Information Divide in Climate Sciences Banner
Home | Report in HTML | Report in PDF | Relevant Links | Appendices | Photo Gallery | Contact Information

The Information Divide in the Climate Sciences

by Andrew Gettelman

May 2003


Table of Contents

Executive Summary
Introduction
Survey results
Country case studies
Synthesis and recommendations
Other issues raised
Recommendations
References
Acknowledgments

Executive Summary

There exists a broad "information divide" in climate and meteorological science, between researchers and forecasters in highly developed countries and those in less-developed countries. Information in this context includes forecasts, meteorological data or analyses, research data, journals, and educational materials and textbooks. Barriers to the flow of information to and from developing countries create adverse effects for the global community, limiting the ability of developing countries to deal with extreme weather or climate events. The lack of in situ data for weather and climate studies for large sections of the planet also impedes research in developed countries into the global climate system, inhibiting identification of climate trends and shifts, as well as limiting forecast skill. This study combines visits to eight developing countries with a survey of 50 scientists and journal editors around the world to characterize the information divide and provides some suggestions to bridge it.

Statistics from journal editors and publishers indicate that there are very few published articles from developing-country authors in international journals. While typical acceptance ratios (the ratio of papers published to papers submitted) are 75% for papers with first authors from the United States, Canada, the European Union, Japan, Australia or New Zealand, for other countries the acceptance ratio drops to 25%. Editors and some contributors identified the major cause of higher rejection rates as inappropriate or poor methodologies in articles. Language difficulties sometimes mask these problems, but they were not the main concern.

There are large collections of weather and climate data in many locations in the developing world. These observations may go back 100 years or more. These data are often archived and kept current (even digitized), but are rarely shared or made available to researchers in developing or developed countries. Often even domestic researchers cannot get access to climate data.

The causes for the information divide are not surprising. Significant costs to access information exist in most countries, and few researchers are able to pay for information. About 3/4 of researchers surveyed noted that the cost of scientific journals was the most significant barrier, and about half noted that cost was a barrier for accessing data. Cost is also an issue for electronic communication. A third of those desiring more data indicated that limitations on Internet usage reduced their ability to get information. The cost of Internet access in relative terms can be 10 times as high in many developing countries as in developed countries. In many places, bandwidth is also a significant limitation. Email is generally accessible about 3/4 of the time. More complex functions are available less often. From many places in Africa and Asia, connections are impractical for downloading even small electronic articles or datasets.

Several interesting results of the survey dispel commonly held beliefs about the information divide. A moderate delay in receiving information was not found to be a problem for most scientists. Most scientists have a rudimentary level of Internet access.

Several simple things would go a long way toward bridging the information divide. In the short term, the most effective action would be to open up electronic access to scientific journals and indexes in the field, along with distributing back issues of journals on CD-ROM to developing countries. In the medium term, scientists would benefit from mentoring for publications and from education and training for new techniques (such as numerical weather prediction) and data (advanced satellite sensors). In the long term, a transition to electronic publishing for developing-country journals should be encouraged, and the Global Telecommunications System (GTS) for weather data should be enhanced to be more interactive and include research data, reports, and journals.


Introduction

Statement of issues & hypotheses

Anecdotal evidence suggests that an "information divide" exists in the climate sciences between developed and developing countries. Basic weather and forecast information may not reach developing countries in a timely manner. In addition, important advances in scientific understanding of our earth's climate system may not reach those researchers in the developing world who might benefit from the information. This creates an information divide. The types of information for which a gap exist include short- and long-term forecasts, journal articles and reports, other climate data, and educational materials. The gap may extend over more than physical science research to include strategies for disaster mitigation or adapting to climate variability. Another consequence of the information divide is that climate researchers in developed countries may not have access to valuable research or data from their counterparts in developing countries.

A vibrant global science base is critical to properly interpret our changing climate and to provide a local base for climate forecasting and impact mitigation. This report details the results of a year-long study that examined in detail the information divide in the climate sciences. The study identifies needs for information and suggests potential ways to meet those needs.

On one side of the information divide is a dearth of available information in developing countries. Analysis of subscription information from major journals (American Meteorological Society, personal communication, 2001) indicates that only a few places in developing countries have access to print journals. Even fewer have access to electronic journals. The lack of information may extend to research data and access to forecast and prediction information.

The other side of the information divide appears as a lack of contributions to the scientific literature from researchers in developing countries. This divide is apparent in general in statistics of published material (Gibbs, 1995; May, 1997) and in anecdotal information from major research journals in climatology and meteorology. Discussions with journal editors indicated fewer submissions from researchers in developing countries, as well as lower acceptance rates.

A case study of Indian climate science (Kandlikar and Sagar, 1999) supports these results and indicates it is likely that the two sides of the information divide are linked: a lack of the same breadth and quality of information that developed-country scientists take for granted, combined with a lack of resources, may result in fewer submissions to the literature. In addition, the information divide may limit the communication that developing country researchers have with colleagues in other developing countries.

With the rise of electronic publishing and communication, there appears to be no physical barrier for communication and access to information. Most scientific journals, indexes, and data are or will be available electronically. There is a danger that the information divide is rapidly growing into a "digital divide" between those with access to electronic information and those without. But technology may also be used to help bridge the information divide at low cost. The marginal cost of delivering a full up-to-date research library anywhere in the world is no longer limited by the cost of the printed page and postage. A full electronic library can be made available with a personal computer and a connection to the global Internet. In addition, digital communities may expand the ability to communicate and collaborate electronically beyond written or oral communication. As we build digital communities of scientists who work collaboratively and virtually across the planet with their colleagues, is there a group with valuable skills and knowledge that we are forgetting?

Several hypotheses were tested: Would scientists value electronic access to information resources? Do they want enhanced access to the community of the Internet, or communities of other scientists built on the Internet? Do they want to publish in journals issued in the developed world but have difficulty doing so? Do they want electronic teaching tools? Or are capabilities so limited that communication enhancements are more basic (a phone line or a photocopy machine)?

Outline of work performed

The survey consisted of three parts. The first portion was a written survey, sent directly by mail to scientists and available on the Internet in electronic form. The second portion of the survey consisted of a series of field visits to eight countries with a developing geoscience base. The third component of the survey was a series of personal interviews with editors of major international journals in climate and meteorology, and discussions with researchers in several scientific societies. The survey was conducted from October 2001 to October 2002.

Survey

The written survey focused on several areas. The English text of the survey is in Appendix A. It was sent to a list of World Meteorological Organization (WMO) members, as well as a list of Intergovernmental Panel on Climate Change (IPCC) contacts in developing countries. The written survey consists of four main sections. Section 1 focuses on access to information resources, including questions on journals read and desired, and sources of data or forecasts. Section 2 focuses on publications written by researchers, including where they were submitted and personal experiences. Section 3 focuses on communications infrastructure, including how researchers communicate with colleagues and their use of the Internet. Finally, general descriptive information about the institutions and areas of expertise was collected, including information about languages spoken and written.

In addition to written form, the survey was made available on the Internet in 4 languages: English, French, Spanish and Portuguese. Forty-nine responses were received from the survey, about half of them electronically. More details about the nature of the responses can be found in Section 2.

Places visited

The country visits focused on developing countries bordering the Indian Ocean, the "Indian rim" countries. These countries represent almost one-third of the earth's population, from Africa to India to Indonesia. The countries share common exposure to several scales of climate variability, dominated by monsoon variability and tropical storms, including effects of the El Niño-Southern Oscillation. Indian rim countries experience similar climate impacts on agriculture and water resources. The country setting varies widely, ranging from highly developed regions of India to very underdeveloped regions of Africa. Many of these countries have significant scientific cultures and active programs that attempt to understand climate variability and its impacts on society. In many of the countries targeted (e.g., India and Kenya), English is widely spoken for historical reasons.

In the course of the survey, 29 different institutes in eight countries were visited. These countries are: Kenya, Uganda, India, Nepal, Myanmar, Thailand, Vietnam, and China. A distribution of the per-capita gross domestic product (GDP) of these countries is presented in Figure 1.1. Most of the countries, with the exception of Thailand and China, have per-capita GDP of less than 500 USD per person.

