=========================================================== Archive of Lunar occultation observations (1623 to the present time) Dave Herald Canberra, Australia Dave Gault Sydney, Australia with the assistance of: Derek Breit Ken Coles David Dunham Brian Loader Jan Manek Kazuhisa Miyashita Robert Sandy Mitsuru Soma Updated September 2015 =========================================================== 1. Abbreviations 2. History of recording lunar occultations 3. Source of the observations 3.1 Ordinary occultations 3.2 Grazing occultations 4. Corrections to the data 5. Algorithm for auto-correcting historical observations 6. Corrections to grazing occultations 6.1 Restoring Lost Data 6.2 Comparing Original reports to the Archive 6.3 New historic Graze Reports 6.4 Updating Geographic Coordinates and Datum 6.5 Correcting Coordinates 6.6 Correcting event times 7. Statistics on the corrections made 8. Basis of the reductions 8.1 Ephemeris basis 8.2 Star positions 8.3 Precession 8.4 Earth rotation 8.4.1 deltaT 8.5 Site coordinates 8.5.1 Conversion to WGS84 latitude/Longitude 8.5.2 Conversion of MSL altitudes 8.5.3 Polar motion 8.5.4 Datums for RGO sites 8.5.5 Atmospheric refraction 8.6 Lunar limb corrections 9. Comments on specific data fields 9.1 WDS double star component identifier Col 26 9.2 Phenomena Col 27 9.3 Personal Equation Col 30-33, 34 9.4 Light level at event Col48-52 9.5 Temperature Col 57-59 9.6 Graze identifier Col 75-78, 79-81 9.7 Vertical datum code 9.8 Closest town or landmark Cols 128-177 9.9 Observer names Cols 189-213 9.9.1 Names in the RGO dataset 10. References This version updates the Archive with recent observations (up until early 2015) and some previously unreported observations dating back to 1983. The reductions have been computed using the DE430 ephemeris, with the lunar limb corrections being derived from the Lunar Reconnaissance Orbiter - Lunar Orbiter Laser Altimeter [LRO LOLA]. =========================================================== 1. Abbreviations =========================================================== RGO - Royal Greenwich Observatory. HMNAO - Her Majesty's Nautical Almanac Office - part of RGO. ILOC - International Lunar Occultation Center, of the Japanese Hydrographic Department IOTA - International Occultation Timing Association. A global organisation of mainly amateur astronomers =========================================================== 2. History of recording lunar occultations =========================================================== This archive is a nearly complete record of all reported lunar occultations. Prior to about 1940, there was no central collection of occultation observations; rather they were reported in the usual astronomical literature. The following, from RGO Bulletin 183 (1978), explains how centralised collection of occultation observations came about: Both E W Brown (1927) and D Brouwer (1938) stressed the importance of observations of occultations of stars by the Moon for the determination of the variation in the Earth's rate of rotation and the improvement of the fundamental constants of the lunar theory. They promoted a campaign of observation which led to a coordinated and systematic compilation and reduction of observations received from a world-wide distribution of observers. H M Nautical Almanac Office agreed to take over responsibility for the centralised compilation and reduction of occultation observations beginning in 1943... Apart from collecting observations made from 1943, HMNAO conducted a literature search. They also established forms for reporting observations, using a set of codes to represent a range of secondary information (such as the observer's estimate of the accuracy of the reported time). In 1981 ILOC took over the responsibility for collecting lunar occultation observations. They introduced a different set of codes for recording secondary data. ILOC also introduced electronic reporting of observations - initially by way of a floppy disk, and later by email. ILOC continued in this role until 2008. In 2008 IOTA took over responsibility for collecting lunar occultation observations. The codes for secondary data were updated to reflect modern observing methods (such as video, and GPS time insertion). Handwritten reports were no longer accepted - with all observations being reported using email. =========================================================== 3. Source of the observations =========================================================== 3.1 Ordinary occultations The observations made before 1981 were published on microfiche in Royal Greenwich Observatory Bulletins 183 (1978), 186 (1981) and 192 (1984). Those microfiche were produced from computer files, with the data having been stored on magnetic tape. Electronic copies of the files on that tape have survived, and almost all observations before 1981 have been derived from those files. However the files that have survived apparently pre-date the production of the microfiche - as the microfiche includes data for several 10s of sites that is missing from the electronic files. As noted in RGO bulletin 183, for the period 1943 to 1970 observer names were not linked to specific observations in