EN中文

Journal of Geographical Sciences >

2022 , Vol. 32 >Issue 10: 2036 - 2052

DOI: https://doi.org/10.1007/s11442-022-2035-0

Research Articles

Spatio-temporal evolution and influencing factors of geopolitical relations among Arctic countries based on news big data

  • LI Meng , 1, 2 ,
  • YUAN Wen 1 ,
  • YUAN Wu 3 ,
  • NIU Fangqu , 1, 2, * ,
  • LI Hanqin 4 ,
  • HU Duanmu 1, 2
Expand
  • 1.Institute of Geographic Sciences and Natural Resources Research, CAS, Beijing 100101, China
  • 2.University of Chinese Academy of Sciences, Beijing 100049, China
  • 3.School of Computer Science & Technology, BIT, Beijing 100081, China
  • 4.First Research Institute of the Ministry of Public Security of PRC, Beijing 100048, China
* Niu Fangqu (1979-), PhD and Associate Professor, specialized in urban and regional sustainable development modeling. E-mail:

Li Meng (1995-), Master, specialized in big data analysis research. E-mail:

Received date: 2022-05-09

  Accepted date: 2022-06-28

  Online published: 2022-12-25

Supported by

National Natural Science Foundation of China(42071153)

The Strategic Priority Research Program of Chinese Academy of Sciences(XDA19040401)

The Strategic Priority Research Program of Chinese Academy of Sciences(XDA20080100)

Fold

Abstract

Global warming has caused the Arctic Ocean ice cover to shrink. This endangers the environment but has made traversing the Arctic channel possible. Therefore, the strategic position of the Arctic has been significantly improved. As a near-Arctic country, China has formulated relevant policies that will be directly impacted by changes in the international relations between the eight Arctic countries (regions). A comprehensive and real-time analysis of the various characteristics of the Arctic geographical relationship is required in China, which helps formulate political, economic, and diplomatic countermeasures. Massive global real-time open databases provide news data from major media in various countries. This makes it possible to monitor geographical relationships in real-time. This paper explores key elements of the social development of eight Arctic countries (regions) over 2013-2019 based on the GDELT database and the method of labeled latent Dirichlet allocation. This paper also constructs the national interaction network and identifies the evolution pattern for the relationships between Arctic countries (regions). The following conclusions are drawn. (1) Arctic news hotspot is now focusing on climate change/ice cap melting which is becoming the main driving factor for changes in geographical relationships in the Arctic. (2) There is a strong correlation between the number of news pieces about ice cap melting and the sea ice area. (3) With the melting of the ice caps, the social, economic, and military activities in the Arctic have been booming, and the competition for dominance is becoming increasingly fierce. In general, there is a pattern of domination by Russia and Canada.

Key words： Arctic; geographical relationship; spatiotemporal data mining; topic model; interactive network; big data

Cite this article

LI Meng , YUAN Wen , YUAN Wu , NIU Fangqu , LI Hanqin , HU Duanmu . Spatio-temporal evolution and influencing factors of geopolitical relations among Arctic countries based on news big data[J]. Journal of Geographical Sciences, 2022 , 32(10) : 2036 -2052 . DOI: 10.1007/s11442-022-2035-0

