por_017.fm Polar Research 26 2007 175 – 180 © 2007 The Author 175 R e s e a r c h n o t e National and institutional productivity and collaboration in Antarctic science: an analysis of 25 years of journal publications (1980–2004) Prabir G. Dastidar Ministry of Earth Sciences, Block No. 9 & 12, CGO Complex, Lodi Road, New Delhi 110003, India Abstract Journal publications on Antarctic science were analysed for a period of 25 years (1980–2004) through a set of scientometrics and network analysis tech- niques. The study is based on 10 942 records (research articles, review articles, letters, etc.) with the word fragment “antarc*” in the title published in 961 international, peer-reviewed journals and retrieved from Thomson Scientific’s Science Citation Index database. During the period under investigation, pro- ductivity increased threefold and there was a 13-fold increase in journal publications co-written by authors from different countries. The five nations with the highest output were the USA (with 26.7% of the total output), the UK (13.8%), Australia (9.7%), Germany (8.8%) and Italy (6.0%). The top five institutions in terms of journal publications were the British Antarctic Survey (972 publications), the Alfred Wegener Institute of Polar and Marine Research, Germany (475), the Australian Antarctic Division (312), the University of Tasmania, Australia (305), and the National Aeronautics and Space Adminis- tration, USA (293). Keywords Antarctic science; journal articles; knowledge mapping; network analysis; Science Citation Index; scientometrics. Antarctica is the fifth largest continent. This coldest, windiest and highest continent is covered with an ice sheet more than 2-km thick, on average. Having remained comparatively undisturbed and unpolluted for millions of years, the Antarctic environment is a treasure trove of information about the Earth’s past. The Indian, Atlantic and Pacific oceans meet around Antarctica, and the mixing process of cold and warm waters contribute to a special regime with unique physical, chemical and bio- logical characteristics. These waters constitute one of the world’s richest biological provinces. For all these reasons, Antarctica provides unique opportunities for scientific research in diverse fields. The Antarctic Treaty System— one of the world’s most successful international agree- ments—ensures that Antarctica remains a natural reserve for science. There are 26 countries with seasonal or year- round stations in Antarctica (Table 1). This paper analyses a data set of journal publications for the period 1980–2004. An attempt has been made to identify the major players in Antarctic science by identi- fying the nations and institutions that produced the greatest number of journal publications about Antarctica. As conducting scientific research is central to the Antarc- tic Treaty System, an analysis of scientific output— measured here in terms of journal publications—in part reflects the functioning of this set of international agree- ments. Similar analyses have been previously undertaken for journal publications in ocean science and technology (Dastidar 2004; Dastidar & Ramachandran 2005), along with an earlier effort for Antarctic science (Dastidar & Persson 2005). Materials and methods The data for this study were drawn from Thomson Scien- tific’s Science Citation Index, a database available on a series of annual CD-ROMs. The Science Citation Index covers about 3700 of the world’s interntional, peer- reviewed scientific and technical journals, and includes publication details for research articles, review articles, news items, book reviews, letters, communications, edi- torial material and so forth (see http://scientific.thomson. Correspondence Prabir G. Dastidar, Ministry of Earth Sciences, Block No. 9 & 12, CGO Complex, Lodi Road, New Delhi 110003, India. E-mail: prabirgd11@rediffmail.com doi:10.1111/j.1751-8369.2007.00017.x http://scientific.thomson 176 Polar Research 26 2007 175 – 180 © 2007 The Author National and institutional productivity and collaboration in Antarctic science P.G. Dastidar com/products/sci/ and http://thomsonscientific.com/ free/essays/selectionofmaterial/journalselection/). The database on the CD-ROM for each year from 1980 to 2004 was searched for publications with “antarc*” in the title. This yielded 10 942 records, which formed the basis of the present analysis. To assess publication output for individual countries, each publication was given a value of 1. In the case of a multinational publication—a publication co-written by authors with addresses in two or more different coun- tries—each contributing country was