Description and substantiation

Present routine numerical weather prediction (NWP) models with the horizontal resolution down to 10 km have contributed to significant improvement of the general short-range weather forecast. Bearing in mind this improvement the quality of present day forecasts of local weather a few hours ahead is rather disappointing (see e.,g. Browning, 1998). The most challenging task to the research involves forecasting of the development of precipitating convective clouds. Such clouds can be accompanied by various hazardous phenomena. In the central Europe the most important is the occurrence of local heavy convective rains that can cause flash floods. Another dangerous phenomena are the extreme wind speed or wind shear. The wind events like downbursts can evolve at cold evaporative outflow from cloud circulations. Rare but heavy damaging are tornadic vorticies. Such phenomena can be followed not only by material damage but they are dangerous for humans. Therefore the improvement of the quality of very short range forecasting severe convective events is a topics of intensive present research.

Satellite observation and radar measurement play an important role in analysing the cloud and precipitation fields. Especially there is an extensive potential of utilisation of the radar measurement in local very short range precipitation forecast (Hardaker, 1998). Radar data can be used to verify model performance or they can be assimilated into NWP models to help in determining the initial moisture fields. Model information about development and movement of the precipitation can prolong the time of radar-based forecast over the nowcast range (Golding, 1998). It has to be stressed that all types of synthesizing radar measurements and NWP model results are in the stage of intensive investigation.

Various case studies of convective events showed that the horizontal resolution of routine NWP models as well as corresponding cumulus parameterization are not sufficient to analyze heavy convective precipitation (from European studies see e.g. Senesi et al., 1996; Doswell et al.,1998) as well as larger scale precipitation events (e.g. Buzzi et al.,1998). It is also not easy to compare radar results with routine NWP model because of different horizontal resolution (see e.g. Keil,Volkert,1998). It can be expected that by using smaller grid spacing the deep moist convection as well as the associated feedback mechanism can be resolved more explicitly. It is also easier to compare the radar indicated precipitation with the model results. The smaller scale model can be used as dynamical postprocessing tool when an evolution of heavy precipitation is expected.

Numerous heavy convective rains have been recorded in the CR. The flash flood from July 1998 with precipitation amount of 204mm accumulated in about 10 hours is the extreme example. Radar measurement showed that precipitation developed in a multicell complex of deep convective clouds (Hančárová et al., 1999). This event belongs to the set of 26 daily sums greater than 200mm that were recorded in the Czech territory in 1897-1998 (Kakos, Řezáčová, 1999). The other heavy convective events that occurred in CR in 90´s were documented by radar and ground measurements.

Meteorological data from local extreme precipitation in July 1998 (Hančárová a kol., 1999) as well as from large scale precipitation in July 1997, which caused a large flood in eastern and northeastern CR (Květoň et al.,1997; ČHMÚ,1998), showed a good quality of short range routine forecasts. However, the analysis of forecasting skill also confirmed that the local model forecast was far from being satisfactory (Odstrčilová, 1998). At present the radar and satellite technologies available in CR have a good European level, which is steadily improved. International collaboration is a good assisstance for the improvement of technology utilisation. It is a good potential for analyses of convective phenomena proposed in this project.

The project aims at the study of convective events by remote measurement and mathematical modelling. Historical events, which were recorded over the Czech territory in 90's , will be analyzed together with events, which will occur during the project period. Radar measurement and satellite information about cloud tops will be used to detect convective clouds producing heavy local rains (parts II and III). The same situations will be analyzed by local NWP model with horizontal resolution of 1-3km (see part I), which corresponds to the dimension of radar pixel 2x2km. This will enable the comparison of model with radar-based precipitation as well as the comparison of radar measured and model simulated radar reflectivity structures.

The "in situ" verification of impacts of extreme convective events has been included in the project (see part IV). We will concentrate on the impacts of strong wind events and the fallout of large hail. These are rare extreme phenomena hardly possible to be detected by professional networks of ground observations. They were only sporadically recorded over the Czech territory and were proved only by "in situ" information (Setvák et al.,1996; Sulan et al.,1998).

The radar data utilization in precipitation forecasting links the proposed project to the new COST717 action. The part of proposer group is going to apply for the participation in COST717 with project aimed at improvement of radar-based estimation of precipitation and the use of radar data in hydrological modelling. This topic is not included in the proposed project. However, we suppose that some of algorithms developed in the international collaboration will be used in the proposed analyses of convective events over the territory of the Czech Republic.