The selection of countries necessarily biased some of the results, particularly with regard to language. All of the interviews and visits were conducted largely in English. Most of the researchers visited are fluent in spoken, and usually written, English. The written survey is less biased in this regard (several of the responses were in French, Spanish, or Portuguese). The bias makes it difficult for the survey to reflect the true prevalence of English in the climate sciences around the world from that in the countries visited.

A list of the institutions visited is presented in Table 1.1. Eleven locations are part of a national weather service engaged in forecasting, predictions, or support. Six institutes are government research centers, and six are primarily educational institutions (universities). Approximately 20% of the locations visited were non-governmental organizations (NGOs), and one (the IBM India Research Laboratory in New Delhi) is run by industry. In addition to these 29 institutes, discussions were conducted with several divisions of the WMO in Switzerland. In the course of all these visits, discussions were conducted with well over 100 researchers, forecasters, students, and managers.

Table 1.1. List of institutes visited

Institute Country Type
African Academy of Sciences Kenya NGO
Asia Disaster Preparedness Center - Bangkok Thailand NGO
Center for Atmospheric Sciences - IIT Delhi India Education
Center for Hydrometeorology of South Vietnam Vietnam Met Service
Disaster Mgmt. Unit, Ministry of Ag. and Rural Development Vietnam Govt Research
Drought Monitoring Center Kenya NGO
Hanoi University, Dept. of Meteorology and Climatology Vietnam Education
Hydro-meteorological Service of Vietnam - Hanoi Vietnam Met Service
IBM India Research Laboratory - IIT Delhi India Industry
IMD Delhi India Met Service
India Meteorological Department - Pune India Met Service
Indian Institute of Tropical Meteorology - Pune India Research Center
Institute of Meteorology and Hydrology - Hanoi Vietnam Met Institute
Kenya Meteorological Service Kenya Met Service
Kenya National Academy of Sciences Kenya NGO
Makere University, Dept. of Computer Science, Kampala Uganda Education
Makere University, Dept. of Geography, Kampala Uganda Education
Myanmar Dept. of Meteorology and Hydrology Myanmar Met Service
National Center for Hydro-meteorological Forecasting - Hanoi Vietnam Met Service
National Center for Medium Range Weather Forecasting - Delhi India Met Institute
National Ozone Centre - IMD Delhi India Met Institute
Nepal Dept. of Hydrology and Meteorology Nepal Met Service
Physical Research Laboratory - Ahemedabad India Research Center
START Southeast Asia Secretariat - Bangkok Thailand NGO
START Temperate East Asia Secretariat - IAP Beijing China Education/Research
Thailand Meteorological Department Thailand Met Service
Uganda Dept. of Meteorology Uganda Met Service
Uganda National Meteorological Center, Entebbe Uganda Met Service
University of Nairobi Dept of Meteorology Kenya Education

The personal interviews covered similar topics as the written survey, and in some cases those interviewed personally filled out survey questionnaires as well. In addition to the written survey, the site visits included tours of libraries and computer facilities, as well as laboratories and offices. The field visits complement the survey results in a fairly unique way and allow some of the biases in the survey to be examined in more detail. In addition, personal interviews touched on cultural issues that could not be fully anticipated in the survey. Those issues are discussed in detail in this report. A summary of these visits is contained in Section 3, and more details appear in Appendix B.

Internet access, availability, and cost were also tested in the course of the survey. A standard set of files and websites was accessed from 25 locations in the eight countries visited. This set included accessing a website and downloading a file located in the United States and accessing a website located in Kenya. The times for access (or failure of access) were recorded and tallied. The results are presented in Section 2.

Interviews with journal editors

In the summer and fall of 2001, several journal editors were interviewed to gauge their perceptions of the information divide. These editors represent several established and well-regarded journals in climate and meteorology. The editor of a developing-country journal was also interviewed in early 2002. The interviews attempted to understand the perceptions of editors toward article submissions from researchers in developing countries, to gather statistics or anecdotes, and to garner ideas and recommendations from the editors. A summary of the discussions is also presented in Section 2.


Survey results

The sample size from the survey of 49 responses does not permit statistical rigor, but common and consistent answers from the survey and field responses (which add another 100+ responses from 8 countries) do allow a broad picture of the information divide at the beginning of the twenty-first century.

Respondents

As stated earlier, 49 responses to the survey were received by mail or over the Internet. These responses came from 23 different countries: 10 from South America, 7 from Africa, 5 from the Middle East, Russia and the Former Soviet Union, 11 from India, 14 from Southeast Asia, and 1 from China. The distribution is skewed toward those regions in which the field survey was conducted. This distribution is likely due to surveys completed by researchers at institutes visited, or colleagues of those researchers.

Of the researchers responding to the survey, 60% work at government institutions, 25% at educational institutions, and 15% at NGOs. Approximately 25% of respondents are primarily concerned with meteorology and forecasting. A further 25% are concerned with climate and climate change. The remainder represent a mix of fields from hydrology to energy and environment to oceanography. Nearly two-thirds are engaged in research and a third in teaching, with 25% also having some management responsibilities. Ten percent of the respondents indicated forecasting is their primary job, or one of their primary jobs.

Of the researchers surveyed, 60% had doctoral (PhD, D.Phil or equivalent) degrees. Approximately one-half of those degrees were received in developing countries, and the other half in developed countries. English is a native language for only 3 of the 49 respondents. Despite this, over 80% (42 of 49) speak and write English. Spanish was the next most common language, spoken by 20% of respondents.

Access to resources

Survey respondents were asked several questions about what journals or publications they use for their research, and how they access these publications.

Respondents were asked to name up to four periodicals that are most important to their work and whether they had easy access to them. In addition, respondents were asked to name up to four additional periodicals they would most like to receive. The top four replies are in Table 2.1. About 2/3 of the respondents did not have timely access to the publications they use in their work. The average delay for receipt of journals was about 6 months, with 10% of respondents facing a delay of a year or more to receive needed journals (if at all). However, 60% did not find this delay detrimental to their work.

Table 2.1

Top 4 Journals Read (25% of total requests)
1. Journal of Climate
2. Monthly Weather Review
3. Journal of the Atmospheric Sciences
4. IPCC publications

Top 4 Journals Desired (30% of total requests)
1. Journal of Climate
2. Monthly Weather Review
3. Boundary Layer Meteorology
4. Journal of the Atmospheric Sciences

Note that most of the journals listed in Table 2.1 are publications of the American Meteorological Society (AMS), and these rank high on the ISI citation lists (ISI Essential Science Indicators, 2003). However, Boundary Layer Meteorology, from a private publisher, costs 1200 USD per year for an individual and 2200 USD a year for a library subscription. While most of the other journals mentioned are "high-impact" journals (in the top 20 of 168 journals) for citations or citations per article), Boundary Layer Meteorology is not (49th in citations and 72nd in citations per paper). The lack of access may have to do with cost.

The most common way in which respondents find out about research is from a reference in another publication, followed by locating and reading articles directly. Internet browsing was the next most common way to find articles. One-half of the respondents used databases to find information, and most of these were electronic (Internet or CD-ROM). It appears from discussions that broad Internet searches (using search engines such as www.google.com) on topics or names, combined with searches of known research websites, are an increasingly common way to access information. These Internet tools have surpassed databases, conferences, and conversations with colleagues as ways to hear about new results and articles.

Research or forecast data was also the subject of several survey questions. The most common data source for respondents was self-generated data (used by 80%), followed by international analyses and other domestic data. The international analyses, typically from the National Centers for Environmental Prediction (NCEP) and the European Center for Medium-range Weather Forecasting (ECMWF), were highly valued by forecasters.

