This document, which describes the meteorological surface data in the Met Office Database - MIDAS, is an abridged version of "MIDAS Data Users Guide", provided by the Met Office. The BADC is not responsible for its content.
Since the early days of this century the Met Office has been responsible for maintaining the public memory of the weather. All meteorological observations made in the UK and over neighbouring sea areas have been carefully recorded and placed in an archive where they may be accessed today by those with an interest in the weather and where they will also be available to those in future generations. Up to fairly recently all observations have been recorded on paper forms which, when completed and checked, have been stored in secure archive buildings. Today there are archive sites at Bracknell, Edinburgh and Belfast. In this way the Met Office has met the requirements of the Publics Record Act for public access to the Met Offices technical records.
In the age of computers, paper records are an unsuitable format for meteorological data and in 1959 a new system was introduced whereby all current climate data received at the Met Office was manually keyed and stored in digital format. At the same time a limited number of records for earlier years were also transferred into digital format so that some continuous climate sequences were available immediately. In the 40 years that have elapsed since those days the digital climate record has been maintained and, as more automation has occurred, especially within observing systems themselves, the role of the paper record has declined. Today fewer and fewer paper records arrive at the archives and their principal purpose is a historical one.
The storage of climate data at the Met Office has slowly developed over time, first as a collection of data on paper or magnetic tape, then, from the early 1970s as an organised database, the CDB, which served many users until its final demise in July 1998. The current climate database is MIDAS (Met Office Integrated Data Archive System) which has a relational structure.
The MIDAS database contains the following general types of meteorological data:
MIDAS does not contain any remote sensing data such as radar estimates of precipitation, thunderstorm locations or satellite imagery. Such data are stored by the Met Office but they are not part of this archive.
The purpose of this guide is to describe the basic features of the observing systems used over the years, the way the observations have been processed and stored in the database and the general characteristics of the data. It is aimed at those with little familiarity with observing methods or instrumentation.
The location of the site should be selected in such a way that the observations are representative on a scale required from the station; a station in the synoptic network should make observations to meet synoptic scale requirements, a rainfall station should measure the impact of local orography on the rainfall amount, while an aviation station should observe the local conditions at the aerodrome. As far as it is possible, stations in the synoptic and climatological networks meet the following requirements for site location:
It is unavoidable that some sites do not meet all these requirements, particularly where a station set up for one purpose gradually takes on a different role, for example an airport site originally established for aviation observing may become a key synoptic or climate station while suffering the effects of urbanisation. A few sites are in city centres and may be unsuitably located close to large obstacles or even on the roof of a building.
The horizontal location of the station and its height above mean sea level are determined from the position of the principal raingauge.
Most instruments at the station are located close together in the enclosure, a flat area of ground approximately 10 m by 7 m covered by short grass and surrounded by fencing. Typically the enclosure contains one or more raingauges, a screen containing the thermometers, soil thermometers, a concrete slab with a concrete minimum thermometer, a grass minimum thermometer, and a sunshine recorder. Where there is a 10 m anemometer mast it is often at a separate location. The barometer may be located in a building, often some distance from the enclosure, such as the forecast office if the station has a forecasting function.
Until recent years all surface observing has been performed manually; the observer reads the instruments, makes visual estimates of visibility, cloud, present weather and the state of ground, records the observation on paper, and, if the station produces SYNOP and NCM messages, encodes the report and transmits it. An observation pad or pocket register is provided for noting the observation as it is made, a permanent record being made once the observation is complete. At synoptic stations observations are recorded in the daily register in permanent ink, which then becomes the official record for the station. Daily observations at climate stations are recorded on climate forms which are returned to the Met Office at the end of each month. Metform 3208b is in use at present but there have been a variety of different form types in the past. The observer may also have other climate returns to complete such as the analysis of hourly wind from anemograms, hourly sunshine from sunshine cards, and hourly rainfall from a rain recorder. A return of instrumentation in use at the station should also be completed once a year.
With the introduction of automated systems the role of the human observer has diminished. Even where a station remains manned 24 hours a day the process of reading the instrument, processing the result, encoding the message, and transmitting the data is largely automated today by systems such as SAMOS. The observer is still required to make the an observation of those parameters for which automated systems of measurement either do not exist or are insufficiently accurate, for example, snow depth, state of ground, cloud and present weather. He also has the capability of modifying the automated input if it is considered to be in error. At such stations the Daily Register is discontinued and the only permanent record of the observation is the electronic one contained in the MIDAS database.
Automation, which began in the 1970s, now affects virtually all aspects of synoptic observing, and to a lesser extent the observations from the climate and rainfall networks. Automated systems have been developed as:
An unavoidable consequence of automation has been the increased number of missing observations in the climate record. The human can improvise by using an alternative means if the primary system fails, the machine can not. The data loss at some fully automatic stations can be as high as 5%. A second consequence of automation has been the gradual replacement of one type of instrumentation by another; the Electrical Resistance Thermometer has replaced the liquid-in-glass thermometer, and the tipping bucket raingauge has replaced the 5 inch raingauge. This may have significant consequences for the user of the climate record interested in small departures from the long term average.
SAWS, ESAWS, SIESAWS and CDL
The Synoptic Automatic Weather Station (SAWS) was the first fully automated observing system deployed in any number by the Met Office. The system operated at many remotes sites and provided hourly SYNOP messages from the mid 1980s. It was later replaced by the Enhanced SAWS (ESAWS) which is capable of producing hourly climate messages (HCM) and 12-hourly national climate messages (NCM) in addition to the SYNOP. A version designed to operate in severe icing environments, SIESAWS, has been installed at a few high level locations. Climate Data Loggers (CDL) of a number of different designs have been installed at a number of sites during the 1990s; most record the main synoptic and climate parameters and there is a facility for polling the data remotely. There is a project underway to develop their use as real time observing systems.
SAMOS and CODET
The Met Office has developed SAMOS (Semi Automatic Meteorological Observing System) and CODETS (Coded Observation Data Entry Transmission System) to aid the manual production of observations. Both are PC based systems which can generate the full range of messages for a synoptic station (SYNOP, NCM, SREW, HCM), the CODET system being specially designed to operate at auxiliary stations. The observer can add to or modify any part of the observation through the PC. Quality control checks are incorporated into the interface. The SAMOS system can operate in automatic mode in the following ways:
Rainfall and wind loggers
A number of rainfall and wind logging systems have been installed at Met Office stations since the 1970s. The data are recorded on magnetic medium at the site and transferred at monthly intervals to MIDAS. The SSER and MTER loggers for rainfall are covered in section 5.5 and the DALE logger for wind is covered in section 5.4. The Environment Agency operates a number of rainfall loggers, but only derived daily values from selected stations are stored in MIDAS.
The standard of time for all UK observations is GMT. This was not always the case in the early days of observing, but it is thought that any observations recorded by clock time have been entered against the correct hour GMT in the MIDAS database. The meteorological convention for midnight is 0000 hours on the new day.
All non-UK Observation times are in UTC.
WMO recommendations state that the observation should be made in as short a time as possible just prior to the nominal time of observation and that the pressure reading should be taken last at the exact hour. The UK practice since the 1960s, and possibly from an even earlier date, has differed from this recommendation in the case of real time observations for reasons connected with transmission schedules which are no longer valid. These observations are typically completed by HH-10 and transmitted before the hour is complete. This practice has continued to the present day, even where SAMOS and ESAWS automatic systems are installed, and applies not only to the SYNOP report, but also to the NCM, SREW and HCM messages. CDL systems, by contrast, record observations on the synoptic hour.
