Bibliometric Study of the International Journal of Remote Sensing

By Chandra Curtin

Submitted May 10, 2004
Note: I give permission for my final project to be made available through the LIS Learning Showcase web server.

INTRODUCTION

PUBLICATION HISTORY

First published in 1980, the journal is largely studied by those in the engineering and environmental studies fields. IJRS is published through Taylor & Francis Ltd., based from Abingdon, Oxfordshire. The journal can be found through its publisher's website at http://www.tandf.co.uk/journals . If you are a subscriber to the journal, you can access all the issues online. If not, you are still allowed to browse the issues, that is, the table of contents can be viewed for every issue. According to the publisher's website, IJRS is, "concerned with the science and technology of remote sensing and the applications of remotely sensed data in all major disciplines" (Taylor & Francis). This journal is largely read by those involved in engineering and environmental studies.

The journal began publication in 1980 and was distributed on both a bi-monthly and monthly basis. January, March, May, July, September, and November had bi-monthly editions whereas the other months of the year just had one a month. This pattern continues until January 2002, where the journal became completely bi-monthly.

Full-text versions of IJRS can be accesses through several mediums. EBSCO carries full text versions of the journal from the years 1998-Current year. In this case, current year is 12 months prior to today, being that the publishers have imposed a restriction causing a 12-month delay. This is so that researchers who are interested in the most up-to-date information on remote sensing from the IJRS must buy the journal directly. The University of Arizona also carries this journal in the Main Library with the call number G70.4 I56. The university carries bound copies of all volumes with the exception of volumes 19 and 20. In the years 1998 and 1999, the university seems to have canceled its subscription, and then restarted it in 2000.

DOCUMENTS

IJRS publishes mostly research articles, but within the last five years, (2003-1999), they have also published bibliographical items, corrections, editorials, letters, and reviews. Table 1.1 shows the frequency at which these items were published in the last five years.

Document Types	2003	2002	2001	2000	1999
Articles	97.19%	96.25%	95.50%	96.0%	95.45%
Bibliographical Items	.26%	0	0	.36%	0
Corrections	.77%	.75%	1.04%	.73%	.35%
Editorials	1.28%	1.87%	1.38%	2.18%	3.50%
Letters	0	0	.35%	0	0
Reviews	.51%	.37%	1.73%	.73%	.70%

This table shows that a majority of the journal is concerned with research articles. For example, in 2002, there were 267 documents that were published and 257 of those were research articles. This can be compared to the amount of letters that the journal publishes. In 2001, only 1 letter was published out of 289 total documents. This data shows readers that the journal is mainly concerned with the communication of the current research and will most likely be of interest to those in the field.

AUTHORS

The top five authors were found by taking each year of the journal and performing a Most Times Cited sort. After looking at each year, the list was compiled for the last five years, showing the top five cited authors within those years. They were found to be; Hansen MC (99 times cited), Defries RS (99 times cited), Townshend JRG (99 times cited), Sohlberg R (99 times cited), and Loveland TR (79 times cited). WOS will not let a list be sorted if there are more than 300 items. In 2003, IJRS had 392 items that were published, 381 of those being journal articles. Therefore, in order to find the highest number of citations, I created a marked items list with the titles of the articles, their authors, and the citation numbers. This list could then be exported into email where I was able to sort the list myself to find the highest cited authors.

These authors have all published articles within IJRS. Townshend JRG has published the most with 13 articles followed by Defries RS with 5 articles, Hansen MC with 4, Sohlberg R with 3, and Loveland TR with 2. Considering the fact that there have been 1451 articles published within the last 5 years, this data seems to show that there is little Matthew effect in the journal. The likelihood that a researcher will be published is not based on past publication history.

CITIVITY

I chose to analyze the September 20, 2002, (V.23, N.18) issue of the journal. Within this issue there was 1 editorial and 24 research articles. The editorial contained no cited references. The articles ranged from having 11 references to having 50, and there were a total of 578 references within the issue. The mean number of references was 24.08 and the median was 20.5. These numbers are difficult to interpret given that they are from one journal and not being compared to another. From my experience in the biological science literature, these numbers are low, but given that the field of remote sensing is not as large, these numbers seen comparable. This seems to show that these articles are the products of major studies and that this journal is quite significant in the communication of remote sensing researchers.

JOURNAL CITATION REPORTS

The impact factor of this journal is 1.154 and the immediacy index is 0.135. The immediacy index seems to show that this journal does not contain articles that are quickly cited. The impact factor shows that the average article is cited about one time. As to tell if this is truly significant, we would first need to compare this with journals in the same discipline. Also, it should be noted that these are dynamic numbers and will change as usage of the journal changes each year. The immediacy index measures the rate at which it takes a journal article to be cited be another researcher. Would this number be higher if the full text was available on EBSCO immediately instead of a 12 month delay? Probably not, as scientists in this field would be subscribers to the journal directly to stay on top of the latest researcher. But it does pose the question of how significant the immediacy rate will be in the future with more full-text databases online.

