FIFTH INTERNATIONAL WORKSHOP ON TROPICAL CYCLONES
25 September 2002


Topic 2.2 Data Assimilation and Numerical Prediction of Precipitation

Rapporteur: Woo-Jin Lee
Forecast Bureau, Korea Meteorological Administration
Seoul, Republic of Korea

e-mail: wjlee@kma.go.kr or heseong@dreamx.net
fax: 02-836-3157

Working Group:
Dr. Scott Braun (NASA-Goddard ) braun@agnes.gsfc.nasa.gov
Dr. Shuyi S. Chen (Univ. of Miami) schen@rsmas.miami.edu
Dr. Woo-Jin Lee (Korea Meteorological Administration) wjlee@kma.go.kr
Dr. Melinda Peng (Naval Research Laboratory) peng@nrlmry.navy.mil
Dr. Naomi Surgi (EMC/ NCEP) naomi.surgi@noaa.gov
Dr. Hideo Tada (NPD/ JMA) hidetada@npd.kishou.go.jp
Dr. Chun-Chieh Wu (National Taiwan University) cwu@typhoon.as.ntu.edu.tw
Dr. Meng Zhiyong (Chinese Academy Meteorological Sciences) tipex@public.bta.net.cn


Abstract


The prediction of precipitation associated with landfalling tropical cyclone is discussed, including data assimilation, model physics, skills in quantitative precipitation and rainfall climatology, with a special focus on the data assimilation of moisture fields and motion fields retrieved from satellite-borne instruments and other asynoptic observations. The status and problems of vortex generation process or bogusing are briefly reviewed. Several points are proposed for recommendations to IWTC-V in the viewpoint of both operation and research, considering the progress of numerical models and associated data assimilation techniques, and voluminous amount of data available over the next decade.



2.2.1: Introduction

Flooding associated with tropical cyclone landfall is one of the major elements of destructive force for the local community along the cyclone path. Little information is available to the forecaster on the convective activity and rainfall intensity in a tropical cyclone over the ocean, except for the satellite-derived wind and precipitable water. A radar provides rich information on the rainfall intensity and three-dimensional wind component when it approaches a coast. Dense ground networks become efficient to measure surface wind, pressure, and rainfall amount once the cyclone landfalls. However, the interaction of landfalling cyclone with local topography and pre-existing multi-scale dynamic environment make the forecasting of severe weather more complicated. Forecasters on the operation bench normally have to digest vast amounts of information in a very short time (never sufficient) and to decide how much rain will come and where severe convective storms will develop.

One of the regional concerns on tropical cyclone is associated with the heavy rain forecasting, as discussed in the Workshop on Typhoon Forecasting Research (Jeju Island, Republic of Korea, 25 to 28 September 2001). The operational forecasting of heavy rain in tropical cyclone situations has remained an intractable problem in spite of heavy demand by the public for such information. A concerted effort is urgently required to develop techniques to give quantitative guidance to forecasters. It is recognized that it would have to involve the concurrent use of several approaches including high-resolution mesoscale numerical models, satellite microwave data and radar data applied in a nowcasting mode.
The accuracy of track forecasting has been much improved, which has improved the timeliness of warnings and advisories well ahead of landfall of tropical cyclone. For example, the 12–h ensemble mean forecast, adjusted for forecast difficulty, has improved at a rate of just under 1% per year, and the improvement rate increases to almost 2.4% per year for the 72–h forecasts from the suite of tropical cyclone track forecast models in the Atlantic basin from the 1976 to 2000 hurricane seasons (Aberson, 2001). It is important to note that the rate of improvement in skill of quantitative precipitation forecasts is rather slow, as compared with other advancements in observations and numerical models (Fritsch et al. 1998). The forecasting of tropical cyclone rainfall shares the same difficulties as involved in the quantitative precipitation forecasting. Based on the evaluation of the skill for landfalling tropical cyclones in the Atlantic basin during 1976 to 2000, landfall timing uncertainty is ~11 h at 24 and 36 h (Powell and Aberson, 2001), which suggests that the accompanied rainfall errors might be considerable due to the position and timing errors for landfalling. Since the state of art of the data assimilation system and numerical models is incomplete for the prediction of track and intensity of a tropical cyclone, it has to be supplemented by the subjective interpretation of raw observation based on conceptual models.


