THE OPERATIONAL WEATHER RADAR NETWORK IN EUROPE [Bulletin of the American Meteorological Society]
(Bulletin of the American Meteorological Society Via Acquire Media NewsEdge) The European Operational Program for Exchange of Weather Radar Information (OPER A) has worked to improve harmonization of radar systems and measurements since 1999 and has recently star ted production of network-wide radar mosaics.
W eather crosses national borders, and hence the exchange of weather observations is most important. The World Meteorological Organization (WMO) has advanced the exchange since its founding in the 1870s. Europe belongs to the WMO regional area VI, and in this area 30 meteorological services are members of the Euro- pean National Meteorological Services Network (EUMETNET). EUMETNET members come from countries within the European Union and Balkans, mostly lying west of 30°E. In the following, we refer to this area as Europe, noting that Europe as a geo- graphical continent is larger.
EUMETNET is active in observations, forecasting, and climate issues. In Europe, the national weather services take care of surface observations, radio soundings, weather radar observations, and alike within their territory. EUMETNET has responsi- bility of observations that are of benefit to weather forecasting in general, such as observations made in airplanes, radio soundings on the Atlantic and remote islands, floating and moored buoys, and observations on board of ships. These observations are carried out by EUMETNET members or by, for example, f light companies on behalf of EUMETNET. The weather radar work within EUMETNET is coordinated by the Operational Program for Exchange of Weather Radar Information (OPERA; www.eumetnet.eu/opera), which was established in 1999. See Huuskonen et al. (2010) for the background of OPERA and its relation to other European activities related to weather radars.
OPERA has been dealing with all aspects of weath- er radars since its founding. The activities include making recommendations on radar hardware as well as software issues, making recommendations on the best practices of doing weather radar measurements, discussing the advances of the meteorological radar science and their relevance to operational radar work, specifying data models for the exchange of weather radar data and products, and providing software for encoding and decoding of these products. Within OPERA, an operational data center, Odyssey, is now operated that collects radar volumetric data from 21 OPERA members and produces composites of radar ref lectivity and rain rate in near-real time. The aim of this paper is to explain and promote the activities of OPERA.
The OPERA members fund the activities through EUMETNET and one of the members is chosen to coordinate the work for a predefined period. The present coordinating member is the Finnish Meteorological Institute (2013-17). The coordinating member and the project manager1 have the overall responsibility in the delivery of the agreed goals. The budget covered by the members is about 300,000euro annually; the members also use their own funds and add a significant in-kind contribution to the benefit of the work. In practice, the work is carried out by a consortium of several members and is supported by an active expert team of members, which meets twice a year to give advice on the project work. Simultaneously, these meetings form a regular meet- ing of the weather radars experts of the European meteorological services, in which experiences of running and building weather radars and networks are exchanged on a regular basis. Some 30-40 radar experts from around 20 members participate typically to the meetings. The team members work in a wide range of positions in their organizations, as scien- tists, software and hardware engineers, operators, forecasters, and network managers. The membership also guarantees a live and continuous connection to the ongoing technical and scientific radar work so that the new results can be the operational work.
T H E E U R O P E A N WEATHER RADAR NETWORK. During the last 50 years the European weather radar network has been developed via a bottom-up approach. At first instance, the national meteorological services of the European countries have independently built their own operational radar networks and in the 1990s the exchange of experi- ences, the harmonization, and the data exchange were started. As a result, the European radar network is extremely heteroge- neous in installation date, manufacturers, scanning strategy, signal process- ing, and product genera- tion. In this respect, the European radar network is fundamentally different from the Next Generation Weather Radar (NEXRAD) network (Crum and Alberty 1993; Serafin and Wilson 2000). For run- ning a continental network, this imposes challenges but also provides a wealth of knowledge on different techniques and algorithms. Because of the increased collaboration and information exchange between the nations, the hardware, processing, software, and operation of the radar network are gradually converg- ing, but differences will continue to exist as the radar networks continue to be owned and operated by the national services.