Most measurements are made with full traceability to national or international standards; in other words, it can be demonstrated that a series of calibrations have been performed which link the instrument used for the measurement to some national or international standard instrument. This practice, which for temperature and pressure has been unbroken since 1851, ensures a uniformity of measurement over time. Instruments are calibrated after receipt from the manufacturer and in some cases at intervals after that (aneriod barometers and some wind, radiation and humidity sensors are returned for recalibration every few years, but not thermometers). Some of the other checks performed to ensure a properly functioning observing system are as follows:
The term metadata, as applied to observations, is taken to mean those data which describe the contents of the climate archive and which allow the user to understand the origins of the meteorological values themselves. They describe the characteristics of the observational networks and typically include details about the station, its location, environment and observing programme as well as information about the instruments, their relative locations and exposure at present and at times in the past. Observational metadata for UK stations have been built up over the years from several sources:
Only a small part of all the information is stored in digital format. A metadata database is under development which will ensure that all recent station metadata will be available electronically.
Surface observations over the UK meet many different requirements in such areas as forecasting, civil aviation, defence, commerce, industry, agriculture and research. Stations are organised into networks which are designed to meet particular user requirements, the details of which are contained in a series of UKON (United Kingdom Observation Network) documents; UKON1 deals with the synoptic network, UKON2 with the climate network, UKON4 with the wind network, UKON5 with the rainfall network and UKON8 with the sunshine and radiation network. The organisation of the present day networks is described in the rest of section 3 below.
The synoptic network meets the requirements of forecasting, nowcasting, NWP and international exchange for real time observations taken at intervals between 1 and 3 hours. The observed elements include weather, cloud, temperature, humidity, wind, visibility, pressure etc. contained in the SYNOP message (see 4.1). Most synoptic stations also report hourly rain in the SREW message (see 4.2) as well as observations which meet more general climate requirements in the HCM and NCM messages (see 4.4 and 4.5). The current synoptic network has an average station spacing of less than 50 km; it is made up of the following sub-networks:
The purpose of the climate network is to enable the climate of the United Kingdom to be determined and monitored and thereby meeting the requirements for international exchange, commercial applications and research. The UKON2 document specifies how requirements for network density, site representativity, observing schedules and accuracy are met at the present time. The minimum observing requirement at a climate station is the daily measurement of maximum temperature, minimum temperature and rainfall amount though many stations make a wider range of observations such as those elements listed in sections 4.5 4.8. The current climate network is made up of the following sub-networks:
Observations of wind are required from the synoptic network to meet the real time needs of forecasting and NWP and have usually always been based on 10-minute averages. The climatological requirement for wind measurement is more varied and observations using a wider range of averaging periods have been made. Although some wind-only sites exist, observations of wind are usually collocated with observations of other meteorological elements. The current wind network is made up of the following sub-networks:
Although many of the requirements for the measurement of precipitation amount may be met by the stations in the climate network, it is important for some applications to be able to resolve detail on smaller space scales. Point to point variability, and therefore the interpolation error between two neighbouring stations, is larger for rainfall than for most other climate elements. In terms of station numbers, the rainfall network is largest having some 5000 stations currently open across the UK. Radar rainfall sites are not included in this network and their data are not yet part of the MIDAS database. The current rainfall network is made up of the following sub-networks:.
Stations in this network provide measurements of solar and terrestrial electromagnetic radiation and the duration of bright sunshine.
The routine measurement of solar radiation was started in the UK at Kew Observatory in 1908. Daily measurements of the direct solar irradiance were taken at noon when conditions allowed and this practice was continued for the next 40 years. Regular measurements of global and diffuse irradiance began at Kew in 1946, at Lerwick and Eskdalemuir in 1952, and at Aberporth, Cambridge and London in 1957. The network in UK was further enlarged by the addition of co-operating stations making solar radiation measurements for their own purposes and by 1975 the network consisted of 29 stations with at least 5 years continuous recording of global radiation. At the present time (1996) the network consists of 11 Met Office manned stations and 20 unmanned stations (ESAWS) plus 17 co-operating stations.
Stations at present measuring radiation and sunshine may be assigned to any of 5 sub networks designed to meet specific user requirements:
A station may be defined as a collection of instruments and observing systems, existing today or for periods of time in the past, which together may be considered as a common source of observations associated with a given location. At a typical station the thermometers and rain gauges are located in an enclosure, which is a small area of level ground suitably exposed to the weather. Not all instruments at a station are necessarily collocated; for example, the anemometer mast may be at the end of a runway some distance from the enclosure, the pressure sensor may be in a separate building and the estimate of visibility made be made on a roof-top. Observations from all these sources may, nevertheless, be regarded as representative of a single location represented by the station. The concept of a station is reinforced by the way meteorological observations are grouped together into reports: the SYNOP, NCM, SREW or F3208 reports (see section 4) represent groupings of observations from a common general location. Each report requires an identifying number or character string which links it to the originating station. Several identifying systems that have evolved over the years are described in more detail in the following sub-sections.
Over time certain instruments, or the whole enclosure, may be relocated some distance away from the original site. Where the distance moved is small, the observations obtained from the new site may have exactly the same climatological characteristics as previously and it makes sense to regard them as coming from the same source or station distinguished by certain identifiers. Where the distance moved is large, or, where the exposure at the new site is sufficiently different that a detectable impact on the measured climatology is judged likely, it is appropriate that observations from the new site are labelled by a different set of identifiers.
The purpose of the station name is to give an unambiguous label to the source of observations such that no two stations have the same name. It should be meaningful, at least to those who know the locality.
The format of the station name is as follows:
: : : Principal Name [,Qualifying Name]
where the square brackets indicate an optional field.
The Principal Name relates to the general location of the station; it may take the name of the nearest village, town or city, or, the name of the (small) island on which the station is located. When naming a station, Met Office staff have always tried to use a name which appears on an OS map. To avoid many stations having the same Principal Name, the names of large cities are often avoided.
A Qualifying Name, separated from the Principal Name by a comma, may often be required to ensure that each station name is unique; it is however optional. It is generally the name of some local feature such as a building, road or small geographical feature; e.g. Walsingham, Hill House and Blackshiels, Fala Dam. Where an observing site moves a small distance, but sufficiently far that that the new site is regarded as a new station, there might be no reason for wishing to change the station name. In such cases differing stations are identified by No.1, No2, etc. added on to the end of the Qualifying name; e.g. Coldwell Resr, No.2.
Synoptic stations selected as suitable for possible international exchange are give a 5-figure WMO number which is used as the identifier for all SYNOP and SREW reports. It is also used as identifier for climate reports, NCM and HCM, exchanged within the UK in real time. The first 2 figures of the WMO number are 03 which is the block number for the UK and Ireland. Numbering runs from north to south across Britain followed by Ireland.
All stations that are part of the climate network have a 4-figure DCNN (district county number). In general, if a station moves more than 800m in a region of homogeneous terrain it is allocated a new DCNN, while a lesser distance will justify a new DCNN only if it is considered that the exposure at the new site is sufficiently different to affect the measured climatology.
The principal purpose of the rainfall number is to identify a rainfall observing site. Where rainfall is the sole observation made at the station, the number doubles up as the station identifier. Where there is more than one rainfall observing site at the station, each site has a different rainfall number. Where the rainfall observing site moves, even a small distance, the new site is given a new rainfall number while the station may retain its original name, WMO number and DCNN; it is therefore very common for a station to be associated with several rainfall numbers, some identifying rainfall sites open today and some that existed in the past. It is also possible for more than one gauge to be attached to one rainfall number, for instance, 2 gauges may have existed in the enclosure, one used for the hourly reading and one for the 12-hourly reading.