2.2.2: Measurement of precipitation

Satellites provide a homogeneous distribution of precipitation in both space and time. In addition to conventional estimates from InfraRed images, new estimates of precipitation from satellite-borne instruments such as passive microwave radiometers (SSM/I) and Tropical Rainfall Measuring Mission (TRMM) Microwave imager (TMI) and an active precipitation radar (TRMM PR) have been successful in the cross-validation with radar reflectivity and rain gauges (Barrett, 1999). Recently, Weng and Grody (2000) show that a reliable precipitation data also can be derived from the NOAA/AMSU-B observation.

There are attempts to produce a climatology based on satellite rainfall estimates for tropical cyclones (Marks, 2002). This climatology is a basis for the development of a persistence model, and can be used to validate numerical models and other precipitation estimates.

Since those sensors are rather new, it takes time to get a reasonable climatology of precipitation. Furthermore, the microwave sensors are mainly onboard the polar orbiting satellite resulting in the poor temporal resolution. However, this limitation will be much mitigated when the planned Global Precipitation Measurement (GPM) is implemented in 2007 or 2008. The estimates of precipitation from satellite, radar and gauges have various spatial scales and time frequencies. Even though those various estimates are more or less consistent when computed over large areas and /or long time periods (6-24 hours), a scheme has to be developed to integrate various information to make a homogeneous map of probability distribution for a given spatial scale.




2.2.3: Satellite data assimilation

Substantial recent improvements in the numerical forecasts have been attributed to the use of 3-dimensional variational data assimilation (3D-VAR) analysis and direct assimilation of TIROS Operational Vertical Sounder (TOVS) and ATOVS radiances (English et al. 2000; McNally et al. 2000). Recent improvements in short-range ECMWF forecasts are in part linked with the progress in the data assimilation system, including assimilation of raw microwave radiances from the TOVS and ATOVS satellite-borne instruments, and retrievals of humidity and surface wind speed from the SSM/I satellite-borne instrument (Mahfouf and Rabier, 2000; Simmons and Hollingsworth, 2002). Recent observing system experiments demonstrate that there is a beneficial impact particularly in the Southern Hemisphere from using TOVS radiances with 4-dimensional variational data assimilation (4D-VAR) (Bouttier and Kelly, 2001). The SATOBs and SSM/I retrievals also contribute to the improvement on the medium-range forecasts.

Those operational data assimilation systems are based on time independent background error statistics. Since its dominant scale is a few hundred kilometers, data at much higher horizontal resolution are effectively smoothed. It is a pre-requisite to use flow-dependent background error statistics to fully utilize the potential from the high resolution data for mesoscale models. Currently, there are several efforts to generate flow-dependent error statistics. These include several kinds of ensemble, suboptimal, and parameterized Kalman filter techniques (Errico et al. 2000, Houtekamer and Mitchell, 2001).

