Retrospecting the researches and efforts on Lancang-Mekong water issues: a bibliometric perspective

As one of the most prominent waterways in the world, the Lancang-Mekong River is the longest river in Southeast Asia and the most crucial transboundary river, flowing through six countries, namely China, Myanmar, Laos, Thailand, Cambodia, and Vietnam. The development and utilization of the Lancang-Mekong water resources and hydropower resources and their possible impacts have always been the focus of close attention of scholars in and outside the region, as well as the central topic of frequent debates in the international communities and civil organizations (Hecht et al., 2019). Among various actors involved in water governance in the Lancang-Mekong Region, including the United Nations, countries' governments in the Lancang-Mekong Basin, news media, NGOs, financial institutions, and enterprises, of which scientific research institutions or researchers can take a relatively neutral position based on technical analysis results and facts (Dore et al., 2012). As an objective record of human understanding of the Lancang-Mekong River and the main form of research results, scientific literature can help us sort out the research context, hotspots, important discoveries, and development frontiers of the Lancang-Mekong River water issue.

Scientometrics is a powerful tool for the quantitative evaluation of scientific literature. Recently, a series of software for knowledge graphs and bibliometrics were figured out, such as CiteSpace (Chen, 2006) and VOSviewer (van Eck & Waltman, 2010). Based on this, scholars can conduct scientific measurement, graph construction, and visual display of knowledge units such as titles, authors, institutions, countries/regions, abstract, keywords, references, published journals, and other index information, and co-word analysis, cooperation network analysis, review of research history, and capture of research hotspots in the areas of interest. Scientometrics has been used in many research fields for co-word analysis, cooperative network analysis, review of research history, and capture of research hotspots, early mainly concentrated in the fields of management and humanities and social sciences. In recent years, it has been rapidly applied in the fields of natural science and engineering technology, but there is still relatively little research in the field of water resources and environment.

Based on scientometrics-related theories, methods, and tools, we retrieved and analyzed relevant literature in the water field of Lancang-Mekong River and explored the main research topics, research hotspots at different stages, and the context of important achievements in the water field of Lancang-Mekong River.

Literature retrieval is one of the key links in the whole knowledge graph research, which is directly related to the quality of the analysis results. We used TOPIC: (‘Mekong’) AND TOPIC: (‘WATER’), without limitation of time span to retrieve the Mekong water-related literature in Web of Science (WoS) core collection database, and a total of 2,061 papers were obtained. The literature data were last updated on April 28, 2021. Since this study takes journal articles as its main research object, conference papers, and book chapters, as well as other editorial materials, book reviews, letters, and other literatures that do not belong to the scope of this study were excluded. After preliminary screening of the search results, a total of 1,776 literatures were selected as the research data of this paper.

In this paper, HistCite (Garfield et al., 2006) was used to conduct statistical analysis research on citation and construct citation history network based on citation relationships of local documents. In HistCite, citation scores include Local Citation Score (LCS) and Global Citation Score (GCS), which represent the number of times a paper has been cited by the other articles in the local dataset and in the WoS Database, respectively. Document Co-citation Analysis (DCA) was adopted for constructing citation network relationship, and spectral clustering algorithm derived from spectral Graph theory was used to divide the overall citation network into several non-overlapping clusters in CiteSpace II V5.8.R1.

From the perspective of the number of published papers and citations in this field, the research process can be divided into five stages: the seed stage (before 1994), the budding stage (1995–2002), the fluctuation stage (2003–2010), the stabilization stage (2010–2016), and the rapid growth stage (2017–present). Before 1994, the research outputs were sporadic. During the budding period (1995–2002), the number of publications increased slowly with the annual publications of 4 in 1995 increasing to 12 in 2002, and it could be attributed to the 1995 Mekong agreement signed by the four member countries. From 2003 to 2010, the research outputs increased rapidly and then decreased sharply, which could be related to the widespread controversy over the planned dams in the Lower Mekong mainstream and the progress of the multi-party games.