The main goals of the project consist of:

. developing of techniques to detect the evolution of deep convective clouds based on radar and satellite measurements; the techniques will be applicable in warning service of the Czech Hydrometeorological Institute (CHMI).

. validating of the usefulness of the local model as a tool of dynamic postprocessing (the model is intended to be applied in the case of severe convection detection)

. documenting the local damages caused by severe convection, especially wind damages not detectable by standard ground networks

The research was divided in the blocks I-IV. The inidividual topics, the schedule, and the participation of proposers are described in the following sections that correspond to the four blocks.

Present stage: Physics of convective phenomena and forecasting of their occurrence have been the topic of investigation in IAP ASCR for several years. Simple convective cloud model and simple NWP model prepared in IAP ASCR were used (Řezáčová,Motl,1992; Řezáčová,Sokol,1995; Řezáčová et al.,1995; Sokol et al.1995; Řezáčová et al. 1996). In the last period the use of statistical posprocessing NWP model output has been investigated by using EM/DM NWP model (Sokol, Řezáčová, 1998; Sokol, Řezáčová, 1999). In those studies output of the EM/DM model has been used. The last studies have been devoted to the output of the model ALADIN-CZ. In the joint project supported by Swiss NSF the iprovement of quantitative radar-based estimation of precipitation was investigated (Řezáčová et al. 1997, 1998).

Project goals and proposed methods: This part of the project aims at using NWP model to analyse severe convective events with horizontal resolution of 1-3 km. The investigation should answer the question whether the application of the model can contribute to the forecast of severe convection over the Czech territory several hours ahead. At present, there is not any model of that type available in the Czech Republic. However, a few models are obtainable for research from abroad at no costs. Therefore the most effective way consists in adaptation of one of foreign models to the Czech data. Among suitable candidates belong the U.S. models MM5 (e.g. Grell et al.,1994) and ARPS (e.g. Xue et al.,1995) and the German lokal model LM DWD (Saito et al., 1998). The last was developed on the basis of MM5. All models mentioned are nonhydrostatic. There is good experience with German model package EM/DM workstation version in IAP as well as with consultative assistance of DWD experts. The EM/DM was used to investigate statistical adaptation of short-range precipitation forecasting for Czech territory (Sokol,Řezáčová, 1998; Sokol,Řezáčová, 1999). Therefore, we propose application of LM DWD model. Based on consultation with partners from German Weather Service the LM DWD model version can be available for IAP this year.

Input data will be prepared by using EM-DM model. The other potential source of input fields is the NWP model ALADIN-CZ routinely run at CHMI. The basic prognostic fileds are archived. However to use the ALADIN outputs the more comprehensive version of prognostic fields will be necessary. Ground data from Czech synoptic stations will be implemented into initial fields to test their effect on the model result. Radar data and ground data will verify the model precipitation. The agreement of model precipitation with radar result will be investigated by studies of model sensitivity to initial fields. Such studies are of high importance because of model sensitivity to small changes in initial conditions (Brooks et al.,1993). Especially the effect of orography modifications and of the changes of initial wind shear conditions in lower levels will be investigated during sensitivity studies. We have not met the ensemble approach to the prediction by smaller-scale models so far. In this approach a few variants of initial conditions based on a disturbance of the basic state will be used. We intend to perform experiments devoted to such technique in the last year of the project.

2000: LM model implementation at IAP, first case studies. Comparison of model precipitation with radar-based precipitation. Computation of simulated reflectivity fields.

2001: Sensitivity of model precipitation to initial conditions (inclusion of detailed ground data, changes of model orography and initial wind field structure).

2002: Investigation of ensemble treatment, tests with ALADIN-CZ model.

Researchers: Daniela Řezáčová (selection and discussion of cases to be studied, meteorological analysis of model results, influence of vertical wind shear on cloud circulation, coordinating work), Zbyněk Sokol (installation of LM DWD in IAP, preparation of input fields from EM/DM model, collaboration on studies of sensitivity to initial conditions), Petr Pešice (collaboration on the LM installation and model computation, graphic output preparation, simulation of the influence of orography on precipitation fields), Ondřej Fišer (verification of model precipitation with radar data. Computation of simulated radar reflectivity fields), Karel Dejmal (evaluation of case studies, comparison with radar reflectivity fields), Marek Kašpar (evaluation of case studies, effect of wind field on cloud development, assessment of model vertical velocities).