The accuracy of quantitative precipitation forecasts are critically dependent on the initial moisture field. Information of humidity vertical structure can be constructed from the precipitable water (Kuo et al. 1993). The convective heating terms or vertical profile of moisture and temperature in the model’s thermodynamic equations can be determined from the observed rainfall rate (Donner, 1988; Krishnamurti and Bedi, 1988). Those moisture equivalents can be assimilated in the model through dynamic initialization or variational procedures. Recently, Xiao et al. (2000) demonstrated that the SSM/I-measured precipitable water and rain rate had great potential to improve the initial conditions for the mesoscale model. The variational assimilation of those moisture estimates significantly improved the rapidly intensifying cyclone, which is reflected in the cyclone track, frontal structure and precipitation along the front. It was also shown that the assimilation of both precipitable water and rain rate produce more satisfactory improvement than the assimilation of either single estimate. The sensitivity to the assimilation of rain rate depends on the moist parameterization scheme. Pu et al. (2002) demonstrate that the TMI rainfall data leads to an improved simulation of Super typhoon Paka in terms of its intensity and kinematical and precipitation structures through the modification of environment of the storm favorable for development. Marecal and Mahfouf (2002) studied the impact of assimilating rain-derived information with 4D-VAR at ECMWF. They used two-step approach with 1-dimensional variational data assimilation (1D-VAR) and 4D-VAR. First the temperature and humidity profiles in the model are balanced with the rainfall estimates from TRMM microwave imager through 1D-VAR. Then 1D-VAR total column water vapor estimates are assimilated in 4D-VAR. The model precipitation spindown over tropical oceans was reduced with the assimilated rainfall estimates. In the sensitivity experiment for the Hurricane Bonnie, it was found that the rainfall assimilation produced positive impact on the improvement of storm tracks and intensity with the favorable modification of environmental wind and pressure.

Over the past several years, NCEP has made significant advancements in the assimilation of satellite data that has led to improvements in the forecasts of the environment of tropical cyclones. In combination with other numerical weather prediction (NWP) modeling advances, these improvements have led to improved tropical cyclone track forecasts. These advances include:

- Direct assimilation of the radiance data from Advanced Microwave Sounding Unit (AMSU) A/B , HIRS, and the Geostationary Orbiting Environmental Satellite (GOES) 8,9,10 IR;
- Physical initialization using the SSM/I and TRMM precipitation estimates;
- Inclusion of cloud liquid water effects in the satellite radiative transfer;
- Improved bias correction and quality control;
- Improved radiative transfer model; and
- Improvements in the model physics and background error covariances.

During 2004, research and development for several advanced instruments will continue: Cross- track Infrared Sounder (CrIS), Advanced Technology Microwave Sounder (ATMS), Infrared Atmospheric Sounder Interferometer (IASI), and advanced sounders on NPP and METOP and initial research for GIFTS. During 2005-2006, the preparations for GIFTS, CrIS, ATMS, IASI and the Global Positioning System (GPS) should be reaching maturity and data will be tested as they arrive. GIFTS data should become available and development will focus on the data as a prototype for the Advanced Baseline Sounder (ABS) and Advanced Baseline Imager (ABI). COSMIC should be launched and begin producing data. NCEP will concentrate on a multi-sensor cloud analysis, use of satellite data in cloudy and rainy situations, and the use of radiance data for the SST and the NCEP Global Land Surface Data Assimilation System. Between 2007-2008, the implementation for CrIS, ATMS, IASI, and COSMIC are anticipated. The ABS which will likely to be a GIFTS-like instrument will be a revolutionary advancement in soundings and will require a very large science effort. This will begin in 2007 and continue in 2008 and beyond.


The Japan Meteorological Agency (JMA) also experienced improvement in performance of their global spectral model after introduction of 3D-VAR with assimilation of TOVS/ATOVS radiances. The improvement in the analysis of the tropical atmosphere further contributed to the reduction of track errors of tropical cyclone. Currently experiments of 4D-VAR with the JMA mesoscale model are going on, assimilating precipitation retrieved from TRMM/TMI and DMSP/SSMI satellite-borne instruments.

Advanced meteorological forecast centers such as JMA and NCEP have also begun assimilating the QuikScat data which to date has showed some improvement over the tropics at the near–surface. However, the NCEP experience is that optimal use of scatterometer data will require substantial development of a quality control; and superobbing algorithms over the next several years. It is also true for the assimilation of satellite derived atmospheric motion vector. LeMarshall et al. (1998, and 2000) have shown that the inclusion of high quality atmospheric motion vector in the 4D-VAR systems could increase the accuracy in both the intensity and track of the tropical cyclone.


