Combined and parallel use of MSG composite images and SAFNWC/MSG products at the Hungarian Meteorological Service Mária Putsay, Kornél Kolláth and Ildikó Szenyán Hungarian Meteorological Service H-1525 Budapest, P. O. Bo 38, Hungary Abstract At the Hungarian Meteorological Service (HMS) single channel data and different RGBs are visualized for the forecasters. We create the composite images recommended by EUMETSAT. The SAFNWC/MSG program package is also running operationally at HMS. We can visualize all the products and several of them are sent to the forecasters. The duty forecasters visualize the MSG data by the HAWK software (developed in Hungary) where they can handle the satellite data together with all available meteorological data. In many cases the use of the appropriate composite image gives a good tool for testing the SAFNWC products. RGBs and SAFNWC products together are a perfect tool to diagnose the actual situation, they complement each other, for eample convective storm RGB and Rapid Development Thunderstorm product, Air mass RGB and Air Mass Analysis product, Microphysical day or night RGB and Cloud Type product etc. Overlaying the vector type product on the appropriate RGB is very informative. INTRODUCTION At the Hungarian Meteorological Service 4 single channel METEOSAT-8 images (VIS0.6, HRV, IR10.8 and WV6.2) and 8 types of RGB composite images (recommended by EUMETSAT) are visualized for the forecasters. Name of RGB Channel combination Time interval resolution RGB_natural (NIR1.6,VIS0.8,VIS0.6) day 3km RGB_day_microphysical (VIS0.6,IR3.9,IR10.8) day 3km RGB_storm (WV6.2-WV7.3,IR3.9-IR10.8,NIR1.6-VIS0.6) day 3km RGB_night_microphysical (IR12.0-IR10.8, IR10.8-IR3.9, IR10.8) night 3km RGB_dust (IR12.0-IR10.8,IR10.8-IR8.7,IR10.8) day & night 3km RGB_airmass (WV6.2-WV7.3,IR9.7-IR10.8,WV6.2) day & night 3km RGB HRV_cloud (HRV,HRV,IR10.8) day 1km RGB HRV_fog (NIR1.6,HRV,HRV) day 1km Table 1: METEOSAT-8 RGB images sent operationally for duty forecasters The satellite images are visualised in the Hungarian Advanced Weather workstation (HAWK). This software was developed in Hungary for duty forecasters to visualize several types of meteorological data: NWP fields, SYNOP and automatic weather station measurements, radar data, satellite data, radiosonde data, lightning data etc. In this tool the user can handle all available meteorological data together creating loops, overlaying different types of data, etc. The Hungarian Meteorological Service is an early user of the SAFNWC/MSG program package as a former beta tester. We produce all SAFNWC/MSG products and we can visualize all products in the HAWK software. Some of the SAFNWC products are operationally sent for duty forecasters (see Table 2). The satellite images and products are used not only by visualisation but as input data into a program as well. The MEANDER program (Mesoscale Analysis, Nowcasting and Decision Routines) is an
automatic nowcasting system for helping the work of duty forecasters. The input data are radiosonde, SYNOP and automatic weather station measurements, mesoscale numerical model outputs, radar, satellite images, lightning data, etc. The satellite based input parameters are the SAFNWC/MSG cloud mask (CMa), cloud types (CT), cloud top height products (CTTH). The MEANDER program could use other SAFNWC products as well. Name of product yes no Cloud detection CMa Dust, volcanic plume CMa Cloud types CT Cloud top temperature, pressure and height CTTH Effective cloudiness CTTH Precipitating clouds PC, (probability of precipitation) Convective rain rate CRR Precipitable water TPW, LPW, SAI High resolution wind HRW Automatic Satellite Image Interpretation ASII-SAT, ASII-NWP, AMW-IR AMV-WV Rapid Developing Thunderstorms RDT Air Mass Analyses AMA Air mass classification Ridge Line,T gradient, WV dark stripe Table 2: SAFNWC products sent operationally for duty forecasters An automatic warning dispatch system is part of the MEANDER service, where the forecasters can determine the warning conditions. They also use the SAFNWC CT product (together with derived low visibility) for assessing warnings of potential night-time fog. EXPERIENCES Both satellite meteorologists and forecasters participated in the verification of the SAFNWC products. Case studies were investigated and for some products statistics were calculated. The RGB images are effective tools for verifying them by case studies. In this paper some eperiences are shared. Eamples are shown how the satellite image products are used in Hungary, what eperiences we have about the quality and usefulness of some of the SAFNWC/MSG products. CLOUD MASK, CLOUD TYPES, CLOUD TOP TEMPERATURE, PRESSURE AND HEIGHT These product groups are developed at Meteo-France in Lannion. Aviation meteorologists use regularly the SAFNWC Cloud top height product (CTTH) to estimate the cloud top height. Forecasters are interested in the cloud top height especially in convective situations. Our forecasters use the Cloud Types product intensively especially in the winter period to see the fog/low cloud regions. They use it in parallel with RGB images (RGB_night, RGB_natural, RGB_day_microphysical). In Fig 1 a nighttime case is shown. The fog/low cloud area can be well seen in both CT and the RGB images (fog/low cloud area is orange in CT, greenish, yellowish in RGB_night, while the cloud free area is green in CT and rose in RGB_night). In Fig 2 the fog/low cloud is covered by thin Cirrus cloud. The RGB image shows the low cloud under the Cirrus while the cloud types product shows only the thin Cirrus clouds.