Since 2011, the number of papers published on the Lancang-Mekong water issue has reversed the downward trend and continued to increase. Scholars used new theoretical frameworks, such as the ‘Water-Energy-Food (WEF) Nexus’, to evaluate and analyze the possible impacts of natural and anthropogenic factors on the Mekong Region from a more comprehensive and comprehensive perspective. From 2017 to now, the number of published articles has accelerated, with more than 100 published annually and even reaching 244 in 2020. The number of publications is closely related to the international financial assistance provided, and there are various launches of cooperation and dialogue mechanisms which can be seen in the significant increase in studies on Lancang-Mekong water issues since the Lancang-Mekong Mechanism was proposed.

The top 9 largest clusters with more than 20 members are Clu#0 (Vietnamese Mekong Delta/Land Subsidence), Clu#1 (Lancang-Mekong River Basin/Lancang River), Clu#2 (Mekong River Basin/Flood Pulse), Clu#3 (Mekong River/Dam Construction), Clu#4 (Arsenic Contamination/Risk Assessment), Clu#5 (Water-Energy-Food Nexus/Food Security), Clu#6 (Striped Catfish/Integrated Rice-Fish Farming), Clu#7 (Cropping Pattern/Sentinel-1a), and Clu#9 (Deltaic Social-ecological System/Water Security). The other five clusters with relatively small sizes include Clu#12 (Monsoonal Asia/water resources), Clu#19 (Flood-Prone Area/Winter-Spring Rice), Clu#27 (Seabed Morphology/Drag coefficients), Clu#28 (Aquaculture Environment Interaction/Nutritional Inputs), and Clu#40 (Redox Control/Organic Carbon).

For each cluster, predominant members of a cluster as the intellectual base and themes identified in the citers of this cluster as research front were analyzed. The top-cited cluster members and their structural, temporal, and saliency metrics including citation counts (φ), citation burstness (τ), degree centrality (Deg), betweenness centrality (σ), and sigma (Σ) are summarized in Table 2. The top citers to each of the nine largest clusters are listed in Table 2.

The most-cited references in the Vietnamese Mekong Delta (Clu#0) are Smajgl_2015 on comprehensive adaptation measures to sea-level rise (Smajgl et al., 2015), Hoang_2016 on the evaluation of hydrological impacts of climate change in Mekong Region (Hoang et al., 2016), and Anthony_2015 on the linkage of the large-scale shoreline erosion and land loss with the sediment retention by dams, large-scale commercial sand mining in the river and delta channels and land subsidence caused by groundwater exploitation (Anthony et al., 2015). The article Manh_2015 that combined the comprehensive impact of hydropower development, climate change, and sea level rise on sediment dynamics had the highest betweenness centrality value of 0.04 and sigma value of 1.27 within Clu#0. The reference with the highest degree value within Clu#0 was Darby_2016 that proposed changes in tropical-cyclone climatology but not upstream reservoir trapping altered the trends in fluvial suspended sediment loads.

The first four most-cited references in the Mekong River Basin (Clu#1) are all hydropower development-related research, including Rasanen_2017 on river discharge change, Winemiller_2016 on biodiversity, and two review papers, Hecht_2019 and Fan_2015, that comprehensively summarized the hydrological impact and environmental consequences of hydropower dams of the Mekong River Basin, respectively. The Clu#2 was mainly composed of references related to but not limited to hydropower development including Ziv_2012 on tradeoff among fish biodiversity, food security, and hydropower, Lauri_2012 and Arias_2014 on integrated influence of climate change and hydropower. The dominant references in Clu#3 are Kummu_2007, Lu_2006, and Fu_2008 that all focused on the impact of upstream reservoir especially Manwan Reservoir on sediment with the State of Basin Report (SoBR) published by Mekong River Commission (MRC) in 2010 as its most-cited reference.