1 Introduction

As early as 1942, George Rainer noted the importance of the Arctic region (Ye et al., 2019). In recent years, the sea ice area of the Arctic has gradually decreased. Although this brings about environmental problems, it has increased the availability of Arctic resources to a certain extent, hence, the strategic position of the Arctic has significantly increased (Du et al., 2020). This has significantly impacted China’s Arctic policies due to changes in geopolitical relations in the Arctic region (He, 2010). A comprehensive real-time analysis of the geopolitical relationships of the Arctic region and its changing characteristics is an important reference for China’s politics, economy, and diplomacy (Xiong et al., 2018; Guan and Li, 2021). Scholars have focused their research on the geopolitical relations of the Arctic region on the following three aspects. (1) Studying the development of geo-relations in the Arctic region based on major events (Michael, 2017); (2) studying changing geopolitical relationships in the Arctic region by analyzing countries’ Arctic policies (Li, 2015; Zhu, 2016; Guo and Wang, 2017; Xiao, 2018; Ma, 2019; Xie and Cheng, 2019); (3) studying patterns of Arctic geopolitical relations by analyzing the involvement of countries in Arctic hotspots (Huang et al., 2020; Ye, 2021). However, international relations are complex and change rapidly, which involves politics, economy, culture, and other aspects. The comprehensiveness and real-time data directly affect the accuracy of international relations analyses. Static analysis methods, such as case studies of major events, policy studies, and single-topic analysis, cannot solely provide an accurate and comprehensive analysis of changes in Arctic geopolitical relations.
With the coming of big data, the emergence and rapid developments in new technologies have made it possible to acquire, store, and compute massive amounts of data to provide new opportunities to study the spatial and temporal evolutionary patterns of geo-relations (Cheng et al., 2018). Big data in news has gained the favor of many scholars because of the advantages of wide content coverage, easy access, and high update frequency. Su et al. (2016) studied changes in the relevant news coverage and aid from different countries after the Nepal earthquake based on news big data to provide new ideas to understand the work of international and local organizations and governments in post-earthquake relief. Ma et al. (2019) used a theme model to extract corresponding news theme words of 25 countries and regions along the Belt and Road, and then quantified the stability and social development trends for each country. Chen et al. (2019) analyzed the geopolitical relations between China and neighboring countries in time and space based on a constructed network of “cooperation-conflict” events. However, most current studies are based on the event class code of news databases, which analyze news data in fixed categories such as riots, protests, conflicts, and diplomacy. These inflexible topics do not reflect the driving factors of changes in international relations.
This paper is based on the GDELT (Global Database of Events, Language, and Tone) database. We collected the full text of the news from GDELT based on the original links to construct a database of news from eight Arctic countries. The key elements of social development contained in the news for the Arctic Circle from April 2013 to December 2019 were analyzed using the labeled latent Dirichlet allocation (LDA) thematic analysis method, and a country (region) interaction network was built based on the interaction behaviors between countries in news events. The results of the analysis were correlated with the Arctic Sea ice extent monitoring data to verify the feasibility of news data instead of real monitoring data to study the interaction network between countries (regions). Ultimately, analyzing the characteristics of national (regional) interaction networks explores the development of international interaction themes and geopolitical relations in the Arctic region in the two dimensions of time and space.

2 Data sources and research methods

2.1 Data sources

This study covers eight countries and regions within the Arctic Circle: Canada, Russia, Norway, Sweden, Denmark, Finland, Iceland, and Greenland. Among them, Greenland has been officially self-governed since June 21, 2009. As a strategic point between the Atlantic Ocean and Arctic Ocean, Greenland is the only pivotal area connecting Asia and Europe and Europe and America in the circum-Arctic shipping lanes and has an important impact on the geopolitical relationships of the Arctic region (Guo and Wang, 2017). Of note, the Alaska state of United States is also located in the Arctic Circle. At present, the United States, as the world’s only superpower, pursues a global strategy and is deeply involved in Arctic affairs. However, its Arctic strategy has recently undergone a major strategic turn. The United States believes that the Arctic has entered a “new era of strategic competition,” and the Arctic is a “potential corridor” for strategic competition with China and Russia (Xu, 2021). The “American factor” in the Arctic geopolitical context, its impact on the region, and its changing relationship with neighboring countries is difficult to generalize in a single paper. Thus, the United States is not included here to control the scope of this study and not to complicate the issue.
The news data used in this paper were obtained from the GDELT Global Knowledge Graph (GKG) database. The GDELT is a real-time, open-source global news event database that is updated every 15 min and monitors broadcasts, print, and online news data from countries around the world in more than 100 languages. It analyzes text and extracts key information such as people, places, events, and sources. The main types of GDELT events are shown in Table 1.
Table 1 Event types in GDELT
Event type
Make public statement Appeal Express intent to cooperate Consult Engage in diplomatic cooperation
Demand Provide aid Yield Investigate Engage in material cooperation
Disapprove Reject Threaten Protest Exhibit military posture
Reduce relations Coerce Assault Fight Engage in unconventional mass violence
In addition, the GDELT GKG database contains sentiment analysis of news content as represented by the TONE field with scores ranging from -100 (negative sentiment) to +100 (positive sentiment). This provides data support for the article’s sentiment-oriented analysis of international relations in the Arctic region. As the GDELT database has been updated daily with global news since April 1, 2013, the amount of news data before 2013 is small and insufficient to support regional international relations analyses based on textual topic mining, so 2013 is the starting point of this study. On the other hand, the GDELT event categories are mainly political, and there are fewer categories related to Arctic hotspot issues, such as ice cap melting, animal protection, and shipping traffic. Therefore, this study only uses the GDELT database as a source to extract the links of news events and then uses crawler technology to obtain the full text of web pages. This helps mine the news topic as well as the event time, location, and other information to build the Arctic region news database.