Over the years, the observation branches of the Met Office have followed an agreed set of rules to determine whether measurements of rainfall should be considered as originating from a new source or rainfall number when the conditions under which rainfall observations are made changes. These are specified as the occurrence of one or more of the following:
The last criterion is included because the rainfall number acts as an instrument and logging system identifier as well as a station identifier. This can lead to a number of confusing situations, for example, it is often the case that when a SAMOS system has been installed at a SSER site; the SSER retains its original rainfall number while the SREW and NCM reports from the same TBR are identified by using a different number (see section 5.5). Rainfall numbers run from north to south and uniquely identify the river basin and tributary nearest which the site lies.
Like rainfall number the wind number identifies the characteristics of the observing site. Each anemometer mast has a 6-figure wind number of the format DCNNnn, where nn=01,02, is the anemometer site number. A new site number is allocated to the observations if the mast is repositioned but not if the height of the anemometer on the mast is changed. A station may therefore be associated with several wind numbers.
Within the MIDAS database, stations are uniquely determined by a source identifier which is an integer running from 1 upwards. All observations in the database are stored with their source identifier.
It can be seen that under the present organisation of the UK observing networks there is no unambiguous definition of a station. At any one time there is a network of synoptic stations identified by WMO numbers, a network of climate stations identified by DCNN and rainfall stations identified by rainfall number. However, the longer the historical record, the more complex the picture becomes and the less easy it is to define what collection of observations should be regarded as originating from a single station. The following rules define the stations stored in MIDAS labelled by their unique source identifier:
The following sections list the parameters contained in various messages and reports. The second column of each list gives the precision to which each parameter is observed and reported. It must be noted that not all elements are reported from every station.
The international SYNOP message format has been used for the real time transmission of synoptic weather observations for about 50 years. Today it is used at some 200 or more Met Office or auxiliary UK stations for observations made at hourly, 3-hourly, 6-hourly or irregular intervals. The hour of the observation is contained in the message header, rather than in the message itself, and there is no facility to specify the time of the observation to the nearest minute. As noted in Section 2 above, the UK practice for many years has regarded the actual observation time as HH-10.
The following from the SYNOP message are stored in MIDAS:
|Mean sea level pressure||0.1 hPa|
|Station level pressure||0.1 hPa||High level stations only|
|Air temperature||0.1 C||Whole degrees pre 1982|
|Dew point||0.1 C||Whole degrees pre 1982|
|Wind speed||1 knot||10-minute average, HH-20 to HH-10|
|Max. gust speed||1 knot||10-minute max. gust speed, HH-20 to HH-10|
|Wind direction||10 degrees||10-minute average, HH-20 to HH-10|
|Visibility||1 m||Conversion from code|
|Cloud amount||1/8||Low/medium/high/total. 9 = sky obscured|
|Cloud height||1 dam||Low/medium/high/total. Conversion from code|
|Past weather 1||Code||Post 1982.|
|Past weather 2||Code||Post 1982.|
|State of ground||Code|
The real time exchange of hourly rainfall accumulations between European countries is achieved using the SREW message format (Synoptic Rainfall Europe West). As for SYNOPs, the UK practice regards the observation period as ending at HH-10, therefore the reported rainfall is for the period HH-70 to HH-10. Accumulations are reported to the nearest 0.1 mm and there is an indicator for trace.
Synoptic observations for aviation purposes use the METAR code. Most stations report hourly.
The following from the METAR message are stored in MIDAS:
|Altimeter pressure (QNH)||1 hPa|
|Air temperature||1 C|
|Dew point||1 C|
|Wind speed||1 knot||10-minute average|
|Wind direction||10 degrees||10-minute average|
|Runway visual range|
|Cloud amount||1/8||Low/medium/high. 9 = sky obscured|
|Cloud height||1 m||Low/medium/high|
|Maximum gust speed||1 knot||10-minute, HH-20 to HH-10|
Hourly Climate Messages (NCM), produced from automated systems such as ESAWS, SAMOS and CDL, are transmitted in real time in the same way as SYNOPs and SREWs, and like those observations the HCM covers the period HH-70 to HH-10. The following from the HCM message are stored in MIDAS:
|Mean hourly wind direction||10 degrees|
|Mean hourly wind speed||1 knot|
|Direction of maximum gust||10 degrees|
|Speed of maximum gust||1 knot|
|Time of maximum gust||HHmm||To the nearest minute|
|10 cm soil temperature||0.1C|
|Global irradiation||W hr/m2|
National Climate Messages (NCM), produced from all Met Office and many auxiliary stations, are transmitted in the same way as SYNOPs and SREWs. A full message is sent at the nominal time of 0900 each day and most stations also send an abbreviated message at 2100. Note that, as for SYNOP and SREW, the periods of the observations start and end at 10 minutes to the main hour, i.e. 0850 and 2050.
The following from the NCM message are stored in MIDAS:
|Grass minimum temperature||0.1 C||0900-0900 or 1800-0900 or one hour before sunset to 0900, depending on local practice|
|Concrete minimum temperature||0.1 C||0900-0900 or 1800-0900 or one hour before sunset to 0900, depending on local practice|
|State of ground||Code||At 0000, 0300, 0600, .. 2100|
|State of concrete slab||Code||At 0900|
|Maximum temperature||0.1 C||0900-2100 & 2100-0900, or 0900-0900|
|Minimum temperature||0.1 C||0900-2100 & 2100-0900, or 0900-0900|
|Rainfall accumulation||0.1 mm||0900-2100 & 2100-0900, or 0900-0900|
|Sunshine||0.1 hr||Reported at 0900 for previous 00-24 day|
|Soil temperature||0.1 C||At 30 cm and 100 cm. At 0900|
|Days of hail, thunder, gale, snow||Code||Reported at 0900 for previous 00-24 day|
|Depth of snow||1 cm||At 0900|
|Depth of fresh snow||1 cm||0900-0900|
Climate stations normally make one observation a day, though over the years practices have changed and twice or three times a day observing was more common than it is today. The paper form, covering a months observations, has always been the usual method of recording though electronic methods are beginning to become more widespread. The Form type in current use is Met Form 3208b which contains up to 31 daily observations. The observation time at most stations is 0900, though a few observe at 1000.
The following from the 3208b form are stored in MIDAS:
|Cloud amount||1/8||At 0900. 9 = sky obscured|
|Wind direction||10 degrees||At 0900|
|Wind speed||1 knot||At 0900|
|Present weather||Code||At 0900|
|Air temperature||0.1 C||At 0900|
|Wet bulb temperature||0.1 C||At 0900|
|Minimum temperature||0.1 C||0900-0900|
|Maximum temperature||0.1 C||0900-0900|
|Grass minimum temperature||0.1 C||Usually 0900-0900, but may be 1 hour before sunset to 0900, depending on local practice|
|Concrete minimum temperature||0.1 C||Usually 0900-0900, but may be 1 hour before sunset to 0900, depending on local practice|
|Soil temperature||0.1 C||At 10, 20, 30, 50 100 cm. At 0900|
|State of ground||Code||At 0900|
|Depth of snow||1 cm||At 0900|
|Rainfall accumulation||0.1 mm||0900-0900|
|Run of wind||1 knot||0900-0900. Integrated speed*time reported in km|
|Days of snow, hail, thunder, gale||Code||0000-2400. Note: though reported many climate stations do not keep 24 hour watch.|
Where the station observes at a time different from 0900 (0700, 0800 or 1000) read that time in the table above in place of 0900. Where a daily observation or a sequence of daily observations are missing, for example because the observer is unavailable or the station is not manned 7 days a week, the rainfall accumulation for the first day after the break will represent a multi-day total; similarly the maximum and minimum values may represent extremes over several days. The correct interpretation of the readings will be made clear on the form and stored in MIDAS, where appropriate, with a day count greater than 1.