Figure 1 shows a map of the current operational weather radar network in Europe. In October 2013 the network consisted of 202 operational radars of which 184 had Doppler capability. The number of dual-polarization radars in Europe is growing steadily and the current count is 48. The major- ity of the radars operate at C band (168 sites), but a number of S-band radars are installed in the south of Europe (33 sites). The most northern radar is located in Hasvik (Norway) at 70.6°N, and the most southern one is on Gran Canaria (Spanish island off the African coast, southwest of Europe; not shown on the map) at 28.0°N. The most western radar is located in Keflavik (Iceland) at 22.6°W, and the most eastern one in Berlevåg (Norway) at 29.0°E.
Although the location of radar sites is primarily a national responsibility, it is evident from the radar map that the sites are distributed rather evenly, also across the national borders. The median distance between two neighboring radars in the network is 128 km, and the smallest distance between a pair is just 7 km (a C-band and an S-band radar in Oradea, Romania). The smallest distance across a national border is 50 km (radars in Vienna and Maly Javornik close to Bratislava), but most distances across the borders are similar to those within the national networks. The density of the network is beneficial for estimation of the surface rain rate. For most of the network area, the distance to the closest radar is less than 100 km, and hence the lowest radar beam is close to the surface, making the difference between the radar ref lectivity aloft and at ground level small. The most remote radars in the network are those in Iceland and Gran Canarias, and their nearest neigh- bors are more than 1,000 km away. When these three radars are discarded, the largest distance between two neighboring radars in the European network is only 276 km.
The lowest radar in the network is located in Røst (island in Norwegian Sea) with an antenna height of only 14.5 m above mean sea level, while the highest radar is located in Valluga (Austria; Alps Mountains), with an antenna height of 2824 m above mean sea level. The median height of the radar antennas in the European network is 256 m.
The European countries are at different stages of the development of their operational radar networks and have varying strategies with respect to refurbish- ment or renewal of radars. As a result the time span between the oldest and the newest radars in the net- work is more than 40 years. Half of the radars were installed before 1998, and the other half are more recent. Typically the radars have an antenna diameter of 4.2 m, a beamwidth of 1.0°, and a one-way antenna gain of 44 dBi. Digital receivers and digital signal pro- cessing are the current operational standard. Some national meteorological services have developed their own signal processors, and some are even building the whole radar systems themselves, but generally these are bought "off the shelf." This is also the case for the radar product generation and compositing software. An interactive tool to study the network and a table of radar properties is available at the OPERA website (www.eumetnet.eu/radar-network).
In summary, Europe has an up-to-date but heterogeneous network of operational weather radars. The diversity of the network imposes a true challenge for generation of high quality radar products, but it also offers great potential for further development of robust and widely applicable algorithms.
THE EUROPEAN RADAR MOSAIC. Between 2006 and 2011 the so-called pilot radar data hub was developed and operated in Europe. This data hub collected national radar composites and single-site two-dimensional products from the meteorological services and combined them into a continental- scale radar mosaic covering the whole of Europe. These radar mosaics had a horizontal grid spacing of 4 km and were generated every 15 min. The pilot data hub was very successful in demonstrating the potential of the continental radar mosaics for many applications, but it also revealed a number of data quality issues that had to be addressed. Lopez (2008) compared the radar mosaics of the pilot data hub with European Centre for Medium-Range Weather Forecasts (ECMWF) model data and other obser- vations and found several regions with gross over/ underestimation of the accumulated precipitation. The experiences with the pilot data hub, both posi- tive and negative, were crucial for obtaining the "go ahead" for the OPERA Radar Data Centre, named Odyssey, and for defining its specifications.