the electronic records. However in many cases that linkage could be deduced - and where this has occurred the names are flagged as having been included in this manner. The source of the historical observations is described in RGO Bulletin 186 as follows: Observations of occultations of stars by the moon have several important applications in dynamical astronomy. One of these is the determination of fluctuations in the earth's rotation over the past 300 years; another is the determination of the rotation of the fundamental stellar reference frame with respect to the ecliptic. For these two reasons, occultations from the 17th to 20th century were compiled and analysed... Before about 1670 the accuracy of timing occultations was severely limited by the use of altitudes of stars, rather than pendulum clocks... The sources of the observations are various, the occultations before 1861 were taken mainly from Newcomb's (1912) compilation which was supplemented by Martin (1969) from Newcomb's unpublished manuscripts. About 550 previously unpublished observations made at San Fernando, Spain, in the years 1773 to 1860 were added to this list. ... A further 110 occultation observations in the years 1700 to 1760 were gleaned from Historie de l'Acadamie Royale des Sciences avec les Memoires de Mathematique et de Physique, Paris and the Philosophical Transactions of the Royal Society, London, and about 70 observations in the 17th century were taken from the Annals Celestes du Dix-Septieme Siecle by A G Pingre [Ed by M G Bigourdan, Paris, 1901]. The work of locating and transcribing the occultation observations for the years 1861 to 1942 was divided between the Nautical Almanac Offices of the Royal Greenwich Observatory and the US Naval Observatory. For the years 1861 to 1898 the subject indexes of astronomical journals and annals of observatories were searched, and from 1899 onwards the Astronomischer Jahresbericht was used as the reference source. All the times have been converted to UT where this had not been done in the original publication. It may be noted that the observations collected by RGO was effectively limited to stars which were in the SAO catalogue, and did not include any occultations of planets. The earliest occultation observations are those listed in Newcomb's 'Researches on the Motion of the Moon'. The times of these events in the RGO dataset differ slightly from those in Newcomb; this is because RGO re-determined all the times, which were generally derived from altitude measurements. (Gordon Taylor - personal communication). The source of observations for the period 1981 to 2008 was electronic files from ILOC. For much of this period ILOC keypunched handwritten reports into a computer system. The reports were then converted into an archive format for storage and analysis. The data was held in three files - the observations, the site details, and the observer name details - linked by a system of site and observer codes. In around 2006 ILOC recognised that the system of codes to link the observations, sites and observers had become unreliable. This resulted in a reformatting of all past observations, with site and name details being included in each record (which is the approach adopted in the current archive.) The data for observations from 1981 to about 2001 was taken from these archive files. However for observations submitted after 2002 a computer error resulted in certain secondary information not being included in those files. For all such reports the observations in the report format were provided by ILOC, and converted into the current archive format. For observations in 2008 and later - observations have been submitted direct to IOTA in electronic form. 3.2 Grazing occultations Regular observations of grazing occultations commenced in the 1960's. These observations provided significant recording challenges. Traditionally occultations were observed from a relatively small number of fixed observatory locations - with it being appropriate to allocate a code for the observing location and the observer. However grazing occultations involve a relatively large number of observers, with no commonality of site locations between events. RGO maintained grazing occultations observations in a completely different dataset - although retaining their system of codes for sites and observers. Regrettably the magnetic tape containing the RGO graze dataset has been lost. However all the observations were published on microfiche in RGO Bulletin 192, and these were OCR'd by members of IOTA - European Section. ILOC incorporated grazing occultations in the same files as ordinary occultations. However the large number of sites required for grazing occultations created difficulties in maintaining the codes to link the observations with the site and observer. An additional source of grazing occultations was a collection by Mitsuru Soma (Tokyo Observatory). This collection mainly contained grazes observed from 1995. Another source of grazing occultations was a collection by Dave Herald as part of the software OCCULT. This was initially derived from the RGO, ILOC and Soma datasets, but also included a small number of events not reported elsewhere. T