Different from the former largest four clusters, the stars in the arsenic contamination (Clu#4), Berg_2001 and JICA_2001, are two of the earliest research achievements on arsenic contamination of groundwater in Vietnam and Central Cambodia, respectively. The publication time of the top-ranked references in Clu#4 was relatively earlier compared with other clusters and concentrated during 2001–2002. For Clu#5 labeled as Water-Energy-Food Nexus, its research focus shifts from purely discussing the positive and negative effects of dam construction to how to tradeoff hydropower development, food security, and water security in the context of hydropower development in the Lower Mekong. The most-cited papers, Grumbine_2012 and Orr_2012, discussed the drivers of dam construction in the Lower Mekong Basin and its potentially huge impact on local fisheries. The book Mirumachi_2015 proposed the Transboundary Waters Interaction NexuS (TWINS) framework to examine the coexistence of conflict and cooperation among riparian countries that share international transboundary rivers, Keskinen_2015 applied the Water-Energy-Food Nexus theory in the Tonle Sap Lake, the most important wetland in the Mekong Basin. The top-cited reference of stripped catfish cluster (Clu#6) is Phan_2009 which traced the catfish farming history and current harvesting and marketing procedures of the farming community in the Mekong Delta. The Nhan_2007 on integrated freshwater aquaculture, crop, and livestock production ranked fourth in Clu#6 in addition to MRC_2005 on hydrological overview report and Kummu_2008 on flood pulse changes caused by dams in the Tonle Sap Region. Clu#7 labeled as cropping pattern is mainly about the application of remote sensing technology in mapping flood (Kuenzer_2013), rice paddy extent (Kontgis_2015), and forecasting rice production (Mosleh_2015). The last one of the nine largest clusters is a vulnerability indicator, with the core paper of Toan_2013 which systematically investigated the pesticide residues in sediments, surface, and drinking water in the Mekong Delta. Climate and environmental change (Kuenzer_2012) and flood disaster (Auerbach_2015) are also components of vulnerability.

The coverage value is the number of references in the cluster that the citing article cited. In this paper, the articles with the most-coverage values were used to indicate the research front. The most frequent citers to each of the nine largest DCA clusters are listed in Table 3.

For the largest Clu#1 and Clu#2, the first four citers are exactly the same, including three review articles (Hecht_2019, Pokhrel_2018, and Arias_2019) and one numerical simulation paper (Shin_2020). The three review papers are all about the impact of hydropower development and dam construction on the Mekong River and its interaction with other driving factors such as climate change and land use conversion on hydrology regime and environmental conditions. The numerical simulation paper used a new model of river hydrodynamic-reservoir operation to comprehensively evaluate and verify the impact of dam and climate change on the flood dynamics of the Mekong River Basin on a finer scale, finding that with 2010 as the dividing line, the dam has little impact on the flood dynamics of the alluvial plain before 2009, and after 2010 its impact has been increased significantly due to doubled reservoir capacity contributed by newly built dams, and the impact of dams on the downstream is largely determined by the reservoirs' operation strategy.

Both of the two most frequently cited literatures of Clu#2 were written by Mauricio E. Arias of the University of Canterbury, New Zealand, who studied the Tonle Sap River floodplain, the most important wetland in Southeast Asia, from the perspective of natural environment change and human activities. Another document with high citation frequency is about the long-term impact of local water conservancy infrastructure on the water level of Mekong River and Tonle Sap Lake system, which is co-authored by Thomas A. Cocharane, and Mauricio E. Arias and Thanapon Piman. These three documents were all published in 2014 and were written by scholars from the University of Canterbury, New Zealand, showing that this institution is leading the research frontier in this field. In addition, the study on Water-Energy-Food Nexus (Keskinen Marko_2016) is a new research trend in the flood pulse-affected areas, especially the flood plain of Tonle Sap Lake.

Lebel_2010 on social learning and Bearden_2012 on legal system were the publications that cited the most literatures within Clu#3, followed by Xue_2011 on the relationship between changes in hydrology and sediment load in the Lower Mekong, and dam construction, monsoon climate and El Niño Southern Oscillation (ENSO). Scholars argue that it is necessary to strengthen interactive learning, build trust, and mutual understanding among social communities to establish community adaptability, and promote the construction of legal systems to adapt and alleviate the impact of earlier dam construction and dam operation on the downstream, as well as analyze the impact of dams from a broader, comprehensive and more objective standpoint. These most frequent citers of Clu#3 were concentrated in the years 2010–2012, indicating that the study on the impact of dam construction, particularly the dams on the Upper Mekong (Lancang River), and the mitigation and adaptation to the dam construction has reached a general consensus or that some newer and more important issues have piqued the interest and enthusiasm scholars in the field around 2010.