The OPERA Radar Data Center (Odyssey) has been operational since 2011. Polar volume data are collected from 134 radar sites of 21 countries (Belgium, Croatia, Czech Republic, Denmark, Estonia, Finland, France, Germany, Hungary, Iceland, Ireland, the Netherlands, Norway, Poland, Romania, Serbia, Slovakia, Slovenia, Spain, Sweden, and United Kingdom), and continen- tal-scale mosaic products (surface rain rate, rainfall accumulation, and maximum reflectivity) are gener- ated in real time (Matthews et al. 2012). The number of radars ingested in Odyssey is smaller than the total number of radars in the network. This is caused by various reasons, such as delays in national decision making and delays in making the required changes to the national production systems. Many of the qual- ity issues apparent in the products of the pilot data hub could be traced back to the use of national radar composites and two-dimensional products, because different algorithms and settings were used to produce them and this led to inhomogeneity. Therefore, it was decided that Odyssey will collect polar volumetric data from individual radar sites and that the products would be generated centrally. This was a major step forward in the harmonization of the product genera- tion and the resulting quality of the European radar mosaics. Moreover, the collected polar volumetric data can now be redistributed for other applications, such as data assimilation for numerical weather predic- tion. Odyssey has been developed with four priority user groups in mind: core services forecasting and nowcasting, numerical weather prediction (assimila- tion and verification), civil and military aviation, and hydrology and water management.
The operational Odyssey runs in parallel at both the Met Office (Exeter, United Kingdom) and Météo France (Toulouse) to ensure a continuous production of the mosaic products, and the data communication between the two centers is done over the Regional Meteorological Data Communication Network, which has a very high level of guaranteed service. The data transfer from members to the centers is mainly done using Internet connections. Internet provides a better capacity and ease of use but is less reliable. Odyssey is established as a Data Collec- tion and Production Centre (DCPC) of the WMO Information System (WIS). Three European radar mosaic products are produced operationally by the data center (Matthews et al. 2012):
* Surface rain-rate composite (every 15 min): Each composite pixel is a weighted average of the lowest valid pixels of the contributing radars, weighted by the inverse of the beam altitude. Polar volume data cells within a search radius of 2.5 km of the composite pixel are considered. Data measured below 200 m altitude are not used.
* Maximum ref lectivity composite (every 15 min): Each composite pixel contains the maximum of all polar volume data cell values of all elevations avail- able of the contributing radars at that location.
* Rainfall accumulation (every hour): A sum of the previous four 15-min surface rain-rate products is produced.
These mosaic products are generated on a Cartesian grid covering the whole of Europe (area: 3,800 × 4,400 km2) with a horizontal grid spacing of 2.0 km. Examples of a rain-rate product and of a maximum ref lectivity product are shown in the top row of Fig. 2. Details of the geographical projection, covered area, and information model (Michelson et al. 2011) and data formats are given in Table 1. The products are available for official-duty use and for research and education. Arrangements to make the products available for commercial use are ongo- ing. More information on licenses is available at the OPERA website (http://eumetnet.eu/odyssey-opera -data-centre).
During the first year of Odyssey operation, sev- eral issues with clutter, anomalous propagation, and other spurious echoes were apparent, and therefore an anomaly-removal module has been implemented in early 2013. The anomaly-removal module ingests polar volumetric data and employs a combination of anomaly detection and removal (Peura 2002) and hit- accumulation clutter filtering (Scovell et al. 2013) to clean the radar data. The bottom row of Fig. 2 shows a pair of product images with and without these al- gorithms in use. It is evident that the amount of radio interferences and permanent scatter has decreased but that they still exist. Comparing the panels of Fig. 2 in the top right (end product; thresholded at -6 dBZ) and the bottom right (same data; no thresholding) indicates that much of the remaining nonmeteoro- logical echoes can be removed by simple thresholding. A quantitative measure of the improvement can be obtained by gauge-radar comparisons but such work has not yet been done.
Since Odyssey started running in 2011, the timeli- ness and availability of the composite products has been maintained at a very high level. The perfor- mance monitoring of the radar data center takes place at the EUMETNET Composite Observing System (EUCOS) Quality Monitoring Portal, hosted at the German Weather Service [Deutscher Wetterdienst (DWD)]. This provides an independent monitoring of Odyssey including a "typical" delivery path of the products to a national meteorological service, in this case to the DWD observation database. During the 2 years of operation, the availability as observed at the quality monitoring portal has varied between 99% and 100% (target 99.0%) and the timeliness (delivery within 20 min after data time) has fluctuated between 96% and 100% (target 95%).
