Research on glacial lake outburst floods (GLOFs) – specific low-frequency, high-magnitude floods originating in glacial lakes, including jökulhlaups – is well justified in the context of glacier ice loss and glacial lake evolution in glacierized areas all over the world. Increasing GLOF research activities, which are documented by the increasing number of published research items, have been observed in the past few decades; however, comprehensive insight into the GLOF research community, its global bibliometrics, geographies and trends in research is missing. To fill this gap, a set of 892 GLOF research items published in the Web of Science database covering the period 1979–2016 was analysed. General bibliometric characteristics, citations and references were analysed, revealing a certain change in the publishing paradigm over time. Furthermore, the global geographies of research on GLOFs were studied, focusing on (i) where GLOFs are studied, (ii) who studies GLOFs, (iii) the export of research on GLOFs and (iv) international collaboration. The observed trends and links to the challenges ahead are discussed and placed in a broader context.
According to the Web of Science (WOS) Core Collection database
(see Sect. 2.1), the term
Numerous studies exist that focus on a scientometric/bibliometric analysis of various research fields and their geographies (e.g. Small and Garfield, 1985), including those in geomorphology (e.g. Dorn, 2002), river research (e.g. Vugteveen et al., 2014) and natural hazard science (e.g. Chiu and Ho, 2007; Liu et al., 2012; Wu et al., 2015). While the amount of research on GLOFs has gradually increased over the past few decades (see Emmer et al., 2016), detailed insight into the GLOF research community, its global bibliometrics and geographies of research is missing. Therefore, the main objective of this study is to map GLOF research and the research community by analysing GLOF research items published in the WOS Core Collection database from the point of view of (i) general bibliometric and scientometric characteristics (temporal analysis, journals, document types, citations and references) and (ii) geographies (where GLOFs are studied and by whom in relation to where the GLOFs occur). This study is expected to provide a comprehensive global overview, targeting primarily the GLOF research community but also a broader audience such as GLOF risk management practitioners and policymakers (see, e.g., Carey et al., 2012).
Focusing on research results published in highly credible scientific
journals and proceedings, the search was performed on items published in
journals/proceedings indexed in the Clarivate Analytics Web of Science
Core Collection database (
The proper definition of the search formula is a key requisite for
successfully searching for a comprehensive set of items. After several
iterations, sighting shots and cross-checks with the control set of 30 papers
focusing on diverse aspects of research on GLOFs, a basic searching formula
was defined as follows:
Analysed characteristics describing each GLOF item.
The basic analysis was performed in the WOS environment, using the WOS Analyze Results tool. This tool allows the analysis of the entire set of 892 items, including an overview of (i) authors, (ii) institutions, (iii) source titles, (iv) WOS categories, (v) research areas and (vi) document types (see Sect. 3.1). Journals publishing GLOF research items (see Sect. 3.1.1) were further analysed using the WOS Journal Citation Reports tool, and citations and references (see Sect. 3.1.2) were analysed using the WOS Citation Report tool. The set of 892 analysed GLOF research items (see Sect. 2.1) was exported from the WOS platform, and subsequently processed and analysed in a Microsoft Excel environment. Some of the data needed to be repaired (e.g. the institutions of all of the authors involved in each research item are not available for 21 (2.4 %) research items written by more than one author; in which case, the country of the affiliation of the reprint author was used).
In addition to the characteristics assigned from the WOS database, each item was manually described by certain other characteristics, based on the search in the titles, abstracts (available for 843 out of the 892 items: 94.5 %) and keywords (available for 517 out of the 892 items: 58.0 %), using the content analysis method (e.g. Hsieh and Shannon, 2005). These include geographical characteristics, characteristics of authors (teams) contributing to individual research items and mono/inter-nationality of the research teams. These characteristics are analysed in Sect. 3.2, focusing on (i) geographical focus of GLOF research (see Sect. 3.2.1), (ii) geographies of the researchers involved (see Sect. 3.2.2), (iii) the “export” of research on GLOFs (see Sect. 3.2.3) and (iv) international collaboration (see Sect. 3.2.4). The content of GLOF research items is further analysed regarding selected aspects of their focus (see Sect. 3.3), using the WOS Advanced Search tool. This analysis is based on searching among the titles, abstracts and keywords; thus, it only provides a basic but nonetheless meaningful view on the given aspect. The observed trends are discussed and placed into a broader context in Sect. 4.
A total of 892 GLOF research items were published in the WOS Core Collection
database as of the end of 2016. At least one GLOF research item was
published each year since 1979 with the exception of 1983 and 1987 (see Fig. 1).