Van Flandern of the US Naval Observatory collected graze observations up to 1977. Some critical data files have been lost. The only extant file gives accurate times, but it does not provide accurate site co-ordinates (nor observer names or telescope details). The data from these sources was not always in agreement. For example, there are instances of an initial report being sent to one location, and a revised report being sent to another. This was compounded by the reports of many observed grazes having never been sent to either RGO or ILOC (or anyone else). Much time was spent resolving these issues. =========================================================== 4. Corrections to the data =========================================================== As might be expected of a dataset extending over almost 400 years, a range of errors were found to exist. Routines were developed to automatically identify and correct a range of errors. Many other errors were corrected after a manual review of all observations having a large residual. The manner of recording and reporting occultation observations has varied considerably over the almost 400 years of observations. For example, some of the old observations were reported in sidereal time; many observations were reported in local time (including up to present times). Before about 1940, observations were published in astronomical literature, with no central collection and analysis. When RGO became responsible for collecting the observations, they were submitted on handwritten forms, with subsequent key-punching. This continued when ILOC took over responsibility. A consequence of how the data has been collected is the inevitable presence of a range of data errors - errors which result in residuals much greater than would be expected from a 'missed' observation. Typical of such errors are: - Time errors of 1 minute are very common. Errors of 2, 5, 10 and 15 minutes are not unusual - Time errors of 1 hour - frequently associated with summer- time issues, or time zone corrections - Time errors of 1 day - sometimes associated with the local day being different to the UT day - Time errors of integral months and years - Star number errors, such as transposed digits, or specifying the incorrect catalogue identifier against the correct star number. Or observer misidentification of the star - Site coordinate errors - such as whether the longitude was East or West While many of these errors are 'observer' reporting errors, there is an additional source of error when the observations were data-keyed. Typical errors are confusion between 1, 2 and 7; 3 and 8; 6 and 8; 8 and 9; 0 and 8 - as well as transposed digits, and data keyed into an incorrect column. The largest time correction applied (1000 years) was almost certainly a data keying issue (with the millennium digit being dropped from the year 1943). For observations before about 1960 there is also the issue of matching the star number as used in the report to an entry in the SAO catalogue. Where an observation is infected with a _single_ data error, it is usually possible to deduce the correction with a high degree of certainty. In the case of an error in the time, the correction will have the effect of changing the residual from many 10's or 1000's of arcsecs to something typically less than 1 arcsec - with there being only one correction that will have this effect. And when considering the possibility that a different star was involved - it is rare for more than 1 star brighter than mag 9 to be within 5 arcsecs of the lunar limb. For these reasons the identification and correction of a single error was capable of automation, and a routine was developed for this purpose. The algorithm is set out below. Having processed the observations for corrections, all observations with large residuals were reviewed. In practice, the review mainly considered events with residuals greater than 100" - as those with smaller residuals usually had no obvious basis for correction (and if the residual was less than 10", the observation was almost certainly a 'bad' observation. This review looked for possible single corrections outside those considered in the automated check. It also considered the possibility of corrections to both time and star number. This frequently entailed running predictions to establish when the identified star was occulted at the particular site; or running predictions for the specified date (or a range of dates) and looking for a match between the event times and the predictions. Typically, when a 'solution' was found the source of the error also became apparent (such as transcription errors). Many corrections to site locations were determined by plotting in GoogleEarth the projected lunar limb onto the Earth's surface for a range of events observed from a location. The various limbs intersect in the region of the actual observing site. The site could then be identified if there was an established observatory (or previous observing location) at that location. For a small number of sites, the correct location was identified by contacting the observer, or someone who knew the observer. For example, RGO site 384 01 was active in the 1930's and 