BIBLIOMETRIC PROFILE

The author I chose to study was Townshend, JRG. Within the last five years, (1999-2003), Townshend was the most published author with ten articles published. For the following data, it should be noted taht any journals mentioned as written by Tonwshend were done so within the last five years and published in the International Journal of Remote Sensing. This author has published in several other journals, and using that data would give significantly different results. Table 2.1 below shows the author's citation set, that is, the number of authors that cited Townshend's work.

Publication Year	Authors who cited Townshend, JRG
1999	19
2000	30
2001	27
2002	28
2003	27

This table seems to indicate that Towshend plays a sigficant role in the field of remote sensing. Each year he is referenced by his peers, even in the years that he does not published in IJRS, (1999 and 2001). By further analysis of the names of these authors, this citation set could show who these peers are and with what frequency they cite Townshend.

The next resource for building a bibliometric profile is to look at Townshend's Citation Identity. These are the authors that Townshend cites. Table 2.2 shows the number of citations that Townshend references in the years he publishes, as well as a count of the number of citations that come from IJRS.

Year of Publication	Number of Authors Cited	Number from IJRS
1999	NA	NA
2000	379	94
2001	NA	NA
2002	54	11
2003	31	10

Further analysis of the names of these authors would show who Townshend frequently cited. This table also shows that many other journals are used in the field of remote sensing, not just IJRS. In 2000, only about 25% of the articles that Townshend cites came from IJRS. This shows that many different journals have significant in the field.

The next table, Table 2.3, shows two things. First, it shows Townshend's citation image, the authors who are co-cited with him, and second, it shows the frequency at which these authors cite themselves.

Year Published	Co-Author	Number of Self-Citations
1999	No articles published	NA
2000	Bader, DA	3
2000	Defries, RS	13
2000	Dimiceli, C	0
2000	Eidenshink, JC	1
2000	El Saleous, NZ	1
2000	Fallah-Adl, H	0
2000	Goward, SN	0
2000	Hansen, MC	6
2000	Huang, C	1
2000	Jaja, J	0
2000	Justice, CO	3
2000	Kalluri, SNV	1
2000	Liang, S	0
2000	Schlberg, R	0
2000	Teillet, PM	10
2000	Tucker, CJ	1
2000	Vermote, EF	0
2000	Yang, K	0
2000	Zhan, X	1
2000	Zhang, Z	1
2001	No articles published	NA
2002	Davis, CS	0
2002	Huang, C	0
2002	Huang, C	0

This shows that there are some authors Townshend frequently collaborates with, Huang C for example. This table also gives readers an idea of how often researchers cite themselves. Table 2.4 shows how often Townshend cites himself in his work.

Year Published	Number of Self Citations
1999	NA
2000	23
2001	NA
2002	2
2003	1

This shows that for each paper, Townshend cites himself at least once. Data gathered from citing oneself can often skew the anaylsis for productivity. For example, if a frequently published author cites himself, this changes the immediacy index. The author does not need to wait to read the latest journal to find the research, as he has already done it.

KEYWORD ANALYSIS

Of the 25 documents that were published in the September 20, 2002 issue, (v.23, N. 18), 21 of them listed one or more keywords. The most frequently used were; model(4), reflectance(4), vegetation(4), area estimation(2), classification(2), satellite(2), spectroscopy(2), surface(2), and water(2). The entire keyword set for these 21 documents is listed below:

absorption, accuracy assessment, Albedo, algorithm, analyzer, architecture, area estimation (2), atmospheric aerosols, AVHRR, AVHRR data, canopy architecture, circulation, classification (2). Color, Conifer forests, digital camera, Douglas-Fir, elevation, error, fires, fluorescence, forest, GIS, hemispherical photographs, hydrography, imagery, images, index, Indonesia, LAI-2000, land-cover classification, LandSAT TM, landscape, leaf-area index, leaves, light scattering, light transmission, Markov-models, matter, membership, Meris, model(4), multiresolution, ocean, optical-properties, orthogonal-function analysis, Pacific Ocean, particles, patterns, precipitation, polarization measurements, reflectance(4), remote-sensing data, rustle, Santa Barbara channel, SAR, satellite(2), shape, sky snow surface, soil color, speckle, spectroscopy(2), surface(2), temperature, transmission, validation, variability, vegetation index, vegetation(4), and water(2).

