date: miš., 11 jan. 2006 11:29:58 +0000 from: Trausti Jonsson subject: Re: latest draft of the Greenland paper to: Phil Jones Dear Phil, I have now browsed through the paper and I have a few notes. I do not think that they will affect the results in any material way. The thing most sorely missed is a table of the manuscript origins of the=20 pre-1873 data, but I know that you can not do anything about that because you can not find the original journals - so it has to be this way. The figure 2 is very difficult to read (at an earlier stage I would have suggested tha addition of a vertical grid at every 20 to 30 years). But when you have actually added the monthly values to your website I can see which years you have and which not (I did not find the data already on the website). I can not e.g. see if you include the=20 following data: 1. The Giesecke observations from Nuuk from 1806 to 1813 (which I have got from the Royal Society) 2. Frederikshaab (Paamiut) data of Aug 1841-April 1842 and Oct 1828 to April 1829. The 1841 to 1842 part the DMet Office=20 has calculated the averages as the mean of the two observing=20 times at 9am and 9pm, wich I can not compare with your monthly values until I have access to the monthly table. The 1828- 1829 mean values are according to the Met.Office a mean of 9-12- 15-18 and 21 (all local times). 3. It seems as if you have the Paamiut 1851 to 1858 data. Here I know that during the 1852 to 1853 winters there were=20 sign (+ or -) problems in the original data, according to the Danish Met. Office. That is the reason for a gap during this period. 4. It also seems that you have the Julianehaab 1807 to 1812 data. There the observations used in the average are only two (with the diffuse timing of morning and evening). On you figure 3 you refer to the observation hours in GMT, but this is a bit confusing because elsewhere the observation times mentioned during the 19th century are the average local sun time, the pre-1830 observation times are not fixed by a clock - only morning and evening, in the case of Giesecke there is also a mid-day observation. =09 As you correctly note, the daily range is small during a large part=20 of the year, but south of Nuuk it is not nonexistent, even in December and January, it is more correct to say that it is negilible. I don't know why you choose Upernivik as a typical station for the illustration of the daily range, because it is larger at the more southerly stations during the summer (4.1=B0C at Julianehaab in July during the last few years and 3.2=B0C in Nuuk in July during the same period). But I don't think that these details will change your results, but the (hopefully erroneously) feeling creeps in that the main author does not know the=20 annual cycle of daily range in Greeland as well as he should do.=20 On inspection of the Danish yearbooks I found there is a wrong citation on page 7 in the draft. The quotation is: "The yearbook 1874 states that (1) is based on hourly observations from Greenland ..... polar expedition, and that (1) has bee verified to give the true daily mean temperature for Nuuk as well." What actually is said in the 1874 yearbook (p.V) is: In the case of Godthaab a direct mean of the actual observation times gives a correct mean. So, in this case Godthaap (Nuuk) is an EXCEPTION to the use of (1). If you then check the Nuuk table a few pages later, you will see that the observation hours at Nuuk during 1874 are the abnormal: 5 am, 1 pm and 9 pm and as a matter of fact the number in the mean column is not a result of (1). This state of affairs persists until June 1875 and in August that year Godthaab finally reverts to the use of (1). You will also note that the Ivittuut means are in these years based on only on observation per day so the formula is not in use there=20 either, for the first few years. The 1875 change in Nuuk could mean that during 1866 to 1875 the observation hours there were probably=20 5, 13 and 21 (local time).=20 You do not mention that the Observations before ca. 1860 are=20 probably made without screens, but it has certainly some influence on the daily range. This is mentioned in the correspondance of the Danish= Scientific=20 Society with the observers in Reykjavik in the 1870. The DSS is particularly= =20 worried about the early morning and late night observations when the=20 instruments are colder than the air, because of radiation effects. There are= =20 also evaporation effects as well on unscreened instruments. You don't= mention=20 this either. =20 As I said before I don't think that these small points change anything in the paper but you should get the Godthaab (1874)=20 quotation corrected.=20 Your search for pre-1873 data confirms that the older observation journals are not to be found at the DMI, the pre-1873 Icelandic data did not come from them, but the archives of the Icelandic Society of Letters in Copenhagen. The Greenland originials are probably found at the archives of the Danish Scientific Society. I have been looking at the ice core data and comparing it to Icelandic records with very encouraging results (it was a part of my Stykkisholmur lecture).=20 But I hope that the CRU will continue the Greenland work with the danes and congratulate you again on the paper. I also hope that the monthly values will be available at your website soon. Best wishes, Trausti