At least 10 GLOF research items were published every single year since
1993, at least 20 items since 2002 and at least 50 items since 2012. A
record number of papers was published in 2015 (
Yearly and cumulative number of GLOF research items published in the WOS Core Collection database (the first item was published in 1979).
GLOF research has been published under 256 diverse
The analysed GLOF research items have been published in journals indexed in 48
diverse
The 892 published GLOF research items have attracted a total of 18 570 citations (as of the end of 2016) and around 13 000 citations without self-citations. The number of citations has gradually increased during the analysed period, from tens of citations per year in the 1980s, hundreds of citations per year in the late 1990s and 2000s, and thousands of citations per year in the 2010s to up to 2514 citations in 2016, which basically reflects the increasing number of GLOF research items published (see Sect. 3.1). Each item has an average of 51.63 records in the list of references, rarely exceeding 100 references.
Each GLOF research item has obtained an average of 20.82 citations (14.57 without
self-citations). The most cited paper (Hemming, 2004) has obtained
581 citations (as of the end of 2016), and the 108 most cited papers
(12.1 % of all) have obtained more than 50 % of all citations.
Approximately one quarter (
Hotspots of GLOF occurrence, number of documented GLOFs and GLOF research items focusing on a given hotspot.
The occurrence of GLOFs is closely tied to the retreating glaciers and is, therefore, geographically clustered. Based on the previous GLOF inventories and overviews (Carrivick and Tweed, 2016; Emmer et al., 2016), 11 non-overlapping hotspots of GLOF occurrence are distinguished around the globe (see Table 2; Fig. 2), of which three are located in Asia (HKH, CAS, KRK), in Europe (ALP, ICL, SCA) and North America (ALA, GRL, NAC) each, and two are located in South America (CAN, PAN). ICL, NAC and HKH are the most prominent GLOF research hotspots with 180, 144 and 142 research items, respectively. The documented numbers of GLOFs are, however, highly disproportionate (see Sect. 4.2). GLOFs have been studied on all continents, including Africa (e.g. Girard et al., 2012), Australia (e.g. Goodsell et al., 2005) and Antarctica (e.g. Margerison et al., 2005), and extreme outburst floods are also thought to shape the landscape of Mars (e.g. Lapotre et al., 2016).
Hotspots of GLOF occurrence (yellow stars; see Table 2 for hotspot acronyms) and the geographical distribution of published GLOF research items by country.
Three types of studies are distinguished among the GLOF research items in
terms of geographical focus: (i) regionally focused items, (ii) multi-regionally focused
items and (iii) items with no regional focus
(e.g. theoretical or model studies). It was determined that almost three
quarters of all GLOF research items (
While the total amount of published GLOF research items is gradually increasing globally over time (see Sect. 3.1.1, Fig. 1), significant differences exist between individual GLOF research hotspots (see Fig. 3). The amount of research preformed in GLOF research hotspots located in developing regions (e.g. CAS, KRK, CAN, PAN) has generally increased, while the amount of research performed in developed regions has been stagnant (e.g. SCA, NAC) or has even decreased (ICL) over the past few decades. The exceptional position of the HKH region in GLOF research can also be seen in Fig. 3. Furthermore, it is clear that GLOFs were studied in hotspots located in Europe and North America before 1991, and expanded to Asia in the 1990s and South America in the 2000s (not considering local publications not indexed in the WOS; see Sect. 2.1).
The amount of GLOF research items focusing on individual GLOF research hotspots (see Fig. 2) over time.
From the point of view of the geographical focus on individual lakes, 37 items are focused on repeated Late Pleistocene outbursts from the proglacial Lake Missoula, likely being the most researched glacial lake outburst flood (e.g. Benito and O'Connor, 2003), followed by the 8.2 ka outburst flood from Lake Agassiz, which is mentioned in 28 research items (e.g. Clarke et al., 2004). Various aspects of jökulhlaups on the Katla volcano, Iceland, are elaborated in 27 research items (e.g. Duller et al., 2014). The most famous glacial lakes in GLOF hazard/risk studies nowadays – Imja Tsho, Tsho Rolpa (both Nepal Himalayas) and Palcacocha (Peruvian Andes) – have received the attention of 17, 9 and 7 research items, respectively (e.g. Rounce et al., 2016; Klimeš et al., 2016).
