I. Introduction

The impetus for the workshop and, fundamentally, for this area of research stems both from the significant societal impact of weather, such as personal property loss and human casualties, but also national economic impact, which amounts to $300 million lost weekly in the US alone (Pielke 1997). This figure accounts significantly for the industrial sector that experiences costly disruption to their communication and transportation links. For example, commercial airlines are dependent on accurate regional forecasts, both enroute and in the terminal area, to avoid costly flight delays and traffic re-routing. Power and communication utility companies require region-specific weather information to insure uninterrupted service. Also, maintaining a free-flowing transportation network of roads and railways, which is aided by precise weather information, is vital for interstate as well as international trade and transport.

The distinctive goal of the workshop was to improve, through the medium of international scientific cooperation, the forecast capabilities in the US and Korea for small-scale weather events (heavy rainfall, fog, terminal winds, etc.) that affect specific localities. The workshop format, which included a series of scientific presentations, plenary sessions and working groups, was designed to encourage fruitful discussion concerning the issues involved in understanding the dynamics and numerical modeling of mesoscale systems. To concentrate on a specific (but widely representative) problem, an emphasis was given to storm complex types that often produce locally heavy rainfall and severe flooding in Korea. By the end of the workshop, a list of key research objectives was identified that would enhance the regional forecasting and warning capabilities of the US and Korea. These research objectives are presented in this report and serve as a basis for future international collaborations and a foundation for the submission of joint research proposals.

The event was co-funded by the National Science Foundation (NSF) and the Korean Science and Engineering Foundation (KOSEF) with supplemental funding by the Center for Analysis and Prediction of Storms (CAPS), Seoul National University (SNU), the Korea Meteorological Administration (KMA), and the Korea Meteorological Society (KMS). The workshop was coordinated in Korea by Prof. Dong-Kyou Lee of SNU and in the US by Prof. Kelvin Droegemeier and David Jahn of CAPS at the University of Oklahoma.

II. Background and impetus for scientific collaboration with Korea

Korea was chosen as a natural research partner due to its depth of mesoscale weather research and support of an extensive weather observation network. KMA maintains an automatic surface weather observations system (AWS) consisting of 400 sites across Korea as well as a network of Doppler radars that provide reflectivity (precipitation) measurements and will be upgraded soon to provide also radial wind information.

Korea also represents an excellent location to test mesoscale analysis and prediction techniques because it is affected by a broad range of interesting meteorological phenomena within a relatively small region, from local severe spring and winter storms to sea-breeze effects and typhoons, the causes of which are complicated by the varied topography of the region made up of both coastlands and mountainous areas. In particular, flooding is a major concern for Korea. At KMA, the Korean Meteorological Research Institute (METRI), and various Korean universities, there has been significant study on the dynamics of convective processes and the effect of topography in the production of heavy rainfall. Already in use by Korean researchers are several US mesoscale models: the ARPS of CAPS (University of Oklahoma), MM5 of the National Center for Atmospheric Research/Pennsylvania State University, and RAMS of Colorado State University.

In February 1996, CAPS hosted a preliminary workshop in Norman, Oklahoma, which included more than 30 participants from 8 Korean and 12 US organizations, in order to explore the possibilities for collaboration with counterpart Korean agencies. Prof. Dong-Kyou Lee of Seoul National University served as the counterpart Korean coordinator. The topics of this initial workshop concentrated on the technology needed to support operational storm- or mesoscale NWP such as data observation and analysis techniques and the logistics for creating a numerical forecast in real-time.

Since early 1996, scientific interaction between US and Korea in the area of small-scale weather analysis and prediction has increased and resulted in several scientific exchange visits and a collaborative project. Namely, KMA has initiated a 4-year evaluation study of CAPS' data analysis (ADAS) and mesoscale model (ARPS) for use in their daily forecast operations: TAKE (Test of ARPS in the Korean Environment).