Seventy percent of those surveyed did not have access to all the data they need for their work. This includes forecasters and researchers. The most common reason cited was cost (nearly half of respondents). Insufficient Internet access was also cited by 1/3 of those who need more data. In particular, there were unprompted complaints, in both written remarks and respondents at visited institutes, about the lack of access to ECMWF global analyses. Many were confused that a publicly funded entity charges other publicly funded entities for data. A small number of respondents also had problems with either domestic bureaucracy or their own computer and data processing facilities. Other issues included problems with data formatting that are pervasive everywhere, as well as a desire for software or training materials for the use of larger data sets.

Much of the problem with the cost of data or journals stems from the fact that most nonprofit entities in developing countries are unable to pay for information, whether journals or data. There are some small budgets for journals, but virtually no budgets for data collected outside a country. The cost of a journal in foreign currency may be a significant fraction of a researcher's salary in many countries. This is also true of Internet access.

The survey results were, however, not entirely bleak. Several researchers noted the increasing availability of data on the Internet and their increasing ability to access certain data sets and forecast products. In some weather services and forecast offices visited, forecasts and analyses from the Internet are used more frequently than information from the Global Telecommuni-cations System (GTS) for forecasts and prediction.

Other efforts

Several existing efforts seek to develop capacity for climate science in the developing world. Some of these efforts include the educational and training efforts of the WMO, outreach and education efforts of the global change system for analysis, research and training (START) project, and the outreach efforts of the IPCC. The International Research Institute for climate prediction (IRI) also has an outreach program to train developing-country scientists in climate forecasting. These efforts are quite rightly focused on specific scientific objectives related to each organization's mission.

Several efforts are currently under way on a formal and informal basis to deliver journals to researchers in developing countries in a more timely fashion. The National Oceanic and Atmospheric Administration (NOAA) has for several years provided free journal subscriptions from the AMS to approximately 100 locations worldwide. In addition, several major libraries in developed countries send extra collections to developing countries. These programs, in conjunction with other donations from individuals, have been a major lifeline for researchers in developing countries. Most libraries visited depend on donated or free journal subscriptions for their collections. This makes the journal collections uneven. Journals only arrive when external funding is available or donations are found. The continuity of collections is lacking in many places, and issues are missing because either donors didn't have that issue to send, or a program didn't have a subscription for a particular year. Most of the journals sent free of charge are publications from not-for-profit societies. Very few recent copies of commercially published journals were seen.

Additional programs are under way to make electronic access available to researchers in developing countries. Some of these initiatives are specific to the climate and meteorology community, and some are more general across scientific disciplines. Several efforts are under way to promote electronic access to journals. These efforts rely on existing websites and tools built for existing subscribers. The AMS is undergoing a transition to open access for back issues of their journals. The American Association for the Advancement of Science (AAAS) is doing the same for the journal Science.

These programs have advantages and disadvantages. Advantages stem from the low marginal cost of supplying electronic information to developing countries, which makes the programs much cheaper, and easier to scale for donors and suppliers of information. The disadvantages of electronic distribution programs are that they require some computer capacity, and most require Internet access faster than a dial-up modem (>56 kilobytes sec-1) to be practical (see discussion below under Communications for more details). A further disadvantage of Internet-based programs is that while they often permit access to an entire and complete catalog of back issues, this access ends if the program terminates or if the connection or computer is inaccessible, while print journals are always available. Note that this latter concern is an inherent problem with networked electronic libraries. If the network is down the entire library is closed. Personal communication with colleagues in the United States indicates that this is a growing concern of research libraries and researchers in developed countries as well.

In addition to these specific activities in the meteorological and climate sciences, there are several larger efforts occurring around the world to try to promote and enhance access to information. The Third World Academy of Sciences is promoting an open electronic archive of journals across the sciences. This archive is targeted to developing countries. In addition, an effort called the Budapest Open Access Initiative has been formed to attempt to promote open access to journals and to encourage public scientific research organizations not to designate copyright to other parties, particularly for-profit entities. These are only two examples of many efforts to bridge the "digital divide" and the information divide in the physical and biological sciences. The number of efforts is increasing with time.


Publishing

Perceptions from authors

According to the survey, papers in climatology and meteorology are published overwhelmingly in English. Seventy-five percent of the developing-country scientists polled submitted their papers in English, though it is a native language for only 3 of 49 respondents. Despite the overwhelming prevalence of English, 45% of the papers recently submitted by these researchers went to emerging country journals (15% of the total to India).

Some of these statistics are biased by the sample chosen. Part of the concentration on English is probably due to the significant fraction of replies from India, where English is the common language. In parts of West Africa, several countries still publish locally in French. And there are significant publications in Spanish from Latin America, though there is little doubt that much of the international literature is in English. In addition, there are other trends and incentives (mostly economic) in many countries to learn English. English usage was higher among younger scientists in rapidly developing countries in Asia, notably China and Vietnam.

Interestingly, language was not considered a major barrier by many of the survey respondents. Only 1/6 of respondents indicated that language was an important issue for them. Translation help was important for a small percentage of respondents. A greater proportion (perhaps 1/3) of the researchers visited thought language is an issue. Difficulties in reading and writing English varied dramatically country by country. Former colonies with English as a colonial language (Kenya, Uganda, India, Myanmar), a large number of those countries sampled, had less trouble with English, and hence contribute to a bias in the survey.

The average time reported by these authors from submission to publication was 11 months. This is comparable to the time required in developed countries. However, the tail of the distribution is skewed toward longer time periods, with some papers taking well over 18 months to publication.

Authors complained about several things in the publishing process in developed-country journals. Page charges for articles are prohibitive for most developing countries. The charge for an article, typically on the order of 1000 USD per paper, is easily the monthly salary of a researcher in many countries. While most journals will waive the fees for researchers in developing countries, many authors are not aware of this. This results in the self-selection of certain journals without page charges. Many authors send papers to specific journals based on the cost (or lack of cost) of publication. Cost was a barrier for 2/3 of those surveyed, as well as a similar proportion of those interviewed during field visits. Page charges in particular were mentioned by half of the respondents. Other desired publication assistance included more access to articles and data.

The type of submission to international journals from developing countries is often different than from developed countries, and is different from that submitted to domestic journals in developing countries. Authors know that certain types of work (reanalysis of an old data set, publishing a limited data set, or research that may be more "applied") is not suitable for certain journals, and they appear to be very selective about sending work to international journals. This is really a difference in priorities between international journals and developing-country scientists. It is perhaps expected in some areas of the field (dealing with advanced techniques or instrumentation). But for climate research, for example, standard data products in new locations or historical records are themselves new research. For forecasting, there are approaches from other developing countries that might be useful. International journals for developing countries are necessary to cover some of these topics, and these journals are rare.

In addition, some respondents and researchers indicated they detected a "reviewer bias" toward developing-country submissions. Part of the bias may be cultural. The scientific method of questioning previous results and hypotheses is culturally different than the deference to authority found in many countries. While it is true that senior scientists in developed countries may also face less scrutiny, in principle all work is evaluated on its merits during peer review. Criticism of the work of a senior person in many developing countries is often not acceptable. This cultural deference to authority may also explain a perceived resistance to using and adopting new methodologies and techniques. These issues were raised by developing-country scientists (often younger ones) during several of the field visits.

Perceptions from editors

Editors of five major journals were interviewed in the fall of 2001 in order to ascertain their perceptions and experiences with submissions from developing-country researchers. In general, editors identified a difference in submissions between developing and developed countries, as well as a difference in acceptance rates, which comes out of the statistics for all the major journals examined. However, there are not many submissions to some of these major journals from developing countries. The editors suppose that many authors will self-select away from these journals either due to cost, time, or subject matter. All of these suppositions are correct based on the discussions with authors. Cost is an additional consideration, as noted above.

Similar to the perceptions of authors described above, language was not cited by many editors as a problem, and by none as the major problem. One editor noted that problems with language were usually correctable; however, problems with methodology or subject matter were more serious. Another editor of an English language journal noted that the paper with the most language problems, received in the last few years, came from Western Europe and had a co-author who was a native English speaker. Formatting issues were also a concern to editors. Submissions did not follow conventions (double-spaced, containing an abstract). While the editors considered these issues minor, it was noted that not conforming to norms was symptomatic of larger issues.