In total, 1885 authors from more than 750 institutions in 45 countries have
contributed the 892 analysed GLOF research items. Considering the number
of items published by each author, it emerges that a research item on GLOF
is written on average by 3.49 authors, indicating generally small teams
executing GLOF research. Almost one-sixth of the 892 research items (
More than 750
Fourteen out of 1885 individual researchers (0.7 % in all) have published
10
Research exports from the top 10 countries to 11 hotspots of GLOF occurrence. Only the reprint address of the first author is considered in this table (each item is counted only once). The number of items from overlapping countries (OCs; e.g. Switzerland and ALP; Norway and SCA) is shown in brackets.
TG – total number of items geographically focused on a given region. TI10 – total number of items authored by researchers from the top 10 GLOF research countries (excluding overlapping countries).
It is obvious that a certain geographical disproportion exists between the countries where GLOFs are studied (11 hotspots of GLOF occurrence; see Sect. 3.2.1) and the top 10 GLOF research countries (see Sect. 3.2.2; see Fig. 2), which have contributed a total of 740 (83.0 %) GLOF research items. The “export” of research from the top 10 GLOF research countries to 11 hotspots of GLOF occurrence is described in this section (see Tables 3 and 4). Considering the country of the first author only (i.e. each item is counted only once in the country of the first author even if it was written by an international research team; Table 3), the following patterns are observed: (i) more than three quarters (78.3 %) of research items focusing on 11 hotspots of GLOF occurrence were authored by researchers from the top 10 GLOF research countries; (ii) 43.5 % of imported research items were authored by researchers from the 10 GLOF research countries; (iii) researchers from the USA and UK have performed research in all 11 hotspots of GLOF occurrence; (iv) researchers from Iceland, China and Norway have focused almost exclusively on hotspots overlapping with their countries (ICL; CAS, HKH and KRK; and SCA, respectively); (v) the research in ALA and NAC has almost exclusively been performed by the researchers from overlapping countries (74.5 % of research items focusing on ALA have been elaborated by researchers from the USA, and 84.7 % of research items focusing on NAC have been elaborated by researchers from the USA and Canada).
Considering the countries of all of the institutions contributing to individual GLOF research items (see Table 4), it is seen that (i) researchers from the USA, UK and Switzerland have contributed to research in all 11 hotspots of GLOF occurrence; (ii) the most prominent GLOF research exporter is the UK (154 items), followed by the USA (84 items) and Switzerland (53 items); (iii) the researchers from Iceland and China focus almost exclusively on hotspots overlapping with their countries (ICL, and CAS, HKH and KRK, respectively); (iv) the share of imported research from the top 10 countries is very high in GRL, KRK, CAS and HKH, while it is very low in ALP, ALA and NAC; and (v) the research in hotspots located in developing countries is dominated by researchers from the USA and Japan (HKH); Germany, Switzerland and the UK (CAS); Switzerland, Germany and the USA (KRK), Switzerland and the USA (CAN); and the UK (PAN).
Geographical focus of research items of the top 10 GLOF research countries (11 hotspots of GLOF occurrence). All items (co-)authored by researchers from the given country are considered in this table (i.e. research items written by international research teams are counted for each country separately). The number of items from overlapping countries (OCs; e.g. Switzerland and ALP; Norway and SCA) is shown in brackets.
TG – total number of items geographically focusing on a given region.
Local researchers in hotspots located in developing regions have generally become more active in publishing their research in journals indexed in the WOS over the past few decades (see Fig. 4), which is in line with the increasing amount of GLOF research items published (see Sect. 3.1.1, Fig. 1). The share of items co-authored by local researchers (items written by local and foreign researchers) and authored by local researchers (items written solely by local researchers) varies significantly among the regions. While the share of items authored or co-authored by local researchers is generally increasing (< 40 % in the period 2007–2011 and > 50 % in 2012–2016), the share of research items published by foreign authors is only slightly increasing in some regions (CAS, PAN). A clear trend is also seen in the share of items authored by local researchers, which is increasing in all three regions located in Asia but is stagnant in PAN. No research items indexed in WOS were authored by researchers from CAN (see Fig. 4).
GLOF research items geographically focusing on five research
hotspots located in developing regions (see also Figs. 2, 3). The
involvement of local researchers is shown. Note the different
Section 3.2.2 shows that items written by international research teams represent 29.1 % of all GLOF research items, with a gradually increasing trend. However, there are significant differences between individual countries (see Table 5). The highest number of international research items was authored (first author) by researchers from the UK (56 items), followed by the researchers from the USA (38). The researchers from the USA contributed 91 international research items (36.1 % of all), followed by researchers from the UK, who contributed 84 international research items (33.3 % of all). The share of international items on the overall number of items published among the top 10 GLOF research countries varies from 34.8 % (Canada) to 71.4 % (Norway).
