilation technique to a cloud-scale model with a warm rain parameterization scheme. The 3D wind, thermodynamical, and microphysical fields are determined by minimizing a cost function, defined by the difference between both radar observed radial velocities and reflectivities (or rainwater mixing ratio) and their model predictions. The adjoint of the numerical model is used to provide the sensitivity of the cost function with respect to the control variables.Experiments using data from a simulated convective storm demonstrated that the variational analysis system is able to retrieve the detailed structure of wind, thermodynamics, and microphysics using either dual-Doppler or single-Doppler information. However, less accurate velocity fields are obtained when single-Doppler data were used. In both cases, retrieving the temperature field is more difficult than the retrieval of the other fields. Results also show that assimilating the rainwater mixing ratio obtained from the reflectivity data results in a better performance of the retrieval procedure than directly assimilating the reflectivity. It is also found that the system is robust to variations in the Z-qrelation, but the microphysical retrieval is quite sensitive to parameters in the warm rain scheme. The technique is robust to random errors in radial velocity and calibration errors in reflectivity.
    Sun J. Z., N. A. Crook, 1998: Dynamical and microphysical retrieval from Doppler radar observations using a cloud model and its adjoint. Part II: Retrieval experiments of an observed Florida convective storm. J. Atmos. Sci., 55, 835- 852.10.1175/1520-0469(1998)0552.0.CO;21cebba5079f4d1ed94947fe25caff633http%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F1998JAtS...55..835Shttp://adsabs.harvard.edu/abs/1998JAtS...55..835SAbstract The variational Doppler radar analysis system developed in part I of this study is tested on a Florida airmass storm observed during the Convection and Precipitation/ Electrification Experiment. The 3D wind, temperature, and microphysical structure of this storm are obtained by minimizing the difference between the radar-observed radial velocities and rainwater mixing ratios (derived from reflectivity) and their model predictions. Retrieval experiments are carried out to assimilate information from one or two radars. The retrieved fields are compared with measurements of two aircraft penetrating the storm at different heights. The retrieved wind, thermodynamical, and microphysical fields indicate that the minimization converges to a solution consistent with the input velocity and rainwater fields. The primary difference between using single-Doppler and dual-Doppler information is the reduction of the peak strength of the storm on the order of 10% when information from only one radar is provided. ...
    Sun J. Z., N. A. Crook, 2001: Real-time low-level wind and temperature analysis using single WSR-88D data. Wea.Forecasting, 16, 117- 132.10.1175/1520-0434(2001)0162.0.CO;2643be36421889bb9613fd0314c6b382fhttp%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F2001WtFor..16..117Shttp://adsabs.harvard.edu/abs/2001WtFor..16..117SAbstract A four-dimensional variational Doppler radar analysis system (VDRAS) has been developed and implemented at a weather forecast office to produce real-time boundary layer wind and temperature analyses using WSR-88D radar data. This paper describes significant changes made to convert VDRAS from a research tool to a real-time analysis system and presents results of low-level wind and temperature analysis using operational radar data. In order to produce continuous analyses with time, VDRAS was implemented with a cycling procedure, in which the analysis from the previous cycle is used as a first guess and background for the next cycle. Other enhancements in this real-time system include direct assimilation of data on constant elevation angle levels, addition of mesonet observations, inclusion of an analysis background term, and continuous updating of lateral boundary conditions. An observed case of a line of storms and strong outflow is used to examine the performance of the real-time analysis system ...
    Sun J. Z., Y. Zhang, 2008: Analysis and prediction of a squall line observed during IHOP using multiple WSR-88D observations. Mon. Wea. Rev., 136, 2364- 2388.10.1175/2007MWR2205.1117507e0fb93cf08d933443bebb23f23http%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F2008mwrv..136.2364shttp://adsabs.harvard.edu/abs/2008mwrv..136.2364sAbstract This paper presents a case study on the assimilation of observations from multiple Doppler radars of the Next Generation Weather Radar (NEXRAD) network. A squall-line case documented during the International H 2 O Project (IHOP_2002) is used for the study. Radar radial velocity and reflectivity observations from four NEXRADs are assimilated into a convection-permitting model using a four-dimensional variational data assimilation (4DVAR) scheme. A mesoscale analysis using a supplementary sounding, velocity-揳zimuth display (VAD) profiles, and surface observations from Meteorological Aerodrome Reports (METAR) are produced and used to provide a background and boundary conditions for the 4DVAR radar data assimilation. Impact of the radar data assimilation is assessed by verifying the skill of the subsequent very short-term (5 h) forecasts. Assimilation and forecasting experiments are conducted to examine the impact of radar data assimilation on the subsequent precipitation forecasts. It is found that the 4DVAR radar data assimilation significantly reduces the model spinup required in the experiments without radar data assimilation, resulting in significantly improved 5-h forecasts. Additional experiments are conducted to study the sensitivity of the precipitation forecasts with respect to 4DVAR cycling configurations. Results from these experiments suggest that the forecasts with three 4DVAR cycles are improved over those with cold start, but the cycling impact seems to diminish with more cycles. The impact of observations from each of the individual radars is also examined by conducting a set of experiments in which data from each radar are alternately excluded. It is found that the accurate analysis of the environmental wind surrounding the convective cells is important in successfully predicting the squall line.
    Sun J. Z., D. W. Flicker, and D. K. Lilly, 1991: Recovery of three-dimensional wind and temperature fields from simulated single-Doppler radar data. J. Atmos. Sci., 48, 876- 890.10.1175/1520-0469(1991)048<0876:ROTDWA>2.0.CO;2d40b6c10161ffd938b403025b94950a0http%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F1991jats...48..876shttp://adsabs.harvard.edu/abs/1991jats...48..876sA method for recovering the full three-dimensional wind and temperature fields from single-Doppler radar data is developed and demonstrated. This method uses a dynamic model, and attempts to determine the initial conditions of the unobserved flow and thermodynamic fields that have generated the time sequence of observed fields. A cost function is defined to measure the difference between the model solutions of the observed variables and the observations. A set of adjoint equations is constructed to determine the sensitivity of the cost function to initial state errors in those not observed. The initial state variables are then adjusted to minimize those errors.Several experiments are conducted using simulated observations produced by a control run of a dry convection model. It is shown that the method is able to determine the spatial structures of the unobserved velocity components and temperature effectively; and the performance is enhanced by the use of a temporal smoothness constraint. The method is not sensitive to moderate amplitude random observational errors.