2.2.4: Radar and other observation platforms

Frank and Ritchie (1999) indicated that the pattern of convection in the storm’s core is strongly influence by vertical wind shear. A recent intercomparison of mesoscale models implies that the asymmetric wind distribution has close relationship with the precipitation distribution (Nagata et al. 2001). The study further suggests that better prediction of the wind distribution is crucial for better prediction of the precipitation distribution. Doppler radar network, airborne Doppler radar, and TRMM PR have produced a new generation of tropical cyclone data to understand the severe weather events associated with landfalling tropical cyclones, including boundary layer wind structures, and the spatial and temporal changes in the storm rain distribution (Marks, 2002).

To maintain the storm after an initial adjustment stage and to predict storm-scale precipitation, detailed boundary wind structure such as convergence and heating profiles have to be observed and assimilated in a very high resolution model with a very rapid update cycle. The 4D-VAR experiments with rainfall estimates from radar provide promising results (Wilson et al. 1998). Sun and Crook (2001) show that the analysis assimilated with the inclusion of boundary layer wind and temperature analysis using WSR-88D radar data outperforms the persistence forecast and a forecast using mesoscale analysis in comparison with observed radial velocity. Lightning data have been tuned with space-borne microwave radiometer data through a probability matching technique, which as a whole yielded improved forecasts of precipitation distributions and vertical motion fields (Chang et al. 2001). It was also shown that the assimilation of latent heating in the correct location on the forecast model was more important than an accurate determination of the rainfall intensity. MacDonald and Xie (2002) developed a 3D-VAR that recovers the moisture fields from the slant-integrated water vapor between ground-based Global Positioning System (GPS) receivers and GPS satellites combined with surface moisture observations and a limited number of moisture soundings, which might have potential for the diagnosis of three-dimensional water vapor with applications for both positioning and mesoscale weather prediction.


Although satellite observations have improved the larger scale tropical circulation in NWP models, in-situ observations from aircraft and dropsondes in the tropical cyclone environment have shown the most promising impacts on improvements to track forecasts by NWP models. Studies at the NOAA Hurricane Research Division over the past several years have shown that the assimilation of dropsondes in the global data assimilation system (GDAS) has, on average improved forecasts in the Global Forecast System (GFS) and the Geophysical Fluid Dynamics Laboratory (GFDL) model by 15-25%.


2.2.5: Asymmetric vortex generation

Recent model intercomparisons for Typhoon Flo (9019) demonstrate the sensitivity of the track and intensity prediction to the choice of analysis and synthetic tropical cyclone vortex for the initial field (Nagata et al. 2001). Several attempts have been made to incorporate a dynamically consistent tropical cyclone vortex with reasonable asymmetries into the initial conditions of a NWP model. Pu and Braun (2001) and Zou and Xiao (2000) developed a 4-dimensional bogus data assimilation scheme to find the optimal state that minimizes a cost function incorporating pre-specified vortex and satellite-derived water vapor wind vectors as observations. Zhu et al. (2002) made use of the retrieved temperature profiles and the total precipitable water from the AMSU satellite-borne instrument. The wind and geopotential fields are derived from the temperature profile with the specified sea-surface pressure distribution. The 48-h simulation with the Hurricane Bonnie captures reasonably well the track and rapid deepening stage of the storm. The JMA also adopted an observation-type of typhoon bogusing scheme to incorporate with the 4D-VAR in their mesoscale model.

It is noted that there is a trend to avoid the artificial vortex generation and to purely rely on the data assimilation of surrounding observations to define the vortex. For instance, NCEP has abandoned bogusing in the GFS, except in certain situations where the cyclone circulation is particularly diffuse. They have found that their prior procedure is inherently detrimental to the larger scale tropical environment, and hence the environmental steering flow. Over the past several years, they have greatly improved their track forecasts in the NCEP GFS with a relocation procedure inserted in the first-guess field based on the real-time storm center location provided by the US National Hurricane Center.