The vast majority of the 45 GLOF research countries (
Selected aspects of the content and focus of GLOF research items published
in the WOS Core Collection database are addressed in this section. A total
of 823 GLOF research items (92.3 %) are indexed as research articles (731
articles, 149 proceeding papers), and 50 items as reviews (5.6 %). An
additional 20 items contain “review*” within the title or abstract. A few
items are also classified as letters, editorial materials or notes.
Considering the different lake types that may be the source of GLOFs
(moraine-dammed, bedrock-dammed, ice-dammed), it is shown that studies
focusing on GLOFs originating in ice-dammed lakes (
International research items authored and co-authored by researchers from the top 10 GLOF research countries.
In terms of distinguishing between studies focusing on palaeo-GLOFs
(Pleistocene and Holocene) and recent ones (i.e. post-Little Ice Age), it is
found that 258 GLOF research items (28.9 %) explicitly focus on
palaeo-events such as Lake Missoula and Lake Agassiz (see also Sect. 3.2.1). It is
further stated in titles, abstracts and keywords that 39.8 % of GLOF
research items (
International cooperation between GLOF research countries. The number of international items (co-)authored is proportionally indicated by the size of a sector (see also Table 5).
Research on GLOFs represents a dynamic research field, reflecting the needs and challenges brought upon by the rapidly changing environment (e.g. Huss et al., 2017). A certain change in the publishing paradigm is observed among the published GLOF research items and documented by the analysed characteristics (see Sect. 3.1). Firstly, the amount of research on GLOFs had an exponentially increasing trend in the WOS database between 1979 and 2016. From the field of natural hazard science, a similar trend is observed in research on tsunamis (Chiu and Ho, 2007), a more remarkable trend is observed in research on landslides (Xu et al., 2015) and a less remarkable trend is observed in research on earthquakes (Liu et al., 2012). This trend is similar despite the fact that the GLOF research field is an order of magnitude smaller than the other above-mentioned fields and is in concordance with the generally observed trends across scientific disciplines (e.g. Sandstrom and Van den Besselaar, 2016). Secondly, the increasing amount of published research items within this research field is directly tied to the increasing amount of citations obtained by individual items within the field (e.g. geomorphology; Dorn, 2002), which is also valid in the case of GLOF research (see Sect. 3.1.2). Thirdly, the share of GLOF research items written by individuals is observed to have dramatically declined over the past few decades (see Sect. 3.2.2). This observation is not surprising considering the general trends in research and science, and it may indicate the increasing average extent and, hence, multidisciplinarity of the research teams involved (see also Skilton, 2009). Last but not least – the changing paradigm in GLOF research is also seen in the thematical content of GLOF research items. While hazards and risks, geomorphology and hydrology have traditionally dominated in GLOF research, recently other aspects such as climate justice have come to the forefront (e.g. Huggel et al., 2016; see Sect. 4.3).
A significant disproportion is observed between the number of documented
GLOFs from specific regions and the number of GLOF research items
geographically focused on a given region (see Sect. 3.2.1). While 270 GLOFs are
documented from ICL (180 research items), only 47 are documented from HKH
(142 research items; see Table 2). This disproportion can be explained as a
result of (i) differences in the causes and mechanisms of GLOFs (repeated
GLOFs from ice-dammed lakes in ICL, one-off GLOFs from moraine-dammed lakes
in HKH); (ii) incomplete GLOF inventories in less researched and/or less
settled regions such as remote areas of HKH (see Veh et al., 2018); and (iii) the
expected future increase in GLOF occurrence in HKH (Harrison et al., in
review). On the other hand, the highest number of GLOFs is documented from
ALP (
It has also been shown that an apparent geographical disproportionateness exists between the hotspots of GLOF occurrence (see Sect. 3.2.1) and top countries performing GLOF research (see Sect. 3.2.2), leading to the geographical “export” of research (see Sect. 3.2.3) – a phenomenon not yet fully captured within the context of the GLOF research field. Considering the items published in the WOS Core Collection database (see Sect. 2.1), the research on GLOFs is traditionally dominated by researchers from Europe and North America (see Sect. 3.2.2), while researchers from other countries – especially those overlapping with GLOF hotspots located in developing regions – have mostly focused on producing local publications and reports (see Sect. 2.1) and have come into play more recently, frequently as members of international research teams, but also as the first authors (Zaginaev et al., 2016; Colonia et al., 2017; Gherardini and Nucciotti, 2017; Prakash and Nagaraja, 2017). This trend can be explained by the increasing interest of local researchers in publishing research items indexed in WOS, instead of traditionally produced local publications and reports (see Sect. 2.1).