These keywords do not seem to fit the specific discipline of geographic information science. With the exception of a few, these keywords could fit any discipline of science. Also, some keywords are more specific than others. Vegetation was listed as a keyword four times, and vegetation index was listed once. A keyword search would bring up all five of these articles. In addition, keywords are subjective, entered by the authors themselves. With all these significant incongruence, using keywords to analyze a journal's productivity may not be the best way to understand scholarly communication in this discipline.

Scholarly Communication and the Future of Webmetrices

Citation reports and other bibliometric methods can be used to evaluate patterns of scholarly communication, but these methods must be critically evaluated for their significance. For example, a citation set may list all of the authors that cite a particular author, but this number would also include the times that the author cites himself. Is this number of importance? Perhaps not, but by looking at the individual authors, a sense could be made on the type of resources an author looks for. Does he often use foreign researchers? Do references often come from journals that the author himself often publishes in? For example, how effective is the immediacy index in determining a journal\rquote s effectiveness in communication? A journal that offers its full text online immediately may have a higher index number than one that has a 12-month delay, such as IJRS. The answer to these questions may help to understand how effective journals are in a particular field, but journals are only one part of scholarly communication. Before the advent of computers and email, journals were often the only way that scientists could communicate research to each other. This is no longer the case thanks to globalization and the internet. Evaluating these resources will be discussed in the next paragraph. And some disciplines communicate differently with each other. Scientists have always used journals, such as the International Journal of Remote Sensing, to communicate the latest research. These papers are often the products of years of research, collaboration, (which means several authors), and with many references. Thanks to the ISI, these journals are indexed and broken down, ready for evaluation on many different angles. Cited references, cited authors, immediacy index, all of these can help determine a journal's significance. However, there are other academic disciplines that are not as rigid. Take, for example, library science. This is not a field that relies on journals primarily for scholarly communication. Though there are journals that rely the latest information to librarians, such as College and Research Libraries, librarians often use American Libraries or ALA's Hotline. These are not the scholarly journals, more like news magazines. The same bibliometic methods could not be used to analyze these fields. To truly determine this, the intellectual pychology of the academic field would need to be examined. Librarianship is a people profession and librarians are much more likely to communicate through conferences than through research journals. This is simply the nature of the work.

In addition to evaluation by bibliometric means, evaluation can also be done using webmetric means. Comparing the number of hits a site gets is equivalent to using subscription rates to test a journal\rquote s effectiveness in communication. Though it may give a researcher a sense of how much something is being read, it offers little insight in how significant the journal or site is. To measure the impact of online resources on research, similar statistics could be gathered. Instead of looking at cited author lists, researchers could examine the number of websites an author cites. Other resources that are now available are online subscription databases. Databases, such as EBSCO and InfoTrac, allow researchers to access journals online. These are dependent upon the journals themselves and what they will allow access to. For example, IJRS is full text in EBSCO with the exception of the latest 12 months. This ensures that researchers must purchase the journal directly through the publisher if they wish to have the latest research. How would this compare to a journal that sells all its rights for the current year. Provided that the author has access to this database, the most current information could be found be just going online. Statistics on the use of the online databases may also be used to evaluate how effective a journal is for scholarly communication.

Another issue that arises with the evaluation of online materials is that of authenticity. Scientific journals go through a peer review process, the data is verified, and the journal goes into publication. The double-edged sword of the internet allows anyone to post to it. The site may list a ton of references, but are these valid and should they be used to evaluate an author's effectiveness? Yes. Good or bad, an author still looked and them and thought that it was something that should be included. With the future technology and using websites in research, there will be more responsibility on the part of the researcher to verify information. Only time will tell which websites will play a significant role in scholarly communication.

Another resource that should be evaluated in its role in scholarly communication are web logs, or blogs. These give academia a chance to discuss information on a day to day basis. The use of these blogs and their significance on SC will depend on the academic disciplines themselves. Librarians often communicate ideas to each other through the use of newsletters and email. Scientists, though they often work collaboratively, may be more reserved about putting their incomplete research out there for fear of intellectual piracy. When using blogs as a resource for SC, there would also need to be some agreement as how to cite these resources in a paper. Once this is decided, the data can be analyzed.

CONCLUSION

The future of scholarly communication will depend largely on the technology that academia decides to embrace as significant. Each discipline will evaluate their rescources a little differently, which makes across-the-board evaluation difficult. The use of bibliometic and webmetric tools can be used to evaluate a system's effectiveness as long as those methods are constantly evaluated themselves. At one point, most of SC was done through journals, but with online technology, SC can happen rapidly without waiting for publication times. Tools such as ISI can be used to analyze aspects of an author's citation identity and this can be used to predict future activity. Data can also be gathered to compare online subscription rates with paper subscription rates to determine if there is, in fact, a trend towards more online resources. The field of scholarly communication is a dynamic one, and will change as the professionals in the field change. Evaluation of these resources is important to validate which are significant in chosen discipline.