40's. The site location in the RGO files was located at sea off the coast from Dundee in Scotland. Contact with the Dundee Astronomical Society located a person who actually knew the observer and where he had lived - and was able to provide GPS coordinates for that location (which resulted in good residuals). For a small number of events, corrections were made to all of site location, star number and event time. Such instances were limited to situations where the site error was identified by having regard to all observations from a single site. Having made that correction, corrections to time and/or star number could be determined as for other events. This type of data correction inevitably involves the possibility of making false corrections. To minimise this possibility, the corrections to the time was limited to integral changes in the units of years, months, days, hours and minutes. Unless the correction was to convert sidereal time to UT, the seconds of an event have not been corrected. It must be noted that many observers in the latter part of the 20th century used 30-second stopwatches to record the event - with this giving rise to a real possibility of 30- second errors in their reported time. Indeed there are many unsatisfactory events in the dataset where a 30-second time error is strongly indicated. However those events have not been corrected, as there is too much uncertainty about the validity of such a correction. There were also a number of events where a 'desired' correction was consistent with some form of compliment to the time eg 19m 38.2s => 21m 21.8s. However no such compliment corrections were applied. A practical check of the reliability of the corrections arose late in the process. An observer in Brazil was able to compare the uncorrected ILOC data against original reports, and identify corrections - and these were then able to be compared to the corrections that had been previously applied to the dataset. Of the 50 events the observer identified as requiring correction: 45 were correctly corrected 2 were incorrectly corrected 2 were identified as being in error, but no correction found; and 1 event was not identified as being in error. From this it may be inferred that the rate of _incorrect_ corrections is of the order of 5%. Alternatively, about 95% of corrections made are valid corrections. To maintain the integrity of the dataset, the format includes both the corrected and uncorrected data. The nature of any corrections made is also indicated by a code. =========================================================== 5. Algorithm for auto-correcting historical observations =========================================================== All observations were reduced and scanned for possible errors, using the following algorithm: i. Correction is only investigated if the limb corrected residual is >4" (events marked as a graze were excluded.) ii. When searching for possibilities, compute the altitudes of the Moon and Sun for the date/time/location. If the Moon altitude is <0, or sun altitude>0, exclude the particular date/time/location combination. iii. Assume the listed time is correct. Identify all stars within 2" of the limb-corrected limb with a phase [D/R] corresponding to the event. Magnitude of stars limited as: elong Mag < 126 10 < 137 9 < 160 8 < 177 7 > 177 10 (to pick up events during a lunar eclipse) Bright Limb events - Mag 4. If more than one star is listed, take the brightest star. iv. If no match at (iii), vary the time of the event as follows: Minute: -20 to +20 minutes (Day and hour as recorded) Hour: -12 to +12 hrs (Day and Minute as recorded) Day: -6 to +6 days (Hour and minute as recorded) If the limb corrected residual is less that 2", correct the time accordingly. v. If no match at (iv), check for the major planets, Plus Jupiter satellites I-IV, and Saturn VI. Treat as a match if the limb-corrected residual is <5" [Note - for the major planets, the reduction algorithm assumes the reported time is corresponds to the 'interior contact' between the planet disk and the lunar limb.] vi. If still no match, redo a match to other stars, using the following magnitude limits: elong Mag < 126 12 < 137 11 < 160 10 < 177 9 > 177 13 (to pick up eclipses) Bright Limb events - Mag 5. =========================================================== 6. Corrections to grazing occultations =========================================================== The review process for grazing occultations was much more extensive than for ordinary occultations. This arose because of the various sources of the observations. Additionally, each graze was individually reviewed to flag observations that were clearly inconsistent with the known lunar profile. This review was initially conducted on the basis of the Watts charts for the lunar limb. However when the data from the Kaguya mission became available, the dataset was reviewed for a second time. [The extra detail available in the LRO LOLA data has not justified a further review.] 