Present stage: Weather radar is the basic tool for operational detection and monitoring of convective storms (Šálek, 1994, Šálek et al., 1997, Sulan et al., 1998). Detection of the potentially intense (thus dangerous) convective storms belongs to important tasks of present radar meteorology and its solution will significantly improve the nowcasting of convective storms.

Basic product of the weather radar is a spatial distribution of radar reflectivity Z, which is proportional to intensity of precipitation (or size and density of cloud particles). These data are an operational output of the Czech weather radar network, operated by CHMI, that covers whole territory of the Czech republic and near surroundings (Kráčmar et al., 1999, www: CHMI Radar Department). Doppler data of radial velocities, suitable for the detection of internal flow structure inside the convective storms, are available preoperationally (Novák, 1998).

Project goals and proposed methods: The main goal of this sub-project will be mainly the implementation of methods, used by National Weather Services in developed countries, in the condition of the Central Europe. Further on, testing and enhancement of these methods for their operatinal usage in the Czech Republic.

We suppose the use of the operational radar reflectivity data for detection of presence of intense convective storms (real-time and ex-post) and for its comparison with other data sources. Especially, detailed study of 3D-structure of convective storms, based on volume data of radar reflectivity and Doppler velocity, will be carried out in selected cases. Knowledge about storms structure (conceptual models: multicell, supercell, squall line, mesoscale convective system) will be connected with the characteristics of convective environment. Work will be focused in appearance of dangerous phenomena connected with convection: intense precipitation causing flash floods, hail-storms, damages caused by wind (gust-fronts, tornadoes, downbursts) and lightning. The existing detection algorithms, based on radar reflectivity data (Verticaly Integrated Liquid, Y-algoritmus) and Doppler velocities (Shear Wind Products, Mesocyclone Detection, Tornado Vortex Signature), will be tested to assess their possible future applicability in nowcasting. Eventually, algorithms will be modified for the conditions in the Central Europe

Using these data sources, a preliminary climatology of intense convection characteristics will be processed and guidelines for operational detection and nowcasting of intense convective phenomena with exploitation of weather radar data will be proposed.

- 2000: To carry out detailed studies of 3D-structure of convective storms using volume radar reflectivity data for archive cases.

- 2000 - 2001: To study and test the Doppler radar velocities algorithms

- 2000 - 2002: Analysis of operational radar data and finding of potentially intense convective storms. For selected cases, to carry out detailed studies of 3D-structure of convective storms using volume radar reflectivity data and Doppler velocities.

- 2001 - 2002: To test warning algorithms in preliminary operation.

- 2002: Elaboration of a preliminary climatology of intense convection characteristics and guidelines for operational detection and nowcasting of intense convective phenomena with exploitation of weather radar data. Publication of results.

Researchers: Petr Novák, Jan Kráčmar

Present stage: Weather satellite data represent another remote-sensing type of information on convective storms. Though storm severity can be inferred from storm cloud top appearance in various spectral bands, these "severity features" and their link to storm internal structure still remain in research rather than applied science levels.

Recent weather satellites observe cloud tops of convective storms in three basic spectral bands. In visible and near infrared bands (approx. 0.5 - 1.1 ?m) the information about storm cloud top morphology is available due to shadows, cast by cloud tops at surrounding cloud layers. The imagery in thermal bands (approx. 10 - 12.5 ?m) provides us with information on cloud top thermal field. Microphysical composition of storm cloud tops can be retrieved from the 3.7 ?m band during daytime hours (Setvak et al, 1996).

The most vigorous form of convective storms - supercell storm or supercell - exhibits in satellite imagery certain features, which seem to be linked very closely to this form of storms. In thermal bands it is a pair structure, described as "cold-U" and "embedded warm spot" (McCann, 1983). In visible and near infrared, as well as in the 3.7 ?m band, it is a structure named after its appearance as a "plume" above storm cloud top (Setvak and Doswell, 1991; Setvak and Levizzani, 1996). However, it is not quite clear whether all of the supercells do exhibit these features, in which stage of storm life cycle do these develop and what is the mechanism that builds them up.