EVALUATION OF NEW TECHNOLOGIES. The most important element of the European co- operation is the exchange of information on radar systems, software, operations, and products. New developments in all radar fields are discussed at the meetings and the most important ones are also presented as documents. In this way, all members are able to take advantage of the experience of those who take the new technology first in use or who are more advanced in the use of their radar systems. Over the years, documents have been prepared on the data acquisition methods, on production practices of volumetric data and products (Huuskonen et al. 2008), on radar site selection and site management (Dombai 2010), and recently on the use of polari- metric radar (Tabary et al. 2012) and X-band radar systems (Cremonini et al. 2012).
During the past decade, the emphasis has been on studies dealing with polarimetry. A clear change of paradigm has occurred and at present nearly all new installations and radar upgrades in Europe are polarimetric. By October 2013 altogether 48 radars in the network were polarimetric, and given the ongoing network upgrades this number will be close to 100 in a few years from now. Therefore, we will highlight two recent OPERA studies on polarimetric radars.
A study on polarimetry in C and S bands (Tabary et al. 2012) presents results of the research and devel- opment work by the French, German, and U.K. me- teorological services, who tested polarimetry in their research installations before decid- ing on the upgrade of their networks. The document presents methods for use in the calibration and monitor- ing of the radars. Careful monitoring of the system properties is important in order to detect possible hard- ware failures at an early stage and to monitor the calibration of the polarimetric quantities, especially the differential reflectivity between the horizontal and vertical polariza- tion channels. A precise knowledge of the differential reflectivity is paramount in many applications (e.g., in hydrometeor classification), and hence its calibration needs to be monitored. Measurements of rain at vertical pointing and solar observations are used for the calibration. Polarimetry enables new methods for the estimation and correction of the attenuation due to heavy rain, which can be significant in the European network because it consists mostly of C-band radars. Tabary et al. (2012) also compare various methods for quan- titative precipitation estimation and for improving its accuracy using polarimetric radar data.
The other study (Cremonini et al. 2012) concen- trates on X-band radars, and it includes an overview of the X-band radars in the OPERA countries. The X-band radars operate at higher frequencies and shorter wavelengths (~10 GHz; 3 cm) than the C-band radars (~6 GHz; 5 cm) and S-band radars (~3 GHz; 10 cm). They are smaller in size and better for mobile applications and have a lower power demand. On the other hand the attenuation due to rain is much more severe, and hence the operational range is shorter compared to the radars using lower frequencies. Many of the X-band radars are operated by non- OPERA members, and those operated within OPERA are more used for research and development than for operations. In X band, the scope and type of the radars vary considerably. Some radars are modified ship radars, with a fan beam antenna, and aimed at urban hydrology measurements, while others are "full scale" polarimetric installations used for the same applications as their C-band counterparts, albeit for shorter range of operations. The study indicates that there is an increased interest to use also the X band of frequencies.
COLLABORATIVE DEVELOPMENTS. ODIM. Exchange of radar data in a heterogeneous network requires that an exchange format is jointly agreed upon, just as for any meteorological or other data. The exchange of volumetric data has been ongoing in legacy formats of software vendors and was often restricted to radars running the same software. A major achievement of OPERA is the introduction of the OPERA Data Information Model (ODIM) by Michelson et al. (2011). This information model describes in general terms radar volume data, radar products, and the associated metadata. ODIM con- tains descriptions from classical radar quantities (radar reflectivity and Doppler velocity) to quan- tities from polarimetric radars (e.g., differential reflectivity). A large selection of radar-based products is also supported.
ODIM has representations in two formats, Hi- erarchical Data Format (HDF5) and WMO Binary Universal Form for the Representation of meteo- rological data (BUFR). Encoders to produce files in these formats from the native formats exist and are also included in several commercial radar software systems. Hence data with ODIM specification are readily obtained for exchange. OPERA is also provid- ing software tools for conversion of radar data be- tween these two formats. The ODIM representations in HDF5 and BUFR are described in Michelson et al. (2011) and Urban (2011), respectively. The formats and the related software are updated when ODIM is expanded. OPERA continues to support the two formats as long as they are used. Presently, HDF5 is more popular than BUFR within OPERA.