Access to current literature and methodology were the two most frequent criticisms about submissions received from emerging countries. Editors noted that authors were not aware of the current literature from the last 5 to 10 years, and that their choice of methodology was often outdated. They often repeated analyses with a slightly different data set, or used older and outdated data products. Papers were often thought to be more like a "textbook," repeating what is known, documenting observations without interpretation, or a model exercise with little evaluation. These also were the most difficult problems to address, and in many cases these issues made papers unpublishable. Editors also noted that postal communication with developing countries is often difficult and unreliable. Electronic communication was therefore viewed as a great leveler.

There are some systematic differences among countries in the developing world noticed by editors: some good, some bad. There were some favorable comments about papers and programs in Brazil and the "Southern Cone" of South America. India was also noted for the quantity, as well as quality, of manuscripts. However, India also appeared to the editors to have the biggest problem with the review process and accepting criticism. The editors noted that this was more of a problem for senior scientists from India than younger ones. In addition, editors are receiving increasing numbers of papers sent in electronic format from some locations (South America was noted especially by one editor).

Editors also provided suggestions for assisting publications from developing countries. Electronic communication was viewed as a big help to the editorial process for developing countries, reducing time delays and postage charges. Enhancing on-line access to papers would be another way to increase access to the literature.

Another suggestion was to enhance co-authorship of papers with Western researchers. Many scientists from developing countries who are publishing successfully have collaborations with institutions in developed countries and have spent time working in developed countries. Also, exposure to current science at meetings would help, rather than simply being exposed to the literature.

Several editors suggested that more financial assistance to send journals and information to developing countries would be beneficial, as well as enhancing electronic access and communication for editors and contributors to journals.

Statistics

Preliminary statistics available from several major US and European journals confirm the perceptions of authors and editors. The simplest metric to use is to look at acceptance rates: the number of papers accepted divided by the total submitted. The acceptance rates for manuscripts for all publications from the AMS are approximately 80% for those whose corresponding author is in the United States, and nearly 50% for all submissions outside the US (including other developed countries in Europe and Japan). For a single AMS journal with available data, the acceptance rates are nearly 75% for manuscripts from the United States, and 15-20% for manuscripts from developing countries. A major European journal has similar numbers (70% acceptance from developed countries and 20% for manuscripts from developing countries).

Communications

Several aspects of communications were explored. Basic information about how and with whom researchers communicate was obtained. In addition, based on the hypothesis that information is increasingly electronic, more detailed information about computing and network capabilities was examined. In addition to the survey responses, data were obtained by visiting various facilities and by using researchers' communications infrastructure. Objective tests of Internet access from various locations were also conducted.

Survey responses

The most common means of communication between researchers and their colleagues was by electronic mail. This applies to domestic colleagues as well as those from other countries. The results are summarized in Table 2.2. Conventional mail has dropped in use among researchers for communication with colleagues. For over 80% of researchers, the ranking was their preferred ranking.

Table 2.2: Means of communication

Rank Domestic Foreign
1 Email Email
2 Face Fax
3 Phone Face
4 Fax Mail
5 Mail Phone

Nearly 80% of those surveyed, and a slightly lower proportion of those visited, work with international networks (such as the IPCC) or collaborate with scientists in other countries. Of the collaborations noted (respondents were encouraged to name up to three countries), more than 2/3 were with a developed country (US, Europe, or Japan) and less than 1/3 of the collaborations were with other developing countries. This situation is perhaps expected, given the desires of developing-country scientists, but one may ask whether the lower level of collaboration with other countries either in the "neighborhood" or with similar conditions is a concern. It would appear at first glance that countries may suffer from similar types of weather patterns, similar climate events and extremes, and would have much greater need for collaboration on research projects with each other. The communication between developing countries appears to be higher in Southeast Asia, with certain projects such as research on the Mekong river system, spanning borders. Some communication between countries in East Africa, which face common drought problems, was also noted. However, there appeared to be little communication between India and either Africa or Southeast Asia.

Computer & Internet access

All of the scientists surveyed have some access to a computer. For most, their primary use is in the office. Most of the computer systems seen in the field survey were models from the mid- to late 1990s. Most of them run some version of Microsoft Windows software and have CD-ROM drives. In Southeast Asia and India, the computers in many places were less than five years old. Only about 20% of those surveyed are required to pay for computer usage. In some places, notably East Africa, most of the computers are donated. According to the scientists in East Africa, the most difficult thing is not to get donations of equipment, but to get the equipment through customs. Several places in Kenya have received computers from the WMO in Switzerland, which had to be brought into the country in diplomatic pouches.

Ninety percent of those surveyed have electronic mail and/or access to the Internet. Most people use the Internet at the office. In many locations these connections are shared. Internet access to the desktop for all but senior researchers is only found in more developed countries, such as Thailand or China. Similar to computer use, 25% pay a marginal cost for access every time they use the Internet. One-third of survey respondents, and close to half of those visited, indicated that cost limits Internet access (see below). Cost was a significant barrier especially in Africa and the Former Soviet Union. In addition to email and the Internet, nearly 3/4 use file transfers for downloading information, and less than one-half use the Internet for remote access to other computers.

Net tests

Several metrics of Internet use were obtained for the countries visited. These indicators help to put the comparisons between countries into perspective. These data come from international sources, as well as from data collected in various countries as part of the survey.

Figure 2.1 illustrates the level of Internet and personal computer (PC) use per capita in various countries based on data from the International Telecommunications Union (ITU, 2002). Of the countries surveyed, Thailand has the highest proportion of Internet users (5%) and personal computers (nearly three computers for every 100 people). For comparison, in the United States 50% of the population uses the Internet, and there are 62 computers for every 100 people. Developing countries are still nearly an order of magnitude behind most European countries (Germany is given as an example, with 1/3 of the population using the Internet).

More striking is the location of Internet hosts, indicated in Figure 2.2. There may exist a global Internet, but in 2001, 75% of all the host computers on the network reside in the United States (with 5% of the worlds population). With 13% of the world's population, Africa has only 2% of the world's host computers (mostly in South Africa). This is dramatic evidence of the "digital divide" between developed and developing countries.

Internet access and cost was examined directly in the countries visited. To test Internet access, the access time for three different locations on the Internet was tested. One test measured the time to load a small (0.25MB) article published in a journal and stored in Portable Document Format (PDF) on a web page at the National Center for Atmospheric Research in Boulder, USA. A typical PDF article from a journal runs 0.5-3MB. The second test calculated the time to load the Internet home page of the National Oceanic and Atmospheric Administration's Climate Diagnostic Center (www.cdc.noaa.gov), also located in Boulder, USA. The third test was the time to load the Internet home page of the Drought Monitoring Center for East Africa (www.meteo.go.ke/dmc/), located in Nairobi Kenya. The results for accessing these three locations are indicated in Figure 2.3. Internet access in most countries, either at dedicated Internet cafés or in offices with mostly dedicated lines, is usually reasonable for most applications. For activities that take longer than a minute or so, however, these speeds can be limiting. Keep in mind that these statistics measure the time to load a small article and a single web page with more complicated links and photos. For an average article or session, access in Africa is prohibitively slow from most locations tested. It might take 5-10 minutes for an average-sized article, assuming that transfer rates were sustained, and the server or phone line did not disconnect. Add to this the time to search for the article, with nearly 2-5 minutes per page viewed, and the time to access a single article may run 30 minutes or more. Also, since charges are a function of time, access gets expensive, particularly with frequent disconnects.

These numbers are in line with the success rates that were reported by survey respondents.

Figure 2.4 illustrates the approximate percentage of time that researchers are able to access various functions on the Internet. The success rate was calculated by assigning an average success rate to verbal categories: almost never (5%), sometimes (30%), usually (70%), almost always (95%), and averaging across all 49 responses. Success rates for email are the highest at 75%, with the web "usually" accessible (70% of the time), and remote access to other machines available only half the time. FTP (file transfers or downloads) may give some indication of the potential for downloading articles and other information. It appears to be successful only 60% of the time.