2.2 Research methodology

2.2.1 LDA and labeled LDA topic models

The LDA is an unsupervised document topic generation model that contains a three-layer structure of words, topics, and documents that can be used to identify latent topic information in large-scale document collections or corpus (Wang et al., 2015). This considers that each word in an article is generated by “selecting a topic with a certain probability and selecting a word from that topic with a certain probability.” The probability that a word appears in a news document is:
$\text{p}\left( \text{word}|\text{document} \right)=\sum\nolimits_{\text{topic}}^{{}}{\text{p}(\text{word}|\text{topic})\text{p}(\text{topic}|\text{document})}$
where the theme vector is $\overrightarrow{{{\theta }_{m}}}$, the word vectors are $\overrightarrow{{{\varphi }_{k}}}$ and obey the Dirichlet prior distributions $\Delta (\vec{\alpha })$ and $\Delta (\vec{\beta })$ of $\vec{\alpha }$ and $\vec{\beta }$, w indicates vocabulary, and z indicates the subject. The joint probability distribution of the above equation can then be expressed as:
$p(\vec{w},\vec{z},|\vec{\alpha },\vec{\beta })=\text{p}(\vec{w}\text{ }\!\!|\!\!\text{ }\vec{z},\vec{\beta })p(\vec{z},\vec{\alpha })$
The probability distribution shape is obtained from the Gibbs sampling as:
$P({{z}_{i}}=\text{k }\!\!|\!\!\text{ w},\text{z})=\left( \frac{n_{m,\neg i}^{k}+{{\alpha }_{k}}}{\mathop{\sum }_{k=1}^{K}n_{m,\neg i}^{(k)}+{{\alpha }_{k}}} \right)*\left( \frac{n_{k,\neg i}^{t}+{{\beta }_{t}}}{\mathop{\sum }_{t=1}^{V}n_{k,\neg i}^{(t)}+\beta } \right)$
where $n_{m,\neg i}^{k}$ denotes the number of feature words in news document m that are not assigned to topic k, $n_{k,\neg i}^{t}$ denotes the number of times feature word t is not assigned to topic k. The $\sum\nolimits_{k=1}^{K}{n_{m,\neg i}^{(k)}}$ indicates the number of unassigned feature words per topic in the news document, and $\sum\nolimits_{t=1}^{V}{n_{k,\neg i}^{(t)}}$ denotes the number of times each feature term of topic k is not assigned. With known data and document generation rules, parameter estimation can help find the values of the maximum hyperparameters α and β in Eq. (3), which mines the implied themes in the news space-time database. The LDA topic model is simple to use, efficient to execute, and widely utilized in the fields of topic recognition and text classification. However, this also suffers from difficulty interpreting topic vectors, parameter calibrations, and limited accuracy (Li et al., 2008).
Labeled LDA models introduce labeled data and use supervised learning to establish the mapping relationship between labels and topics, which avoids the problem of difficult interpretation of topic vectors and significantly improves the classification accuracy (Wang et al., 2015). However, the acquisition of labeled data is labor-intensive and not suitable for situations where the data cover a wide range and have large sample sizes.