As a result of the preliminary workshop and subsequent scientific interaction, there existed a growing interest in and impetus for holding a follow-up workshop. The emphasis of this second workshop was slightly different from the first giving more prominence to the "hard" science involved in numerically forecasting mesoscale phenomena. It also served to avail an increased number of research groups across the US to the opportunities that exist for collaboration with Korea.

III. Summary of workshop results

Prior to the workshop, 6 major areas of research were identified to serve as the focal point for a set of individual research presentations, plenary sessions, and working groups. These topic areas are:

1. Sensitivity and targeted observations;
2. Assimilation and modeling;
3. Orographic effects;
4. Microphysics of heavy rainfall;
5. Dynamics of mesoscale convective systems;
6. Operational testing.

Through the course of the 4-day workshop, scientists, took advantage of the congenial and yet professional atmosphere of the workshop, engaging themselves frequently in open and very productive discussions. In the context of improving the forecasting of mesoscale phenomena and especially the onset of heavy rainfall in Korea, participants sought corporately to identify the major limitations in each of the 6 topic areas as a basis for proposing further research opportunities. In general, there is need to improve the scientific understanding of basic weather mechanisms and their interaction between different scales, to enhance the availability of observational data on these scales, and to determine and ameliorate constraints in numerical modeling of pertinent dynamical processes. The major points for consideration are detailed by topic area in section IV, but are also consolidated below into an overall list of primary research objectives:

1. Identify the effects of topography on convective initiation and morphology including related orographic effects on the microphysical processes and large-scale forcing mechanisms. Consider also the transport of CCN particles and the influence on boundary-layer flow and transport of low-level moisture;

2. Increase the basic understanding of microphysical processes over Korea (there is very little observational data) including the time and spatial scales of the processes and their interaction with larger scale forcings to support the development of appropriate microphysical parameterizations for use in regional models;

3. Investigate the role of scale-interaction in the development of mesoscale systems with the mechanisms of the larger-scale forcings (as in the case of precipitation induced by the Changma front or MCS's that propogate over the Yellow Sea from China);

4. Introduce and evaluate 4DVAR data assimilation techniques (including microphysical variables) to model studies over Korea;

5. Investigate means of obtaining data over nearly data-void areas (Yellow Sea, N. Korea) and measure relative impact on model accuracy of data at various resolutions and from different observational platforms.

6. Assess the viability of high resolution, storm-resolving models in the operational prediction of intense weather systems and events.

IV. Research objectives by topic area

A. Modeling and assimilation

Numerical modeling is a vital research avenue that will enable in-depth investigation of the mechanics of mesoscale weather and improve operational forecasting techniques. Modeling, however, in itself is a huge study area involving several factors such as the appropriateness of model physics, impact of data type and resolution on forecast accuracy, sensitivity issues, sub-grid parameterization schemes of moisture and turbulence processes, etc. To remain consistent with the focus of the workshop, participants narrowed their scope to issues affecting the numerical prediction of highly precipitative mesoscale systems over the Korean region.

In particular, a key focus was the means of storm initiation over the Korean peninsula. Considerable amount is known about the synoptic influences, but little is understood about the actual triggering mechanism(s). Areas of focus are orographic influences, adequate model characterization of boundary layer moisture evolution, resolution of small-scale wind convergence regions (sea breeze effects), as well as multi-scale interaction.

For the most part, there is a lack of observational data on a resolution sufficient to sense mesoscale processes. This is especially true over North Korea (a virtual data void) as well as over the Yellow Sea, which is an origin of much of the convective systems that propogate west to east over the peninsula. Although KMA does operate a finely-resolved automatic surface observations network across South Korea, these data are ineffectual in characterizing systems upstream from the Korean peninsula that greatly determine the weather over the region but are not adequately sensed over the Yellow Sea.

In Korea 4DVAR data assimilation techniques are currently under consideration, but more work is needed to achieve an optimal use of observational data in modeling studies. There is a need to evaluate the appropriate means of representing surface observations in the data assimilation as well as techniques that properly handle data analysis and assimilation in the presence of complex terrain.