In addition, the "street" cost of Internet access was examined in various countries. The relative cost was calculated by dividing the cost of an hour of Internet usage by the cost of a typical meal in a restaurant. This metric is presented in Figure 2.5. For example, in the Netherlands, Internet access costs 1.25 EUR/hour, and a meal typically 10 EUR/meal, so the cost is 0.125 meals per hour of Internet access. Also note that while Myanmar does not have access available to residents, they can send a single email for about the cost of a restaurant meal (this would probably work out to something like 10-15 meals/hour on the scale in Figure 2.5).

Figure 2.5 has a similar structure to Figure 2.3. Higher costs are often associated with slower access speeds (note the speed and cost in China and Uganda on either ends of the scale). These numbers do not resemble GDP per capita (Figure 1.1). Rather, with the exception of India, they are much closer to the relative relationships of total GDP. Larger and richer countries have higher levels of Internet access. The relatively high cost of access in India is a result of extremely low costs for food, and lower concentration of Internet users (Figure 2.1). The physical bandwidth in India was quite fast when tested either at cafés or in researchers' offices.

The results were confirmed in discussions with researchers. Africa in particular suffers from physical bandwidth problems (it is a relatively poor place with low population density). A similar perception can be gathered from looking at physical transportation infrastructure in Africa. The road networks are thinner and smaller than in many places in Asia. Myanmar, with a lower GDP per capita than Uganda or Kenya (Figure 1.1) and with far less foreign assistance, has about the same level of transportation infrastructure as East Africa, largely due to population density.


Country Case Studies

Eight countries were visited during the course of the survey, and more detailed information is available on the state of these countries. As indicated in the introduction, these countries form a sample of several different levels of capability and size.

Countries: Summary

Comparisons among the countries visited are very revealing. There is tremendous variety in the issues facing these countries, and an information divide that ranges from a crevasse, easily bridged, to a canyon. Some of the broad themes and issues for each country are noted here, with more detailed notes in Appendix B. In general, certain countries are struggling to maintain a basic level of weather services and forecasting (Uganda, Myanmar, and Nepal); some are attempting to do more seasonal prediction (Kenya, Nepal, and Vietnam); some are already running numerical prediction models (Vietnam, India, China, and Thailand);and some are doing extensive fundamental research (China and India).

East Africa

Two countries in East Africa were visited in the course of the survey: Kenya and Uganda. Interviews were conducted with meteorological services, NGOs, and universities in both countries. Common issues of concern include basic access to the Internet and inadequate levels of funding for research. As indicated in the results, Africa suffers from a generally low level of development in all infrastructure, whether physical or electronic. In East Africa, this is due to poverty and low population density, which affects telecommunications and Internet access as well. The physical bandwidth simply does not exist in the region to make high-speed Internet access available, even for government entities or researchers.

The levels of poverty in the region significantly limit the funding available to support research and services across the board. Ability to pay for anything is limited. Many researchers are forced to use donated computer equipment or purchase it themselves. Funding for information is nonexistent for research and very limited even for operational meteorology. Research conducted in climate and meteorology is usually funded by international agencies or donor institutions.

However, education levels are high, and the widespread use of English as a second language is a benefit to many in the region. There is an active scientific research community, particularly in Kenya.

Kenya

Kenya has a large and historically active meteorology community, dating from colonial times. However, along with the condition of the general economy, that community is eroding. Kenya's research community fared better 20 years ago than it does today. Gaps between Kenya and developed countries have widened over this period of time. This is a general condition of the economy and not localized in the scientific community.

Currently, infrastructure is poor for meteorology. The number of domestic reporting stations has declined over the past 10 years, largely due to a lack of funds to continue programs to support voluntary reporting. The meteorology community uses a donated satellite link from France as its sole Internet connection. This provides good service to the Kenya Meteorology Department, the WMO regional headquarters, and Drought Monitoring Center co-located within it, as well as the Department of Meteorology at the University of Nairobi, and even the Department of Meteorology and Hydrology in Kampala, Uganda (see below).

The general education level in the meteorology community is high. There is a good tradition of higher education in meteorology at the University of Nairobi, which also trains large numbers of students from other African countries.

The influx of funds from donors has helped in a number of areas. In particular, Kenya serves as a regional training center for East Africa, with educational and training facilities, as well as a large meteorology library. There is an active Drought Monitoring Center, sponsored in part by the WMO, which helps to issue seasonal forecasts and predict extreme events, working with researchers and data from throughout the region. This program is well funded and has been successful.

Uganda

Uganda is struggling to maintain a basic level of meteorological services against very difficult odds. The general situation in the country is that it is struggling to recover from a civil war and to maintain its borders against disputes with neighbors. The situation has been getting better, but at least 25%, and possibly as much as 50%, of the government budget consists of donations from developed countries.

Communications facilities are even more limited than in Kenya, and there are even fewer available resources. The entire meteorology service gets its information through one modem line located in Nairobi and a backup satellite receiver. No data are received from the neighbor to the west (Democratic Republic of the Congo), and this makes forecasting difficult. The cost of Internet access is quite high, almost unaffordable for the Meteorology and Hydrology Service. There is no funding to purchase journals and little for Internet access. Downloading a single 1MB article takes about 40 minutes and costs 4 USD, which is prohibitive.

Uganda has several examples of innovative as well as failed programs to bridge the information divide. On the good side, they are using satellite radio receivers (Worldspace) to obtain weather and forecast data, as well as to broadcast warnings to remote regions. This system is well suited for hydrology, sending warnings to far-flung regions of the country. There is a climate records rescue project to digitize old climate records, but several older projects (such as microfilming the records) were unsuccessful.

The major university in the capital does work with the Meteorology Department to train forecasters. The university has recently received an influx of funds to their computer science department, but maintaining campus-wide Internet access is, again, very expensive.

South Asia

Two countries in South Asia, India and Nepal, were visited. Several major centers in India were visited and one center in Nepal. In general, South Asia, particularly India, has a very different situation than East Africa; the level of development in India is much higher. This is apparent in the general infrastructure (roads, rail, etc.) as well as in electronic communication. Internet bandwidth is not really an issue. Resources are an issue in Nepal. In India, it is simply a question of how to allocate resources. Because of the large and capable science base in India, some interesting cultural issues arise for scientists trying to publish in developed-country journals, which they do on a regular basis.

India

India is a country of tremendous contrasts, with world-class research centers, supercomputing facilities, and locations with Internet access faster than the WMO offices in Switzerland. These facilities coexist with tremendous poverty and a much larger number of substandard research and education facilities. There is a class of top-tier universities and research centers with good facilities, which were mostly the class of center visited in this survey. Researchers discussed the problems faced by non-world-class centers in India, which were more typical of East Africa, lack of resources, poor computers and communications infrastructure, and a lack of access to basic textbooks and educational materials in the field.

English is a de facto second or third language among a large percentage of the educated population and is widely used throughout all society. All the researchers encountered were fluent in spoken English, and few had trouble with written English, though some students were not totally fluent.

The Indian Meteorology Department (IMD) is a large institution with many forecast centers and an active research program. The Indian government operates its own weather satellite. Several active groups are working operationally or experimentally with numerical models. Much of the cutting-edge research work focuses on the dynamics and predictability of monsoon circulations.

Access to data and research information is quite good. Most major journals can be found in a few major locations, the world-class centers. Beyond these, the information is not available. Several libraries were visited, with comprehensive collections of most common journals and textbooks, and extensive archives. An active data digitization project is under way, but it is not clear if the data will be made accessible or migrated to a usable form.

India also has an active scientific publishing community. There are a few world-class journals in meteorology, including Climate Change, and Mausam. Mausam, the official journal of the IMD, has a long history, dating back almost 100 years. A broader set of scientific publications exists as well. Many of these publications receive submissions from all over the world.