Strong collaboration is observed between the local institutions in developing regions and foreign researchers, for example, in the Peruvian Andes, where none of the 11 research items (published until 2016) (co-)authored by Peruvian researchers were led by them; all were led by foreigner researchers. A similar but not-so-strong trend is also observed among the research items (co-)authored by researchers from Central Asian countries (Kyrgyzstan, Kazachstan). On the other hand, researchers from Nepal tend to publish their GLOF research items on their own – out of 20 GLOF research items (co-)authored, 11 items were authored solely by the authors affiliated with Nepalese institutions (e.g. ICIMOD, Tribhuvan University). A similar trend is also observed for researchers in the PAN region (Argentina, Chile). The number of research items (co-)authored by local researchers, however, significantly differs considering the total number of items focused on the given region (see Sect. 3.2.3, Fig. 4). A general trend of larger involvement of local authors and the internationalization of GLOF research teams has also been observed over the past few decades (see Sect. 3.2.4), which is in line with the general trends observed in global environmental change research (e.g. Jappe, 2007).
The research on GLOFs struggles with numerous challenges brought upon by the complexity of generic processes and the general characteristics of these events (low frequency, high magnitude, complicated predictability). One of the greatest challenges in the basic understanding of the spatio-temporal occurrence of GLOFs is to compile comprehensive databases of these events (see Carrivick and Tweed, 2016; Emmer et al., 2016). While past GLOFs are well described in some of the regions (e.g. Alps; see also Sect. 3.2.1), others suffer from data scarcity (e.g. Veh et al., 2018). This different level of detail makes any assessment of GLOF frequencies among the regions (on global level) a rather challenging task (Harrison et al., in review). Compiling regional databases of GLOFs aiming at a globally comprehensive database is, thus, considered one of the greatest challenges in GLOF research.
Since GLOFs have claimed thousands of lives and have caused considerable material damage in the past (Carrivick and Tweed, 2016), and further GLOFs are expected in the future, reflecting the trend of retreating glaciers (Huss et al., 2017; Harrison et al., 2017), hazard and risk assessment represents another highly challenging task and a recurrent topic among GLOF research items (see also Sect. 3.3). While hazard (susceptibility) assessment is traditionally dominant in GLOF risk studies (see the overview of methods by Kougkoulos et al., 2018), vulnerability assessment and adaptation in the broader context of retreating glaciers and water resources still remain brand new topics, which need to be further addressed in most regions and especially those located in developing countries (e.g. Vuille et al., 2018). Moreover, the communication of scientific results with local authorities and integration into the decision-making process is an extremely challenging task, which is far beyond the scope of geoscientific research (e.g. Gagne et al., 2014), promoting the need for broader interdisciplinary collaboration.
Various modelling tools are being developed and implemented hand in hand with the efforts to reliably assess the impacts of potential future GLOFs (e.g. Worni et al., 2014; Schaub et al., 2016; Chisolm and McKinney, 2017). Certain challenges are still seen ahead in this regard (e.g. Mergili et al., 2018; Westoby et al., 2014) despite rapid progress in this direction and improvements to technical capabilities as well as data availability and acquirability (e.g. Mallalieu et al., 2017; Wigmore and Mark, 2017).
This study shows how research on GLOFs
published in the Web of Science Core Collection database has become topical
over the past few decades (analysed period 1979–2016), how the publishing
culture and paradigm have changed over time and what the trends and
disproportions in geographies of research on GLOFs are. A significant
exponential increase in the number of GLOF research items published in the
WOS database is revealed, with > 50 % of the research items
being published since 2008. While 1885 researchers from more than 750 institutions
in 45 countries have contributed 892 of the analysed GLOF
research items, a relatively small number of 90 leading researchers (4.8 %
of all) have published 5
Data are accessible in the Supplement.
The author declares that he has no conflict of interest.
I would like to thank Jonathan L. Carrivick for his insights into the early draft of this study, Petr Bašta for discussion on processing Fig. 5 and Craig Hampson for language revision. I would also like to thank two anonymous referees for their comments and Sven Fuchs (NHESS editor) for handling the manuscript. This work was supported by the Ministry of Education, Youth and Sports of the Czech Republic within the National Sustainability Programme I (NPU I), grant number LO1415. Edited by: Sven Fuchs Reviewed by: two anonymous referees