    Sun J. Z., M. X. Chen, and Y. C. Wang, 2010: A frequent-updating analysis system based on radar, surface, and mesoscale model data for the Beijing 2008 forecast demonstration project. Wea.Forecasting, 25, 1715- 1735.10.1175/2010WAF2222336.1ee57e36d81b1db1db9c86fef11064778http%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F2010WtFor..25.1715Shttp://adsabs.harvard.edu/abs/2010WtFor..25.1715SThe Variational Doppler Radar Analysis System (VDRAS) was implemented in Beijing, China, and contributed to the Beijing 2008 Forecast Demonstration Project (B08FDP) in support of the Beijing Summer Olympics. VDRAS is a four-dimensional variational data assimilation system that produces frequently updated analyses using Doppler radar radial velocities and reflectivities, surface observations, and mesoscale model data. The system was tested in real time during the B08FDP pretrials in the summers of 2006 and 2007 and run during the Olympics to assist the 0-6-h convective weather nowcasting. This paper provides a description of the upgraded system and its Beijing implementation, an evaluation of the system performance using data collected during the pretrials, and its utility on convective weather nowcasting through two case studies. Verification of VDRAS wind against a wind profiler shows that the analyzed wind is reasonably accurate with a smaller RMS difference for 2006 than for 2007 due to better radar data coverage in 2006. The analyzed cold pools in three convective episodes are compared with surface observations at selected stations. The result shows good agreement between the analysis and the observations. The two case studies demonstrate the role that VDRAS could play in nowcasting convective initiation.
    Tai S.-L., Y.-C. Liou, J.-Z. Sun, S.-F. Chang, and M.-C. Kuo, 2011: Precipitation forecasting using Doppler radar data, a cloud model with adjoint, and the weather research and forecasting model: Real case studies during SoWMEX in Taiwan. Wea.Forecasting, 26, 975- 992.10.1175/WAF-D-11-00019.148687b7823d8a78580f1ac1c0174a43chttp%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F2011WtFor..26..975Thttp://adsabs.harvard.edu/abs/2011WtFor..26..975TNot Available
    Tompkins A. M., 2001: Organization of tropical convection in low vertical wind shears: The role of cold pools. J. Atmos. Sci., 58, 1650- 1672.10.1175/1520-0469(2001)0582.0.CO;2febb366452ed64d8ed59949b796bb94fhttp%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F2001JAtS...58..529Thttp://adsabs.harvard.edu/abs/2001JAtS...58..529TAbstract A modeling study is conducted to gain insight into the factors that control the intensity and organization of tropical convection, and in particular to examine if organization occurs in the absence of factors such as vertical wind shear or underlying sea surface temperature (SST) gradient. The control experiment integrates a cloud-resolving model for 15 days using a 3D domain exceeding 1000 km in length, with no imposed winds, and horizontally uniform SST and forcing for convection. After 2 days of random activity, the convection organizes into clusters with dimensions of approximately 200 km. Convective systems propagate through the clusters at speeds of 2–3 m s611, while the clusters themselves propagate at minimal speeds of around 0.5 m s611. Examining the thermodynamic structure of the model domain, it is found that the convective free bands separating the clusters are very dry throughout the troposphere, and due to virtual temperature effects, are correspondingly warmer in the lower tropospher...
    Weisman M. L., J. B. Klemp, and R. Rotunno, 1988: Structure and evolution of numerically simulated squall lines. J. Atmos. Sci., 45, 1990- 2013.10.1175/1520-0469(1988)0452.0.CO;29ce3b85dddf3c2045aed8899e15ae785http%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F1988JAtS...45.1990Whttp://adsabs.harvard.edu/abs/1988JAtS...45.1990WUsing a three-dimensional numerical cloud model, we investigate the effects of vertical wind shear on squall-line structure and evolution over a wide range of shear magnitudes, depths, and orientations relative to the line. We find that the simulated squall lines are most sensitive to the magnitude of the component of shear perpendicular to the line, and that we may reproduce much of the range of observed structures by varying this single parameter. For weak shear, a line of initially upright-to-downshear-tilted short-lived cells quickly tilts upshear, producing a wide band of weaker cells extending behind the surface outflow boundary. For moderate-to-strong shear, the circulation remains upright-to-downshear tilted for longer periods of time, with vigorous, short-lived cells confined to a relatively narrow band along the system's leading edge. At later times, however, these systems may also weaken as the circulation tilts upshear. For strong, deep shears oriented obliquely to the line, the squall line may be composed of quasi-steady, three-dimensional supercells. The squall-line lifecyle that occurs in most of the simulations is dependent on both the strength of the developing cold pool, which induces an upshear-tilted circulation, and the strength of the ambient low-level shear ahead of the line, which promotes a circulation tilting the system downshear. When these two factors are in balance, the overall system circulation remains upright, and we obtain the optimal conditions for deep lifting that promotes the regeneration of strong cells along the outflow boundary. In the current experiments, this optimal state occurs with 15-25 m sof velocity change over the lowest 2.5 km AGL.
    Weygand t, S. S., A. Shapiro, K. K. Droegemeier, 2002a: Retrieval of model initial fields from single-Doppler observations of a supercell thunderstorm. Part II: Thermodynamic retrieval and numerical prediction. Mon. Wea. Rev., 130, 454- 476.10.1175/1520-0493(2002)1302.0.CO;2f7d8d20cc0910e4e99f6e6172e756d34http%3A%2F%2Fonlinelibrary.wiley.com%2Fresolve%2Freference%2FADS%3Fid%3D2002MWRv..130..454Whttp://onlinelibrary.wiley.com/resolve/reference/ADS?id=2002MWRv..130..454WIn this two-part study, a single-Doppler parameter retrieval technique is developed and applied to a real-data case to provide model initial conditions for a short-range prediction of a supercell thunderstorm. The technique consists of the sequential application of a single-Doppler velocity retrieval (SDVR), followed by a variational velocity adjustment, a thermodynamic retrieval, and a moisture specification step. In Part I, the SDVR procedure is described and results from its application to a supercell thunderstorm are presented. In Part II, results from the thermodynamic retrieval and the numerical model prediction for this same case are presented. For comparison, results from parallel sets of experiments using dual-Doppler-derived winds and winds obtained from the simplified velocity retrieval described in Part I are also shown. Following the SDVR, the retrieved wind fields (available only within the storm volume) are blended with a base-state background field obtained from a proximity sounding. The blended fields are then variationally adjusted to preserve anelastic mass conservation and the observed radial velocity. A Gal-Chen type thermodynamic retrieval procedure is then applied to the adjusted wind fields. For all experiments (full retrieval, simplified retrieval, and dual Doppler), the resultant perturbation pressure and potential temperature fields agree qualitatively with expectations for a deep-convective storm. An analysis of the magnitude of the various terms in the vertical momentum equation for both the full retrieval and dual-Doppler experiments indicates a reasonable agreement with predictions from linear theory. In addition, the perturbation pressure and vorticity fields for both the full retrieval and dual-Doppler experiments are in reasonable agreement with linear theory predictions for deep convection in sheared flow. Following a simple moisture specification step, short-range numerical predictions are initiated for both retrieval experiments and the dual-Doppler experiment. In the full single-Doppler retrieval and dual-Doppler cases, the general storm evolution and deviant storm motion are reasonably well predicted for a period of about 35 minutes. In contrast, the storm initialized using the simplified wind retrieval decays too rapidly, indicating that the additional information obtained by the full wind retrieval (primarily low-level polar vorticity) is vital to the success of the numerical prediction. Sensitivity experiments using the initial fields from the full retrieval indicate that the predicted storm evolution is strongly dependent on the initial moisture fields. Overall, the numerical prediction results suggest at least some degree of short-term predictability for this storm and provide an impetus for continued development of single-Doppler retrieval procedures.