The development of ODIM and the associated software tools have removed a major obstacle in the exchange of radar data. A WMO task team is cur- rently considering ODIM as the global standard sup- porting the exchange of weather radar data. ODIM is definitely one of the major achievements of OPERA during the last 5 years. The format specification and tools produced by OPERA are available freely for everyone (www.eumetnet.eu/opera-software)
Quality information. At present, most of the radar data and products are distributed without any quality information included. The volumetric data are often high-pass filtered to remove ground clutter and thres- holded to remove pixels close to the noise level. The aim has been to provide the best possible data for the users, but users are not all alike. The quantitative use of radar data and products is increasing in numerical weather prediction, nowcasting, and hydrology, and moreover the users want to know more and more what kind of processing the data have undergone. There is a general need for quality information to be included with the radar data and products. The qual- ity information is also crucial when the compositing algorithms are improved from simple ones selecting data by using distance or beam height only. The type of quality information to go with data and the question of how the quality information is then used are not simple issues. During the OPERA lifetime, a number of projects have been carried out, and the most recent one includes testing of a proposed quality indicator in the operation context of the Met Office (Sandford and Gaussiat 2011).
Weather radar wind profiles. Volumetric data from Doppler radars can be processed to provide verti- cal wind profiles at the radar location. Altogether 184 radars in the OPERA network are Doppler radars, and from about 110 radars the wind pro- files are available in the WMO GTS network. The work to improve the quality of the wind profiles is ongoing in close cooperation of the EUMETNET Profiling Program (E-PROFILE; see www.eumetnet .eu/e-profile) and the comparison statistics against the ECMWF numerical model is an important monitoring tool. It is clear that a discrepancy does not automatically imply that the wind profile is in error, but in general a gross deviation with the model results can be considered as an indication of a radar error. EUMETNET has set 5 m s-1 root-mean-square (rms) difference as a target for acceptable upper-air wind observations. About half of the 100 wind profile sites are monitored by ECMWF. Figure 3 shows monthly rms differences for these radars, calculated by com- paring observations to the first-guess model forecasts (Haseler 2004). It is seen that some networks (Finland, United Kingdom, the Netherlands, and Hungary) meet consistently the acceptance criterion of 5 m s-1, while other networks also show larger differences. A large rms difference may be caused by hardware problems or configuration problems but is mostly related to flaws in the algorithm. The performance of the radar network can also be compared to the cor- responding rms value reached by the European radio soundings, for which the median value is 3 m s-1 and the overall minimum value is 2.3 m s-1. Some net- works (Finland and the Netherlands) reach this level, and hence we can conclude that the quality of their wind profiles is comparable to the quality of the radio soundings. Recommendations on producing good quality wind profiles are given in Holleman (2005). Each member produces the wind profiles of their radars themselves and hence the implementation of the recommendations is members' responsibility. OPERA participates to the work by giving recom- mendations and help in their implementation.
Solar monitoring. Solar monitoring is a method based on the analysis of sun signals in the polar volume data produced during operational scanning of weather radars (Huuskonen and Holleman 2007; Holleman et al. 2010a,b). A continuous reflectivity signal is searched along radials in the operational scan data and, depending on the hardware of the radar, the volume coverage pattern, the season, and the latitude of the radar, several tens of sun hits are found per day. Least squares fitting will give information on the antenna pointing bias and on the stability and calibra- tion of the receiver. Several OPERA members have suc- cessfully implemented this method as an operational monitoring tool. This method will be implemented at the operational data center so that the solar monitoring results will become available for all OPERA members.