Figure 1: SAFNWC Cloud Types product and the corresponding RGB_night_microphysical image taken on 20.10.2004. at 02:00UTC. Figure 2: SAFNWC Cloud Types product and the corresponding RGB_night_microphysical image taken on 20.10.2004. at 21:00UTC. The fog/low cloud is covered by thin Cirrus clouds. Figure 3: The SAFNWC Cloud Type categories RAPID DEVELOPING THUNDERSTORMS This product is developed at Meteo-France, Toulouse. The basic objectives of the product are the identification, tracking and monitoring of intense convective systems and the detection of rapidly developing convective cells. RDT indicate the cloud by a contour if it was detected as convective. Several useful parameters are retrieved for these clouds to characterise their development, such as cooling rate, area epanded rate, minimum temperature etc. Unfortunately the visualisation of the product is very difficult. The RDT product was verified by investigating several case studies, by overlaying it on different types of RGB images.
Problems were found in the convective cloud discrimination. Many problems are most probably caused by the fact that the algorithm uses only one channel data. RDT uses only IR10.8 channel data and as optional input lightning data. We have found that without lightning data the RDT does not detect all convective cells, thunderstorms, sometimes not even the very marked typical Cb, or only rather late in its life cycle. On the other hand several misdetections were found: thin Ci clouds, low level water clouds, fog, huge parts of cold front. In Fig 4 some eamples for the problematic cases are shown. Thin Ci are detected as convective clouds and severe storms are not detected by the RDT product. Figure 4: left: RGB_storm image taken on 19.04.2006. at 11:45UTC overlaid by SAFNWC RDT product. Red contour indicate the convective cloud detected by RDT. The severe storms over Bulgaria indicating by yellow color on the composite image are not detected but some thin Ci over Croatia was misdetected. Right: RGB_natural image taken on 28.09.2005. at 10:00UTC over France overlaid by SAFNWC RDT product. Cyan contours indicate the convective cloud detected by RDT. Thin Ci cloud is detected as convective cell. Comparison was performed between the RDT products of two users: Netherlands and Hungary. All input data were compared. The only differences were the region boundaries but there was a large overlap. Investigating only this overlap, differences were found between the outputs of the two users (see yellow ellipses). There are some clouds detected only by one of the users and the system directions are different. Figure 5: RGB_day_microphysical taken on 15.06.2005. at 16:45UTC overlaid by SAFNWC RDT product run in Hungary (left) and in Netherlands (right). The region boundaries are also marked (black lines). Black and white contours indicate the convective clouds detected by RDT. Clouds detected only by one of the users are indicated by yellow ellipses. The developers eplained that the differences are due to the different region boundaries because the discrimination algorithm depends on the life age of the cloud. However, if the boundaries can cause such differences, then perhaps we should use a much bigger region and trust only the inner part.
AIR MASS ANALYSIS This Air Mass Analysis (AMA) products of SAFNWC are developed at ZAMG, Austria. The goal of AMA products is to indicate areas with high probability of initiation of convection. It marks the dark stripes on the WV6.2 channel, the ridge line and the gradient zone of the equivalent-potential temperature at pressure level 850hPa. The gradient zone is the area on the north and the west side of the ridge line, close to it, with high gradient values and high IR10.8 brightness temperature values (large coherent areas of bright piels in IR10.8 are cleared, because they are likely to represent frontal cloud bands, and we are interested in clear areas where the future initiation of convective clouds are high). If the gradient zone is in a prefrontal region then there is a high probability of the initiation of convection. We have not performed yet an etended verification of this product. It is not so evident how to do it, because the high gradient value is only one of the conditions for initiation of convection. The usefulness of the product can be judged only by investigating case studies. Figure 6: RGB day and night microphysical images overlaid with the SAFNWC AMA product: ridge line (green solid line) gradient zone of the equivalent potential temperature at pressure level 850hPa (black dots). The black dots th indicate the area with high probability of initiation of convection. The images were taken on the 24 of May 2006 at 14:45, 17:45, 20:45, 23:45UTC. We demonstrate the usefulness of this product in a case study. On the 24 th of May 2006, in the afternoon and at night severe thunderstorms developed along a stationary warm conveyor belt, in the warm sector in a pre-frontal region. Close to the Hungarian border in Romania very intensive elevated deep convective cells could develop that evening. Microphysical RGB images in 3-hour steps are shown overlaid by the ridge line and the gradient zone of the equivalent-potential temperature at pressure level 850hPa (i.e. with the SAFNWC AMA products). In the 14:45UTC image the gradient
zone dots indicate high probability of initiation of convection in the pre-frontal region over Ukraine and East part of Romania. The images at 17:45,20:45 23:45UTC prove that severe thunderstorms really developed afterwards on the previously marked area. AUTOMATIC SATELLITE IMAGE INTERPRETATION To investigate the synoptic situation in the above case we include a traditional surface front analysis, the manual SETREP analysis and an RGB_airmass image overlaid with the SAFNWC Automatic Satellite Image Interpretation product. This product is developed at ZAMG, Austria. The goal of the product is to produce automatic weather analysis. (The overlaid ASII-NWP product was retrieved using both satellite and Numerical Weather Prediction (ECMWF)). Figure 7: Surface front analysis at 24.05.2006 00UTC (above left), the manual SETREP analysis for 25.05.2006 00UTC (above right) and RGB_airmass image taken at 25.05.2006 00UTC overlaid with the SAFNWC Automatic Satellite Image Interpretation product.