    Weygand t, S. S., A. Shapiro, K. K. Droegemeier, 2002b: Retrieval of model initial fields from single-Doppler observations of a supercell thunderstorm. Part I: Single-Doppler velocity retrieval. Mon. Wea. Rev., 130, 433- 453.10.1175/1520-0493(2002)1302.0.CO;2269363ce6f1a21cc73ae9283253293f8http%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F2002MWRv..130..433Whttp://adsabs.harvard.edu/abs/2002MWRv..130..433WAbstract In this two-part study, a single-Doppler parameter retrieval technique is developed and applied to a real-data case to provide initial conditions for a short-range prediction of a supercell thunderstorm. The technique consists of the sequential application of a single-Doppler velocity retrieval (SDVR), followed by a variational velocity adjustment, a thermodynamic retrieval, and a moisture specification step. By utilizing a sequence of retrievals in this manner, some of the difficulties associated with full-model adjoints (possible solution nonuniqueness and large computational expense) can be circumvented. In Part I, the SDVR procedure and present results from its application to a deep-convective storm are discussed. Part II focuses on the thermodynamic retrieval and subsequent numerical prediction. For the SDVR, Shapiro's reflectivity conservation-based method is adapted by applying it in a moving reference frame. Verification of the retrieved wind fields against corresponding dual-Doppler anal...
    Yussouf N., Dowell D. C., Wicker L. J., Knopfmeier K. H., & Wheatley D. M., 2015: Storm-scale data assimilation and ensemble forecasts for the 27 April 2011 severe weather outbreak in Alabama. Mon. Wea. Rev., 143( 8), 3044- 3066.10.1175/MWR-D-14-00268.1cf1d2afc418811798f2295ae856758f3http%3A%2F%2Fadsabs.harvard.edu%2Fabs%2F2015MWRv..143.3044Yhttp://adsabs.harvard.edu/abs/2015MWRv..143.3044YNot Available
    Zhang F. Q., C. Snyder, and J. Z. Sun, 2004: Impacts of initial estimate and observation availability on convective-scale data assimilation with an ensemble Kalman filter. Mon. Wea. Rev., 132, 1238- 1253.10.1007/BF016081165e0c2171e0056b80191c93ca77628631http%3A%2F%2Fciteseer.ist.psu.edu%2Fshowciting%3Fcid%3D3855060http://citeseer.ist.psu.edu/showciting?cid=3855060Peer reviewed
  • [1] JIANG Yuan, LIU Liping, 2014: A Test Pattern Identification Algorithm and Its Application to CINRAD/SA(B) Data, ADVANCES IN ATMOSPHERIC SCIENCES, 31, 331-343.  doi: 10.1007/s00376-013-2315-9
    [2] Xin LI, Mingjian ZENG, Yuan WANG, Wenlan WANG, Haiying WU, Haixia MEI, 2016: Evaluation of Two Momentum Control Variable Schemes and Their Impact on the Variational Assimilation of Radar Wind Data: Case Study of a Squall Line, ADVANCES IN ATMOSPHERIC SCIENCES, 33, 1143-1157.  doi: 10.1007/s00376-016-5255-3
    [3] DING Jincai, YANG Yinming, YE Qixin, HUANG Yan, MA Xiaoxing, MA Leiming, Y. R. GUO, 2007: Moisture Analysis of a Squall Line Case Based on Precipitable Water Vapor Data from a Ground-Based GPS Network in the Yangtze River Delta, ADVANCES IN ATMOSPHERIC SCIENCES, 24, 409-420.  doi: 10.1007/s00376-007-0409-y
    [4] Jo-Han LEE, Dong-Kyou LEE, Hyun-Ha LEE, Yonghan CHOI, Hyung-Woo KIM, 2010: Radar Data Assimilation for the Simulation of Mesoscale Convective Systems, ADVANCES IN ATMOSPHERIC SCIENCES, 27, 1025-1042.  doi: 10.1007/s00376-010-9162-8
    [5] Ji-Hyun HA, Dong-Kyou LEE, 2012: Effect of Length Scale Tuning of Background Error in WRF-3DVAR System on Assimilation of High-Resolution Surface Data for Heavy Rainfall Simulation, ADVANCES IN ATMOSPHERIC SCIENCES, 29, 1142-1158.  doi: 10.1007/s00376-012-1183-z
    [6] HOU Tuanjie, Fanyou KONG, CHEN Xunlai, LEI Hengchi, HU Zhaoxia, 2015: Evaluation of Radar and Automatic Weather Station Data Assimilation for a Heavy Rainfall Event in Southern China, ADVANCES IN ATMOSPHERIC SCIENCES, 32, 967-978.  doi: 10.1007/s00376-014-4155-7
    [7] LIU Liping, ZHUANG Wei, ZHANG Pengfei, MU Rong, 2010: Convective Scale Structure and Evolution of a Squall Line Observed by C-Band Dual Doppler Radar in an Arid Region of Northwestern China, ADVANCES IN ATMOSPHERIC SCIENCES, 27, 1099-1109.  doi: 10.1007/s00376-009-8217-1
    [8] Sun Tingkai, Tan Zhemin, 2001: Numerical Simulation Study for the Structure and Evolution of Tropical Squall Line, ADVANCES IN ATMOSPHERIC SCIENCES, 18, 117-138.  doi: 10.1007/s00376-001-0008-2
    [9] WU Duochang, MENG Zhiyong, YAN Dachun, 2013: The Predictability of a Squall Line in South China on 23 April 2007, ADVANCES IN ATMOSPHERIC SCIENCES, 30, 485-502.  doi: 10.1007/s00376-012-2076-x
    [10] LIU Lu, RAN Lingkun, SUN Xiaogong, 2015: Analysis of the Structure and Propagation of a Simulated Squall Line on 14 June 2009, ADVANCES IN ATMOSPHERIC SCIENCES, 32, 1049-1062.  doi: 10.1007/s00376-014-4100-9
    [11] Shaowu BAO, Lian XIE, Sethu RAMAN, 2004: A Numerical Study of a TOGA-COARE Squall-Line Using a Coupled Mesoscale Atmosphere-Ocean Model, ADVANCES IN ATMOSPHERIC SCIENCES, 21, 708-716.  doi: 10.1007/BF02916368
    [12] Yujie PAN, Mingjun WANG, 2019: Impact of the Assimilation Frequency of Radar Data with the ARPS 3DVar and Cloud Analysis System on Forecasts of a Squall Line in Southern China, ADVANCES IN ATMOSPHERIC SCIENCES, 36, 160-172.  doi: 10.1007/s00376-018-8087-5
    [13] Lingkun RAN, Changsheng CHEN, 2016: Diagnosis of the Forcing of Inertial-gravity Waves in a Severe Convection System, ADVANCES IN ATMOSPHERIC SCIENCES, 33, 1271-1284.  doi: 10.1007/s00376-016-5292-y
    [14] Zhu Keyun, 2001: On the 4D Variational Data Assimilation with Constraint Conditions, ADVANCES IN ATMOSPHERIC SCIENCES, 18, 1131-1145.  doi: 10.1007/s00376-001-0028-y
    [15] Banglin ZHANG, Vijay TALLAPRAGADA, Fuzhong WENG, Jason SIPPEL, Zaizhong MA, 2015: Use of Incremental Analysis Updates in 4D-Var Data Assimilation, ADVANCES IN ATMOSPHERIC SCIENCES, 32, 1575-1582.  doi: 10.1007/s00376-015-5041-7
    [16] Jidong GAO, Keith BREWSTER, Ming XUE, 2006: A Comparison of the Radar Ray Path Equations and Approximations for Use in Radar Data Assimilation, ADVANCES IN ATMOSPHERIC SCIENCES, 23, 190-198.  doi: 10.1007/s00376-006-0190-3
    [17] Ji-Hyun HA, Hyung-Woo KIM, Dong-Kyou LEE, 2011: Observation and Numerical Simulations with Radar and Surface Data Assimilation for Heavy Rainfall over Central Korea, ADVANCES IN ATMOSPHERIC SCIENCES, 28, 573-590.  doi: 10.1007/s00376-010-0035-y
    [18] LIU Hongya, XUE Jishan, GU Jianfeng, XU Haiming, 2012: Radar Data Assimilation of the GRAPES Model and Experimental Results in a Typhoon Case, ADVANCES IN ATMOSPHERIC SCIENCES, 29, 344-358.  doi: 10.1007/s00376-011-1063-y
    [19] LIU Xiaoyang, MAO Jietai, ZHU Yuanjing, LI Jiren, 2003: Runoff Simulation Using Radar and Rain Gauge Data, ADVANCES IN ATMOSPHERIC SCIENCES, 20, 213-218.  doi: 10.1007/s00376-003-0006-7
    [20] Xu Yumao, J.C. Fankhauser, 1989: Correlations of the Evolution of a CCOPE Squall Line with Surface Thermodynamics and Kinematic Fields, ADVANCES IN ATMOSPHERIC SCIENCES, 6, 99-112.  doi: 10.1007/BF02656921