Similar to Clu#3, Clu#4 also has relatively concentrated publication years of frequent citing literature, which indicates that the research on arsenic pollution of groundwater in the Mekong River Basin has made important breakthroughs in 2008–2009. Rodriguez_2008 and Winkel_2008 simulated and predicted arsenic pollution, respectively, and Kocar_2008 studied the mechanism of arsenic release from sediment to groundwater driven by biogeochemical and hydrological processes. Buschmann_2009 analyzed the effect of sulfate reduction on the scale of arsenic pollution in groundwater, and basically made clear the scope, degree, geographical causes of arsenic pollution in the Mekong River Basin, especially in some arsenic pollution study areas in Cambodia and Vietnam, as well as the possible effects of long-term arsenic intake on human body.

Study on the Water-Energy-Food Nexus is the research frontier of Clu#5, with Carl Middleton and Marko Keskinen are the most representative scholars in this field, and their representative works were published in 2015–2016. Integrated rice-fish farming and striped catfish farming are research hotspots of Clu#6, which were concentrated in 2011. Integrated rice-fish farming was one of the most important ways and means to cope with food security challenges and improve comprehensive production efficiency at that time. In Clu#7, Rudiyanto 2019, Minh_2019, Clauss_2019, and Fikriyah_2019, the top 4 citing documents in terms of citing frequency, are all related to the application of remote sensing satellites. From the publication year of the above citing documents, it can be seen that the research in this field is still developing rapidly. Sebesvari_2016, the most citing core paper in Clu#9, provided a structured review of potential vulnerability indicators for delta socio-ecological assessments and proposes a socio-ecological system-centric framework for research on risk and vulnerability.

In terms of time span, Clu#4 has the longest time span of 7.8 years, while the time spans between the research front and knowledge intellectual of all the other clusters are less than 4.5 years. Of which, Clu#2 and Clu#6 have the shortest time spans of 2.7 and 2.6 years, respectively. In addition to the time span, the average year of intellectual base and research front of Clu#4 are significantly earlier than those of other clusters. The average publication year of references cited that constitute the knowledge basis for arsenic contamination (Clu#4) is 2001.8, and the average year of publication of its heavyweight cutting-edge outputs is 2008.6. The second earlier formed intellectual bases formed are Clu#3 (2008.3) and Clu#6 (2008.6), and the average years of knowledge base formation of other clusters are all after 2013, among which Clu#1 has the most recent formation year of 2016.3. In terms of the average years of research frontiers, in addition to Clu#4 being the earliest to obtain research frontiers, the frontier studies of Clu#6, Clu#3, and Clu#2 were also formed in 2010.2, 2010.7, and 2013.8. Clu#7 and Clu#1 and Clu#0, which are the largest clusters, are still in the process of research, and the average years of frontier studies are 2018.1, 2018.6, and 2019.1, respectively.

Of the top 50 articles with highest LCS, Polizzotto_2008 and Berg_2007 have the highest GCS of 399 and 321, both of which are all about groundwater arsenic pollution in the Mekong River Basin, especially in the Mekong Delta of Cambodia and Vietnam. The article with the highest GCS is a study on arsenic pollution in the Mekong Delta of Cambodia published in the journal Nature by Matthew L. Pollizzotoo in 2008, proposing that arsenic in groundwater comes from sediments of near-surface, river-derived sources and will return to rivers again through subsurface aquifers on a 100-year time scale and the release and transport of arsenic is very sensitive to continuous anthropogenic disturbance including groundwater extraction, changes in agricultural practices, sediment excavation, and construction of dykes and dams (Polizzotto et al., 2008), which is also the only article with LCS/GCS ratio less than 20%. Other arsenic-related research outputs, such as Berg_2007, Buschmann_2008, and Buschmann_2007 ranked 2nd, 5th, and 6th, respectively, in GCS, while their ratios of LCS and GCS are 21.2, 25.0, and 20.3%, respectively, indicating that a considerable part of the citations come from the contributions of scholars in other fields.