CHALLENGES TO RADAR NETWORK OPERATION. Weather radars operate in an en- vironment that is in a constant change because of human activities. These activities may impact the data quality of the radar network and may even pose a threat to its viability and include encroachment of manmade buildings and towers toward radar sites, increased pressure upon weather radar spectrum, and the provision of sufficient resources to sustain the op- eration of the radars. At present, there are two issues related to the radar working environment, which are in importance way ahead of any other: namely, the reflections and blockages caused by wind turbines and the radio interferences caused by wireless radio networks and radio links.
Wind turbine impacts. The number of wind turbines and wind farms is increasing rapidly in Europe. Many locations that are suited for weather radar siting are also good for wind farms, and hence there can be a conflict of interests. The disturbances caused by wind turbines are two-fold. The towers and moving blades create partial beam blockages, just as any building in the radar main beam, but these blockages vary as the turbines are oriented according to the wind direction. Another effect is that the blades cause clutter signal, which cannot be removed by using traditional clutter filtering because the blades are moving. The clutter is seen from wind turbines several tens of kilometers away. As a way ahead, OPERA formulated a recom- mendation that states that within 5 km of a C-band radar and 10 km of an S-band radar no wind turbines should be built and that within 20 km (30 km for S band) the potential development should be assessed before proceeding. This recommendation has been endorsed by both EUMETNET and WMO. The aim of the recommendation is to provide a common support to the European weather radar community when our members are dealing with individual wind farm cases. A number of disturbance cases and the recommendation are shown in one of the OPERA deliverables (Chèze and Haase 2010). The ability of the national meteorological services (NMSs) to pre- vent wind farm development close to radar sites or to force changes to the plans differs greatly from country to country but is rather limited in many countries. OPERA has started an activity to collect the national practices into a consolidated document.
Radio interferences. The interferences by other users of the C band of frequencies are a major problem in some European countries. These users include wireless local area networks [WLANs; or radio LANs (RLANs)] and radio links. In principle, the radio regulations give suf- ficient protection to weather radars because the other users operate at C-band radar frequencies on a second- ary basis and are not allowed to disturb radars. Techni- cally, this is accomplished by dynamic frequency selec- tion (DFS), in which a device first listens to the radio traffic and uses the band only if no radars are operating there. In practice, not all devices obey the regulations or rogue users may willingly disable the DFS feature. Hence, the DFS mitigation concept has turned out to be inadequate in certain countries. Within EUMETNET, the work against radio disturbances is coordinated by the EUMETNET Radio Frequency Manager (www .eumetnet.eu/secretariat), and OPERA supports this work, for example, by providing new cases of distur- bances and increasing the awareness of the issue among its members. Several interferences caused by RLANs are visible in the bottom panels of Fig. 2 as spokes pointing toward radar locations. The radio interfer- ences in the C band, if not solved, may push the market toward using S band, which would increase the cost of radar systems. But similar interferences affect also S band [e.g., Worldwide Interoperability for Microwave Access (WiMAX)], and S band has also been named as a candidate band for future mobile communication networks.
FUTURE. The world around us is changing and the international collaboration has to evolve as well. During the coming 5 years, the OPERA members will continue to meet twice a year. As many members are in the process of upgrading their national radar networks, there will be a lot of experience to share. Open discussions between experts of the field (scien- tists, hardware engineers, software specialist, product developers, forecasting meteorologists, and customer contact persons) and learning from others' experi- ences have always been OPERA's inherent strengths. The funding from EUMETNET is crucial, as it guar- antees ongoing project management and funding of the project work. But the work can only be completed because the members are willing to add significant in-kind contributions: for example, by funding all travel to meetings and allowing the projects use the members' infrastructure free. There will be more pressure than before on the frequency management and on the protection of radar sites against wind farm development. OPERA will continue to communicate with the stakeholders on these issues.
The OPERA Radar Data Center, Odyssey, has become operational in 2011, and the European countries will continue the development and the operation of this data center. OPERA members agree that producing European composites is beneficial and have decided to share their data for the purpose. Data ownership is an important and delicate issue, and discussions on a wider and freer distribution of data are ongoing. Some members already distribute all their data for free to all. Data quality is another key issue, as well as a timely distribution of the products and volumetric data to many different users. The European radar mosaics will be made available to var- ious users in meteorology, hydrology and transport, and volumetric data will be used for assimilation of numerical weather prediction models. The dialog with these (new) users is actively pursued in order to obtain their experiences and emerging needs.