The Automatic Satellite Image Interpretation product detects the Conceptual Models and if any is found a corresponding character is overlaid on the satellite image. The legend is shown in Fig 8. Figure 8: The conceptual models in the SAFNWC Automatic Satellite Image Interpretation product (using both satellite and Numerical Weather Prediction data). On the airmass RGB the jet stream is well seen on the back edge of the cold front (brownish stripe). Related Conceptual Models are indicated alongside: front intensification by jet streak crossing (ji). Wave is indicated on the product (~) like on the manual SATREP analysis. The severe thunderstorms were detected as embedded Cb (dots). SUMMARY At the Hungarian Meteorological Service the forecasters can visualize single channel METEOSAT-8 images, RGB images and SAFNWC products. They can use them separately, in parallel, or combined, overlaying vector type products on satellite images. They use them the more and more often. Mainly the RGB images and the SAFNWC cloud type, Cloud Top Height products are used the more and more regularly. Satellite meteorologists and forecasters participate in the verification of the SAFNWC products qualitatively by investigating case studies. Some products are validated quantitatively as well. The Cloud Type, Cloud Top Temperature and Height and the Rapid Developing Thunderstorm and the precipitation products were thoughtfully verified. The verification of the Air Mass Analysis and Automatic Satellite Image Interpretation products have just started. In this paper our eperiences were discussed and some eamples for four SAFNWC products are shown. In conclusion the RGB images are very effective tools in the investigation of case studies for verification purposes. ACKNOWLEDGEMENT This study was partially supported by the Hungarian Space Office (TP250). REFERENCES Kerkmann J., Lutz H. J., König M., Prieto J., Pylkko P., Roesli H. P., Rosenfeld D., Zwatz-Meise V. And Schmetz J. (2004) MSG channels Interpretation Guide, weather, surface conditions and atmospheric constituents, on CD and html://www.eumetsat.int. Kertész S. (2000) The HAWK system: Recent developments at HMS. Proceedings of the 11 th EGOWS meeting held in Helsinki, 5-8 June, 2000. pp:13-14. SAFNWC/MSG program package documentation, (2003), http://nwcsaf.inm.es.
Putsay M., Diószeghy M. and Rajnai M. (2004a) Preparations and First Results of MSG Data Processing at the Hungarian Meteorological Service, Eperiences of the Beta Tests of SAFNWC/MSG Program, Proceedings of The 2003 EUMETSAT Meteorological Satellite Data Users' Conference, Weimar, Germany, 29 September 03 October 2003, pp:286-290. M. Putsay, Rajnai M., Diószeghy M., Kerényi J, Szenyán I.G. and Kertész S. (2004b) Interpretation of MSG Images, Products and SAFNWC Outputs for Duty Forecasters, Proceedings of The 2004 EUMETSAT Meteorological Satellite Data Users' Conference, Praga, pp:162-168. http://www.eumetsat.de/en/area2/proceedings/eump41/2004docs/2_24_rajnai_p.pdf M. Putsay and Diószeghy M (2004): Report on Validation of MSG Precipitating Clouds Product with Hungarian Radar Data and Data Echange for Tuning of the PC Product, Nowcasting-SAF, Visiting Scientist Activity. http://nwcsaf.inm.es/vsa/hms_report_2004.pdf M. Putsay and Diószeghy M. (2005) Report on Validation of NWCSAF/MSG Precipitation Products with Hungarian Radar and Tipping Bucket Rain Gauge Data. Nowcasting-SAF, Visiting Scientist Activity. http://nwcsaf.inm.es/vsa/safnwc_vs_2005_uj.pdf M. Putsay, Szenyán I.G and Diószeghy M (2005) Validation of SAFNWC/MSG precipitation products with Hungarian Radar and surface measured rain gauge data, Proceedings of The 2005 EUMETSAT Meteorological Satellite Data Users' Conference, Dubrovnik, pp:193-200. http://www.eumetsat.int/groups/cps/documents/document/pdf_conf_p46_s2_18_putsay_p.pdf Putsay M. (2005) Eperiences on the interpretation and some test results of the SAFNWC/MSG program package products at the Hungarian Meteorological Service, Product Assessment Review Workshop Madrid October 2005, http://nwcsaf.inm.es/parwshop/presentation_par_oms_putsay_hungary.pdf