On the Met Form each line represents the observation for the given day; values of rainfall, maximum temperature, run of wind and sunshine are "thrown back" to the previous day, in the sense that, for example, the 24 hour accumulation of rainfall read at 0900 on the 10th is entered on the 9th because it is assumed that most of the rain fell on that day. This practice does not affect the way the data are stored in MIDAS as the database indicates the observation period by the two parameters: observation end time and hour count.
There were a great many different form types that preceded the 3208b, each allowing for the different observing practices that were in force at the time. MIDAS only contains the complete set of observed parameters since 1959; before that date only the main climate parameters (e.g. maximum, minimum and rainfall) are stored for selected stations and periods.
Stations of the rainfall network, for which rainfall accumulation is the only observation produce, return daily amounts on rain cards (Met Forms 7133, 7135). Where observations are made less frequently than once a day, multi-day accumulations are marked on the card and stored in MIDAS with a day count greater than 1. Many stations in England and Wales are operated by the Environment Agency who may also return keyed data on tape to the Met Office.
Many stations with autographic or other recording instruments provide analyses of hourly values for climate purposes. The data are usually returned on Met Forms at the end of the month. The following are stored in MIDAS:
|Hourly rainfall from Met Form 7113|
|Rainfall accumulation||0.1 mm||HH-60 to HH-0|
|Rainfall duration||0.1 hour||HH-60 to HH-0|
|Hourly wind from Met Form 6910|
|Mean wind direction||10 degrees||HH-60 to HH-0|
|Mean wind speed||1 knot||HH-60 to HH-0|
|Direction maximum gust||10 degrees||HH-60 to HH-0|
|Speed maximum gust||1 knot||HH-60 to HH-0|
|Time maximum gust||HH-60 to HH-0, in tenths of hour|
|Hourly sunshine from Met Form 3445|
|Sunshine duration||0.1 hr||HH-60 to HH-0|
|Radiation from Met Form|
Various data logging systems are in use or have been used in the past to record meteorological data, and the recording media are returned at intervals (usually monthly) to Bracknell for downloading, processing and loading into MIDAS. Such systems include:
Surface CLIMAT messages are produced each month by WMO Climate Reference Stations of which there are currently 20 in the UK and some 400 world wide. The messages contain climatological summaries for the month made by the observing station or the responsible National Meteorological Service using all data available locally. The following is stored in MIDAS
|Monthly mean msl pressure||0.1 hPa||Using 3-hourly observations 00,03|
|Monthly mean station level pressure||0.1 hPa||Using 3-hourly observations 00,03|
|Monthly mean air temperature||0.1 C||Using 3-hourly observations 00,03|
|Monthly mean vapour pressure||0.1 hPa||Using 3-hourly observations 00,03|
|Monthly mean daily max. temperature||0.1 C|
|Monthly mean daily min. temperature||0.1 C|
|Total precipitation for month||mm|
|Total sunshine for month||Hours|
|Days precipitation > various limits||No.|
|Days max. temperature >/< various limits||No.|
|Days min. temperature < various limits||No.|
|Days snow depth > various limits||No.|
|Days wind speed > various limits||No.|
|Days visibility < various limits||No.|
|Days of other phenomena||No.|
|Monthly extremes and dates|
|Days of missing data|
At manned or part-manned stations where an automated system such as SAMOS has been installed, Check NCM messages are normally transmitted in real time at 0900 each day. The message contains observations of air maximum, air minimum, grass minimum and concrete minimum from liquid-in-glass thermometers, and 24-hour rainfall accumulation from a 5 inch gauge. These observations, made with instruments used from climate observing for over 100 years, allow long-term comparisons to be made with the modern instrumentation, such as ERT and tipping bucket gauges, which are required for automation. A paper record of the Check NCM messages is kept at each station and returned at the end of the month since 1997.
Routine maintenance at fully automated sites is undertaken by caretakers, perhaps once or twice a week, who in addition to cutting the grass and checking and cleaning the instrumentation, make check readings of air maximum, air minimum, dry bulb and wet bulb minimum from liquid-in-glass thermometers, and rainfall accumulation from a 5 inch gauge. Readings are recorded on paper and returned at the end of each month.
Water equivalent of snow
Designated stations, mostly manned Met Office stations, make daily measurements of the water equivalent of any snow that is lying. 3 samples are normally taken at different locations around the enclosure from which the mean is calculated. All measurements are recorded on paper and returned at the end of the month between November and April.
There is a convention, followed by the human observer, that where a value is read as half way between two points on the observing scale it is "rounded to the odd", except for pressure which is rounded down. A value falling half way between 13.4 and 13.5 would be reported as 13.5.
Units, accuracy and precision
Temperature was recorded in degrees Fahrenheit before 1961 and in degrees Celsius after that date. The precision of the measurement in the meteorological report has varied over the years:
All temperatures in MIDAS have been converted to Celsius and are stored with a precision of 0.1C. The J-descriptor indicates the original units of observation.
Instruments are calibrated on receipt from the manufacturer and at intervals after deployment by the QA Lab at Bracknell with the aim of ensuring that the accuracy of the measurement is better than 0.2C. Traceability to national standards has been maintained since 1858 and corrections determined from the calibration are applied to most observations made with the instrument. For the large majority of observations, MIDAS contains only the corrected value, however, at some climate stations in the past the normal practice was to record the original measurement on the meteorological form and only to apply the correction to the monthly mean. Reference to the original records stored in the archives is necessary to determine the observing practice at any particular station.
A liquid-in-glass thermometer, having a surrounding glass sheath, has been the normal means of measuring temperature since the earliest days of observing. The liquid is generally mercury, except for minimum thermometers in which ethanol is used. Maximum and minimum thermometers are reset after the reading has been taken, by a vigorous shake in the case of the maximum, and by tilting upright in the case of minimum.
Electrical resistance thermometers
Electrical thermometry is now in widespread use; its main virtue lies in the ability to automate the measurement process. Electrical resistance thermometers (ERT) first came into regular use in the early 1980s at the fully automatic SAWS stations, and since then have been introduced at all Met Office synoptic stations. The instrument measures the resistance of platinum which depends on temperature according to a quadratic relationship. As is the case for liquid-in-glass thermometers, calibration is performed at regular intervals by the QA Lab. There is no means of determining from the data stored within MIDAS whether an ERT or a liquid-in-glass thermometer was used to make the measurement, though this information may be found in the archived station documentation.
Screen level temperature
The standard exposure for thermometers for measuring the dry, wet-bulb, maximum and minimum temperatures is at 1.25m above the ground in a louvered white screen of wooden construction. This design allows a free circulation of air around the thermometers while shielding them from precipitation and external radiation.
Grass and concrete minimum
The grass minimum temperature is the lowest temperature reached overnight by a thermometer freely exposed to the sky with its bulb just touching the tips of short grass (25 to 50 mm above the ground). Normally the thermometer is exposed at the last hour before sunset and the reading is taken next morning. However, at stations where an observer is not available near sunset, such as at Ordinary Climatological Stations, the thermometer is often exposed throughout the day. When snow covers the ground the thermometer should be supported immediately above the surface of the snow without actually touching it, though this is only possible at manned stations. Doubtful readings, such as might occur where snow falls overnight, are not normally reported. Long grass and other characteristics of a poorly maintained site will cause inaccurate measurements.
The concrete minimum thermometer is exposed at the centre of, and in contact with, a concrete slab which should be cleared of any snow. At some climate stations in the first half of this century readings were taken over bare soil but these values have never been stored in MIDAS.