Even though the European radar mosaics are the most visible products of Odyssey, the data center will also act as a collection, cleaning, and redistribution station for single-radar volumetric data. In numeri- cal weather prediction, assimilation of global models and verification of models tends to have more use for the composite, while the limited area models need data in its original measurement format (i.e., as polar volumes). The latter is especially true for Doppler radial velocity data.
The resolution of the limited area models for nu- merical weather prediction has grown so much that the old assumptions used when assimilating radar observations are no longer valid. We are used to think about radar measurements as point observations in regular spherical coordinates that can be converted to Cartesian coordinates using simple equations. However, the present models can use their informa- tion on temperature and humidity along the radar beam and then forecast the actual propagation of the microwaves. This and other advances in data assimila- tion on the one hand create more needs for the radar metadata to be exchanged but on the other hand create completely new opportunities for work on data quality.
For use of weather radars, the challenges and ad- vantages have been related to the prevailing climate and applications: some need to detect tornadoes, and others need to improve their quantitative snowfall estimates. The optimal use of weather radar strongly depends on the local infrastructure: cost of electricity and bandwidth of data transfer. The local infrastruc- ture is improving globally, and therefore the use of radars is also getting more uniform on a global scale. The new regions with emerging radar networks can benefit from the work done in Europe and elsewhere.
ACKNOWLEDGMENTS. The authors are greatly indebted to the close and warm cooperation within the OPERA program since its start in 1999. The present paper summarizes the outcome of the work, which is a result of the joint effort of all participants. The OPERA work has been financed by its member organizations. The authors wish to thank the reviewers (Tim Crum, Paul Joe, and David Schultz) for careful reading of the manuscript and valuable comments, which have improved the paper considerably.
1 Authors are present or past project/program managers of OPERA. Asko Huuskonen was OPERA PM during 2004- 06 and again 2010-12, Iwan Holleman was PM during 2007-09, and Elena Saltikoff is the present PM for the 5-yr period of 2013-17.
Chèze, J.-L., and G. Haase, 2010: Site protection (wind turbines). OPERA Deliverable OPERA_2010_05, 17 pp. [Available online at www.eumetnet.eu/sites/default /files/OPERA_2010_05_Wind_turbines.pdf.]
Cremonini, R., P. Tabary, J. Sugier, M. Frech, P. Alberoni, and L. Baldini, 2012: Evaluation of new radar tech- nologies. OPERA Deliverable OPERA_2012_04, 46 pp. [Available online at www.eumetnet.eu /sites/default/files/OPERA_2012_04_Operational _polarimetry_in_X_band.pdf.]
Crum, T., and R. Alberty, 1993: The WSR-88D and the WSR-88D Operational Support Facility. Bull. Amer. Meteor. Soc., 74, 1669-1687.
Dombai, F., 2010: Weather radar site selection and protection. OPERA Deliverable OPERA_2010_04, 60 pp. [Available online at www.eumetnet.eu /sites/default/files/OPERA_2010_04_Radar_site _selection_and_protection.pdf.]
Haseler, J., 2004: Early delivery suite. ECMWF Tech. Memo. 454, 26 pp.
Holleman, I., 2005: Quality control and verification of weather radar wind profiles. J. Atmos. Oceanic Technol., 22, 1541-1550.
-, A. Huuskonen, R. Gill, and P. Tabary, 2010a: Operational monitoring of radar differential ref lec- tivity using the sun. J. Atmos. Oceanic Technol., 27, 881-887.
-, -, M. Kurri, and H. Beekhuis, 2010b: Opera- tional monitoring of weather radar receiving chain using the sun. J. Atmos. Oceanic Technol., 27, 159-166.