Get Citation+

Export:  

Share Article

Manuscript History

Manuscript received: 30 December 2015
Manuscript revised: 24 March 2016
Manuscript accepted: 08 April 2016
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

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

Assimilating Surface Observations in a Four-Dimensional Variational Doppler Radar Data Assimilation System to Improve the Analysis and Forecast of a Squall Line Case

  • 1. Key Laboratory for Mesoscale Severe Weather/MOE and School of Atmospheric Science, Nanjing University, Nanjing 210046, China
  • 2. National Center for Atmospheric Research, Boulder, Colorado 80301, USA

Abstract: This paper examines how assimilating surface observations can improve the analysis and forecast ability of a four-dimensional Variational Doppler Radar Analysis System (VDRAS). Observed surface temperature and winds are assimilated together with radar radial velocity and reflectivity into a convection-permitting model using the VDRAS four-dimensional variational (4DVAR) data assimilation system. A squall-line case observed during a field campaign is selected to investigate the performance of the technique. A single observation experiment shows that assimilating surface observations can influence the analyzed fields in both the horizontal and vertical directions. The surface-based cold pool, divergence and gust front of the squall line are all strengthened through the assimilation of the single surface observation. Three experiments——assimilating radar data only, assimilating radar data with surface data blended in a mesoscale background, and assimilating both radar and surface observations with a 4DVAR cost function——are conducted to examine the impact of the surface data assimilation. Independent surface and wind profiler observations are used for verification. The result shows that the analysis and forecast are improved when surface observations are assimilated in addition to radar observations. It is also shown that the additional surface data can help improve the analysis and forecast at low levels. Surface and low-level features of the squall line——including the surface warm inflow, cold pool, gust front, and low-level wind——are much closer to the observations after assimilating the surface data in VDRAS.

1. Introduction
  • With the increases in computing power, it is possible to resolve and forecast the features of severe convective systems through a convection-permitting numerical model. One of the major challenges of storm-resolving numerical weather prediction (NWP) models is the accuracy of initial conditions. However, the forecast of meso- or convective-scale weather systems is a highly nonlinear initial value problem; the inaccuracy of the analyzed initial state becomes the fundamental limitation for the prediction of severe convective systems (Schenkman et al., 2011a). The high temporal (<10 min) and spatial (<1.0 km) resolution of Doppler radar makes it the best remote sensing device to probe the variation of hydrometeors and the airflow structures inside the convective system (Lilly, 1990; Sun et al., 1991). Numerous studies have shown that the assimilation of radar reflectivity and radial velocity data can improve the analysis and subsequent forecast of convective systems (e.g., Sun and Crook, 1998; Weygandt et al., 2002a, 2002b; Dawson and Xue, 2006).

    Considering the influence of Earth's curvature, the height between the lowest radar beam (if the lowest radar elevation is 0.5°) and the surface can exceed 1.0 km at 80 km away from the radar site, and could be even more extended under sub-refraction conditions. This blind region to the radar largely lies in the planetary boundary layer (PBL), in which PBL processes such as horizontal convective rolls, dry-lines, and cold-pools can trigger convection under a favorable large-scale environment (Sobash and Stensrud, 2015). In particular, surface-based cold pools produced by the evaporation of precipitation can influence the initiation, development and propagation of convective systems (e.g., Tompkins, 2001; Lima and Wilson, 2008). The intensity of cold pool and environmental vertical wind shear are critical factors in determining a convective system's structure and evolution (e.g., Rotunno et al., 1988; Weisman et al., 1988; Bryan et al., 2006; Parker, 2010). These PBL processes cannot be resolved well if only radar data are assimilated in NWP. Data from surface observations have been used in the past to detect these processes and will be more valuable if assimilated into NWP models along with radar data.

    In the past several years, with the increase and improvement of surface observation networks [e.g., automatic weather stations (AWSs)], surface data at high spatial (<10 km) and temporal (5 min) resolution have become commonly available in real time. Rapidly evolving surface mesoscale features that are important to the analysis and forecasting of convective systems can be monitored routinely. Therefore, it is expected that assimilating these surface observations into NWP models at frequent intervals can fill the low-level blind region of radar detection and yield a better analysis and forecast, since PBL processes play an essential role in convective systems.

    Several studies have been conducted to assimilate surface observations together with radar observations with a convection-permitting model. For example, (Zhang et al., 2004) found that when low-level radar observations are missing, the EnKF (ensemble Kalman filter) can provide a better estimate of the storm with the incorporation of surface wind and temperature observations. (Dong et al., 2011) indicated that when a radar is far away from a convective system, a positive impact results from assimilating surface observations, as long as the network spacing is less than 20 km. Through assimilating high-resolution Doppler radar radial velocities and in situ surface observations into an EnKF system, (Marquis et al., 2014) successfully simulated mesocyclone-scale processes in the Goshen County, Wyoming, and the tornadic supercell of 5 June 2009. (Yussouf et al., 2015) developed a multiscale ensemble-based assimilation and prediction system that can assimilate the radar, surface, and other observations together. Using this system, they analyzed the storm-scale features of the 27 April 2011 Alabama severe weather outbreak. Through assimilating radar and surface observations with an EnKF in combination with a double-moment microphysics scheme, (Putnam et al., 2014) analyzed the microphysical states and precipitation structures within a mesoscale convective system that passed over western Oklahoma during 8-9 May 2007. (Schenkman et al., 2011b) examined the impact of radar and Oklahoma Mesonet data assimilation on the prediction of mesovortices in a tornadic mesoscale convective system (MCS) through a three-dimensional variational data assimilation (3DVAR) system. Results again confirmed the positive impact of surface data in the absence of radar data. (Snook et al., 2015) analyzed the same tornadic MCS through an EnKF system and found that assimilating the convectional observations (including surface, wind profiler, and upper-air observations) together with radar observations resulted in better meso- and convective-scale features. Similar conclusions, regarding the prediction of heavy rainfall over southern China, were reached in (Hou et al., 2013).

    Despite positive results, surface observations are still underused in operational data assimilation systems (Pu et al., 2013). To the best of the authors' knowledge, there have been no studies that have assessed the impact of assimilating surface observations along with radar data using a four-dimensional variational (4DVAR) data assimilation system. 4DVAR has been proven to be more accurate than 3DVAR. Moreover, when compared with the EnKF, it allows for the assimilation of asynchronous observations with serially correlated errors by including time correlations (Kalnay et al., 2007). However, the 4DVAR method requires an adjoint model, whose numerical structures must closely follow those in the prognostic model (Tai et al., 2011). Therefore, high computational cost is accrued, and greater complexity expected, in 4DVAR systems.

    In this study, we describe a surface data assimilation method in the 4DVAR radar data assimilation system VDRAS (Variational Doppler Radar Analysis System) (Sun and Crook, 1997) developed by NCAR. A squall-line case observed during the Observation, Prediction and Analysis of Severe Convection of China (OPACC) field campaign in 2014 is used to test the impact of the surface data assimilation. The objective is to examine whether additional surface observations can improve the analysis and short-term forecast ability of VDRAS.