At many stations temperatures below the surface are measured at various depths. The depths used today are 5, 10, 20, 30 and 100cm, although measurements are not necessarily made at all these depths at a station and exceptionally measurements may be made at other depths. When imperial units were in general use, typically before 1961, the normal depths of measurement were 4, 8, 12, 24 and 48 inches.
Liquid-in-glass soil thermometers at a depth of 20 cm or less are unsheathed and have a bend in the stem between the bulb and the lowest graduation. At greater depths the thermometer is suspended in a steel tube and has its bulb encased in wax.
Sources of error
Common errors in the measurement of temperature include:
There may be slow changes in the characteristics of the thermometer, but these are reduced to a minimum by regular maintenance and calibration.
To be completed.
Units, accuracy and precision
In MIDAS humidity is stored either as wet bulb temperature or as dew point to a precision of 0.1C. WMO accuracy requirements for wet bulb temperature are +0.2C or an equivalent +3-5% in derived values of humidity. Measurements of dry and wet bulb temperatures are reported to the nearest 0.1K and when units of measurement were Fahrenheit reports were given to the nearest 0.1F.
Wet bulb thermometer
Almost all observations of surface humidity have been made using a wet bulb thermometer though some recent automatic stations have a relative humidity sensor fitted. The wet bulb is exposed alongside the dry bulb in the meteorological screen with no additional ventilation other than that provided by the natural flow of air. The wet bulb is kept moist through the capillary action of water up a muslin wick. Where ERT thermometers are used, such as in any automatic station, the siting of instruments is the same. On occasions when air temperatures fall below zero, the wet bulb should be covered by a thin film of ice, this being ensured by the observer at manned sites each time a reading is taken. Conversion to a dew point uses values of saturated vapour pressure with respect to a ice covered surface.
Sources of error
Common sources of error in the measurement of humidity include:
Units, accuracy and precision
Horizontal wind is a 2-dimensional vector and is usually reported as an averaged direction from which the wind is blowing and a speed. The maximum observed speed over a specified time interval and the time of occurrence may also be reported. The unit of speed used at UK stations is the knot (0.515 ms-1) and the unit of direction the degree. Where observations of wind speed from overseas stations are in ms-1, values are converted to knots before storage. The data are reported to the nearest knot and 10 degrees in the SYNOP and HCM messages and on Form 6910. 1-minute DALE data are recorded to the nearest 0.1 knot and 1 degree.
The accuracy requirements, which are met in most instances by the current synoptic network, are for speed to be measured within 1 knot or 10%, direction within 5 degrees and gust speeds within 10%. There are, however, larger errors in near calm conditions where the Munro anemometer is used (see below).
Speed and direction are averaged separately which introduces a slight overestimate of the mean vector. The calculation of mean direction has to take into account the cyclical nature of the measurement between 0 and 360 degrees. Where the sampling rate is sufficiently high, the maximum gusts are normally calculated from 3-second averages of speed and this is the case at most SAMOS stations today. A 1.5-second averaging period was used in early versions of SAMOS, it is still used at all ESAWS stations and it is the typical averaging period obtained from an anemograph trace. Metadata describing what averaging periods have been used at stations at a given date has been poorly recorded over the years. Where wind speed and direction are recorded on an anemograph, mean values and gust speeds are estimated by human analysis. Estimates of average speed from a continuously changing dial, which are subject to large error, are made at a few stations in the supplementary network. This observing practice was more common in the past.
10-minute averaged winds reported in the SYNOP message are for the period HH-20 to HH-10. Hourly mean winds in the HCM message and gusts reported in the SYNOP message are for the period HH-70 to HH-10. Some Ordinary Climatological Stations report run of wind from an anemometer on a 2 m mast. This is converted to a 24-hour mean wind speed in knots for storage in MIDAS.
The Munro cup anemometer and vane have been the basic instrumental method for measuring wind in the UK for many years. Traditionally it has been connected to an anemograph and/or dial, but with the advent of automatic systems it has been attached various processing devices, e.g. the DALE logger. The cup has a large inertia and therefore has a relatively slow response time, but of more concern is the high start-up speed, which in the case of the Mk4 is about 6 knots. In 1998 a project began for the replacement of Munros by a lightweight anemometer with better response characteristics.
Other wind measuring methods
Visual estimates of wind have been made a number of stations in the supplementary network for many years and the observations have been reported as 10-minute winds in the SYNOP message.
Dines pressure tube anemometers, having a better response at low wind speeds than the Munro, were installed at a number of stations in the early years of observing, but few of these instruments have remained in use in recent years.
SIESAWS systems use an orthogonal arrangement of pressure tubes to measure wind and have not proved reliable in the extreme environment in which they are sited. Light winds were not handled well but most now have a bias reset which the error.
Other methods of wind measurements, such as hot-wire anemometers, are not used in operational observing.
Exposure, effective height and corrections
Correct exposure of the wind instrument is essential for accurate measurement. The standard exposure is over level, open terrain at a height of 10m above the ground. Open terrain is defined as an area where the distance between the anemometer and any obstruction is at least 10 times the height of that obstruction. Measurement of wind in the direct wake of buildings or a row of trees are of little value and this should be borne in mind with city centre or other highly sheltered sites. If standard exposure is unobtainable the anemometer may be installed at a height greater than 10m. Whether or not such an adjustment is made, all anemometers are allocated an "effective height" which is defined as the height above open, level terrain in the vicinity at which mean wind speeds would be the same as those actually recorded by the anemometer. Various methods have been devised for the calculation of effective height. At stations where the effective height differs substantially from the actual height, corrections are applied to the 10-minute wind speed reported in the SYNOP message. No corrections are applied to any gusts measurements or to any hourly mean wind speeds.
Wind analysis and processing methods
Over the years many stations have provided climatological returns on Form 6910 containing hourly mean wind, speed/direction of the maximum gust in the hour and time of the maximum gust. The following rules have been applied for the analysis of the anemograph record:
1-minute speeds and directions from the DALE logging system are processed to provide mean hourly winds for the period HH-60 to HH-0. For a number of years after its installation the processing method attempted to mimic the Form 6910 analysis method, in particular, no instances of 1 knot mean speed were allowed. This practice was discontinued in the mid 1990s after which date simple hourly averages of the 1-minutes data were calculated.
Sources of error
Common sources of error in the measurement of wind include:
Units, accuracy and precision
The total amount of precipitation which reaches the ground over a stated period is expressed as the depth to which it would cover a horizontal surface. The reported accumulation of rainfall is the sum of the amount of liquid precipitation plus the liquid equivalent of any solid precipitation (that is the liquid obtained by melting snow or ice that has fallen). Similarly, the depth of snowfall is expressed by the depth of fresh snow covering an even horizontal surface. The unit of rainfall is the mm and amounts are measured and reported to the nearest 0.2 mm, and where possible, to the nearest 0.1 mm. Snowfall is measured and reported to the nearest cm. Trace is reported where precipitation is observed during the period, but the amount would otherwise be reported as zero.
The inch was the unit of rainfall in the past and amounts were usually measured and reported to the nearest 0.01 in, but other levels of precision have been in use at a few stations (e.g. 0.005 in, 0.1 in). Inches were the sole unit of measurement in the 19th century and mm the sole unit since 1970, but in the long period between these dates the reporting practice varied from station to station. All rainfall accumulations in MIDAS have been converted to mm and are stored with a precision of 0.1 mm. The J-descriptor indicates the original units of observation.
The achievable accuracy of precipitation measurements with raingauges in current use is about 5% though errors will be larger at exposed sites.