From the perspective of bubble size, the larger the bubble indicates that the higher the LCS/GCS ratio, and the more the research content represented by the bubble focuses on the water issue of the Mekong Basin itself, and the smaller the bubble, it means that scholars in other fields are also widely citing research results derived from this field, indicating the scalability and generalizability of the research outputs. In addition to the arsenic contamination studies mentioned above, Sakamoto_2006 and Sakamoto_2007 on mapping the spatial distribution and flood extent of rice cropping system in Cambodia and VMD based on MODIS time-series data also have the low LCS/GCS ratios. It is worth noting that Sneddon & Fox_2006 on water politics in the Mekong River and Bakker_1999 on the hydropower politics have also been widely cited in studies outside the region.

The document with the highest betweenness centrality of 0.10 is Eastham_2008 funded by AusAID on climate change's influence on Mekong River Basin water resources (Eastham et al., 2008), and Buschmann_2008 on the contamination of drinking water in the Mekong Delta floodplains especially arsenic and manganese contamination and their potential threats to human health (Buschmann et al., 2008). Kummu_2008 (Kummu & Sarkkula, 2008) and Lu_2006 (Lu & Siew, 2006) have the second highest betweenness centrality values of 0.07, which links the water discharge variation with Tonle Sap flood pulse and upstream dam construction, respectively, Lu_2006 is one of the earliest papers that linked changes in flow and sediment flux in the Lower Mekong River with the operation of the Manwan Dam upstream in China, which divided the 1962–2000 period into pre-dam and after-dam stages taking the operation of Manwan dam in 1992 as the critical time node, and proposed the river flow and sediment flux in the downstream decreased significantly after the impounding of Manwan dam. Subsequently, Kummu_2007 and Fu_2008 also studied the sediment interception capacity of Manwan Dam and its possible impact on the downstream, and their research results were widely cited by scholars in the field. There were two other papers with a betweenness centrality greater than 0.05, Kuenzer et al. (2013b) mapped the flood dynamics of the Mekong Delta based on time-series analysis of ENVISAT data, and Delgado_2012 studied the correlationship between Western Pacific monsoon and discharge of Mekong River, and proposed the possibility of predicting floods in the Mekong Delta based on the correlation analysis between the Pacific sea surface temperature and monsoon variance (Delgado et al., 2012).

The calculation results of Sigma values suggested that Eastham_2008 on climate change impact on Mekong River Basin water resources has the highest Sigma value of 2.56, Buschmann_2008 that links the arsenic contamination of groundwater with human health risk ranked the second highest Sigma value, Kummu & Sarkkula_2008 that links the flow variation caused by hydropower dam construction, large-scale irrigation plan, and the rapid urban development in the upper stream with ecosystem productivity of the Tonle Sap Lake Plain (Kummu & Sarkkula, 2008), and Lu & Siew_2006 that attribute changes in flow and sediment fluxes in the Lower Mekong River over the past few decades to Chinese dam construction also have the higher Sigma values of 1.74.

However, neither the centrality measure that indicates the role of the bridge nor the citation burstness indicator that measures the novelty of the study were mainly scored by members of Clu#2, and no transformative findings were identified from the members of the two largest clusters, Clu#0 on the Vietnamese Mekong Delta and Clu#1 on the whole Lancang-Mekong River Basin.

The water-related research process in the Mekong Region can be divided into seed stage (before 1994), budding stage (1995–2002), fluctuation stage (2003–2010), stabilization stage (2010–2016), and rapid growth stage (2017–present), which among the 12 identified clusters, the topic of Vietnamese Mekong Delta has the most research results, followed by the Lancang-Mekong River Basin. Their knowledge bases were both formed after 2015, showing their exuberant vitality, and it can be predicted that the research achievements in these two fields will continue to grow rapidly in the future. Clu#4 (Arsenic pollution/risk assessment) has the longest time span between the research frontier and its knowledge base, and articles with the highest GCSs and lowest LCS/GCS ratios are all members of Clu#4, indicating that the arsenic pollution of groundwater in the Mekong Region has been studied for a long time, and its research results have provided support for research in other fields or regions.

During 2009–2010, both the number of articles published and articles with high LCSs declined significantly, which could be related to the progress of widespread controversy and multi-party gaming over the hydropower development and dam construction. The first and highest peak in terms of LCS and GCS is in 2008, and the articles with the top highest betweenness centralities, Eastham_2008, Buschmann_2008, and Kummu_2008 were also concentrated around 2008, suggesting the periodical achievements and consensus were achieved in some fields during this period.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.