6.1 Restoring Lost Data A consequence of the loss of the magnetic tape(s) and subsequent OCR process, as detailed in section 3.2 above, was that for all graze observations in the period 1960 to 1975, not all data was successfully recovered. Lost data included; observer name, site name, and telescope characteristics. However copies of many of the original observations reports that were the source of the lost electronic data had been retained by a number of people. These reports were reviewed for whether they contained any of the lost data. This enabled restoration of most of the missing graze data in the period 1960-1975. 6.2 Comparing Original reports to the Archive The available original reports were also compared with the data in the archive. Any discrepancies were corrected in the Archive - subject to the need to correct data that is erroneous. 6.3 New historic Graze Reports The review of the original observation reports located many grazes that were not included in the Archive. [The most likely explanation is that the reports were never sent to RGO or ILOC.] All such reports were processed and added to the archive. A total of 290 new graze reports were found and created for the period 1960-2007. Lists of observed grazes published in the Occultation Newsletter were compared against the observations in the Archive. Over 100 successfully observed grazes were identified as not being reported. For some of these the observations were able to be retrieved by direct contact with the graze organiser. For the remainder a single-entry place- holder was inserted into the Archive, giving the basic details of the graze. Additionally about 70 successfully observed grazes which had not been reported at all, were uncovered. The data for those grazes, which extend back over more than 10 years, is not yet available. A single-entry place-holder was inserted into the Archive for each such event, giving the basic details of the graze. If in the future the graze is properly and fully reported, the intention is to remove the single-entry place- holder from the archive. 6.4 Updating Geographic Coordinates and Datum In some instances, the graze leader still retained the topographic map used to plan and report the observation. The graze leader was then able to re- determine WGS84 coordinates using Google Earth. If this was done the datum code was changed to code '10'. 6.5 Correcting Coordinates The nature of multiple site graze observations is that (especially before the ready availability of portable GPS units) observers were almost always positioned near road intersections and bends, and other geographic features that could be identified on survey maps. When the residuals of one or more observers appeared to be consistently in error, a plot of the observer's locations on GoogleEarth would often indicate both the existence of an error, and the likely solution. For example: - a multiple observer graze observed in the South African Transvaal region gave the site altitudes 350m. By plotting the sites in GoogleEarth, the correct altitudes were found to be 1350m. - In a multiple observer graze observed on the Florida coastline, one observer had bad residuals. A Google Earth plot showed that observer 40km out to sea. Finding and correcting a typographical error in the site longitude placed him back on the coast, with results consistent with the other observers. - All residuals of a multiple observer graze were found to be bad. The graze leader was contacted. Upon reviewing his original notes and map, it was discovered that the wrong cross-street had been used to determine the site coordinates of all observers. Correcting the coordinates restored the precision of the observations. 6.6 Correcting event times Due to the nature of a graze observation, an observer will typically observe a number of events in succession. An error in the time base means that all events will be in error by an equal amount when comparing the events with the observations of other observers. A time offset can be estimated and applied to all events that will correct the observation. =========================================================== 7. Statistics on the corrections made =========================================================== The following table lists the statistics of the corrections made. The table provides separate data for ordinary occultations and grazing occultations - reflecting the fact that the errors associated with grazing occultations tend to be different to ordinary occultations. The data is grouped into three time periods: before 1981 {the data collected by RGO}; 1981-1990 {the data collected by ILOC before electronic reporting became common); and after 1991 - where electronic data reporting became increasingly common. Ordinary Occults Grazing Occults All 1620 1981 1991 1620 1981 1991 Correction to: -1980 -1990 -2010 -1980 -1990 -2010 Date/time 2162 878 614 217 585 808 5264 Star 1654 2406 2787 79 164 69 7159 Site 450 817 496 1700 1265 3420 8148 Date/time + Star 98 145 103 0 1 0 347 Date/time + Site 0 44 7 13 9 2 76 Star + Site 1 44 10 0 1 0 56 Date + Star + Site 0 3 0 0 1 0 4 Total corrections 4365 4337 4017 2009 2026 4299 21054 Total events 168057 99592 126797 23279 17198 26769 461692 A large proportion of the Date/time corrections for grazes relate to Start and End timings that were recorded for graze events. Many of these times were not actual times reported