An ordinary funnel-type raingauge has been in use for all manual measurements since the earliest days of observing. The design has varied over the years but today the Met Office strongly encourages conformity in order to maximise comparability of readings across the network. The standard design has a rim of diameter 5 in (127 mm) standing 12 in (30 cm) above the ground. Raingauges based on the standard design are adapted to meet specific needs; there is a version having a capacity to hold a large volume of rain which is used in remote sites where readings may only be taken once a month. Exposure of the gauge should be on open ground distant from the effects of sheltering objects. At a few windy sites, established a number of years ago, there may be a surrounding turf wall of diameter 3 m and height 30 cm which shields the gauge from the extreme effects of strong winds. Systematic differences as large as 12% have been noted between an unsheltered gauge and one within a turf wall. It is not the present practice to build turf walls at new stations.
Tipping bucket raingauge
The tipping bucket raingauge is particularly suitable for the automation of rainfall measurement but it does not perform well in freezing conditions. It is based on a small container or bucket which, when filled with rain, tips and empties while recording the event electronically. The size of bucket used widely across the UK network holds 0.2 mm of rain. A larger 0.5 mm bucket is in use at some Environment Agency stations though only those at some remote or high level locations send readings to the Met Office. Bucket tip times may be stored directly by a data logger (such as the Met Office SSER system in use since the early 1980s) or converted locally to a rainfall accumulation by systems such as SAMOS, ESAWS or CDL.
Tilting siphon raingauge
The tilting siphon raingauge produces an autographic record of rainfall accumulation from a pen attached to a float in the rainfall chamber of the instrument. The rainfall chart is analysed to give hourly accumulation as well as duration of precipitation where this exceeds 0.1mm/hour.
The only sub-hourly data stored in MIDAS originate from Met Office SSER systems. The values are the tip times to the nearest minute from a 0.2 mm tipping bucket raingauge and span the period 1982 to the present. Sub hourly data from Environment Agency loggers are archived locally by the Agency.
The earliest hourly accumulations of rainfall stored in MIDAS date back to 1949, but it was not until the 1960s that the network of stations reporting hourly began to increase. Some autographic rainfall records from before the war exist in the archives but these have never been digitised.
Hourly rainfall from a station may be observed and reported in one or more of 4 different ways, and the observation may be made using one of several raingauges at the site:
A number of different practices for observing and storage of the climate data have become established over time depending on the equipment available at the station and the reports received:
12-hourly accumulations of rainfall are contained in the NCM which is sent in real time twice a day at 0900 and 2100 from most synoptic stations. NCMs have been in use since 1982 before which time daily climate readings were returned each month on 3259 forms. The NCM value is used for all climatological studies of rainfall on time scales of a day or more, and in the past the observation derived from the reference 5 inch gauge at the site. With the introduction of automated systems, as noted above, the tipping bucket gauge has become the reference gauge. This also applies to fully manned stations where a SAMOS system has been installed to simplify the manual observing process. All automated sites retain a 5 inch gauge, but it takes the role of a check gauge, which is used by the observer or the caretaker for additional readings, taken once a day, or perhaps less frequently, for checking purposes.
Daily accumulations are the starting point for climatological studies of rainfall. They will originate from:
In addition, where an automated system is in operation, check readings are made by the observer, or if the station is fully automatic by the caretaker, using the 5 inch gauge. The values are not stored in MIDAS.
From the earliest days of observing in the UK, the standard period for measuring daily rainfall has been 0900-0900. However, at any time a few stations have departed from this practice; for example, some climate stations open today take the observation at 1000 and others, particularly in the first half of the 20th century, made observations at clock time.
The current dense network of rainfall stations relies heavily on contributions from observers whose profession is not meteorology; they may work voluntarily or for an employer in a weather related field. There is no guarantee that observations will be taken every day and indeed an infrequent observing schedule may be agreed when a station opens. It is also possible that the observing time of 0900 is not rigorously adhered to, particularly at remotes sites or where several gauges are read by one observer. Many observed rainfall totals are valid for periods of more than one day; a reading may be taken at 0900 on Monday covering 3 days over the weekend, or at remote sites, it might cover a complete month. In each case, readings for multiple-day accumulations of rainfall should be taken at 0900. All original values are stored in MIDAS with the day count set to the appropriate value.
Depth of fresh snow and depth of lying snow are reported daily at 0900 at synoptic and climate stations and hourly at synoptic stations when snow is lying. Such measurements should exclude the effects of drifting or blowing snow. All reports of rainfall accumulation should include the liquid equivalent of any solid precipitation that has fallen in the period. At only a very few sites is the raingauge heated to improvement the measurement in snow. The effects of snow are a major source of error in the measurement of rainfall, see below.
Sources of error
Common errors in the measurement of rainfall accumulations include:
To be completed.
Sources of error
The Campbell-Stokes recorder
The routine measurement of sunshine was started in London, in 1853 by J.F.Campbell. His instrument was modified in 1880 by Stokes into the ubiquitous Campbell-Stokes recorder which is still the official sunshine recording method in UK. The instrument has a mounted glass sphere which focuses the suns rays onto a thick card, burning a hole when the sun is shining. The passage of the sun across the sky translates into a linear burn pattern along the card which may be analysed by the observer to give measurements of sunshine duration. Most stations report hourly or daily (00-24) amounts of sunshine although health resort stations are an exception, reporting for the period 1700-1700.
Sources of error
Units, accuracy and precision
Visibility is reported in m or km and is stored in MIDAS in dam. In the SYNOP message a non-linear code is used giving a reporting precision of 30m (30m to 100 m); 100m (100 m to 5 km); 1km (5 km to 30 km) and 5 km (30 km to 70 km). There is a further coarser reporting code for use where there are few visual reference points which is principally used at sea. The accuracy requirement for observations of visibility from the synoptic network is +10%. Where visibility is measured at climatological stations the accuracy achieved is generally less that this value.
Visibility estimated by the observer
Visibility is defined as the greatest distance at which an object in daylight can be seen and recognised, or at night could be seen and recognised if the illumination was raised to daylight levels. Observations should be made at ground level not from observation towers or roof tops. The long standing method of observation has been estimation by the observer using known fixed reference points, such as trees or buildings, which stand out well against the background. Each reference point should subtend an angle of at least 0.5o at the eye. Estimation of visibility at night is prone to greatest error and should, ideally, by performed with the aid of suitable fixed lights. Visibility estimates on airfields, where accuracy is of particular importance, are often aided in this way. On occasions when the visibility varies in different directions, the minimum value should be reported in the main part of the message and this is the value stored in MIDAS. The guidance to observers at coastal stations states that only visibility over land should be reported; any differing values over the sea being noted in the remarks column of the weather register though it is not clear how closely this practice is followed at voluntary stations.
Vertical visibility may be reported from airfields in the SYNOP and METAR messages to a precision of 100ft. It is measured when the sky is obscured and the horizontal visibility tends to be very low. It is a visual measurement for which a nearby object of known height (and intermediate points of known or estimable height) may still be visible. In daylight the timed ascent of a pilot balloon rising vertically, at known rate, may also be used. It is an optional measurement, not being reported where reference objects are absent.
Runway visual range may be reported from airfields in the METAR message. It is defined as the maximum distance in the direction of take-off or landing at which the runway, or specified lights or markers along the runway, can be seen from a point above its centre line. The height of this point is taken to be 5m above ground level, corresponding to the average eye-level of pilots at touchdown. The observation is usually made by personnel belonging to the airfield authorities such as the Fire Service. Where there are two runways, for instance as at Heathrow, two measurements may be reported.
Hand-held Gold visibility meters have been available for many years to aid the estimation of visibility, but it was not until the 1990s that visiometers have been widely installed at stations in the synoptic network for the automation of the measurement. Visiometers measure the transmissivity of a sample volume of air, expressed as the Meteorological Optical Range (MOR), which is converted for observational use into a visibility and coded directly into the SYNOP message at automatic stations. Where the station is manned the visiometer measurement may be altered by the observer if, in his judgement, it is in error or unrepresentative of the conditions obtaining at the time.