In anticipation of running the high resolution Hurricane Weather Research Forecast model(HWRF), NCEP is developing a local mesoscale data assimilation capability for initializing the hurricane core circulation with real-time wind profiles from the NOAA G-IV (as a candidate platform) to describe the 3-D wind structure of the hurricane core circulation from the outflow layer to the surface. This capability is being developed in concert with the current NCEP 3D-VAR and will make use of future advanced mesoscale variational data assimilation techniques being developed at NCEP. They believe this capability is critical to address the intensity problem by resolving the outflow layer of the hurricane core with detailed vertical resolution to the surface to properly simulate the boundary layer fluxes and resolve the vortex asymmetries.


2.2.6: Relation to physical parameterization

The skill of quantitative precipitation forecasts depends on the quality of the initial analysis of moisture variables, and partially on the physical parameterization of convection. The data assimilation in turn is strongly influenced by the moist processes in the model through the forward projection, and the errors of which can be properly treated with the introduction of a random error source term in the 4D-VAR and Kalman filter techniques. To improve the data assimilation of precipitation and clouds, parallel progress on the following areas is a pre-requisite:


- Improved forward models for specific instruments;
- Observation error statistics;
- Moisture balance constraints and background error formulation; and
- Efficient numerical algorithms.

There have been studies showing significant effects of the coupling of atmosphere with ocean on the intensity of tropical cyclones (Bender and Ginis, 2000). Ocean data assimilation effort is a new effort at NCEP and is important to improve the oceanic/ atmosphere boundary layer in the coupled model systems so as to include important processes in the mixed layer and account for the oceanic heat content. This effort will include the development of data assimilation techniques for satellite data (e.g. altimeter data) as well as asynoptic in-situ data (e.g. AXBTs, XBTs, ARGOS) for improved model representation of SST decreases in the wake of the tropical cyclones and the effects of oceanic warm core eddies, both of which are important to the intensity forecast problem.

The coupling of the land surface model to the GFDL model is underway at NCEP and this technology will be transferred to the coupled Weather and Research Forecast model (WRF) to account for the changes in the surface fluxes for landfalling tropical cyclones. The data assimilation effort will make use of the above remote sensing platforms in addition to the land based current or new in-situ observations over the continental US.


2.2.7: Summary and recommendation

  1. Rain climatology

It is pre-requisite to establish a climatology of precipitation for a tropical cyclone from the satellite and radar data, which will be used as a reference for the subjective forecasting of precipitation amount and to validate the performance of a numerical model.


  1. Satellite data assimilation

In 2000, 87% of data in NCEP models were derived from satellites; currently, it is 99%. It is estimated that five orders of magnitude of new satellite data will become available in the next 10 years, including the advanced sounders such as Atmospheric Infrared Sounder (AIRS) on the NASA EOS PM (Aqua) spacecraft and Infrared Atmospheric Sounding Interferometer (IASI) on the EUMETSAT METOP-1, and the occultation data from the navigation satellites such as GPS. There might be required a radical change to background error statistics to gain the full benefit from the new observations. The assimilation of satellite radiances should be extended to accommodate the effect of clouds and precipitation in the data retrieval. In addition, time-dependent background error covariances have to be evaluated to utilize the high-resolution data available in coming years.


Although the obvious task at hand is in accommodating this exponential increase in satellite observations, a significant part of the scientific challenge at operational NWP centers remains in the development of data assimilation techniques, continuous model development and in improving the background error covariances. This effort is fundamental to advancing NWP and should continue to remain a focus of scientific investment within the global NWP community.


  1. Targeted observations

As demonstrated by Gelaro et al. (2000) for the sensitivity to targeted observations on model performance, upstream of major unstable modes of a baroclinic system is the most sensitive area for targeted observations. In the case of tropical cyclones, the issue is relevant when a tropical cyclone is going to make a turn into the upper-level westerlies. More studies are needed to get a best strategy for the permanent or in-situ observation network for the forecasting of tropical cyclones.


  1. Sharing of forward models

The IWTC-IV recommended sharing forward models for facilitating the adoption of new satellite and radar observations into existing data assimilation platforms. In principle, the idea seems to greatly help each operational center reduce the duplicative workload on developing pre-processors and quality control procedures for a new kind of observation, mostly for satellite data and radar data. Currently, such experience and software are frequently shared among EUMETSAT, ECMWF, and Meteo France in Europe, and NASA, NESDIS, and NCEP in the United States. Radiative transfer codes have been freely shared among many organizations.