by the observer, but instead were added later for various reasons, and many were out of sequence with the actual observations. Start and end times were corrected to restore the logical sequence. =========================================================== 8. Basis of the reductions =========================================================== Included in the dataset is a reduction of the observations, to facilitate the investigation of individual objects. Key elements of the reductions are: 8.1 Ephemeris basis JPL-DE430 8.2 Star positions All positions are taken from the XZ80Q [I/291]. This catalogue is a hierarchical compilation of Hipparcos-1 [I/239], Tycho-2 [I/259] and UCAC2 [I/289] catalogues. The catalogue includes all Hipparcos and Tycho-2 stars brighter than magnitude 12.0. UCAC2 positions are used in preference to Tycho-2 positions. 8.3 Precession Precession is computed using the IAU2006 model, ignoring the small periodic terms (which are insignificant for lunar occultation reductions). 8.4 Earth rotation GMST is computed using the Earth Rotation Angle and Equation of Origin, consistent with the IAU2006 model for precession but ignoring the small periodic terms (which are insignificant for lunar occultation reductions). Reported times were converted to Terrestrial Time by: - Observations after 1972. The relevant offset from TAI (which was adjusted via leap seconds during this period) was applied. - Observations 1961 - 1971. The relevant offset from TAI (which was subject to small frequent adjustments during this period) was applied. - Observations before 1961. TT time was derived by the application of the quantity deltaT. 8.4.1 deltaT The value of deltaT depends upon the lunar ephemeris (including the adopted value of tidal acceleration of the moon used in that ephemeris), the adopted values of the precession constants, and the stellar reference frame. The correct value of deltaT is one where the average residuals in longitude are close to zero. The values calculated using the DE430 ephemeris and the LOLA data for limb corrections resulted in small changes to those previously computed using the DE422 and DE423 ephemerides previously used for the reductions. For observations prior to 1700, deltaT was computed using the expressions of Espenak and Meeus (Five Millennium Canon of Solar Eclipses, NASA Goddard Space Flight Center, 2006), adjusted to correct for the different values of the lunar secular acceleration relevant to those expressions (-26.00 arcsec/century^2) and that inherent in DE422/DE423 & DE430 (-25.85 arcsec/century^2). The following table lists the values of deltaT used in the reductions. In use, the table is linearly interpolated to the date of the event. 1700.0 11.4 1800.0 17.9 1900.0 -1.3 1701.0 11.8 1801.0 17.7 1901.0 -0.1 1702.0 12.4 1802.0 17.5 1902.0 1.3 1703.0 13.0 1803.0 17.4 1903.0 2.7 1704.0 13.7 1804.0 17.2 1904.0 4.0 1705.0 14.3 1805.0 17.1 1905.0 5.3 1706.0 14.9 1806.0 16.9 1906.0 6.5 1707.0 15.6 1807.0 16.8 1907.0 7.5 1708.0 16.2 1808.0 16.7 1908.0 8.8 1709.0 16.8 1809.0 16.5 1909.0 10.2 1710.0 17.0 1810.0 16.4 1910.0 11.6 1711.0 17.0 1811.0 16.3 1911.0 12.9 1712.0 17.1 1812.0 16.2 1912.0 14.2 1713.0 17.1 1813.0 16.0 1913.0 15.4 1714.0 17.2 1814.0 15.9 1914.0 16.7 1715.0 17.2 1815.0 15.7 1915.0 17.9 1716.0 17.2 1816.0 15.6 1916.0 19.0 1717.0 17.1 1817.0 15.5 1917.0 20.0 1718.0 17.0 1818.0 15.3 1918.0 20.9 1719.0 16.9 1819.0 15.2 1919.0 21.6 1620.0 73.8 1720.0 16.8 1820.0 15.1 1920.0 22.2 1621.0 72.5 1721.0 16.7 1821.0 14.9 1921.0 22.8 1622.0 71.1 1722.0 16.6 1822.0 14.7 1922.0 23.3 1623.0 69.8 1723.0 16.5 1823.0 14.5 1923.0 23.6 1624.0 68.5 1724.0 16.4 1824.0 14.1 1924.0 23.9 1625.0 67.1 1725.0 16.3 1825.0 13.7 1925.0 24.1 1626.0 65.7 1726.0 16.2 1826.0 13.3 1926.0 24.3 1627.0 64.4 1727.0 16.1 1827.0 12.7 1927.0 24.4 1628.0 63.0 1728.0 16.0 1828.0 12.1 1928.0 24.4 1629.0 61.6 1729.0 15.9 1829.0 11.5 1929.0 24.3 1630.0 60.2 1730.0 15.8 1830.0 10.8 1930.0 24.2 1631.0 58.8 1731.0 15.7 1831.0 10.1 1931.0 24.2 1632.0 57.4 1732.0 15.7 1832.0 9.5 1932.0 24.2 1633.0 56.0 1733.0 15.8 1833.0 9.0 1933.0 24.2 1634.0 54.6 1734.0 15.8 1834.0 8.6 1934.0 24.2 1635.0 53.2 1735.0 15.9 1835.0 8.3 1935.0 24.1 1636.0 51.8 1736.0 15.9 1836.0 8.0 1936.0 24.1 1637.0 50.3 1737.0 16.0 1837.0 7.9 1937.0 24.0 1638.0 48.9 1738.0 16.1 1838.0 8.0 1938.0 24.0 1639.0 47.5 1739.0 16.1 1839.0 8.0 1939.0 24.2 1640.0 46.1 1740.0 16.2 1840.0 7.9 1940.0 24.5 1641.0 44.7 1741.0 16.3 1841.0 7.9 1941.0 24.9 1642.0 43.3 1742.0 16.5 1842.0 8.0 1942.0 25.4 1643.0 41.9 1743.0 16.7 1843.0 8.2 1943.0 26.0 1644.0 40.5 1744.0 16.9 1844.0 8.2 1944.0 26.5 1645.0 39.1 1745.0 17.1 1845.0 8.3 1945.0 27.0 1646.0 37.7 1746.0 17.3 1846.0 8.6 1946.0 27.5 1647.0 36.3 1747.0 17.5 1847.0 8.8 1947.0 27.9 1648.0 35.0 1748.0 17.7 1848.0 9.1 1948.0 28.3 1649.0 33.6 1749.0 17.9 1849.0 9.3 1949.0 28.7 1650.0 32.3 1750.0 18.0 1850.0 9.3 1950.0 29.0 1651.0 30.9 1751.0 18.2 1851.0 9.5 1951.0 29.4 1652.0 29.6 1752.0 18.3 1852.0 9.8 1952.0 29.8 1653.0 28.3 1753.0 18.5 1853.0 9.9 1953.0 30.2 1654.0 27.0 1754.0 18.6 1854.0 10.0 1954.0 30.6 1655.0 25.7 1755.0 18.8 1855.0 10.1 1955.0 30.8 1656.0 24.4 1756.0 18.9 1856.0 10.3 1956.0 31.1 1657.0 23.2 1757.0 19.0 1857.0 10.4 