given a fraction value where the sum of fractions equalled 1. For exam- ple, in the case of a publication written by one author with a US address and one with an address in the UK, the USA and the UK would each receive a value of 0.5. These fractional values were summed to asssess the productivity of individual countries. It should be emphasized that this method of calculation makes no attempt to assess the actual relative contributions of the co-authors (neither, by extension, the nations they represent) by, for example, weighted scores. Rather, each co-author of an article was assigned an equal fraction. It is a widespread practice to place the co-author who has contributed disproportion- ately to the the work of preparing a scientific article first on the author byline (Yank & Rennie 1999). However, other conventions may be applied and these vary across the international scientific community. Among teams of scientists who frequently publish together, the co-authors may take turns being named as lead authors (O’Connor & Woodford 1978). In some scientific circles it may be cus- tomary to name heads of departments, laboratories or research groups as first authors (O’Connor & Woodford 1978). Some journals have clear policies about how to determine the order of authors; most do not and there is apparently no clear consensus about the meaning of the order of authors among journal editors (Yank & Rennie 1999). For the purposes of the analysis, publications authored by writers in the Federal Republic of Germany and the German Democratic Republic were merged under “Ger- many”, whereas publications from writers with addresses in the USSR and Russia were combined under “Russia”. The country and names of institutions associated with the publications were isolated separately and rank ordered (Persson 2004). The most productive units were chosen to form co-occurrence matrices to which a multi- dimensional scaling algorithm (a SYSTAT subroutine) was applied to produce the network maps. The relative size of the circles in these diagrams indicates the relative produc- tivity of each entity, and the lines between the circles indicate the presence of collaboration links, whereas the line thickness indicates the strength of the interconnec- tions (Dastidar 2004; Dastidar & Persson 2005; Dastidar & Ramachandran 2005). An important limitation of this study is that the data were limited to the journals in Thomson Scientific’s Science Citation Index. Alhough non-English publica- tions are included in the database, Thomson Scientific’s selection criteria make it easier for English-language publications to be included for coverage (see http:// thomsonscientific.com/free/essays/selectionofmaterial/ journalselection/). In addition, journal publications that concerned Antarctica but which did not have “antarc*” in their titles were not included in this study. Results During the study period the output of journal publica- tions with “antarc*” in the title increased threefold, rising from 165 in 1980 to 552 in 2004 (Fig. 1). Output peaked in 2002, with 729 publications. Over 80 nations were represented among the author- ship of the publications being analysed (Table 2). The output is highly skewed, with two countries—USA and Table 1 Number of stations in Antarctica and manpower deployment contributed by the 28 consultative parties to the Antarctic Treaty System (www.comnap.aq/facilities). Country No. of stations (year when first station was established) Manpower (annual peak) Argentina 6 (1904) 417 Australia 3 (1954) 213 Belgium — — Brazil 1 (1984) 40 Bulgaria 1(1988) 15 Chile 5 (1948) 224 Ecuador 2 (1990) 88 Finland 1(1989) 20 France 4 (1956) (one jointly with Italy) 145 Germany 2 (1981) 78 India 1 (1989) 65 Italy 5 (1986) (one jointly with France) 135 Japan 2 (1957) 150 the Netherlands 0 — New Zealand 1 (1957) 85 Norway 2 (1985) 44 China 2 (1985) 70 Peru 1 (1989) 28 Poland 1 (1977) 40 Russia 7 (1956) 429 South Africa 1 (1962) 80 South Korea 1 (1988) 60 Spain 2 (1989) 28 Sweden 1 (1989) 20 UK 5 (1947) 205 Ukraine 1 (1996) 24 Uruguay 1 (1984) 60 USA 3 (1955) 1250 —, Data not available. http://thomsonscientific.com/ http:// P.G. Dastidar National and institutional productivity and collaboration in Antarctic science Polar Research 26 2007 175 – 180 © 2007 The Author 177 UK—each contributing shares larger than 10%, 16 coun- tries each contributing shares between 1 and 10%, and all the remaining countries contributing shares of less than 1% each. The share for the USA was 26.7%, almost double that of the UK, which was the next most produc- tive country, with 13.8%. The five most productive coun- tries (USA, UK, Australia, Germany