    The rest of this paper is organized as follows: In section 2, VDRAS and the new development for surface data assimilation are described. An overview of the squall-line case observed in OPACC is introduced in section 3. The experimental design is described in section 4. Results and verification are presented in section 5. A summary and discussion are given in section 6.

2. Surface data assimilation in VDRAS
  • VDRAS is a convective-scale data assimilation system based on the 4DVAR technique with a cloud model as its constraint. It has been successfully used as the analysis and short-term forecast system of the Beijing Summer Olympics (Sun et al., 2010) and is still in operation at NCAR, the Beijing Meteorological Bureau and Taiwan Central Weather Bureau. The main objective of VDRAS is to produce rapidly updated (in the order of minutes) analysis for severe weather analysis and nowcasting, which is made possible by using a short 4DVAR assimilation window that covers three volumes of radar observations. The main data sources of VDRAS are from Doppler radar and surface networks, while other large-scale data are indirectly incorporated by using a high-resolution mesoscale model's analysis/forecast as the background and boundary conditions. Through the cloud model and the 4DVAR scheme, VDRAS can retrieve the unobserved (by radar) temperature, wind and other microphysical variables by assimilating reflectivity and radial velocity observations from a single or multiple radar networks. Readers are referred to Sun and Crook (1997, 2001) for the system's design and cloud model, (Chang et al., 2015) for the recent addition of an ice physics scheme, and (Sun and Crook, 2001), (Crook and Sun, 2004), and (Sun et al., 2010) for an evaluation of the system's operational performance.

    In this study, the operational WRF forecast during OPACC at 4 km spatial resolution is used to provide the background and boundary conditions. The operational WRF forecast is performed over a single domain that covers most parts of China with 1409× 1081 horizontal grid points and 27 vertical levels. The WRF model employs the ACM2 boundary layer scheme (Pleim, 2007), the Pleim-Xiu surface model (Pleim, 2007), the CAM longwave and shortwave radiation schemes, and the WRF Morrison double-moment microphysics scheme (Morrison et al., 2009). In addition, the VAD (velocity azimuth display) wind profile (Klazura and Imy, 1993) for each radar has been computed and blended with the WRF background to produce a gridded mesoscale analysis using a Barnes interpolation technique (Barnes, 1964). This mesoscale analysis is used as a first guess in the first analysis cycle. Each analysis cycle contains a 15-min assimilation window and a 5-min short forecast to provide background for the next cycle. In each cycle, an optimal initial state between the background field and radar observations is calculated by minimizing a cost function (J), \begin{eqnarray} J&=&\dfrac{1}{2}({x}_0-{x}_{\rm b})^{\rm T}{B}^{-1}({x}_0-{x}_{\rm b})^{\rm T}+\nonumber\\ &&\dfrac{1}{2}\sum_{\sigma,t}[\eta_v(v_{\rm r}-v_{\rm 0,r})^2+\eta_q(q_{\rm r}-q_{\rm 0,r})^2]+J_{\rm p}+J_{\rm mb} . (1)\end{eqnarray} In Eq. (1), the summation is conducted over space (σ) and time (t), x0 represents the model state variables at the beginning of the current cycle, x b stands for the background information, and B is the covariance matrix of background error. The weighting coefficients ηv and ηq stand for the inverse of the observational error variance for radial wind and rainwater, respectively. Uncorrelated observation error is assumed. The model-generated Doppler radial velocity and rainwater mixing ratio are represented by v r and q r, while their observed counterparts are expressed by v 0,r and q 0,r. J p is a penalty term for additional constraints to preserve the temporal and spatial smoothness of the analysis. J mb stands for the background penalty term that measures the difference between the 4DVAR analysis and the mesoscale analysis to ensure that the 4DVAR radar analysis is not too far from the mesoscale analysis in the region outside the radar coverage. The observed rainwater mixing ratio is estimated from radar reflectivity by the following formula, which assumes a Marshall-Palmer raindrop size distribution (Sun and Crook, 1997): \begin{equation} q_{\rm r}=\dfrac{1}{\rho}10^{[(Z-43.1)/17.5]} , (2)\end{equation} where Z is the reflectivity in dBZ and ρ is the density of air. A complete description of VDRAS can be found in (Sun and Crook, 2001), and (Crook and Sun, 2004).

    Many previous studies have shown that VDRAS can retrieve the meteorological variables of convective storms through the depth of the troposphere (e.g., Sun and Zhang, 2008). For real-time applications, however, VDRAS is usually implemented to retrieve only low-level wind and temperature, due to limited computation resources (Sun et al., 2010). In this paper, we retrieve the meteorological variables of the squall line up to 15 km, with a focus on the improvements at the low levels.

  • In its original version, surface observations are used to improve the retrieval capability of VDRAS. However, they are used to correct the WRF analysis/forecast prior to the 4DVAR radar data assimilation using a 3D analysis method in the mesoscale background analysis step (Crook and Sun, 2004). Surface observations are interpolated horizontally onto the VDRAS grid using the Barnes interpolation scheme, and then vertically combined with the VAD-corrected WRF background data (described in section 2.1) using a least-squares fitting method. The reason for such a design is that, in the past, most of the available surface observations were from the conventional synoptic network. The spatial and temporal resolutions of the conventional surface observations are much sparser and less frequent than radar observations, so that they only stand for the large-scale environmental conditions. This two-step approach can provide a more detailed analysis of the convective-scale features from the radar observations by using a small assimilation window and a small length scale for the background error covariance represented in a recursive filter (Hayden and Purser, 1995; Sun and Zhang, 2008).

    With the development of AWS surface observation networks, mesoscale and in some cases convective-scale processes, can be captured. The temporal resolution of surface observations can reach 5 min, which is similar to the duration of radar volume scan. Concurrent assimilation of surface and radar observations in 4DVAR is expected to yield a better mesoscale analysis and short-term forecast than the original two-step approach in which the surface observations are firstly interpreted to the background analysis, and then the radar data assimilated through the 4DVAR scheme. In the current work, surface temperature and wind are assimilated with radar reflectivity and radial velocity simultaneously, with the following cost function: \begin{eqnarray} J&=&\dfrac{1}{2}({x}_0-{x}_{\rm b})^{\rm T}{B}^{-1}({x}_0-{x}_{\rm b})^{\rm T}+\dfrac{1}{2}\sum_{\sigma,t} [\eta_v(v_{\rm r}-v_{\rm 0,r})^2+\nonumber\\ &&\eta_q(q_{\rm r}-q_{\rm 0,r})^2+\eta_{\rm T}(T_{\rm s}-T_{\rm 0,s})^2+\eta_U(U_{\rm s}-U_{0,s})^2+\nonumber\\ &&\eta_V(V_{\rm s}-V_{\rm 0,s})^2]+J_{\rm p}+J_{\rm mb} . (3)\end{eqnarray}

    Three additional terms are included in Eq. (3), representing surface temperature (T s) and the x and y components of the surface winds (U s,V s). The observed surface 10 m winds (U 10m,V 10m) from AWSs are extrapolated to the lowest model level (U 0,s and V 0,s) based on an empirical power law relation in the surface layer (Peterson and Hennessey, 1978), \begin{eqnarray} \dfrac{U_{\rm 10m}}{U_{\rm 0,s}}=\left(\dfrac{10}{Z_{\rm s}}\right)^\alpha , (4)\end{eqnarray} where Z s is the height of the lowest model grid level, set to 100 m in this study, which is around the top of the surface layer. The exponent α is an empirically derived coefficient that depends on the local stability. For neutral conditions, α is approximately 0.143 (Hsu et al., 1994). Estimated temperature at the lowest model level (T 0,s) is extrapolated from the AWS 2 m temperature using a lapse rate of 6.5°C km-1. The weighting coefficients η T U and η V stand for the inverse of the respective observation error variances. In this study, all surface weighting coefficients are taken as constants (set to 1.0) for simplicity. The optimal weighting coefficients will require further study. Corresponding to the new cost function, the adjoint model of VDRAS is modified and the accuracy of the adjoint model verified following the method in (Sun and Crook, 1997).