However, a concern exists that in practice the value of sharing may be rather limited since each forward model is closely tied to the given computer system and data assimilation modules, and database. Therefore, a modular approach and a standard procedure need to be established for the development of a forward model. No less of an important point is the limitation of computer resources in terms of data volume, speed of data processing, and /or parallelization of code. As more data will be ingested in a data assimilation system, more and more efficient hardware and software will be required to meet the operational production schedule. Apart from the collaborative efforts to develop forward operators, it is desirable to share experiences on working-level knowledge such as selection of background, treatment of biases, quality control process, observation data characteristics, etc. (Erico et al. 2000).


  1. Physical parameterization

Parallel improvements are required on the physics of the sea and land surface model and hydrology, and especially for deep and shallow convection along with the advancement of remote sensing technology and associated data assimilation.


  1. Interaction with orography and other circulation system

The landfalling tropical cyclone often interacts with orography and produces persistent rainfall on the upwind side. The sensitivity of model horizontal and vertical resolution versus orographic rainfall and inner core structure of tropical cyclones may provide some indication of the expected precipitation over mountainous areas (Nagata et al. 2001; Mass et al. 2002). When a baroclinic zone with an upper-level cold pool meets the southerly moist air stream from a tropical cyclone, strong convective rainfall occurs. The mechanism for the extratropical transition of tropical cyclone could contribute to the conceptual framework for the rainfall forecasting at higher latitudes (Klein et al. 2002). Further research should be on how the structure of the boundary layer and vertical shear of the tropical cyclone are modified to become unstable with the surrounding environment.


  1. Probability of precipitation

Precipitation is highly localized and intermittent in space and time, and results from a wide range of scale interactions. In parallel with the research on forecasting of quantitative precipitation, the analysis and forecasting of probability of precipitation deserves attention (Fritsch et al. 1998), since it could incorporate in a realistic manner the various sources of errors or uncertainties in both observations and model predictions. Observation errors are involved with measurement, representativeness in space and time, retrieval, and calibration with different observation platforms. The uncertainty of precipitation amounts could be measured from the ensemble of numerical simulations perturbed with initial condition and/ or physical parameterizations (Krishnamurti et al. 1997; Elsberry and Carr, 2000). The probability of precipitation could provide the most efficient means of communication for the decision makers faced with an uncertain environment with landfalling tropical cyclones.


  1. Balance between operations and research

The pre-processing of satellite and radar data ranges from the quality control, forward operation, evaluation of error covariances and other statistics for the validation of the model and assimilation system. The byproduct from the pre-processing, including the precipitation estimates retrieved from the satellite and radar in particular, contains valuable information for the forecaster to understand the current state of atmosphere, while the same data are being used as an input for the data assimilation system. It is desirable to have a balance between the operation tools based on conceptual models and research for the development of forward operators.


  1. Data management

The voluminous amount of satellite data that will become available over the next decade places an increased demand on the planning and design of optimal NWP systems that have to be aligned to upgrades in computing in the various global communities.

Many channels from the advanced sounders will provide detailed structures of temperature and humidity. The optimum selection of channels or thinning will be a hot issue for the future.

The issue of data handling, data storage and overall data management issue has become a huge challenge to the overall global NWP community. Although NOAA and NASA have aligned efforts to meet this challenge with the establishment of the Joint Center for Satellite Data Assimilation (JCSDA), this challenge is global in nature and will benefit from the overall collaboration of the entire global NWP community.


Bibliography


Aberson, S.D., 2001: The ensemble of tropical cyclone track forecasting models in the North Atlantic basin (1976–2000). Bull. Amer. Meteor. Soc.,82, 1895–1904.

Barrett, E.C., ed., 1999: Estimating the amount of rainfall associated with tropical cyclones using satellite techniques. Tropical Cyclone Programme Report No. TCP-42, WMO/TD No. 975, 294 pp.