and Italy) together contributed 65% of the total output. Table 2 Country-by-country output of journal publications with “antarc*” in the title, based on the annual CD-ROMs of the Science Citation Index, 1980– 2004 ( N + 10 942). Each publication was given a value of 1. In the case of a multinational article, each contributing country was given a fractional value where the sum of fractions equalled 1. These fractional values were summed to asssess the productivity of individual countries. Consultative parties to the Antarctic Treaty System are indicated. (Ecuador, also a consultative party to the treaty, is not included in the table because of its very low journal publication output.) Data on per capita gross domestic expenditures on research and development are from the UNESCO Institute for Statistics (http://www.uis.unesco.org/ev.php?ID = 5182 = 201&ID2 = DO_TOPIC). Country Output % of world total Per capita gross domestic expenditure on research and development, in purchasing power parity USD (year of data in parentheses) 1 USA a 2886.90 26.7 954 (2002) 2 UK a 1491.83 13.8 490.6 (2002) 3 Australia a 1051.85 9.7 404.5 (2000) 4 Germany a 948.87 8.8 686 (2002) 5 Italy a 653.21 6.0 288.7 (2001) 6 France a 526.08 4.9 611.2 (2002) 7 Japan a 492.22 4.5 836.6 (2002) 8 New Zealand a 430.22 4.0 246.1 (2001) 9 Russia a 305.86 2.8 102.2 (2002) 10 Spain a 241.37 2.2 222.4 (2002) 11 South Africa a 232.99 2.2 68.7 (2002) 12 Argentina a 188.37 1.7 44 (2002) 13 the Netherlands a 152.82 1.4 536.6 (2001) 14 India a 126.29 1.2 20.5 (2000) 15 Belgium a 117.06 1.1 614.7 (2002) 16 Canada 109.65 1.0 588.4 (2002) 17 Sweden a 107.30 1.0 1082.5 (2001) 18 Poland a 107 1.0 62.7 (2002) 19 China a 90.74 0.8 — 20 Norway a 88.73 0.8 612.2 (2002) 21 Chile a 63.16 0.6 51.9 (2001) 22 Brazil a 58.49 0.5 76.9 (2000) 23 Switzerland 50.78 0.5 740.4 (2000) 24 Denmark 45.97 0.4 777.6 (2002) 25 South Korea a 44.53 0.4 492.3 (2002) 26 Austria 37.84 0.3 645.2 (2002) 27 Finland a 35.99 0.3 905.2 (2002) 28 Czech Republic 13.99 0.1 — 29 Bulgaria a 10.5 0.1 34.9 (2002) 30 Hungary 9.16 0.1 135.3 (2002) 31 Greece 9.08 0.1 115.8 (2001) 32 Ukraine a 9 0.1 57.6 (2002) 33 Ireland 8.91 0.1 369.2 (2001) 34 Taiwan 8.5 0.1 — 35 Israel 7.91 0.1 997.2 (2002) 36 Mexico 6.16 0.1 38.3 (2002) 37 Turkey 4.5 0.0 42.6 (2002) 38 Bermuda 4.33 0.0 27.4 (1997) 39 Iceland 3.90 0.0 925.7 (2002) 40 Jamaica 3.33 0.0 3.0 (2002) 41 UKSSR 3 — — a Consultative parties. —, Data not available. 42 Pakistan 2.5 0.0 5.2 (2002) 43 Estonia 2.25 0.0 98.9 (2002) 44 Philippines 2 0.0 45 Romania 2 0.0 24.9 (2002) 46 Singapore 2 0.0 525.7 (2002) 47 Antarctica 1.83 0.0 — 48 Monaco 1.83 0.0 — 49 Reunion 1.5 0.0 — 50 Ciskei 1 0.0 — 51 Colombia 1 0.0 10.5 (2001) 52 Indonesia 1 0.0 — 53 Ivory Coast 1 0.0 — 54 Kenya 1 0.0 — 55 Papua New Guinea 1 0.0 — 56 Peru a 1 0.0 5.2 (2002) 57 Portugal 1 0.0 170.2 (2002) 58 Saudi Arabia 1 0.0 — 59 Vanuatu 1 0.0 — 60 Zimbabwe 1 0.0 — 61 Bolivia 0.5 0.0 6.9 (2002) 62 Belarus 0.5 0.0 35.1 (2002) 63 Comoros 0.5 0.0 — 64 Costa Rica 0.5 0.0 34.5 (2000) 65 Latvia 0.5 0.0 42.8 (2002) 66 Luxembourg 0.5 0.0 961.1 (2000) 67 Morocco 0.5 0.0 — 68 Namibia 0.5 0.0 — 69 New Caledonia 0.5 0.0 — 70 Niger 0.5 0.0 — 71 Nigeria 0.5 0.0 — 72 Qatar 0.5 0.0 — 73 Slovenia 0.5 0.0 286.2 (2002) 74 Sri Lanka 0.5 0.0 5.1 (1996) 75 Uruguay a 0.5 0.0 20.6 (2002) 76 Venda 0.5 0.0 — 77 Vietnam 0.5 0.0 — 78 Yugoslavia 0.5 0.0 — 79 Fiji 0.33 0.0 — 80 French Polynesia 0.33 0.0 — 81 French Guiana 0.33 0.0 — 82 Venezuela 0.2 0.0 20.7 (2002) Country Output % of world total Per capita gross domestic expenditure on research and development, in purchasing power parity USD (year of data in parentheses) a Consultative parties. —, Data not available. http://www.uis.unesco.org/ev.php?ID 178 Polar Research 26 2007 175 – 180 © 2007 The Author National and institutional productivity and collaboration in Antarctic science P.G. Dastidar As shown in Fig. 1, the number and proportion of publications co-written by contributors from different countries increased over the period being examined. There were 15 multinational publications in 1980, which represented 9.09% of the total for that year. By 2004 there were 190 multinational publications, which consti- tuted 34.42% of the total for that year. Figure 2 gives an indication of the collaborative ties between countries, with the countries with writers who most frequently co- author publications lumped together in the middle of the diagram. American authors most frequently contribute to multinational journal publications, followed by the UK, Australia and Germany. About 2837 organizations contributed to journal publi- cations with “antarc*” in the title during the period stud- ied. The 75 most productive institutions are presented in Table 3. The British Antarctic Survey topped the list, having contributed to 972 publications. In second place was the Alfred Wegener Institute for Polar and Marine Research, with 475 publications. The remaining institu- tions among the top five were the Australian Antarctic Division (312 publications), the University of Tasmania, Australia (305 publications), and the National Aeronau- tics and Space Administration, USA (294 publications). The country with the most institutions appearing in the top 75 was the USA, with 26 institutions. Australia had Fig. 2 Collaboration network of the top 13 countries in terms of the output of multinational journal publications with “antarc*” in the title, based on the annual CD-ROMs of the Science Citation Index, 1980–2004 ( N = 82). Line thick- ness and proximity of the circles indicates col- laboration intensity between the countries. Fig. 1 Overall number of journal publications with “antarc*” in the title, and the number of these publications co-authored by writers from different countries, based on the annual CD- ROMs of the Science Citation Index, 1980–2004. The scale on the left-hand side of the figure indi- cates the overall number of publications and the scale on the right-hand side indicates the num- ber of multinational publications. P.G. Dastidar National and institutional productivity and collaboration in Antarctic science Polar Research 26 2007 175 – 180 © 2007 The Author 179 10 institutions among the top 74, New Zealand had seven, and Argentina and Germany each had four. Data for the ten most productive organizations were used to produce the network map presented in Fig. 3. The diagram indicates frequent co-authorship between the British Antarctic Survey and the Alfred Wegener Institute for Polar and Marine Research. Discussion Scientific journal publications about Antarctica in the period 1980–2004 were dominated by the nations with a strong interest in the continent. Nineteen of the 20 coun- tries producing the most publications are consultative parties to the Antarctic Treaty, and the top 20 countries included all 11 original signatories of the treaty (Table 2). Setting aside Canada and the 11 original signatory nations, the remaining eight countries in the top 20 gained consultative status by having conducted substan- tial research activity there, as provided by Article 9 of the treaty (http://www.ats.aq/). Productivity in science is investment dependent, as is shown by the strength of the contribution to journal publications about Antarctica in 1980–2004 by the USA. Table 3 Rank list of the 74 most productive institutes in terms of journal publications with “antarc*” in the title, based on the annual CD-ROMs of the Science Citation Index, 1980–2004. Ranking Institution No. articles 1 British Antarctic Survey, UK 972 2 Alfred Wegener Inst. for Polar & Marine Res., Germany 475 3 Australian Antarctic Div., Australia 312 4 University of Tasmania, Australia 305 5 NASA, USA 293 6 Ohio State University, USA 244 7 University of California, San Diego, USA 220 8 NERC, UK 216 9 CNR, Italy 214 10 CNRS, France 205 11 National Institute of Polar Research, Japan 196 12 University of Colorado, USA 163 13 CSIRO, Australia 146 14 Russian Academy of Sciences, Russia 148 15 NOAA, USA 137 16 CALTECH, USA 135 17 University of Washington, USA 127 18 CSIC, Spain 124 19 University of Melbourne, Australia 123 20 Columbia University, USA 103 21 University of Wisconsin, USA 102 22 University of Maine, USA 101 23 University of Genoa, Italy 100 24 Antarctic CRC, Australia 98 25 Argentinean Antarctic Institute, Argentina 95 26 University of Tokyo, Japan 94 27 University of Canterbury, New Zealand 94 28 US Geological Survey, USA 92 29 University of Kiel, Germany 89 30 University of Alabama, USA 88 31 University of Otago, New Zealand 88 32 University of Buenos Aires, Argentina 85 33 Australian National University, Australia 84 34 Christian Albrechts University of Kiel, Germany 83 35 DSIR, New Zealand 82 36 Hokkaido University, Japan 82 37 University of Auckland, New Zealand 81 38 University of Utrecht, the Netherlands 76 39 University of Cape Town, South Africa 75 40 Texas A&M University, USA 72 41 University of Cambridge, UK 72 42 University of California, Santa Barbara, USA 71 43 University of Illinois, USA 71 44 University of California, Santa Cruz, USA 71 45 Lab Glaciol & Geophys Environm, France 70 46 University of Texas, USA 69 47 Arctic & Antarctic Research Institute, Russia 69 48 Macquarie University, Australia 68 49 Woods Hole Oceanographic Institution, USA 