3. Overview of the squall-line case
  • The squall-line case studied in this paper is selected from the OPACC field campaign. OPACC is a 5-year project focused on studying the dynamical and microphysical characteristics of convective weather systems over eastern China, as well as the observation techniques and forecasting methods for such systems. The first field campaign of OPACC was conducted from 1 June to 3 August 2014 over the Yangtze River-Huaihe River (YHR) basin. Several quasi-linear convective systems and two squall lines were observed on 30 July 2014 during the intensive observing period (IOP) 10 of OPACC. A tropical cyclone was situated over the South China Sea at that time. Its circulation transported warm and moist air from the ocean inland to the YHR basin, setting up a moist and unstable environment that was favorable for convective initiation. Figure 1 shows the merged composite radar reflectivity from 0830 to 1600 UTC 30 July 2014, using six operational radars (FYRD, BBRD, NJRD, HFRD, TLRD and AQRD in Fig. 2). The merged composite radar reflectivity is computed at each grid point by taking the maximum reflectivity among all six radars. A convective system was initiated near 0330 UTC over the northern part of the observation domain in Fig. 1 (not shown). It organized into a west-east oriented squall line (the first squall line) with extreme rainfall and high surface wind 5 hours later (Fig. 1a). In the next three hours, this squall line moved to the central part of the domain and broke into several north-south oriented quasi-linear MCSs around 1130 UTC (Fig. 1b). By 1430 UTC, a new squall line had been triggered on the east side of these MCSs (Fig. 1c). The second squall line matured at around 1600 UTC and formed a bowing structure with a maximum reflectivity of greater than 55 dBZ (Fig. 1d). Extreme surface wind (up to 19.2 m s-1) and precipitation (up to 56.6 mm h-1) were also observed at this time.

    Figure 1.  Mosaics of composite radar reflectivity at (a) 0830, (b) 1130, (c) 1430 and (d) 1600 UTC. Provincial borders are superimposed as black contours. The location of Hefei radar is represented by the black triangle, and the two surface stations used for independent verification are marked by black dots.

4. Experimental design
  • Two single observation experiments (SP-AWS and SP-NOAWS) were designed to investigate the impact of assimilating single-point surface observations on the VDRAS analysis in both the horizontal and vertical directions. The domain of the single observation experiments is shown by a red dashed box in Fig. 2. Reflectivity and radial velocity from the Hefei radar (HFRD in Fig. 2) are assimilated together with the surface wind and temperature observations from the AWS (red dot in Fig. 2) in SP-AWS, whereas HFRD data alone are assimilated in SP-NOAWS. The domain of the two experiments has 80× 80 horizontal grid points and 50 vertical levels. The horizontal resolution of these two experiments is 3 km and the vertical grid size is 200 m, with the lowest model level at 100 m above the surface. The data assimilation procedure starts at 1300 UTC 30 July 2014, and ends at 1620 UTC. This time period includes the initiation and development stages of the second squall line documented in section 3 (Figs. 1c and d). Altogether, eleven analysis cycles are included. Each cycle has a 15 min 4DVAR time window, followed by a 5-min forecast, to provide background for the next cycle. The short forecasts between two assimilation windows are added such that no observations will be used twice (Sun and Zhang, 2008). The typical volumetric scan for radars is 5-6 min and the time resolution of AWS observations is 5 min; two to three radar volumetric scans and three surface observations are included in each assimilation window. Increments (SP-AWS minus SP-NOAWS) of different model variables (temperature, pressure and horizontal wind) are analyzed in section 5.1.

    Three assimilation experiments (RAD-ONLY, SURF-MESO and SURF-4DVAR) are conducted to examine whether including the AWS observations can improve VDRAS with multiple radar assimilations. The model domain for these three experiments is larger than the single observation experiment, in order to include all seven radars. It has 140× 140 horizontal grid points and 50 vertical levels (see Fig. 2). The grid spacing is the same as the single observation experiments. In RAD-ONLY, surface data are not used either in the background analysis step or in the 4DVAR cost function. Radar reflectivity and radial velocity data from six S-band operational radars (FYRD, BBRD, NJRD, HFRD, TLRD and AQRD) and one C-band research radar developed by Nanjing University (NJU-CPOL in Fig. 2) are assimilated through the 4DVAR cost function shown in Eq. (1), using the VAD-corrected WRF 4-km data as the mesoscale background. In the experiment SURF-MESO, the original surface data assimilation method is applied, i.e., the AWS data are used in the background analysis prior to the 4DVAR radar data assimilation, as described in section 2. In SURF-4DVAR, the background analysis is the same as RAD-ONLY but additional AWS data are assimilated in the 4DVAR cycles using the cost function in Eq. (3).

    The three assimilation experiments start at 0700 UTC 30 July 2014, and end at 1940 UTC. The period extends from the initiation of the first quasi-linear MCS to the dissipation of the second bowing squall line (Fig. 1). Time lengths of assimilation windows are 15 min, followed by a 5-min forecast——the same as in the single observation experiments. A 1-h forecast is performed at the end of each analysis cycle to further investigate the influence of surface data assimilation on the forecast. Data from two AWSs (green dots in Fig. 2) are not assimilated in SURF-MESO or SURF-4DVAR, and are used for verification. A wind profiler (green square in Fig. 2) is also used to examine the effects of surface data assimilation on low-level wind fields.

    Figure 2.  Analysis and forecast domains for the real-data assimilation experiments (black dashed box) and the single observation experiment (red dashed box). The locations of the six S-band operational radars and one C-band research radar (CPol) are marked by the black triangles. The blue dots represent AWSs. Two AWSs (ID: 58203 and 58236) and a wind profiler used for verification are marked by the green dots and square. The AWS used in the single observation experiment is marked by the red dot. Orography (units: m) is shown in gray-scale, and provincial borders by black lines.

5. Results and verification
  • Figure 3 shows the analysis increments of temperature, pressure and horizontal wind fields on the lowest model level (100 m) between SP-AWS and SP-NOAWS at 1500 UTC 31 July 2014. The second squall line evolved into a bowing structure on the east side of the dissipating quasi-linear MCS at this time. The assimilation of a single surface observation produces near-surface cooling around the AWS station (Fig. 3a) and intensifies the surface-based cold pool strength of the old MCS. The second squall line is located at the leading edge of this intensified cold pool. Spatial inhomogeneity can be found with a stronger cooling over the region covered by the new squall line (around 117.8°E), which represents the intensification of the cold pool produced by the second squall line after surface data are assimilated. Though the pressure field has not been assimilated through the 4DVAR scheme directly, it is adjusted through the cloud model and its adjoint model in the 4DVAR system during the assimilation cycles, wherever the pressure responds to the analyzed temperature. The analysis increment of the pressure field has a similar spatial distribution pattern as the temperature field. Increases in pressure are found over the old and new cold pool area produced by the original MCSs and the second squall line, respectively. Analysis increments in U and V wind components are shown in Figs. 3b and c. The spatial inhomogeneity of the wind fields is much clearer than for the temperature and pressure fields.