1957.0 31.4 1658.0 21.9 1758.0 19.1 1858.0 10.4 1958.0 31.9 1659.0 20.7 1759.0 19.2 1859.0 10.4 1959.0 32.5 1660.0 19.5 1760.0 19.3 1860.0 10.3 1960.0 32.9 1661.0 18.3 1761.0 19.3 1861.0 10.1 1662.0 17.2 1762.0 19.4 1862.0 9.7 1663.0 16.0 1763.0 19.5 1863.0 9.3 1664.0 14.9 1764.0 19.5 1864.0 8.8 1665.0 13.8 1765.0 19.6 1865.0 8.1 1666.0 12.8 1766.0 19.7 1866.0 7.5 1667.0 11.7 1767.0 19.7 1867.0 6.6 1668.0 10.7 1768.0 19.8 1868.0 5.3 1669.0 9.7 1769.0 19.9 1869.0 4.1 1670.0 8.7 1770.0 19.9 1870.0 3.1 1671.0 7.8 1771.0 20.0 1871.0 2.2 1672.0 6.9 1772.0 20.0 1872.0 1.2 1673.0 6.0 1773.0 20.0 1873.0 0.4 1674.0 5.1 1774.0 20.1 1874.0 -0.3 1675.0 4.3 1775.0 20.1 1875.0 -0.9 1676.0 3.5 1776.0 20.1 1876.0 -1.5 1677.0 2.8 1777.0 20.1 1877.0 -2.1 1678.0 2.1 1778.0 20.2 1878.0 -2.6 1679.0 1.4 1779.0 20.2 1879.0 -3.1 1680.0 0.7 1780.0 20.2 1880.0 -3.6 1681.0 0.1 1781.0 20.2 1881.0 -3.7 1682.0 -0.5 1782.0 20.2 1882.0 -3.7 1683.0 -1.0 1783.0 20.1 1883.0 -3.7 1684.0 -1.5 1784.0 20.0 1884.0 -3.8 1685.0 -2.0 1785.0 19.9 1885.0 -3.9 1686.0 -2.4 1786.0 19.8 1886.0 -4.0 1687.0 -2.8 1787.0 19.7 1887.0 -4.1 1688.0 -3.1 1788.0 19.6 1888.0 -4.3 1689.0 -3.4 1789.0 19.5 1889.0 -4.3 1690.0 -3.6 1790.0 19.4 1890.0 -4.3 1691.0 -3.8 1791.0 19.3 1891.0 -4.4 1692.0 -4.0 1792.0 19.2 1892.0 -4.7 1693.0 -4.1 1793.0 19.1 1893.0 -5.0 1694.0 -4.2 1794.0 18.9 1894.0 -4.9 1695.0 -4.2 1795.0 18.7 1895.0 -4.7 1696.0 -4.2 1796.0 18.6 1896.0 -4.6 1697.0 -4.1 1797.0 18.4 1897.0 -4.0 1698.0 -3.9 1798.0 18.2 1898.0 -3.3 1699.0 -3.8 1799.0 18.0 1899.0 -2.4 8.5 Site coordinates 8.5.1 Conversion to WGS84 latitude/Longitude Coordinates referred to the various national datums are converted to the WGS84 geoid using the standard Molodensky transformation, with parameters mainly taken from document TR8350.2 at: http://earth- info.nga.mil/GandG/publications/tr8350.2/tr8350_2.html. 8.5.2 Conversion of MSL altitudes Altitudes referred to mean sea level are converted to elevations referred to the WGS84 geoid using a standard look- up table. 8.5.3 Polar motion Site coordinates are corrected for Polar motion using the following series from Earth Orientation Center at: http://hpiers.obspm.fr/eop-pc/ 1846 - 1961: C01 1962 - now: C04_05 8.5.4 Datums for RGO sites The RGO dataset does not identify the national datum for the coordinates. However when putting the dataset together, wherever possible RGO re-determined the coordinates on national datums current at around 1970. Codes 1 or 2 apply to those sites. Codes 3 to 5 ('probably geodetic' [466 sites, 11315 events], 'probably astronomical' [166 sites, 6815 events], and 'astronomical' [323 sites, 18884 events]) generally apply to sites where RGO could not determine 'modern' site coordinates. The relevant national or regional datum for coordinates under codes 1 and 2 are identified using an algorithm developed by US Naval Observatory in the 1970's. Coordinates under codes 3 to 5 are treated (inadequately for codes 4 and 5) the same as code 2 (that is, as if they were referred to a national datum.) 8.5.5 Atmospheric refraction If the elevation of the moon is less than 30 degrees, the topocentric position of the Moon is computed using an increased site altitude - to incorporate the differential displacement of the moon with respect to the star that is caused by atmospheric refraction. 8.6 Lunar limb corrections Limb corrections were derived from the Lunar Reconnaissance Orbiter - Lunar Orbiter Laser Altimeter [LOLA]. The 1/512 gridded data set was converted to the [P,D] reference frame to facilitate the determination of the apparent limb height at any point around the limb. It is believed that the limb heights derived from this data set are accurate at about the 0.01" level. =========================================================== 9. Comments on specific data fields =========================================================== 9.1 WDS double star component identifier Col 26 This field allows the component identifier of the star in a double star system to be specified - so that the separation and position angle of the component from the catalogue position for the star can be extracted from the WDS or the 6th Catalogue of Orbits of Binary stars. While this field is completed for new observations, the data for this field did not exist in either the RGO or ILOC datasets. It is hoped to add data to this field in the future. 9.2 Phenomena Col 27 This records the event type as reported by the observer. In processing the corrections for ordinary occultations this field was not reviewed or changed. Thus there will be instances where this field indicates an event type that is inconsistent with the remainder of the record. However for grazing occultations this field was reviewed and changed where appropriate. 9.3 Personal Equation Col 30-33, 34 If personal equation is not specified for a visual observation [Code N in column 34], the reductions have been based on the values derived by Morrison & Appleby (1981). That is, a PE of 0.48s for Disappearances, 0.99s for Reappearances. If the star is magnitude 4.0 or brighter the values used are 0.28 and 0.79 respectively. 