Some interesting cultural issues were highlighted in the course of the interviews, which shed some light on publication concerns identified above. India suffers or survives under the weight of a large bureaucracy, which sometimes does things only for its own sake. Researchers indicated a strong cultural deference to authority. This deference, for example, prevents junior staff from critiquing the work of senior staff. It also means that new methodologies and techniques are not necessarily encouraged, and that it is safer to work on more established projects. This is particularly the case at smaller institutions. These cultural traits appear stronger in India than in Western countries, and may explain why there appears to be consistent cultural differences and problems with papers submitted from India to Western journals.

Nepal

Nepal is a much smaller and less-developed country than India, with more significant physical and development barriers. The Meteorology and Hydrology Service has a higher level of infrastructure than in East Africa (Internet access, computers, etc.), and a competent staff, but fewer resources than in India. Due to geography, the service struggles to maintain a network and conduct forecasting in the presence of highly variable conditions. Significant donor assistance exists for all research conducted. In most cases, research topics appear to be chosen by the donor countries.

Data needs are met through a satellite reception link and through use of connections with India. Public websites are also used to supplement data received through the Global Telecommunications System (GTS) or via direct satellite links. Again, there are no resources to pay for publications, and few researchers try to publish their research in international journals. Several programs are under way with researchers who study mountain meteorology (these foreigners publish their research in international journals). Internet access is improving, and sometimes researchers are able to get copies of articles on the Internet.

Southeast Asia

Southeast Asia has a wide variety of capacities, ranging from Thailand, which is relatively highly developed, to Vietnam, which is developing rapidly, to Myanmar, which has a low development level for the region and is struggling to move forward. There are some interesting issues of language and education in the various countries, and some interesting insights into different problems faced by countries, even those with significant resources.

Thailand

Thailand is one of the most developed countries in this survey. It has the highest per-capita GDP (Figure 1.1), as well as the highest percentage of Internet use (Figure 2.1) and servers of any country surveyed. The speed of Internet access and the penetration of computers into offices is high. However, there are some interesting aspects to this development. Many universities and several active NGOs are working on disaster management, predictability, and climate change. Yet there is no faculty of atmospheric science or meteorology in any university, and the strongest tradition in the field is an active research program on weather modification, which is a pet project of the King of Thailand. The first professorship in meteorology or climatology is due to be filled in late 2002 or 2003.

The NGOs are quite active and often based within universities. As a communications and economic center, Bangkok has a number of regional organizations (as well as a large United Nations presence) that work throughout Southeast Asia.

The Thailand Meteorology Department (TMD) has a relatively new and spacious campus outside of Bangkok with good facilities and a technically advanced forecast room. There are 10 Doppler radar sites in the country and a well-developed network. Yet only two of the sites appeared to be working when the visit was conducted. TMD has several research divisions, including one for climate. However, training must occur in-house without university support. In addition, the general level of English proficiency was low. This could be seen in the TMD Library, which has a relatively large collection, but most of the books are old, and it contains few journals. Most of the researchers at the TMD do not use the library, and many have difficulty reading English. Though Thailand has facilities and an advanced level of development, the general education level in climate and meteorology is low. This is paradoxically due to Thailand's historical independence, wealth and use of a native language other than English. It illustrates some of the problems faced by successful developing countries that have their own strong traditions and cultures.

Myanmar

Myanmar is a very different case from Thailand. Its level of development is more in line with East Africa than with other countries in Southeast Asia. The country has virtually no Internet access. This is largely by government choice, which restricts the flow of information and keeps the country insulated from the outside world. Even the Department of Meteorology and Hydrology as of May 2002 was unable to gain approval to access the World Wide Web.

Interestingly, the Meteorology and Hydrology Department recently invested a great deal of money on a sophisticated computer system to produce television weather forecasts and has hired and trained staff to be television presenters for the government network. The system comes with some expensive computer hardware to render the presenter's graphics over a map of the country showing weather systems. This same system contains the software and capacity for running a numerical weather prediction model which the department is not currently using. This system has raised the Department's visibility in the government, as well as in the general population and is judged to be a valuable resource and a great success, despite its cost.

As the monsoon season started in late May 2002, the Department had no satellite link to the Global Telecommunications System (GTS) because of local line problems between the government receiving station and the Department forecast center (about 10 km apart). Satellite data were viewed from a secondary system donated by China that could receive information directly (using a small dish) from either the Chinese or the Japanese weather satellites. Forecasts were constructed with limited satellite pictures and charts, since there was virtually no GTS data. The department was considering investing in world space receivers (similar to those in Africa) to attempt a remedy to the situation. It is not clear whether permission from the government for such a device would be forthcoming.

In general, the senior staff at the Department is well educated, often receiving advanced degrees in a developed country. There is a desire to send more staff abroad (to India, Europe, or the United States) for training. Some funding exists for this effort, but political considerations make this effort difficult. The level of English proficiency is good among most of the staff. There is a well-developed and maintained domestic network of meteorology and hydrology stations. Forecasting is done using empirical or statistical methods (such as a statistical hydrology model for streamflow). Myanmar is rarely able to access to Indian meteorological data, which can be detrimental to forecast skill.

Vietnam

Vietnam is a large country that is moving forward rapidly after a long period of several wars throughout the latter half of the twentieth century. The pace of development increased after the collapse of the Soviet Union and the influx of foreign investment. The biggest change, according to researchers, came in 1995 following the normalization of relations with the United States. While the government still restricts the flow of information (keeping Internet rates high and international phone rates the highest in the world), things appear to be improving rapidly. It is still difficult for researchers to get high-speed or dedicated Internet access. The meteorological service itself appears, however, to have good resources.

Several different locations were visited in North and South Vietnam. An active research community in climate, meteorology, and hydrology works largely under the umbrella of the Department of Meteorology and Hydrology (DMH), with large regional forecast centers in the south and north of the country. Hanoi University also has a small but significant Department of Meteorology, which serves as a training ground. Both the university and researchers in the DMH are working to get numerical weather prediction models running.

Vietnam is also actively collaborating with its neighbors, particularly in the southern part of the country, where the Mekong River Delta is located. There are several basin-wide projects that include Cambodia, Laos, Thailand, and China. In addition, Vietnam is now helping to train forecasters in Cambodia and Laos.

Vietnam is unique in having an incredible variety of "second" languages. The library of the Hydro-Meteorological Center for South Vietnam in Ho Chi Minh City contains publications in French, English, Russian, and Vietnamese. Senior scientists were trained largely in Russian, whereas younger scientists are increasingly learning English. Rapid development in the country has made it difficult to attract staff, because technical expertise is usually valued more highly in other sectors of the economy rather than government service.


Northeast Asia: China

China was the only country visited in Northeast Asia. It is perhaps not a "developing" country in the same sense as the others in this survey, but it is certainly a country with a developing and emerging geoscience base. Like India, China is a large country with tremendous contrasts, is developing rapidly (perhaps even more rapidly than India), with many world-class centers in the country. The level of development in information technology is notable. Internet access is widespread and inexpensive. As a large manufacturing center, computer equipment is also cheap. Resources exist to purchase capital equipment as well in the top-tier centers.
Journals are widely available, and like India, a number of domestic science journals are produced, although they do not receive as much exposure or foreign submissions, because most of them are in Chinese. Again, availability is limited in many cases to top-tier centers. As an example, the Chinese Academy of Sciences has a very extensive on-line catalog and information search tools on its website (www.las.ac.cn) in both Chinese and English.

The availability of Internet access gives researchers the ability to search and find papers on-line, and a large percentage of researchers visited at the Institute for Atmospheric Physics in Beijing make use of this.

The use of English is improving in China. In many cases, younger researchers have a greater command of English than do more senior researchers. English as a second language is spreading, as mentioned earlier. In addition, Chinese researchers can afford to hire editors to help write papers in English if they choose to send them to foreign journals.


Synthesis & Recommendations

Synthesis of results

The results of the field and written survey permit a number of observations. Despite the individual conditions encountered, several common issues emerged. Some are not surprising and validate the initial hypotheses that there is a significant information divide, and that it exists as a problem and limitation for many. Some of the results are counter to initial assumptions. Some of the reasons for problems with developing-country acceptance rates are easier to understand in light of the survey results.