    Figure 3.  Horizontal distributions of the analysis increments (SP-AWS minus SP-NOAWS) of (a) temperature, (b) pressure, (c) $U$ wind and (d) $V$ wind, on the lowest model level (100 m) at 1500 UTC 31 July 2014. Composite radar reflectivity is shown in black contours, starting from 30 dB$Z$ and with an interval of 10 dB$Z$.

    Incremental divergence can be found under the quasi-linear MCS, corresponding to the region with decreasing temperature and increasing pressure. Assimilation of the single-point surface observations also increases the strength of the outflow to the east of the new squall line.

    Vertical cross sections along the black dashed lines in Fig. 3 are presented in Fig. 4. Two convective systems are seen from the radar reflectivity contours, the old quasi-linear MCS in the center and the newly developed squall line on the right. As in Fig. 3, strengthening of the cold pool under the MCS is indicated by decreased temperature in Fig. 4a, and increased pressure in Fig. 4b, below 1.5 km. Meanwhile, the incremental strengthening of the squall line's cold pool is not clear over this cross section. This squall line was formed shortly before 1500 UTC; its cold pool is still very shallow at this time and can only be found at the lowest levels (Fig. 3b). Incremental low-level and high-level divergence of the MCS in the center can also be found in Figs. 4c and d. The gust front produced by the squall line is also intensified at around 118°E (Fig. 4c).

    Figure 4.  Vertical cross sections of the analysis increments of (a) temperature, (b) pressure, (c) $U$ wind and (d) $V$ wind, along the black dashed line in Fig. 3. Observed radar reflectivity is shown in black contours, starting from 30 dB$Z$ and with an interval of 10 dB$Z$.

    Figure 5.  Surface temperature (color-shaded) and wind vectors (black arrows) at 1500 UTC 30 July 2014 from (a) AWS observations, (b) RAD-ONLY, (c) SURF-MESO and (d) SURF-4DVAR. The position of the gust front is shown as a purple dashed line, and the observed composite radar reflectivity——starting from 30 dB$Z$ and with an interval of 10 dB$Z$——is superimposed in all subfigures as black solid lines. The surface inflow region on the east side of the squall line is shown by a white box.

  • The AWS observed surface wind and temperature fields at 1500 UTC are shown in Fig. 5a, with the observed composite radar reflectivity superimposed as black contours. As seen in Fig. 1, this is the time when the convective systems have moved from the northern part to the central part of the domain. Heavy rainfall occurred over the path of the system, so the surface temperature over these areas is much colder than in the southern part. Based on the surface wind observations, a clear gust front, marked by the black dashed line in Fig. 5a, and warm surface inflow, occur over the southeast side of the squall line.

    The estimated surface wind and temperature fields at 1500 UTC from the three assimilation experiments are shown in Figs. 5b-d. Surface wind and temperature are extrapolated from the model temperature and wind on the lowest grid level at 100 m. When only radar data are used in the 4DVAR assimilation, large differences can be found between the AWS observation (Fig. 5a) and the RAD-ONLY analysis (Fig. 5b). The surface temperature of the northern and central domain in the RAD-ONLY analysis is much warmer than the AWS observations, especially for the region without the radar coverage. On the other hand, the surface warm inflow on the southeast side of the squall line (shown by the white box in Fig. 5) is colder than observed. Surface-based cold pools of the old quasi-linear MCS and the bowing squall line are weaker. However, surface outflow from the cold pools is much stronger than in the AWS observations, and the position of the gust front is also further ahead of the squall line. In the RAD-ONLY experiment, the surface wind analysis is a combination of the upper-level radar radial velocity observations and the background field through the spatial correlation implied in the cloud model, which also results in the temperature increment in the region with radar data through the cross-variable correlation in the cloud model. However, in the region without radar observations, both the wind and temperature analyses are mainly from the WRF background. Without the surface observations, the errors in the background cannot be corrected. That is the main reason for the large temperature error in the environment surrounding the convective system. It is also noted that in the region with radar observations, the temperature still shows a few differences from the AWS observations, which can be partly attributed to the inadequate sample of radar data at low levels and the uncertainty of the cross-variable correlation in the cloud model.

    In SURF-MESO, the AWS observations are used during the background analysis step. This results in an improvement of the surface temperature analysis compared to RAD-ONLY (Fig. 5c). Nevertheless, significant differences can still be found in the surface temperature, the southeast inflow, the cold-pool outflow and the position of the gust front. When AWS observations are assimilated together with radar data in the SURF-4DVAR experiment, the modeled surface temperature and wind fields are much closer to the observations (Fig. 5d). The northern and central domain is much colder, while southeast surface inflow is much warmer than that in the other two experiments. The strength of the cold pool, southeast inflow, surface outflow and the location of the gust front are also close to the AWS observations.

    In order to quantitatively compare the analysis between the three experiments and the AWS observations, the gridded surface temperature and wind are interpolated to the AWS locations. The mean temperature (MTD) and wind (MVD) differences are calculated: \begin{eqnarray} {\rm MTD}&=&\dfrac{1}{N}\sum_{k=1}^N\sqrt{(T_{\rm aws}-T)^2} ,(5)\\ {\rm MVD}&=&\dfrac{1}{N}\sum_{k=1}^N\sqrt{(U_{\rm aws}-U)^2+(V_{\rm aws}-V)^2} , (6)\end{eqnarray} where N represents the number of AWSs. T aws,U aws and V aws stand for the observed surface temperature and winds from each AWS. The extrapolated model surface temperature and winds at the AWS locations are represented by T, U and V. The MTDs at 1500 UTC in RAD-ONLY, SURF-MESO and SURF-4DVAR are 3.49 K, 2.38 K and 1.23 K, and the MVDs are 3.91 m s-1, 2.52 m s-1 and 2.13 m s-1, respectively. Comparison of SURF-MESO with RAD-ONLY suggests that assimilating surface observations can help improve the near-surface fields of VDRAS. The further improvement of SURF-4DVAR over SURF-MESO demonstrates the advantages of the newly developed 4DVAR surface data assimilation methodology in comparison with the original methodology.

    Two independent AWSs (Nos. 58203 and 58326——marked by green dots in Fig. 2) are chosen to further examine the improvement of VDRAS analysis using the new methodology in SURF-4DVAR. Station 58203 is located in the northwest. It recorded the surface temperature change during the passage of the first west-east oriented squall line at around 0840 UTC. Similarly, station 58326 is chosen for its spatial proximity to the second squall line. In Fig. 6a, a dramatic decrease in surface temperature (13 K) between 0830 and 0930 UTC is found at station 58203 after the passage of the squall line. In RAD-ONLY, the timing of the surface temperature drop is close to the observations, while the amplitude (4.5 K) is much smaller. For SURF-MESO, the amplitude of the temperature drop is 7.3 K——closer to the observations than RAD-ONLY. The timing of the initial temperature drop, however, is 20 minutes earlier than both the observations and the RAD-ONLY experiment. In SURF-4DVAR, the surface temperature decreases by about 10.3 K——even closer to the observed magnitude. The timing of the event also agrees well with the observations.

    Figure 6.  Time series of observed and analyzed surface temperature at AWS No. (a) 58203 and (b) 58326.