Present and past weather are only reported from a manned station, though present weather sensors are under development.
A weather group is reported only when there is or has been significant present or past weather during the period covered by the observation (present weather ww > 03 or past weather W1 > 2 or past weather W2 > 2). Where there is no significant weather to report, the group is omitted and there is a null or missing data indicator in the hourly report in MIDAS.
At stations where manual input is available at all hours of the day special practices apply to the first observation taken after a break. The reported present weather should be taken as best assessment by the observer bearing in mind that he/she may not have been at the station for the whole of the previous hour. Similarly the first of the past weather values will be made in the same way. Where it is not possible to give a second past weather value a null value (/) will be reported.
Units, accuracy and precision
Cloud height is one of the remaining meteorological parameter still measured in the imperial units. It is reported in code form for each of the cloud groups in the SYNOP message to a precision of 100ft (100ft to 5000ft), 1000ft (5000ft to 30000ft) and 5000ft (above 30000ft). Heights are converted to decametres (dam) for storage in MIDAS. Cloud amount has been measured in eighths since 1949. Cloud amounts before 1949 were measured in tenths but have been converted to eighths for storage in MIDAS. The J-descriptor for the observation indicates the observation units and the sign of the rounding error where a conversion has taken place. In this way the original value as observed may be recovered.
Manual observations of cloud
A complete observation of cloud containing amount, type and height of cloud base is available only from manned stations. The first automatic weather stations were not equipped to provide any cloud information, however, more recently some SAMOS stations operating automatically provide measurements of cloud height from a Laser Cloud Base Recorder (LCBR).
Manual observation of cloud height may be purely visual or be made with the assistance of measuring devices. Pilot balloons, with a lantern attached at night, may be timed to the point of entry into the cloud, and reports from aircraft are a valuable additional source of information at airfields. At some stations, especially airfield sites, cloud searchlights were used at night. These throw a spot of light on the base of the cloud at a height which can be measured by triangulation. Various cloud base recorders for measuring the height of cloud directly above the instrument have been used at times in the past. The Mark 3a (nodding beam, and the main one in use) and 3b (vertical beam) could measure heights up to 4000ft but the accuracy above 1000ft was not good. More recently, these have been replaced by LCBRs (see below).
In the report however there is no way of knowing whether the observer measured or visually estimated the cloud height - this information is noted only in the hand written daily register.
(NB Aircraft reports of cloud base and tops are reported in a separate message format but they are not stored in MIDAS.)
Laser Cloud Base Recorder (LCBR)
Where Laser Cloud Base Recorders have since been installed, measurements are available of cloud amount and height at different layers up to a maximum of 25000 ft. Cloud amounts are obtained by processing data over time and are of limited accuracy; orographically forced cloud, not advected over the instrument, will not be recorded. At the manned stations the LCBR is used as an observing aid, while at automatic stations amounts and heights with missing cloud types are reported in the SYNOP message and stored in MIDAS. The ix indicator shows that the observation was made automatically and is also stored in MIDAS.
MIDAS is a relational database residing on the Met Offices General Purpose Computer Server (GPCS) at Bracknell. The description of the structure of MIDAS, the relationships between the tables and the attributes of each table may be found in the MIDAS Handbook. This section gives details of how the observations described in the sections above are stored in the database and should be read in conjunction with the Handbook.
MIDAS is not a real time database but every attempt is made to store data as soon after the time of receipt as is practically possible. Synoptic messages (SYNOP, METAR and SREW) and climate messages (NCM and HCM) arriving in real time at Bracknell are stored in the real time synoptic database, the MetDB. Programs run twice daily to load these data into MIDAS:
0400: SYNOP, METAR, SREW, NCM and HCM for 0000 2300 the previous day plus any late messages or correction to messages for earlier dates
1000: SYNOP and SREW for 0000 0900 plus NCM for 0900
Monthly climate data that derive from forms, disks or tapes are loaded as soon as possible after they have been processed. This may be long after the observation time because of delay by the observing station in making the return. Data may not be received from some rainfall stations until 6 months or more have elapsed. Automatic quality control is applied to real time data once a day after the 0400 loading has completed, similar quality control is applied to climate data as soon as possible after loading. Detail of the manual and automatic quality control procedures are given in section 7.
The synoptic reporting system allows for the transmission of a correction message (COR) by a station if some part of the original message is seen to be in error. Where a COR message is received and the original has already been loaded into MIDAS, the corrected value overwrites the original on the next occasion that data loading takes place and any quality control flags are unset. Where observations are so much in error that it is not possible to store them in MIDAS, the data are sent to a rejects file where they are investigated by quality control staff at a later date.
Each observation record has attributes which identify the station, observation period, observed value, quality flags and instrumentation information. The following describes the attributes in use.
The station source for each observation record is defined by the unique source identifier (SRC_ID) and ID_TYPE, ID where ID_TYPE is WMO, DCNN, ICAO, RAIN, WIND, CLBD or LPMS. See section 3.6 for more details.
Observation time or period
MIDAS does not use the Database Management Systems representation of time. Where an observation is valid at a specific instant, that date/time is defined by the 4 attributes OB_YEAR, OB_MONTH, OB_DAY and OB_TIME. Where an observation is valid over a period of time that period is defined by the date/time at the end of the period (OB_YEAR, OB_MONTH, OB_DAY and OB_END_TIME) plus an hour count or day count (OB_HOUR_COUNT, OB_DAY_COUNT).
The Met Domain specifies the source of the observations in the record. Typically it will identify the message type by which the observations were received (e.g. SYNOP, NCM). Observations under one Met Domain may be spread across a number of tables; for instance, the daily observations from a climate station entered on Form 3208 are stored in up to 6 MIDAS tables under Met Domain DLY3208. Met Domain is a key attribute in the Source Capability table. Met Domains in use for UK surface records are listed below:
AWSDLY*: Elements from an automatic climate logger reporting
AWSHRLY*: Elements from an automatic climate logger reporting hourly
CHECK*: Daily check readings by caretaker (automatic station) or observer
DLY3208: Elements from Metform 3208, or, more generally, a daily climate return
DLY3259: Elements from Metform 3259 (pre 1982 NCM message)
DRADR35: Elements from Met O 1 Form R35 radiation/sunshine
ESAWRADT: Hourly radiation from ESAWS
ESAWSOIL: Hourly soil temperatures from ESAWS
ESAWWIND: Mean hourly wind from ESAWS
HSUN3445: Elements from Metform 3445 analysis of hourly sunshine
HWNDAUTO: Elements from automatic wind loggers
HWND6910: Elements from Metform 6910 analysis anemograms
METAR: Elements from FM 15 METAR message
MODLERAD: Hourly radiation values from MODLE
NCM: Elements from National Climate Message, including climate reports pre 1982
SREW: Hourly rainfall from SREW message
SSER: Rainfall amounts from SSER loggers
SYNOP: Elements from FM 12-VII SYNOP message (and state of ground from NCM)
WADRAIN: Daily rainfall amounts from rainfall network
WAHRAIN: Hourly rainfall values from analysis of TSR autographic records
WAMRAIN: Monthly rainfall amount from rainfall network
Where quality control is performed and changes are made to the observed value to correct for errors it is important that the original value is not lost. When a change is made to any value in a table record, the original record is stored with Version Number 0 and a new record is created for the corrected version having version number 1. Any further changes are made to version number 1, so that at any time no more than 2 versions exist, the original and the current best version. On initial storage all new records are given Version Number 1. A table will not have Version Number as an attribute if no quality control is ever performed on the observations (e.g. CLIMAT data, sub-hourly rainfall).