Project goals and proposed methods: This part of the project will be performed in close relation to the radar part of it (Part II) and to the "ex post" diagnostics of weather, produced by storms over the Czech Republic. (Part IV). Storm cloud top structure will be studied in all available spectral bands, links between satellite-observed cloud top features, storm internal processes (as determined by radar observations) and accompanying weather will be examined. Similar research is being performed for storms over the U.S. region; the purpose of the proposed project is to achieve characteristics of storms, developing over the Central Europe region - thus under somewhat different climatic conditions. Results obtained for Central European storms could be compared to those from North America and thus could lead to generalization the conceptual models of convective storms. These steps are crucial for utilization of satellite data in nowcasting of convective storms.

As a part of this research, a subproject will be performed aimed at the 1.6 ?m spectral band and its potential for severe storm detection. First weather satellite, providing operational data in this band, is the NOAA-15 (NOAA-K) spacecraft; other polar orbiting satellites (NOAA-L,M) with this spectral channel will follow-up in next years. First geostationary satellite, imaging in the 1.6 ?m spectral band, will be Meteosat Second Generation (MSG), which is scheduled for launch in the second half of year 2000. Therefore, this subproject will be carried out after operational data from MSG are available. Appearance of cloud tops in the 1.6 ?m band and its development will be compared to storm structure in all the other available spectral bands of MSG and NOAA-KLM satellites.

All of the satellite data, necessary for the project, are available at the joint applicant's premises (CHMI, Libus). Data from MSG should be available (according to development plans of CHMI) from the first half of 2001. Joint applicant and his collaborators are experienced in the field of satellite data processing and interpretation, thus good and meaningful results can be anticipated.

- Continuously: during the warm season (May - September) of 2000 - 2002, analysis of appearance and structure of cloud tops of all convective storms, which will occur on territory of the Czech Republic;

- Selection of those cases, where storm on satellite imagery exhibits either one of the above mentioned "supercell features", or significant increase of the 1.6 or 3.7 ?m spectral band reflectivity. For these cases examination of storm internal structure (from radar data, Part II) and diagnostics of the accompanying weather (Part IV).

- For the storms where radar depicts supercell nature of these or storms accompanied by severe weather (Part IV), however with no significant satellite "supercell signatures", detailed analysis of their cloud top structure, especially in the 1.6 or 3.7 ?m spectral bands. Goal of this is to seek for presence of some so far unnoticed (or overlooked) cloud top features, linked to supercells.

- Progress of the research will be presented at the EUMETSAT Data Users' Conferences and on meetings (both, European as well as the U.S. ones), aimed at severe convective storms and their accompanying weather. In the end of the project, publication in one of the internationally referenced journals.

Researchers: Martin Setvák, Pavel Hampl, Petr Pešice

Present stage: Remote sensing methods (radar and satellite observations), proposed to detect possibly severe convective storms, need to be tested and verified by diagnostics of the real weather, accompanying storms. Besides that, in order to refine statistics of severe convective events at territory of the Czech Republic, much closer attention must be devoted to these vigorous weather phenomena. To achieve both, all the storms at this territory, either remotely detected as potentially severe or reported as being severe by any other source, will be studied and documented in much as possible detail. Since flash-flood storms are subject of interest to the Part I of this proposal, this part (Part IV) addresses mainly strong damaging wind events (tornado, downburst, microburst) and large hail occurrence. Realistic information on frequency of these at the Czech Republic and knowledge of their local characteristics are crucial for local nowcasting.

Czech meteorologists derive most of information on significant weather, produced by severe convective storms, from two main sources. More or less realistic information can be acquired from synoptic weather stations. Unfortunately, due to limited number of these stations and rather small area covered by their observations, and taking into account typical horizontal dimension of the phenomena, it seems very likely that most of the severe convective events remain unnoticed in the official weather reports. The other information source are reports in mass media. Typically, severe weather event is reported in the news only if a stray witness informs the media and if the event appears "sufficiently interesting" to editor-in-chief. Another drawback stems from the fact that most of our public has only a very low notion of these phenomena and is unable to describe sufficiently the witnessed event. Therefore we assume that significant part of severe convective phenomena remains unnoticed by public, as well as by meteorologists. Moreover - if noticed at all - majority of these events is poorly documented. For these reasons, the frequency of severe convective events over the Czech Republic, as well as their local characteristics, remain uncertain.