Bender, M.A., and I. Ginis, 2000: Real-case simulations of hurricane-ocean interaction using a high resolution coupled model: effects on hurricane intensity. Mon. Wea. Rev., 128, 917-946.

Bouttier, F., and G. Kelly, 2001: Observing-system experiments in the ECMWF 4D-Var data assimilation system. Quart .J. Roy. Meteor. Soc., 127, 1469-1488.

Chang, D.-E., J.A. Weinman, C.A. Morales, and W.S. Olson, 2001: The effect of spaceborne microwave and ground based continuous lightning measurements on forecasts of the 1998 Groundhog day storm. Mon. Wea. Rev., 8, 1809-1833.

Donner, L.J., 1988: An initialization of cumulus convection in numerical weather prediction models. Mon. Wea. Rev., 116, 377-385.

Mass, C. F., D. Ovens, K. Westrick, and B. A. Colle, 2002: Does increasing horizontal resolution produce more skillful forecasts? Bull. Amer. Meteor. Soc., 83, 407–430.

Elsberry, R. L., and L.E. Carr III, 2000: Consensus of dynamical tropical cyclone track forecasts-Errors versus spread. Mon. Wea. Rev. 128, 4131–4138.

English, S.J., R.J. Renshaw, P.C. Dibben, A.J. Smith, P.C. Rayer, C. Poulsen, F.W. Saunders, and J.R. Eyre, 2000: A comparison of the impact of TOVS and ATOVS satellite sounding data on the accuracy of numerical weather forecasts. Quart. J. Roy. Meteor. Soc., 126, 2911-2931.

Errico, R.M., G. Ohring, J. Derber, and J. Joiner, 2000: Meeting summary. NOAA-NASA-DoD workshop on satellite data assimilation. Bull. Amer. Meteor. Soc., 81, 2457-2462.

Frank, W.M., and E.A. Ritchie, 1999: Effects of environmental flow upon tropical cyclone structure. Mon. Wea. Rev., 127, 2044-2061.

Fritsch, J. M., R.A. Houze, Jr, R. Adler, H. Bluestein, L. Bosart, J. Brown, F. Carr, C. Davis, R.H. Johnson, N. Junker, Y.-H. Kuo, S. Rutledge, J. Smith, Z. Toth, J.W. Wilson, E. Zipser, and D. Zrnic, 1998: Meeting summary. Quantative precipitation forecasting: report of the eighth prospectus development team, U.S. Weather Researh Program. Bull. Amer. Meteor. Soc., 79, 285-299.

Gelaro, R., C.A. Reynolds, R.H. Langland, and G.D. Rohaly, 2000: A predictability study using geostationary satellite wind observations during NORPEX. Mon. Wea. Rev., 128, 3789-3807.

Houtekamer, P.L., and H.L. Mitchell, 2001: A sequential Kalman filter for atmospheric data assimilation. Mon. Wea. Rev., 129, 123-137.

Klein, P. M., P.A. Harr, and R.L. Elsberry, 2002: Extratropical transition of western North Pacific tropical cyclones: Midlatitude and tropical cyclone contributions to reintensification. Mon. Wea. Rev.,130, 2240–2259.


Krishnamurti, T. N., R. Correa-Torres, G. Rohaly, and D. Oosterhof, 1997: Physical initialization and hurricane ensemble forecasts. Weather and Forecasting, 12, 503-514.

Krishnamurti, T.N., and H.S. Bedi, 1988: Cumulus parameterization and rainfall rates: Part III. Mon. Wea. Rev., 116, 583-599.

Kuo, Y.-H., Y.-R. Guo, and E.R. Westwater, 1993: Assimilation of precipitable water measurements into a mesoscale numerical model. Mon. Wea. Rev., 121, 1215-1238.

Le Marshall, J., N. Pescod, R. Seaman, G. Mills, and P. Stewart, 1994: An operational system for generating cloud drift winds in the Australian region and their impact on numerical weather prediction, Weather and Forecasting, 9, 361-370.

