AN OVERVIEW OF USING WEATHER RADAR FOR CLIMATOLOGICAL STUDIES Successes, Challenges, and Potential ELENA SALTIKOFF, KATJA FRIEDRICH, JOSHUA SODERHOLM, KATHARINA LENGFELD, BRIAN NELSON, ANDREAS BECKER, RAINER HOLLMANN, BERNARD URBAN, MAIK HEISTERMANN, AND CATERINA TAssONE Old measurements are precious: once lost, they cannot be replaced. But without carefully saved information of how the data were measured, we also create a risk of false conclusions. ver the last century, weather radars have been climate community. Possibilities to improve historical widely used to detect and quantify precipitation datasets are limited, but the data we archive today are O and severe weather. Issuing warnings of se- part of the data to be used by future generations. vere hail, tornadoes, bliz- zards, and flooding has greatly benefited from high-resolution data from the radar networks. These networks are now covering a majority of the dense- ly populated areas of the world (Fig. 1). Now that some of the networks have collected data for up to 50 years, scientists have start- ed to use the rich informa- tion for climate studies. Unfortunately, a lot of the older radar data have been FIG. 1. A map of weather radar coverage in the world (in Robinson projection). lost, and not all of them To compute and map the areas “illuminated” by radar, we used the wradlib are archived even today. library (https://wradlib.org), assuming each radar has a range of 200 km The objective of this paper irrespective of bandwidth, polarization, and local terrain. Most radar locations is, on one hand, to provide included in this map have been retrieved from a WMO database (WMO 2019). an overview about weather Note that not all operational radars are included in the database. Additional radar information that is radar locations have arbitrarily been added for China (manually digitized already available to the cli- from WMO 2013), the Philippines (I. Crisologo 2018, personal communica- tion), Vietnam (locations estimated in 2017 from the webpages of the National mate science community, Centre for Hydro-Meteorological Forecasting in Vietnam, www.nchmf.gov and on the other hand, to .vn/Web/en-US/73/Default.aspx), and Myanmar (locations estimated in 2017 make the radar community from the webpages of the Department of Meteorology and Hydrology in aware of the needs of the Myanmar, www.moezala.gov.mm/radar-image). AMERICAN METEOROLOGICAL SOCIETY SEPTEMBER 2019 | 1739 Unauthenticated | Downloaded 10/01/21 11:43 AM UTC Weather radars have been used to detect precipi- tation since the 1940s and 1950s. The first displays were analog, and the im- age existed only as long as the fluorescent tube was glowing (unless it was saved as a photograph). During the following decades, in- formation technology has been developed, and sev- eral weather services have FIG. 2. Cumulative sum of 45 national weather services archiving radar data started archiving data and since 1975. providing the data to the research community; however, most of the archiving The Global Climate Observing System (GCOS), procedures were and still are not standardized. For a program co-sponsored by the WMO, the Inter- example, some save the data in the original format de- governmental Oceanographic Commission of the fined by the radar software manufacturer while oth- United Nations Educational, Scientific and Cultural ers only save the final products such as precipitation Organization (IOC-UNESCO), the United Nations intensity maps. Furthermore, upgrading the radar Environment Programme (UN Environment), and operating software usually introduces changes in the the International Science Council (ISC), regularly data format. Very few radar operators have executed a assesses the status of global climate observations reanalysis to create a homogeneous time series. While and produces guidance for its improvement. GCOS’s radar data have primarily been used by radar experts vision is for climate observations to be accurate in the past, more and more other research fields are and sustained, and for all users to have free and now using radar data. For instance, weather radar open access to the climate data they need to address data from several data providers are now assimilated climate-related questions. The Atmospheric Obser- into weather forecasting and nowcasting models. As vation Panel for Climate (AOPC) is one of the three a result, this raises the importance of common data scientific panels of GCOS, together with the Ocean formats and well-documented metadata, for example, Observations Panel for Climate and the Terrestrial among the World Meteorological Organization Observation Panel for Climate. It was established (WMO). by the GCOS Steering Committee in recognition of the need for specific scientific and technical input concerning atmospheric observations for climate. AFFILIATIONS: SALTIKOFF—Finnish Meteorological Institute, With the increased usage of radar data around the Helsinki, Finland; FRIEDRICH—Department of Atmospheric and world for diverse applications and increasing demand Oceanic