Though severe convective storms (and their accompanying severe destructive weather) occur significantly less frequently in Europe than over North America (Church et al., 1993), more attention of meteorological community is being devoted to them even here (in Europe) recently than was before. In February 2000 will be held Conference on European Tornadoes and Severe Storms, which will address some of the topics described in this project proposal. We expect that international projects aimed at European severe storms will be proposed (or agreed-on) there, thus presence of Czech specialists at the conference is highly desirable.

Project goals and proposed methods: Whether a storm, detected and evaluated by remote sensing means as "possibly severe", has reached the severe stage or not can be verified only on basis of "ground truth data" - direct observation of the storm during the event ("storm chasing"), diagnosis of the accompanying weather from witness reports and/or from investigation of damage caused by the storm. Direct (real-time) observation of storms is a very difficult task, which in the area of Czech Republic is even further complicated by complex orography of the region, as well as by "chaotic" road network (compared e.g. to the U.S. Great plains). Therefore, the project will rely mostly on "ex post" diagnostics of storm behavior rather than on direct observations (though we do not exclude these under certain favorable conditions).

Acquisition of witness reports and analysis of occurred damage will be performed as soon as possible after the event (if possible, up to 48 hours - any additional delay increases "information distortion"). Primary information will be collected by phone (local city headquarters, fire department, police, forest management, ...). Based on this information, a decisions will be made about sending a crew of specialist to the area. Several tornadoes over the Czech Republic have been documented this way during the last decade (Šálek, 1994; Setvák et al, 1996; Sulan et al, 1998; Setvák, 1998), as well as cases of disastrous hail (Šálek, 1998) - authors of these studies are among the project researchers. This proves that researchers of the proposed project already do have certain (though limited) experience with this topic. Here we expect spending of the project finances mainly to cover (domestic) traveling expenses, for purchase of films and its processing, and purchase of GPS station and GPS-compatible digital maps (for accurate documentation of the location and extend of the damage).

Besides the ground-based damage diagnostics, research by means of aerial photography of the damage swath will be emphasized. By aerial observations it is easier to distinguish between damage caused by tornadoes from damage caused by downbursts/microbursts (especially in grain fields and forests). Forestry specialists, aimed at wind tree damage, will be contacted and offered collaboration. Here we expect spendings on aerial observations and photography, and purchase of scanner for transparent films. All obtained documentation will be archived on CD-ROM's in digital format.

- Continuously: 2000 - 2002, during the convective period of these years (May - September), at once after occurrence of severe storms or potentially severe storms their diagnosis and documentation by all available means;

- Consequently processing of all retrieved information and data;

- Simultaneously: cross-correlation of the obtained information with results of Parts II and III;

- Results will be continuously presented at conferences or international meetings, aimed at severe convective storms and their accompanying weather, and in the end of the project published in one of the referenced journals.

Researchers: Martin Setvák, Petr Novák, Milan Šálek, Jan Sulan, Ondřej Fišer

Buzzi,A.,Tartaglione,N.,Malguzzi,P.(1998): Numerical simulations of the 1994 Piedmont flood: Role of Orography and moist processes. Month.Wea.Rev.126,9,2369-2383.

Brooks,H.E., Doswell,C.A.III,Wicker,L.J.(1993): STORMTIPE: A forecasting experiment using a three-dimensional cloud model. Wea.Forecasting, 9, 352-362.

Browning,K.A.(1998): Use of radar in mesoscale analysis. In "COST75 Advanced Weather Radar Systems. COST75 Final Seminar, Locarno, Switzerland, 23 to 27 March 1998, 239-256.

CHMI Radar Department home-page - http://www.chmi.cz/meteo/rad/main_eng.html

CHMI Radar Department References - http://www.chmi.cz/meteo/rad/erad_lit.html

ČHMÚ (1998): Vyhodnocení povodňové situace v červenci 1997. Zpráva pro MŽP ČR.