    In Fig. 6b, before the initiation of the second squall line, simulated surface temperatures at station 58326 are close to one another, and in general agree with the observations. Better results are found in SURF-4DVAR over the other two. After the passage of the squall line, surface temperature decreases in all three experiments. The magnitude and timing of the surface temperature drop in SURF-4DVAR is the closest to the observations among the three. The MTDs of this surface station are 1.82 K, 1.42 K and 0.91 K for RAD-ONLY, SURF-MESO and SURF-4DVAR, respectively. The comparisons with the two representative stations again confirm the improvement of VDRAS analysis by assimilating surface observations together with radar data.

    Besides the improvements of the near-surface fields, assimilating surface observations may also improve the analysis results at higher levels through the 4DVAR scheme. To investigate whether the surface data assimilation can improve the VDRAS analysis in the PBL and above, observations from a nearby wind profiler (green square in Fig. 2) are compared for further validation. Observed and analyzed vertical profiles of U and V below the 3 km altitude at 1500 UTC are presented in Fig. 7. The vertical profiles of U in SURF-MESO and SURF-4DVAR are closer to the observations than those in RAD-ONLY. A slight improvement is achieved in SURF-4DVAR in the lowest 1 km for U, while a significant improvement is found for V, where the analysis results are closest to the observations. There is little difference between the three experiments above 1 km. The MVDs below 3 km are also calculated for each experiment, and they are 2.69 m s-1, 2.15 m s-1 and 1.82 m s-1 in RAD-ONLY, SURF-MESO and SURF-4DVAR, respectively. The comparison results show that assimilating surface data through the new 4DVAR scheme can improve the analyzed low-level wind profiles in VDRAS, with the most significant improvement found within the PBL.

    Figure 7.  Vertical profiles of $U$ and $V$ from wind profiler observations (black line) and VDRAS analysis at 1500 UTC 30 July 2014.

    Figure 8.  As Fig. 5, but for the 1-h forecasted field at 1600 UTC.

    Figure 9.  As in Fig. 6, but for the observed and 1-h forecasted surface temperature.

  • To validate the improvement of the very short-term forecasting skill of VDRAS through assimilating the surface data, a 1-h nowcast is launched from each analysis cycle in all three assimilation experiments (RAD-ONLY, SURF-MESO and SURF-4DVAR). Observed surface temperature, surface wind and composite radar reflectivity at 1600 UTC are compared with the 1-h nowcast fields starting from the analysis at 1500 UTC in Fig. 8. Overall, the forecasted reflectivity, surface wind and temperature from SURF-4DVAR are the best among three experiments. The one-hour forecasted temperature field in RAD-ONLY at 1600 UTC (Fig. 8b) is much warmer than observed for most parts of the domain, just as the analyzed surface field is at 1500 UTC (Fig. 5b). The cold pool is too weak and the surface outflow on the west side of this squall line is too strong. In SURF-MESO, the surface temperature is closer to the AWS observations than the RAD-ONLY forecast (Fig. 8c), but the cold pool is still too weak and the outflow is again too strong. Compared with the other two experiments, the 1-h forecasted surface fields in SURF-4DVAR are much closer to the observations (Fig. 8d). The northern and central parts of the OPACC domain are colder than the southern domain, the cold pool is stronger, the surface outflow is weaker, and the surface inflow on the east side of the squall line is warmer. The position of the gust front ahead of the squall line is also close to the AWS observations. Quantitatively, the MTDs of the RAD-ONLY, SURF-MESO and SURF-4DVAR experiments are 4.44 K, 2.48 K and 1.26 K, and the MVDs are 4.09 m s-1, 2.42 m s-1 and 2.29 m\;s-1, respectively. All these results indicate improved forecast ability with the additional assimilation of the surface observations in VDRAS. It is noteworthy that the cold pool for the squall line (Fig. 8a) remains close to the leading edge of the convection in the observations at 1600 UTC, while it appears to surge quickly outward in the 1-h forecasts of the three experiments. This disagreement may be induced by the lack of PBL processes in VDRAS, which can result in a stronger cold pool in the 1-h forecast.

    Comparing the temperature time series of the observations at the two AWS stations (Figs. 9a and b) with that of multiple 1-h forecasts from the continuous analyses for the three experiments, SURF-4DVAR performs much better at capturing the surface temperature drop after the passage of the two squall lines. The MTDs of station 58203 (58326) are 5.61 K (2.2 K), 3.96 K (1.81 K) and 1.49 K (1.45 K) in RAD-ONLY, SURF-MESO and SURF-4DVAR respectively. Figure 10 compares the observed and forecasted vertical profiles of U and V at 1600 UTC. Both the U and V components from SURF-4DVAR agree better with the wind profiler observations than the rest. The primary improvement is found again in the PBL below 1 km-similar to the analysis results in Fig. 7. The MVDs below 3 km are 5.35 m s-1, 4.46 m s-1 and 3.46 m s-1 for RAD-ONLY, SURF-MESO and SURF-4DVAR, respectively. These validation results show that assimilating surface data together with radar observations not only improves the representation of the surface winds and temperature, but also the low-level dynamic and thermodynamic fields, especially within the PBL.

    Compared with the temperature and wind fields, the 1-h forecast of the precipitation is not as significantly improved by the surface data assimilation. Quantitatively, the RMSE of composite radar reflectivity is 11.11 dBZ in SURF-4DVAR——slightly smaller than the 11.39 dBZ and 11.35 dBZ in RAD-ONLY and SURF-MESO for the 1-h nowcast from 1500 UTC. The differences between these forecasts are not significant. The reasons for the small differences need to be further investigated.

    Figure 10.  As in Fig. 7, but for the observed and 1-h forecasted wind profiles at 1600 UTC.

6. Summary and discussion
  • This paper explores the effect of using an improved surface observation assimilation scheme on the analysis and nowcasting ability of VDRAS. Different from the original version, where surface temperature and wind observations are mainly integrated into the model background using a simple interpolation scheme, the current method can assimilate high spatial and temporal resolution surface observations together with the radar reflectivity and radial velocity in the 4DVAR scheme. Surface observations are interpolated to the model bottom level and assimilated every 5 min. The objective of this study was to show whether the new scheme, through minimizing the cost function with the additional surface observation terms in the 4DVAR-based VDRAS, can obtain better low-level dynamic and thermodynamic analysis. The squall-line case of 30 July 2014, observed during IOP 10 of the OPACC field campaign, is used to demonstrate the performance of the new scheme. A first set of experiments assimilates observations from a single AWS station. Results show that the surface-based divergence, cold pool and outflow of the squall line are all strengthened towards the observed values. Some improvements are exhibited in the vertical profiles of wind, mostly within the PBL. A second set of three experiments (RAD-ONLY: only radar observations are assimilated; SURF-MESO: radar observations are assimilated with surface data blended with the WRF background; and SURF-4DVAR: radar and surface observations are assimilated together) is conducted to examine the impact of the surface data assimilation on the analysis and forecast of VDRAS. Among the three experiments, the analyzed surface temperature and wind in SURF-4DVAR show the best agreement with the AWS observations. The low-level wind profiles, especially those within the PBL, are also improved upon surface data assimilation. Similar improvements within the PBL are found in the 1-h forecast starting from the analysis fields.

    The current study reveals the potential benefits of assimilating surface data in a convection-permitting model in addition to radar data assimilation. Assimilating surface data can fill in the near-surface observational gap of radar data and improve the representation of near-surface and PBL conditions, which are essential for accurate analysis and forecasting of convective systems. Needless to say, more case studies are still needed to confirm the impact of the surface data assimilation, especially on the precipitation forecast. The idea presented in this study will be applied to mesoscale data assimilation systems, such as WRF data assimilation systems, in the future, to evaluate the applicability of the 4DVAR assimilation combining surface and radar observations to other systems.

Reference

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return