To address these scientific issues, several research opportunities have been identified:

1. Investigate the initiation mechanisms as well as the evolution and sensitivity of storms in different synoptic environments;

2. Procure in situ or remotely sensed observations over the Yellow Sea (radar, satellite, etc.);

3. Tune existing models (ARPS, MM5, and RAMS are all currently being used) to account for the important features and mechanisms in the Korean environment (terrain effects, moisture parameterizations, etc.);

4. Conduct a data impact study to determine forecast sensitivity to the number and type of observational data;

5. Run data assimilation experiments using various 4DVAR techniques, initially with simulated data.

Korea has a diverse topography consisting of coastlands and mountainous terrain that have a significant impact on the evolution of weather over the peninsula. Of major concern is the production of heavy rainfall and flash floods, which, along with typhoons, is a major cause for loss of life and property annually. A full list of local weather phenomena that are greatly influenced by the terrain and for which Korea is wanting to improve their forecast capabilities include:

• initiation mechanisms of heavy rainfall
• trapping and redistrubution of air pollutants
• inter-valley drainage flow
• lee waves, severe downslope wind, and hydraulic jumps to the east of the Taebak Mts.

To forecast the weather over a local region, it is also important to understand the impact of terrain on the larger-scale circulations as well such as:

• deflection on prevailing wind (shear flow)
• modification of frontal boundaries (Changma and polar cold fronts)
• intensification of cyclones (typhoons) and onset of cyclogenesis over adjacent seas
• cold-air damming to east of Taebak Mts.

Several technical issues that need to be addressed in the numerical modeling of these various phenomena are:

• physical representation of terrain in numerical models (terrain parameterizations have been proposed)
• appropriateness of using a hydrostatic vs. nonhydrostatic model
• influence of terrain type on model initiation and morphology

To address these forecasting concerns, several research studies have been prioritized:

1. Investigate the effects of terrain on propagating cyclones and frontal structures. It is proposed that variations in propagation both in speed and in direction relative to a mountain range will impact the precipitation forecasts.

2. In the case of approaching cyclones, establish a database of anticipated rainfall distribution dependent on the direction and speed of the prevailing wind (e.g. 750mb wind) created systematically based on a series of numerical model runs. These statistics would be used as a quick estimation of local precipitation before fine-resolution forecast model results are available.

3. Investigate effects of terrain on low-level moisture transport and concommitant convergence zones, which provide crucial guidance information for heavy precipitation.

4. Develop a reference for terrain data for numerical models at various resolutions, which will help ease the interpretation of model results.

The microphysical processes can have a significant impact on storm morphology and must be considered in numerical forecasting especially if emphasis is placed on quantitative precipitation forecasting (QPF). In Korea, however, there is a lack of observational data that is otherwise needed to define the parameterizations of pertinent precipitation processes, to determine the relative importance of ice versus warm rain processes in clouds and convective systems over Korea, or to determine the time and space scales of heavy rainfall events (whether resolvable or not).

Because of the paucity of observational data in Korea, workshop participants proposed a research program that includes not only modeling efforts to uncover the basic microphysical mechanisms, but also the observational data to guide the development of modeling techniques as well as verify the results.

Specific objectives of collaborative research include:

1. Investigate the appropriate complexity of the microphysical parameterizations for the scope of weather phenomena in Korea. A microphysics model can quickly become complicated and computationally intensive if all phases of water are considered along with their source/sink interactions and the heat exchange associated with the change-of-state. A balance must be determined between using a simple and less computationally expensive model and a model that does not compromise the effect of prominent microphysical processes.

2. Consider the importance of resolvable versus parameterized precipitation. Depending on the grid resolution of the model and the type of weather event under consideration, there must be guidelines to determine the use of parameterizations versus explicit microphysics to generate precipitation in the model or an appropriate combination of the two.

3. Identify the extent to which the pertinent microphysical processes are influenced by larger-scale dynamical forcings or external effects such as orography and the transport of cloud-condensation nuclei (CCN).