ČHMÚ - Tornáda na území ČR. Internetové stránky ČHMÚ - http://www.chmi.cz/meteo/sat/torn/tor_main.html 

Church C. et al., Editors, 1993: The Tornado: Its Structure, Dynamics, Prediction, and Hazards. Geophys. Monogr. 79, Amer. Geophys. Union.

Doswell C.A., D. Zrnic, M.Setvak, V.Levizzani, P.Hampl, 1997: Multispectral Observations of Storm Tops,. Internetové stránky ČHMÚ, družicové oddělení. - http://www.chmi.cz/meteo/sat/most/most.html 

Doswell,Ch.A.,Ramis,C.,Romero,R.,Alonso,S.(1998): A diagnostic study of three heavy precipitation episodes in the western Mediterranean Region. Wea.Forecast.,13,1,102-124.

Golding,B.,V.(1998): Nimrod: A system for generating automated very short range forecasts. Meteorol.Appl.,5, 1-16.

Grell,G.,Dudhia,J.,Stauffer,D.(1994): A description of the Fifth-Generation of the Penn State/NCAR Mesoscale Model (MM5). National Center for Atmospheric Research, 138pp.

Hančárová,E.,Metelka,L.,Mrkvica,Z.,Pavlík,J.,Pozler,R.,Šiftař,Z.,Váchal,P.,Veselý,R. (1999): Catastrophic flood in foothils of Orlické Hory mountains on July 22-25.1998 (in Czech). Meteorol.Zprávy, 52,1,1-12.

Hardaker,P.(1998): Weather radar data and numerical wether prediction models. In "COST75 Advanced Weather Radar Systems. COST75 Final Seminar, Locarno, Switzerland, 23 to 27 March 1998, 451-459.

Kakos,V.,Řezáčová,D.(1999): Flash flood rains as viewed by meteorologists (in Czech). In"Sborník referátů konference Orlice", 25-26.3.1999, v tisku.

Keil,Ch.,Volkert,H.(1998): Radar composite and mesoscale precipitation forecasts; Determination of model deficiences through case studies. In "COST75 Advanced Weather Radar Systems. COST75 Final Seminar, Locarno, Switzerland, 23 to 27 March 1998, 407-417.

Kráčmar, J., Havránek, P., 1995: First results of data processing of digitized MRL-5 weather radar in Prague, In: International Seminar of COST-75: "Advanced Weather Radar Systems" - Brussels, 20.-23.9.1994, European Commission, Luxembourg, 65-72.

Kráčmar J., Joss J., Novák P., Havránek P., Šálek M., 1999: First Steps Towards Quantitative Usage of Data from Weather Radar Network, In: Final Seminar of COST-75: "Advanced Weather Radar Systems" - Locarno, 23.-27. 3. 1998, European Commission, Luxembourg, 91-101. - http://www.chmi.cz/meteo/rad/pub/cost75_fs/

Květoň,V.,Srněnský,R.,Veselý,R.(1997): Precipitation distribution during the July 1997 floods (in Czech). Meteorol.Zprávy, 50,6.

McCann D.W., 1983: The enhanced-V, a satellite observable severe storm signature. Mon.Wea.Rew., 111, 887-894.

Novák P., Kráčmar J., 1997: Vertikální profily radiolokační odrazivosti a jejich vliv na určení radiolokačních intenzit srážek, In: Seminář ČMeS a ČHMÚ "Stoleté výročí extrémních atmosferických srážek", Praha, 186-193. - http://www.chmi.cz/meteo/rad/pub/vo/

Novak P., Kráčmar J., 1998: Radar Skalky Vertical Profile "Climatology", project RadHyd contribution, Praha - http://www.chmi.cz/meteo/rad/pub/rh_mvp/

Novak P., 1998: Dopplerovská radiolokační měření radiálních rychlostí, seminární práce MFF UK, Praha, 22 s. - http://www.chmi.cz/meteo/rad/pub/sem_dop/

Odstrčilová,M.,Šedivka,J.,Šopko,F.(1998): Comparison of precipitation forecasts according to numerical models in the period of floods in July 1997. Meteorol.Zprávy,51,1,10-17.

Řezáčová,D.,V.Motl, 1990: The use of the simple 1D steady-state convective cloud model in the decision tree for determining the probability of thunderstorm occurence., Studia Geoph.et Geod.,34, 147-166.