9.4 Light level at event Col48-52 Data in this field is only included if the event had a duration that was clearly longer than the (for example) video frame rate. 9.5 Temperature Col 57-59 This data item was introduced by ILOC. However the data report format did not permit temperatures lower than -9 deg C to be specified. An informal practice developed of adding 100 deg to such cold temperatures. While these have generally been corrected, any 'odd' temperatures should be considered against the possibility that it is intended to specify a temperature below -9 deg, with some number having been added to the real temperature. 9.6 Graze identifier Col 75-78, 79-81 The graze identifier is provided to allow for easy identification of all events relating to a grazing occultation. For the small number of events where a graze of the same star was observed at both the northern and southern limbs of the moon on the same night, the two grazes have the same identifiers. They can be separated using the flag in column 71. 9.7 Vertical datum code The default datum for vertical elevations is Mean Sea Level. However some observers using GPS have reported geoid (ellipsoid) heights. That practice is generally discouraged. 9.8 Closest town or landmark Cols 128-177 The name of the closest town or landmark is a free-text field that has never been subject to any format control. Its primary value has been to confirm the general correctness of the reported site coordinates. There is considerable variability in the how the town or landmark is selected - particularly with regard to distance from the actual observing site, and whether a suburb is specified when the location is within a large metropolis. There is also wide variation in the degree to which administrative regions (Country, State, Prefecture, etc) are specified. And while there is apparent regularity for many records in the location of a country code within the field, this is not uniform. For the event records which are place-holders for missing or unreported grazes, this field is also used to give any information known about the number of stations, and number of observed events. 9.9 Observer names Cols 189-213 The name of the observer is a free-text field that has never been subject to any format control. Variations in the name style that exist in this field include: - For the majority of names, the name order is given name - family name. However some names are in the order of family name-given name. - For some names the family name is based on father's family name followed by mother's first family name - but this can sometimes be reversed. - Given names can be one initial, several initials, one or more names, or a combination of names and initials. The Archive records the name as reported. No attempt has been made to regularise the name style. There is no assumption of any consistency in the name style used by an observer - particularly after 1981. Any search for observations by a particular observer should be based on characters contained in the family name, with subsequent review to assess whether the names retrieved relate to the same person. 9.9.1 Names in the RGO dataset The names of the observers for the RGO sites were manually added from the lists provided in: - RGO Bulletin No 183 (1943 - 1971) - RGO Bulletin No 192 (1972 - 1980) The 1972-1980 dataset provided specific correspondence between Station/Telescope identifier and the observer name. The exact linking of observers to observation in the period 1943-1971 has been effectively lost (see explanation in RGO Bulletin 183, page 10). However in many instances a probable association of an observer to an observation could be determined. In such instances an asterisk (=) has been added to the name. For a small number of sites the RGO data sets indicate that the observer's name was not known. These instances are marked with '?'. Observations before 1943 did not have any names in the RGO data. The names are almost certainly available in the literature where the observations were originally reported. For the oldest observations, names were obtained where possible from Newcomb (1912). Otherwise no attempt to retrieve that data has been made or planned. =========================================================== 10. References =========================================================== Brouwer D, 1938. The occultation campaign. Outline of revised program. Astr J, 47, 1938 [http://adsabs.harvard.edu/full/1939AJ.....47..191B] Brown, E W, 1927. Request for more observations of occultations. Astr J, 37, 99 [http://adsabs.harvard.edu/full/1927AJ.....37...99B] Martin C F, 1969. A study of the rate of rotation of the Earth from occultations of stars by the Moon 1627-1860, PhD dissertation, Yale University. Morrison L V and Appleby G M. Analysis of lunar occultations - II. Personal Equation. (1981) MNRAS 196, 1005-1011 [http://adsabs.harvard.edu/full/1981MNRAS.196.1005M] Newcomb S, 1912. Researches on the Motion of the Moon and related astronomical elements based on observations extending from the era of the Babylonians until AD 1908. Astronomical Papers prepared for the users of the American Ephemeris, Washington, IX, Part II [ http://adsabs.harvard.edu/abs/1912USNAO...9....1N. See also Part I at http://adsabs.harvard.edu/abs/1878USNOM..15B...1N].