In many cases the observed value is stored as received though some processing of raw data or conversion from outdated measuring units may have been performed. The table below gives detail of the storage units, while all processing is described in section 5 above.
|Met element||Units||Met element||Units|
|Temperature (all)||0.1 deg C||Radiation flux||Watt m-2|
|Pressure||0.1 hPa||Sunshine duration||0.1 hour|
|Wind direction||deg true||Cloud amount||octas|
|Precipitation amount||0.1 mm||Cloud height||dam|
|Precipitation tip amount||0.001 mm||Precipitation duration||minutes|
|Snow depth||cm||Time of max gust||hour/minutes|
The following tables in MIDAS contain observational data. The different Met Domains and ID-types of the records in each table are listed. (assumes some cleaning up and rationalisation of what is in MIDAS at present)
|MIDAS table||Met Domain||ID-TYPE||Comments|
|RADT-EXTRA-OB||MODLERAD||DCNN||Contains optional radiation values that have not been reported since 1994|
|DRADR35||DCNN||Daily radiation from Form R35|
|MODLERAD||DCNN||Hourly radiation from MODLE|
|ESAWRADT||DCNN||Hourly radiation from automatic stations|
|AWSHRLY||DCNN||Hourly radiation from CDLs|
|NCM||RAIN||Daily precipitation amount from synoptic station (no 12-hour values reported)|
|DLY3208||RAIN||Daily precipitation amount from ordinary climatological station|
|DLY3208||LPMS||Daily precipitation from long period climate station|
|WADRAIN||RAIN||Precipitation amount from daily rainfall station|
|WAMRAIN||RAIN||Precipitation amount from monthly rainfall station|
|AWSDLY||RAIN||Daily precipitation from CDLs|
|CHECK||RAIN||Daily check gauge reading|
|SREW||RAIN||Hourly amount received in real time SREW report from synoptic station|
|NCM||RAIN||12-hourly amount received in real time NCM report from synoptic station|
|SSER||RAIN||Hourly amount calculated from SSER tip times|
|WAHRAIN||RAIN||Hourly amount and duration analysed from TSR. Reported on Form 7113|
|AWSDLY||RAIN||12 hour precipitation from CDLs|
|AWSHRLY||RAIN||Hourly precipitation from CDLs|
|SSER||RAIN||Tip times from SSER loggers|
|METAR||ICAO||Runway visual range from METAR report|
|NCM||DCNN||Soil temperature from a Principal Climatological Station reported in a NCM|
|ESAWSOIL||DCNN||Hourly 10cm soil temperature from an automatic station|
|DLY3208||DCNN||Soil temperature from an Ordinary Climatological Station|
|AWSHRLY||DCNN||Hourly soil temperatures from CDLs|
|HSUN3445||DCNN||Hourly sunshine reported on Form 3445|
|NCM||DCNN||12 or 24-hour max/min temperatures from a Principal Climatological Station reported in a NCM|
|DLY3259||DCNN||As above but reported on Form 3259 (only in archive segments)|
|DLY3208||DCNN||24-hour max/min temperatures from an Ordinary Climatological Station|
|DLY3208||LPMS||24-hour max/min temperatures from a long period climate station|
|AWSDLY||DCNN||12 or 24-hour max/min temperatures from CDLs|
|NCM||DCNN||Various daily climate values from a Principal Climatological Station reported in a NCM|
|DLY3259||DCNN||As above but reported on Form 3259 (only in archive segments)|
|DLY3208||DCNN||Various daily climate values from an Ordinary Climatological Station|
|SYNOP||WMO||Hourly synoptic observation received in SYNOP report|
|AWSHRLY||DCNN||Hourly synoptic observations received from CDLs|
|METAR||ICAO||Half-hourly synoptic observation received in METAR report from station in aviation network|
|DLY3208||DCNN||Daily synoptic report at 0900 or other hour from an Ordinary Climatological Station|
|HWND6910||WIND||Mean hourly wind and gust from analysis of anemograph record reported on Form 6910|
|HWNDAUTO||WIND||Mean hourly wind and gust from wind logging equipment (DALE)|
|ESAWWIND||WIND||Mean hourly wind and gust from automatic station|
|DLY3208||DCNN||24 hour run of wind from an Ordinary Climatological Station|
|AWSHRLY||WIND||Hourly mean wind from CDLs|
Where nn is the database segment: LH, CE, SW, NT, SS, SN, IR (Regional tables)
Some CDL stations are operated for commercial purposes only. The data are stored in MIDAS under met domain AWSHRLY or AWSDLY in the same way as for other CDL stations, except that different identifier types are used: CLBD in place of DCNN, CLBR in place of RAIN and CLBW in place of WIND.
Quality control at the point of observation
Basic quality control is performed at each observing site which ensures that some errors are trapped before being transmitted. In the days before automation, the trained observer was required to check all his entries in the weather register, and these practices continue in the few islands of manual observing that exist today. Observing system software applies a range of checks to all reported parameters which ensure that no irregular values leave the site, however, the human observer, if present at the time of observation, may override many of the queries raised.
MetDB quality checks
All real time observations that are destined for MIDAS are stored first in the MetDB. These include SYNOP, NCM, HCM, SREW and METAR reports. Some range and self-consistency checks are performed on the values at the point of receipt in the MetDB and the associated flags are passed to MIDAS on data ingestion by setting the query flag equal to 1.
All data on ingestion to MIDAS undergo basic range checks. These do no more than ensure that the meteorological value does not lie outside long-term climatological extremes (taking no account of time of year or location). Those failing have query flag set equal to 2.
All UK surface data, with the exception of rainfall data, undergo a set of "QCL" checks soon after their ingestion into MIDAS. A level flag set equal to 1 indicates that QCL checks have been performed. These checks ensure:
Checks against neighbours
Automatic algorithms are applied to certain meteorological elements to ensure consistency with neighbours (areal checks). The elements checked are maximum and minimum air temperatures, grass minimum temperatures, 0900 air temperature, rainfall and sunshine. A buddy check of mean hourly wind is also performed Where there are no automatic checks of the climate elements basic manual checks are performed.
Manual quality control
With few exceptions, all quality control flags raised by the checks are scrutinised by trained meteorological staff in the QC Teams at Bracknell and Edinburgh. It is their decision whether the status flag is set or not; where errors are found corrections are supplied and where values are missing estimates are made.
The final sweep
To ensure no spurious values have escaped the processes described above, a final sweep through the observations is made in order to trap any remaining gross errors.
MIDAS Handbook, 1997. Maintained by Desktop and Databases branch.
Observers Handbook, 1982. Published by HMSO.
Guide to Meteorological Instruments and Methods of Observation, 1997. Published by WMO, WMO-No. 8.
Handbook of Weather Messages, Part III, 1979. Published by HMSO, Met O 920c.
Handbook of Meteorological Instruments, 1980. Published by HMSO, Met O 919.
United Kingdom Observing Networks, UKON-1, 1998. A user requirement for the surface synoptic network. Maintained by Observations Plans and Requirements branch.
United Kingdom Observing Networks, UKON-2, 1998. A user requirement for the climate network. Maintained by Observations Plans and Requirements branch.
United Kingdom Observing Networks, UKON-3, 1999. A user requirement for the wind network. Maintained by Observations Plans and Requirements branch.
United Kingdom Observing Networks, UKON-5, 1999. A user requirement for the rainfall network. Maintained by Observations Plans and Requirements branch.
United Kingdom Observing Networks, UKON-6, 1998. A user requirement for the surface radiation and sunshine networks. Maintained by Observations Plans and Requirements branch.