4. Obtain a basic analysis of the climatology of clouds and rainfall.

5. Develop data assimilation methods to ingest microphysical data for model initialization.

Several studies have been conducted in Korea and the US, but the current level of understanding is not as yet entirely sufficient in most cases for predictive purposes. In Korea this is especially true for MCS's in the summer and convective bands associated with baroclinic waves in the winter. Specific mesoscale processes, for which better understanding is needed, include boundary layer processes, surface physics especially related to land-water interface, radiation effects at the surface and cloud interaction, as well as other factors that are considered under separate workshop topics such as microphysics and orographic effects.

Using MCS's and winter-time convection in Korea as a focus, several research objectives are proposed:

1. Investigate reasons for convection initiation and decay looking for possible self-organization by convective elements;

2. Study the role of larger-scale forcing interaction with smaller scales. This could involve, for example, the interaction of a convective region with the Changma front and low-level jet in summer or the developing baroclinic wave in the presence of a cold air mass in the winter;

3. Pursue a better understanding of boundary-layer processes, examine in particular the initial distribution and evolution of moisture while considering the temporal and spatial scales of pertinent processes;

4. Investigate the impact of topography on storm- and mesoscales as well as influences on the larger-scale forcings;

5. Investigate the significance of microphysical processes (warm vs. ice), which can at times be a first-order importance to MCS evolution. The necessary level of complexity of the microphysics scheme should be addressed.

6. Radiation effects in the boundary layer as well as the free atmosphere. Like microphysics, radiation feedback mechanisms (interconnection of absorption and reflection from clouds, surface, and free atmosphere) can become complicated; first- and second-order effects should be identified.

Operational forecasting is dependent on the "hard" science issues such as the degree to which mesoscale dynamics are understood (including microphysics and orographic effects) and the success of modeling these processes numerically. However, the use of a mesoscale model in an operational environment, is itself an area demanding considerable study. In order to evaluate the viability of high resolution, storm-resolving models in the operational prediction of intense weather systems, workshop participants compiled a list of focus areas:

1. Define succinctly what constitutes a good forecast. This includes development of techniques to verify forecast of specific features (e.g. inversions, low-level jet, moisture convergence, surface boundaries, etc.). It is noted that traditional statistical scoring techniques are not applicable to storm-scale phenomena due to phasing errors;

2. Develop communicative forecast products from model grid point data. This could include guidance from a forecast ensemble in the form of a probability chart of a specific weather occurrence (composite reflectivity, cloud top height, etc.);

3. Investigate the optimal mode of operation. Besides determining best choice of model set-up (physics, grid resolution, nesting, etc.), the means of ingesting real-time data and running the model must be considered such as the use of forecast cycles and ensembles. Computer resource and timing will greatly determine what approach is feasible;

4. Assess the best means of promoting operational forecaster use of mesoscale model forecast products and engaging their input in the set-up of operational tests.

V. References

Represented Korean Institutions

K-JIST  Kwangju Institute of Science and Technology
KMA     Korea Meteorological Administration
KNNU    Kangnung National University
KNU     Kyungpook National University
METRI   Meteorological Research Istitute (KMA)
PKNU    Pukyung National University
PNU	Pusan National University
ROKAF	73 Weather Group, Republic of Korea Air Force
SNU     Seoul National University
YU      Yonsei University

Represented U.S. Institutions

CAPS    Center for Analysis and Prediction of Storms (OU)
CIMMS   Cooperative Institute for Mesoscale Meteorological Studies (OU)
EMC	Environmental Modeling Center (NCEP)
FSU     Florida State University
LLNL	Lawrence Livermore National Laboratory
NCAR    National Center for Atmospheric Research
NCEP	National Center for Environmental Prediction
NCSU	North Carolina State University
NSSL    National Severe Storms Laboratory
OU      University of Oklahoma
SOM	School of Meteorology (OU)
SPC     Storm Prediction Center (NCEP)
PSU	Pennsylvania State University
UH	University of Hawaii
USAF    607 Weather Squadron, US Air Force
UW	University of Wisconsin

Represented Japanese Institutions

UT	University of Tsukuba