Le Marshall, J. F. Leslie, L. M. Morison, R. Pescod, N. Seecamp, and R. Spinoso, 2000: Recent developments in the continuous assimilation of satellite winds for tropical cyclone track forecasting. Adv. Space Res., 25, 1077-1080.

MacDonald, A.E., and Y. Xie, 2002: Diagnosis of three-dimensional water vapor using a GPS network. Mon. Wea. Rev., 130, 386-397.

Mahfouf, J.F., and F. Rabier, 2000: The ECMWF operational implementation of four-dimensional variational assimilation. II: Experimental results with improved physics. Quart. J. Roy. Meteor. Soc., 126, 1171-1190.

Marecal, V., and J.-F. Mahfout, 2002: Four-dimensional variational assimilation of total column water vapor in rainy areas. Mon. Wea. Rev., 130, 43-58.

Marks Jr., F. D., 2002: State of the science: Radar view of tropical cyclones. Dave Atlas Symposium, AMS Annual Meeting. 14 Jan. 2002, Orlando, FL, 92pp.

McNally, A.P., and J.C. Derber, W. Wu and B.B. Katz, 2000: The use of TOVS level-1b radiances in the NCEP SSI analysis system. Quart. J. Roy. Meteor. Soc., 126, 1171-1190.

Nagata, M., L. Leslie, H. Kamahori, R. Nomura, H. Mino, Y. Kurihara, E. Rogers, R.L. Elsberry, B.K. Basu, A. Buzzi, J. Calvo, M. Desgagne, M. D’Ididoro, S.-Y. Hong, J. Katzfey, D. Majewski, P. Malguzzi, J. McGregor, A. Murata, J. Nachamkin, M. Roch, and C. Wilson, 2001: A mesoscale model intercomparison: a case of explosive development of a tropical cyclone (COMPARE III). J. Meteor. Soc. Japan, 79, 999-1033.

Powell, M.D., and S. D. Aberson, 2001: Accuracy of United States tropical cyclone landfall forecasts in the Atlantic Basin (1976–2000). Bull. Amer. Meteor. Soc., 82, 2749–2768.

Pu, Z.-X., and S.A. Braun, 2001: Evaluation of bogus vortex techniques with four-dimensional variational data assimilation. Mon. Wea. Rev., 118, 67-93

Pu, Z., W.-K. Tao, S. Braun, J. Simpson, Y. Jia, J. Halverson, W. Olson, and A. Hou, 2002: The impact of TRMM data on mesoscale numerical simulation of Supertyphoon Paka. Mon. Wea. Rev., 130, 2448-2458.

Simmons, A.J., and A. Hollingsworth, 2002: Some aspects of the improvements in skill of numerical weather prediction. Quart. J. Roy. Meteor. Soc., 128, 647-677.

Sun, J., and N.A. Crook, 2001: Real-time low-level wind and temperature analysis using single WSR-88D data. Weather and Forecasting, 16, 117-132.
Weng, F., and N. C. Grody, 2000: Retrieval of ice cloud parameters using a microwave imaging radiometer, J. Atmos. Sci., 15, 1069-1081.

Wilson, J. W., N. A. Crook, C. K. Mueller, J. Sun, and M. Dixon, 1998: Nowcasting thunderstorms: a status report. Bull. Amer. Meteor. Soc., 79, 2079-2099.

Xiao, Q., X. Zou, and Y.-H. Kuo, 2000: Incorporating the SSM/I-derived precipitable water and rainfall rate into a numerical model: A case study for the ERICA IOP-4 cyclone. Mon. Wea. Rev., 128, 87-108.

Zhu, T., D.-L. Zhang, and F. Weng, 2002: Impact of the advanced microwave sounding unit measurements on hurricane prediction. Mon. Wea. Rev., 130, 2416-2432.

Zou, X., and Q. Xiao, 2000: Studies on the initialization and simulation of a mature hurricane using a variational bogus data assimilation scheme. J. Atmos. Sci., 57, 836-860.