Huuskonen, A., and I. Holleman, 2007: Determin- ing weather radar antenna pointing using signals detected from the sun at low antenna elevations. J. Atmos. Oceanic Technol., 24, 476-483.
-, M. Kurri, and J. Koistinen, 2008: Harmonized production practices for volume data, low level re- flectivity, and weather radar wind profile. OPERA Deliverable OPERA_2008_06, 46 pp. [Available online at www.eumetnet.eu/sites/default/files /OPERA_2008_06_ProductionPractices.pdf.]
-, L. Delobbe, and B. Urban, 2010: News on the European weather radar network (OPERA). Proc. Sixth European Conf. on Radar in Meteorology and Hydrology, Sibiu, Romania, Ministry of Environment and Forests, and National Meteorological Adminis- tration, 362-367.
Lopez, P., 2008: Comparison of OPERA precipitation radar composites to CMORPH, SYNOP and ECMWF model data. ECMWF Tech. Memo. 569, 22 pp.
Matthews, S., P. Dupuy, R. Scovell, A. Kergomard, B. Urban, A. Huuskonen, A. Smith, and N. Gaussiat, 2012: EUMETNET OPERA Radar Data Center: Providing operational, homogeneous European radar rainfall composites. Proc. Weather Radar and Hydrology Symp., Exeter, United Kingdom, IAHS, 9-14. Michelson, D., R. Lewandowski, M. Szewczykowski, and H. Beekhuis, 2011: EUMETNET OPERA weather radar information model for implementa- tion with the HDF5 file format. OPERA Deliverable OPERA_2008_03, 36 pp. [Available online at www .eumetnet.eu/sites/default/files/OPERA_2008_03 _WP2.1b_ODIM_H5_v2.1.pdf.]
Peura, M., 2002: Computer vision methods for anomaly removal. Proc. Second European Conf. on Radar in Meteorology and Hydrology, Delft, Netherlands, Copernicus GmbH, 312-317.
Sandford, C., and N. Gaussiat, 2011: Evaluation of an error-based quality index for compositing using UK radar data. OPERA Deliverable OPERA_2011_12, 13 pp. [Available online at www.eumetnet.eu/sites /default/files/OPERA_2011_12_Quality_index_ evaluation.pdf.]
Scovell, R., N. Gaussiat, and M. Mittermaier, 2013: Recent improvements to the quality control of radar data for the OPERA data centre. Proc . 36th Conference on Radar Meteorology, Breck- enridge, CO, Amer. Meteor. Soc., 361. [Available online at https://ams.confex.com/ams/36Radar /webprogram/Manuscript/Paper229061/OPERA_ QC_v1.pdf.]
Serafin, R., and J. M. Wilson, 2000: Operational weather radar in the United States: Progress and opportunity. Bull. Amer. Meteor. Soc., 81, 501-518.
Tabary, P., J. Figueras, J. Sugier, M. Frech, and P. Alberoni, 2012: Operational polarimetry in C- and S-bands. OPERA Deliverable OPERA_2012_03, 57 pp. [Available online at www.eumetnet.eu/sites /default/files/OPERA_2012_03_Operational _polarimetry_in_C_and_S_bands.pdf.]
Urban, B., 2011: BUFR format and graphical format for polar volume data submitted to and composites pro- duced by the OPERA Odyssey. OPERA Deliverable OPERA_2011_13, 14 pp. [Available online at www .eumetnet.eu/sites/default/files/ODIM-bufr-polar -and-compo-and-graphic.pdf.]
AFFILIATIONS: Huuskonen And sAltikoff- Finnish Meteorological Institute, Helsinki, Finland; HollemAn -Institute for Molecules and Materials, Radboud University Nijmegen, Nijmegen, Netherlands
CORRESPONDING AUTHOR: Asko Huuskonen, Finnish Meteorological Institute, Obser vation Services, P.O. Box 503, FI-00101 Helsinki, Finland
The abstrac t for this article can be found in this issue, following the table of contents.
In final form 9 October 2013
©2014 American Meteorological Society
(c) 2014 American Meteorological Society
[ Back To TMCnet.com's Homepage ]