Řezáčová D.,Z.Sokol, 1995: Skill of 6hr and 12hr thunderstorm forecasts based on assesment of larger-scale information. Studia geoph. et geod.39, 60-73.

Řezáčová D.,Z.Sokol,P.Pešice, 1996: Categorical forecasts of convective events using upper air predictors. Meteorological Applications, 3, 19-30.

Řezáčová,D., Z.Sokol, J.Kráčmar, P.Havránek, 1997: Comparison of MRL5 radar-based precipitation amount with precipitation amount reported by ground stations., Annalen der Meteorologie, Proc.3rd ECAM, 293-295.

Řezáčová, D., Sokol, Z., Kráčmar, J., Havránek, P., Joss, J., 1998: Statistical Correction of Radar Based Precipitation Amount. COST-75 Final International Seminar "Advanced Weather Radar Systems",

Senesi,S.,Bougeault,P.,Cheze,J.L.,Cosentino,P.,Thepenier,R.M.,(1996): The Vaison-La-Romaine flash flood: Mesoscale analysis and predictability issues. Wea.Forecast.11,4,417-442.

Setvák M., C.A.Doswell III., 1991: The AVHRR Channel 3 Cloud Top Reflectivity of Convective Storms. Mon.Wea.Rev., 119, 842-847.

Setvák M. and V.Levizzani, 1996: Multispectral, High-Resolution Satellite Observations of Plumes on Top of Convective Strorms. J. Atmos.Sci., 53, 361-369.

Setvák M., R.M:Rabin, V.Levizzani and C.A.Doswell, 1996: Relations between increased 3.7/3.9 ? m reflectivity above U.S. Great Plains thunderstorms from NOAA/AVHRR and GOES-8 and internal storm structure from NEXRAD radar. Proc. 1996 Meteorological Satellite Data Users' Conference, Vienna, Austria, EUM P 19, ISSN 1011-3932, 143-150.

Setvák M., D.Žídek a M.Hradil, 1996: Tornáda na severovýchodní Moravě a Slezsku 8. července 1996? Meteor. zprávy, 49, 143-146.

Setvák, M., 1998: Tornádo v polesí Teplá 21.-22. července 1998. Internetové stránky ČHMÚ, - http://www.chmi.cz/meteo/sat/torn/cases/19980721/tepla98.html 

Sulan J., M. Setvák, P. Novák, 1998: Silná konvektivní bouře v Čechách v noci z 27.-28.6.1997, Meteor. zprávy, 51, 118-125. - http://www.chmi.cz/meteo/sat/torn/cases/19970627/rokyc_97.html

Sokol Z., D.Řezáčová, P.Pešice, 1996: The inclusion of surface synoptic data in very short range predictions of convective events, Studia geoph. et geod., 40, 178-186.

Sokol Z., Řezáčová,D.(1998): Local precipitation forecasting based on the statistical postprocessing of NWP model outputs (in Czech). Meteorol.Zprávy,51,4,97-105.

Sokol Z., Řezáčová,D.(1999): Improvement of local categorical precipitation forecasts from an NWP model by various statistical postprocessing methods, Studia geoph. et geod. 44, p 38 - 56.

Saito K.,Doms, G.,Schaettler,U.,Steppeler,J. (1998): 3-D Mountain vawes by the Lokal-Modell of DWD and the MRI Mesoscale Nonhydrostatic Model, Meteorology and Geophysics, 49, p7-19.

Šálek M., 1994: Silné bouřky na Moravě spojené s výskytem tromby v obci Lanžhot dne 26.5.1994. Meteor. zprávy, 47, 172-177.

Šálek, M., 1998: Silné krupobití na okrese Žďár nad Sázavou ze dne 30.6. 1997. Meteor. Zprávy, 51, 73-80.

Šálek M., Kráčmar J., 1997: Výpočet srážek z meteorologického radiolokátoru Skalky. Meteor. Zprávy, 50, 99-109.

Šálek M., Kráčmar J., Novák P., Setvák M. 1997: Využití distančních metod při povodňových událostech v červenci 1997. Meteor. Zprávy, 50, 177-178, - http://www.chmi.cz/meteo/rad/pub/pov97_mz

Xue, M., Droegmeier, K.K., Wong, V., Shapiro, A.,Brewster,K. (1995): ARPS Version 4.0 User's Guide.