Expected

Several results from the survey are not surprising. Chief among these is the high demand for scientific information of all types in developing countries. The needs of researchers vary, but most do not have access to current journal articles, and many would like better access to data. There also appears to be a significant deterioration of the Global Telecommunication System (GTS) in several regions of the world, so that meteorological services in some of the least developed countries are not receiving information on a daily basis. There is less information available now than in the past. In many locations, the basic protocol of the GTS, that information is provided locally and received from everywhere else, no longer functions effectively, and countries circumvent the GTS to get the information they need.

It is also not surprising that there are few, if any, resources in most developing countries to pay for information. In some cases, there is barely enough to pay salaries. In most countries, information from foreign sources must be paid in hard currencies. These are inordinately expensive for most locations. Even where such resources exist, such as in India, the lack of hard currency makes it difficult to sustain funds to purchase information. This applies to journal articles as well as analyses and data.

The Internet is changing the way people communicate in the developing world, just as it is in the developed world. Despite limitations, almost all of the researchers visited and surveyed have some sort of access to the Internet, usually to electronic mail. Everyone, from researchers to editors of international journals, agrees that the situation will likely improve, access will increase, and that electronic communication can reduce some barriers to sharing information between developing and developed countries. This may operate on a more simple level initially in the least developed areas, where the Internet is already a lifeline for receiving weather data. On a more sophisticated level, higher bandwidth in world-class centers in the developing world may permit them to participate more fully in the global geoscience enterprise and gain access to a wider array of information.

In some places, Internet cost is prohibitive, and bandwidth is still limited. This is most notable in Africa and in the former Soviet Union. It appears to be a limitation based both on government policies (monopolies on information and telecommunications) and on physical costs of providing access. In less developed regions of Africa and in Central Asia, governments control all telecommunications services. In isolated regions (such as Myanmar), extreme controls exist for political reasons. In addition, in many of these regions, the lower population density discourages competition for telecommunication services or limits the physical infrastructure for higher bandwidth. It is hoped that this will change over time, but it is likely to be slower and driven more by governments and donations than by the private sector. Many of the countries that lack Internet infrastructure also lack other basic infrastructure, so the digital divide is a broader expression of a development divide.

With regard to publications, because of the cost of information, many locations do not receive desired information. In addition, there are barriers to publishing. Costs of publication, both in terms of mailing and postage costs, and page charges constitute a significant barrier. There is also a perceived bias against submissions from developing countries. This bias is real, and there are some good reasons for it, but the bias also discourages publication. In general, some of this is expected from the different focus of much research in developing countries, which is often more applied and less theoretical.

Unexpected

Several results from the survey were unexpected. As noted earlier, Internet access is everywhere, and almost 90% of respondents had some access and were reachable by this means. In Asia, there were very few complaints about the quality or speed of access, just its cost. The physical infrastructure has developed rapidly from India to Southeast Asia to China. In more remote areas this may not be true, but at least for major universities, research centers, and meteorological services, the picture is improving.

While few researchers were able to get all the information and journals they needed, several interesting encouraging trends were noted. A delay of up to six months or so to receive journal articles was not as important to many of those surveyed, as long as the information arrived eventually. In addition, language was not perceived as a major barrier to many. To some extent, this survey is biased toward countries and groups of people with a good command of the English language. However, in other countries without a tradition of English as a second language, it is rapidly becoming a second language. This is likely due to what could be called a "Microsoft Effect," after the de facto standard in the personal computer industry. There is a desire for a common means of communication and expression, which gains users not necessarily on its merits, but is based on the number of other users. In many places, a younger generation is learning English and information technology for economic advancement, and this population is filtering into the climate and meteorology fields.

However, language can be a big barrier. Ironically, one of the places where language was found to be the largest barrier is also the most developed country in the survey: Thailand. Thailand was one of only three countries visited not to have been a former European colony (Nepal and China being the others). Perhaps as a result, the strong local language and traditions have led to a lower penetration of English among researchers. This may also reflect the lack of education in the field, as evidenced by the absence of a university program in climate or meteorology. The lack of a higher education programs in meteorology may not be surprising in Nepal or Uganda, but it is quite surprising not to have one in Thailand, a country of 60 million people and many universities.

Rejection rates

Finally, the survey provides some interesting answers as to why developing country submissions are rejected more often than others. The higher rejection rates are likely due to problems with methodology and the focus of research, as well as to cultural differences. Old or out-of-date methodologies could be partially corrected with better access to current literature. The focus of research in many developing countries is more applied than theoretical, and the focus of many journals is geared more toward theoretical research.

In addition, significant cultural differences in many of these countries make adherence to the scientific method, developed in a European culture, alien to many researchers. In particular, younger scientists identified a risk in using new methods over traditional methods. And secondly, there is deference to authority that does not permit or encourage the questioning of senior scientists. Some of these issues are also prevalent in developed countries, but they are supposed to be factored out of the editorial process by peer review, and they are culturally less prominent.

According to editors, and also according to most authors, language is not the major issue with submissions. Difficulty with written English often masks problems with methodology. Most editors see difficulty with English from European or Japanese researchers as well. Problems with language do make it more difficult for reviewers. Language and methodology issues often combine to create a bias on the part of reviewers.


Other issues raised

In the course of the survey, several other interesting and important issues were raised which are largely tangential to the major goals of the survey, but are discussed briefly here.

Sustainability

The concept of sustainability or sustainable development has a very different meaning in many of the locations visited. In developing countries, scientists are often thinking of these ideas not in terms of environmentally friendly policies and development, but in more basic and literal terms: will a particular project or research initiative be sustained after funding from an international source ends? In many cases, research is conducted with donor funds, and projects cease as soon as these funds are exhausted. Capacity may be built in the process, but usually the activity has little lasting local (constrained) benefit. Examples of such projects range from initiatives to develop inventories of greenhouse gases and ozone-depleting substances to data-rescue efforts to microfilm or digitize climate records. Projects that are unsustainable usually do not have immediate and direct or tangible benefits to the affected country, while sustainable projects are those with tangible benefits, and usually low cost, or use domestic technology. This begs the question of what is the best way to ensure the maximum benefit and the maximum level of "sustainability" of efforts in developing countries.

Privatization

In the last decade, while it has become increasingly difficult for weather services in many developing countries to sustain their current level of activity, several experiments to turn weather services into semi-private or for-profit entities have emerged (The WMO calles this process "adjunctization"). In 1992, the New Zealand Weather Service became a semi-private entity. In 1999, Tanzania did the same. In all cases, these entities receive government contracts to provide weather services. In discussions in Africa, both the Kenyan and Ugandan meteorology departments are considering a similar step. Several issues are involved. A key question is whether this will result in an improvement of available services. The basic concept is that users can be expected to pay for services. In Africa (as well as to some extent in New Zealand), civil aviation is a major user of weather services, which can be forced to pay (usually with hard currencies) for services. This enables the weather service to meet basic international aviation requirements. Without a basic weather facility at the Entebbe International Airport in Uganda, for example, international airlines would not be permitted to land. It is not clear that this pay-for-services policy could be extended to other key users of services (such as agriculture). Another question is how privatization or "adjunctization" will affect the availability of data for weather and climate use. New Zealand continues to provide data to the Global Telecommunications System (GTS) as required, and there are existing models to resolve some of these issues (WMO Resolution 40 spells out many of the issues regarding commercial activities. See www.wmo.ch).

Organization

The organization of meteorological services within a government structure may also have significant implications for providing services. These are issues that are felt in many countries and across many levels of development. It has been a big issue recently in Myanmar. Basically it is a question of outlook, culture, and expected users. In some cases, the meteorology and/or hydrology service is under an "Environment" Ministry (usually in a more developed country). In smaller countries it is often linked to a Transport ministry, sometimes very tightly linked with civil aviation, or to an Agriculture Ministry. In other cases the meteorology service may be viewed strategically as part of the Air or Armed Forces. This has significant funding implications and for access to information within the service and with other countries. In other cases, climate and meteorological services might be split (between Agriculture and Transport ministries, for example).