Sciences, University of Colorado Boulder, Boulder, for archiving historic, current, and future radar data, Colorado; SODERHOLM—Meteorological Institute, University the AOPC formed a Task Team on Radar Observations of Bonn, Bonn, Germany; LENGFELD, BECKER, AND HOLLMANN— Deutscher Wetterdienst, Offenbach, Germany; NELSON—NOAA/ for Climate Applications (TTROCA, referred to as National Centers for Environmental Information, Asheville, Radar Task Team hereafter) to define procedures to North Carolina; URBAN—Météo-France, Toulouse, France; use radar data for climate applications. As a first step, HEISTERMANN—Institute of Environmental Sciences and Geography, the Radar Task Team conducted a survey to determine University of Potsdam, Germany; TAssONE—GCOS Secretariat, what operational radar data exist and how weather WMO, Geneva, Switzerland services have archived data in the past. According to CORRESPONDING AUTHOR: Elena Saltikoff, [email protected] a survey, some radar data have been archived since the The abstract for this article can be found in this issue, following the 1970s, for instance, in Hong Kong, while the majority table of contents. of weather services (>35) started archiving around the DOI:10.1175/BAMS-D-18-0166.1 year 2000 (Fig. 2). Therefore, it is becoming increas- In final form 3 May 2019 ingly relevant to review what can be done with existing ©2019 American Meteorological Society For information regarding reuse of this content and general copyright data and how to manage the present datasets to ensure information, consult the AMS Copyright Policy. future climatologists can utilize the full capabilities of the material, which is the goal of the Radar Task Team. 1740 | SEPTEMBER 2019 Unauthenticated | Downloaded 10/01/21 11:43 AM UTC This paper summarizes the work of the Radar Task changed, the coverage has increased, and processing Team and is organized as follows: we first analyze methods have improved, so the time series are by no what data exist, how these data have been used, and means homogeneous. then conclude with recommendations for processing Results from the survey also show that 90% of the historical data and archiving new data. Discussion archives have saved radar reflectivity (in the unit of of these topics concentrates on operational radar dBZ) in native polar coordinates (“Level 2 data,” see networks, most often operated by national meteoro- sidebar). logical services, and it is based on work conducted by In addition to inquiring size and quality of ar- the Radar Task Team of the AOPC. chives, two questions were asked about access to data: “Is the data available for researchers outside of DECADES OF DATA ARE ALREADY AVAIL- your institute?” and “How would you describe the ABLE, BUT THEY ARE OFTEN PATCHY effort of retrieving one year of old data from your AND HETEROGENEOUS. The first impulse archives?” To the first question, 15 national weather of climate experts invited to the Radar Task Team— services selected “Yes, and they use it regularly,” and which was soon proven to be wrong—was that radar 19 selected “Probably yes, but it may be complicated.” data volumes are too large to be archived, and that the Only 7 replied “No,” while 4 selected “I do not know.” radar operators are probably not willing to share their The weather services who replied positively are listed archives. The Radar Task Team executed a survey in in Table 1. Six of these mentioned they have a graphi- 2017 to get an estimate of existing radar archives. The cal interface for selecting the data, while according invitation to the survey was sent to contact points in to most of the other replies the process required different countries suggested by the relevant WMO programmatic interaction with data as opposed to a task teams. To ensure a good response rate, the graphical interface. questions were kept short and straightforward with A list of hydrometeorological or meteorological multiple choice. After the survey the team contacted services and other authorities that operate weather those countries with the longest archives and asked radars, with links to their official websites, is provided additional questions. in the WMO Radar Database (WMO 2019). Out of the 91 weather services which operate weather radars based on WEATHER RADAR LEVELS the WMO’s radar database (WMO 2019), experts from he table below is proposed by the WMO Interprogram Expert Team of Weather Radars as the standard wording to describe “levels” of weather radar data. 45 national weather ser- T vices replied to this survey. While all these 45 nations Definition currently maintain a radar Level 0 Data at full resolution as received at the sampling rate of the receiver. archive, with 40 of them Generally only available internal to the system. Special equipment may be required to measure and record such data. having more than 10 years of data archived already, Level 1 Data in sensor units also known as “time series” or “I/Q” (in-phase and quadrature) data. Produced and processed by the instrument’s only 20 countries have signal processor.