67 50 University of Paris, France 67 51 Polish Academy of Sciences, Poland 66 52 University of Siena, Italy 65 53 University of Alaska, USA 64 54 University of Bremen, Germany 62 55 Nat Inst Water and Atmospher Res, New Zealand 60 56 University of Wyoming, USA 60 57 University of Copenhagen, Denmark 60 58 University of Pretoria, South Africa 60 59 Consejo Nacl Invest Cient & Tecn, Argentina 60 60 University of Hawaii, USA 59 61 Chinese Academy of Sciences, China 58 62 Oregon State University, USA 58 63 Monash University, Australia 56 64 University of California, USA (no specific campus stated) 54 65 University of Bern, Switzerland 54 66 University of California, Los Angeles, USA 53 67 University of Waikato, New Zealand 53 68 Department of Science, Australia 52 69 University of Liege, Belgium 52 70 Montana State University, USA 52 71 Victoria University of Wellington, New Zealand 52 72 University of New South Wales, Australia 49 73 National Center for Atmospheric Research, USA 48 74 University of Bristol, UK 47 Ranking Institution No. articles http://www.ats.aq/ 180 Polar Research 26 2007 175 – 180 © 2007 The Author National and institutional productivity and collaboration in Antarctic science P.G. Dastidar The USA alone contributed more than one quarter of all publications during the period, and, of the most produc- tive individual organizations, one of three institutions in the list was American. The USA deploys the greatest manpower to the continent and maintains the largest research complex there. The USA has one of the highest per capita gross domestic expenditures on research and development, which is almost double that of the UK (Table 2). As was also seen in the case of journal publications in ocean science and engineering (Dastidar 2004), the most productive countries have higher per capita gross domestic expenditures on research and development. International collaboration is an important element in the Antarctic Treaty, which was originally signed in 1959 (http://www.ats.aq/uploaded/treaty_original.pdf). To promote international cooperation, Article 3 of the treaty calls for the exchange of personnel and scientific results between the parties to the treaty. The 16th Antarctic Treaty Consultative Meeting declared 1991–2000 the Decade of International Antarctic Scientific Cooperation (Cohen 2002). This study has shown a marked increase in one indicator of international scientific cooperation in Antarctica: publications authored by writers from differ- ent countries. An important subject beyond the scope of this brief paper is the research topics of the journal publications relating to Antarctica during the period in question. It would be extremely interesting to identify Antarctic research trends in this way, and it is hoped that this will be the subject of further analysis. Acknowledgements The author expresses his gratitude to Prof. Olle Persson, Dept of Sociology, Umeå University, Sweden, for his guid- ance and constructive suggestions. The author is thankful to Dr José Retamales, director of the Chilean Antarctic Institute and chairman of the Council of Managers of National Antarctic Programs, for his suggestions and for showing keen interest in the work. The author is also thankful to Dr B.S. Aggarwal for editing an earlier version of the manuscript. References Cohen H.K. (ed.) 2002. Handbook of the Antarctic Treaty System . 9th edn. Washington DC: US Department of State. Dastidar P.G. 2004. Ocean science & technology research across the countries: a global scenario. Scientometrics 59 , 15– 27. Dastidar P.G. & Persson O. 2005. Mapping the global structure of Antarctic research vis-à-vis Antarctic Treaty System. Current Science 89 , 11552–11554. Dastidar P.G. & Ramachandran S. 2005. Engineering research in ocean sector: an international profile. Scientometrics 65 , 199–213. O’Connor M. & Woodford F. P. 1978. Writing scientific papers in English . London: Pitman Medical. Persson O. 2004. Bibexcel. Downloaded from the internet at http://www.umu.se/inforsk/Bibexcel on 27 February 2007. Yank V. & Rennie D. 1999. Disclosure of researcher contribu- tions: a study of original research articles in The Lancet . Annals of Internal Medicine 130 , 661–670. Fig. 3 Collaboration network among the ten most productive organizations in terms of the output of multinational journal publications with “antarc*” in the title, based on the annual CD- ROMs of the Science Citation Index, 1980–2004 ( N = 2726). Line thickness and proximity of the circles indicates collaboration intensity between the institutions. http://www.ats.aq/uploaded/treaty_original.pdf http://www.umu.se/inforsk/Bibexcel