Electronic publishing

The future of the scientific journal is now in flux. The traditional model of scientific publishing uses a peer-reviewed editorial process to judge the quality of submissions and to regulate how much is appropriate to disseminate in print. However, with electronic publishing the costs of printing and distribution drop dramatically. At the same time, an increasing volume of research makes the role of the editor even more important.

The change in the economics of scientific publishing may have profound implications for the future of the scientific journal. It is not clear that electronic publishing is substantially cheaper to produce in a developed country, where much of the labor costs for editing and layout are fixed. This may be different for developed and developing countries, since the cost structure of a journal may be different. In developing countries labor costs are low and fixed costs of printing are high. Understanding these costs and the appropriate balance for developing country journals in the electronic age will be important for assisting these journals in the future.

Future and role of the GTS

The Global Telecommunications System (GTS) for weather information has been very effective at sharing information for forecasting. However over the last 20 years the availability of station data has decreased, due to the loss of stations on the network. Major international centers have compensated with more complex modeling systems and space-based sensors. However, these systems may not serve us well for monitoring climate. Trying to increase the participation of countries in the network and the data contributed to the GTS will be very important for climate monitoring.

In addition, there are currently discussions on how to best utilize the GTS in conjunction with the Internet. The use of the packet switched global Internet to deliver GTS data will enable a much larger bandwidth and enhanced interactivity with all national centers connected to it. This may enable the system to provide an expanded set of products for climate and environmental monitoring, which may benefit all parties to the system, as the original GTS has done.


Recommendations

The results indicate several obvious avenues for dealing with the information divide. Not all proposed recommendations will be appropriate for every country, but most countries will benefit from these recommendations.

Short Term

In the short term, enhancing the availability of electronic journals would be of great value to most developing countries. This might be done in two different ways. One way would be to permit open electronic access to journal back issues that are available electronically. Given that up-to-the-minute access does not seem to be critical in many developing countries, having a one-year delay in availability, similar to what journals such as Science has done, or what the American Meteorological Society is planning, would be reasonable. This solution would benefit those with well-developed and reliable Internet access.

In addition to this, back issues of electronic journals should be made available on CD-ROM sets to libraries in developing countries. One of the problems with electronic access is that it may have a prohibitive cost, and if Internet access lapses or fails, all information is lost. If access is interrupted, or if a website becomes unavailable, archived copies would be invaluable, and the access is not dependent on continual contact or payment. While many locations struggle with Internet access, almost all locations visited are capable of reading a standard product such as a file in Adobe Portable Document Format (PDF), its open-source derivatives, or HyperText Markup Language (HTML) used on the Internet.

CD-ROM access is also easier to disseminate broadly. While this may not be the aim of publishers in developed countries, it should be their aim in developing countries. CD-ROMs can be easily copied in many places, allowing researchers to share the information with others who currently have no hope of getting the information. Such a solution would strictly violate some of the copyright permissions currently inherent in journal publishing. This starts to get at larger issues of whether information should become free, and the costs of producing information (see below).

In addition to journals themselves, on-line database and searching tools should be made available to scientists from developing countries. This is particularly true for tools that were developed by scientific societies such as the Earth and Space Index from the American Geophysical Union, and perhaps Meteorological and Geophysical Abstracts. This will enable researchers to become aware of the latest research and where to find it when it does become available.

Medium Term

In the medium term, efforts could be focused on trying to better include researchers into the global geoscience community by encouraging publications and dissemination of information from developing countries to one another, as well as to developed countries. There are significant benefits over time for those in developed countries trying to understand global phenomena.

The way that such support might most usefully take place is to find ways to institutionalize informal information networks, reduce the barriers to entry, and make communications and networking in the sciences more open.

This might be accomplished by providing forums in which researchers can learn about what is being done by others in similar areas around the globe. Rather than focusing on expanding the already significant meeting agenda, such activities might be conducted as opt-in electronic communication groups: either mailing lists, websites with information, and/or dedicated climate and meteorological "collaboratories" (virtual collaborations and laboratories), which could be set up in individual subject areas with interested parties.

Focusing specifically on publications, mentoring assistance could be provided to developing-country scientists to help them through the editorial process. This happens in an informal way when developing-country scientists spend time in developing countries, or when collaborations result. But it might be institutionalized by matching up volunteer developed-country scientists with developing-country scientists. One way to do this is to submit a paper with a developed-country scientist as the last co-author and the corresponding author. The developing-country scientist gains exposure, and the developed-country scientist (who might be a junior faculty member, researcher, or post-doc) would gain publication credit for their efforts, without taking credit away from the developing-country researcher. Ideally, volunteers could work on papers in their field of interest and it might lead to future collaborations. It might work in a centralized way, or it might work on a journal-by-journal basis.

Long Term

In the long term, there is a need to improve the capacity of developing countries to use new technologies and methodologies. There are two important avenues for doing this. One is through education, and the other is by enhancing the capability and viability of developing-country scientific journals.

Educational development and access to education can be greatly enhanced by the Internet and information technologies. The WMO is involved in some of these efforts, but programs could be expanded from core WMO competencies into data analysis and other methodologies, particularly for climate research. Regional Meteorological Training Centers (RMTCs) are already using and developing electronic tools to teach forecasting and analysis (for example, CD-ROM courses on how to use satellite data). These programs could be expanded to include new tools and methodologies to examine climate data and trends. Packages might contain sample data sets as well as articles describing the methodologies. In many cases, these could be built from electronic materials developed by faculty to teach graduate students. One example might be objective analysis tools for analyzing climate data, including not just spectral analysis, but singular vector and empirical orthogonal function or wavelet analysis. Similar packages might be built for using numerical prediction models for meteorology or hydrology (including how to develop input data sets from analyses or station data). Packages and tools that assist scientists in learning how to use and search for information electronically would also be useful, as many scientists have not been trained in information searching methods.

Secondly, the transition to electronic publishing also portends fundamental changes to scientific journals around the world. Not only developed countries are affected. In fact, it is likely that developing countries would benefit the most from a transition to electronic publishing. Much of the cost of a journal is fixed by the time and labor needed to communicate between author, reviewer, and editor, as well as for typesetting, copy editing, and formatting. An electronic journal merely dispenses with printing costs in favor of electronic distribution. In developed countries, where labor and time are expensive, the fraction of costs eliminated by printing and postage are likely smaller than in a developing country, where labor is cheap, and many publications are sent though a comparatively expensive mail system. Developing-country journals might save significantly by a transition to electronic publishing, as long as they have the technical capacity to make the switch. The transition might be encouraged through assistance with open-source electronic software and expertise from journals and societies that have already accomplished it.

References

Gibbs, W. W. Lost science in the third world. Scientific American, August:92-99, 1995.

ISI, 2003. ISI Essential Science Indicators (geosciences), www.isiknowledge.com, 2003.

ITU (International Telecommunications Union), Basic Telecommunications Indicators, Geneva, 2002.

Kandlikar, M. and A. Sagar, Climate change research and analysis in India: an integrated assessment of the North-South divide, Global Environmental Change 9:119-138, 1999.

May, R. M. The scientific wealth of nations. Science 275:793-96, 1997.


Acknowledgments

Thanks are due to Al Cooper of the Advanced Study Program at NCAR, and to Robert Harriss of the Environmental and Societal Impacts Group at NCAR for their encouragement and support for this study. Thanks also to Dr. Sundararaman and Dr. Love of the IPCC Secretariat for their assistance with mailing. This report would not have been possible without the excellent editorial assistance of D. Jan Stewart and the graphics expertise of Anne Oman.

 


Home | Report in HTML | Report in PDF | Relevant Links | Appendices | Photo Gallery | Contact Information

NCAR Logo The National Center for Atmospheric Research is managed by the University Corporation for Atmospheric Research and is sponsored by the National Science Foundation. PO Box 3000
Boulder, Colorado, 80307
Tel: 303-497-1887; Fax: 303-497-1492
http://www.esig.ucar.edu