2.2.2 National (regional) interaction network

A country (region) interaction network is a structure chart that describes the interaction patterns between countries (regions) with the countries (regions) as nodes and interaction behaviors as edges. With the rise of the news media, large and small events have been reflected in news coverage as well as interactions between countries. When two countries (regions) appear in a news article at the same time, it is assumed that they are involved in the news event, i.e., they interact in the type of event. In this paper, the relationship between any country-pair that appears in the news at the same time is defined as an “interaction,” and the topic of the news is defined as the event. A non-directional network $\text{G}(\text{V},\ \text{E},\ \text{W})$ for country interactions is constructed based on the built news database as:
$W=\{{{w}_{{{L}_{1}}}},{{w}_{{{L}_{2}}}},\ldots,{{w}_{{{L}_{N}}}}\}$
${{w}_{{{L}_{i}}}}=\frac{\mathop{\sum }_{t}\mathop{\sum }_{{{n}_{i}}}Numart{{s}_{j}}}{\mathop{\sum }_{t}\mathop{\sum }_{N}\mathop{\sum }_{{{n}_{i}}}Numart{{s}_{j}}}$
where the point set V denotes the eight countries (regions) in the Arctic Circle, the edge set E denotes the set of interactions between countries (regions) at time t, and W denotes the weight of the edge, i.e., the strength of the interactions between countries (regions). The interactions between countries (regions) are divided into N categories based on the news content by labeled LDA topic analyses, and the proportion of the news volume for each category in time t is the weight of the topic ${{w}_{{{L}_{i}}}}$. The volume of the news per topic category is obtained by summing the total number of source documents for that category.

3 Analysis of news topics in the Arctic

This paper obtained news data from eight countries and regions in the Arctic Circle from 2013-2019 (Table 2) and used the LDA topic model to conduct an exploratory analysis of the news topics. The eight most popular topics in the Arctic region were then selected as target topics through manual interpretation of each topic term and combined with the results of the Arctic literature analysis. A manual labeling method was used for the news data to obtain more refined target topic training data. The Arctic region news topic analysis model was trained using the labeled LDA method, and the top eight topic words corresponding to the frequency ranking of each topic are shown in Table 3.
Table 2 Statistics on the number of news articles in the Arctic by topic from 2013-2019
Year Animal protection Shipping traffic Climate change Economic
Activities		 Military activities Regional governance Resource development Melting ice caps
2013 324 398 766 625 1803 712 773 499
2014 671 575 977 1336 5422 1120 953 580
2015 956 2381 1625 1045 3616 1342 1434 1199
2016 1447 3915 1377 1195 3923 1476 1028 1716
2017 1029 1603 1326 832 4080 1460 943 1109
2018 1280 1193 2147 933 3244 1270 930 1338
2019 1589 1537 3591 991 4605 3162 1220 1940
Table 3 High frequency words corresponding to each news topic in the Arctic from 2013-2019
Events Top 1 Top 2 Top 3 Top 4 Top 5 Top 6 Top 7 Top 8
Melting ice caps ice climate glacier sheet cap record researcher melting
Animal protection bear polar wildlife climate animal ice conservation study
Shipping traffic ship cruise ice polar route coast expedition Norwegian
Regional governance policy development region oil country paper cooperation state
Climate change climate temperature cold weather record degree polar wind
Resource development energy oil fuel drilling fossil gas company climate
Economic activities oil market price natural industry company car gas
Military activities military defense force state missile Nordic war logistics
The news topics for melting ice caps are mainly “ice”, “climate”, “glacier”, “sheet ice”, “glacier”, “sheet”, and “cap”, and the news content is related to melting ice caps caused by the warming of the Arctic region and the progress of scientific research in the field. Climate change news describes the warming Arctic climate and environmental changes. The word “climate” is also found in the subject line of animal protection news together with the words “bear,” “wildlife,” and “animal,” which form the main content of Arctic animal protection news. The distribution of subject words shows “oil” has a large proportion in the news of regional governance, resource development, and economic activities. Regional governance news focuses on inter-regional political exchanges and cooperation in the Arctic region, while the rich oil and gas resources of the polar regions make resource development an important theme of regional exchanges and cooperation; thus, economic activities in the Arctic region revolve around resource development. In addition, the Arctic news shows that shipping traffic and military activities for territorial expansion, resource seizure, stability maintenance, etc. are important issues. To study the intrinsic connection between these eight types of target topics, this study counted the number of news articles for each topic in different years, and the number of news represented their importance for correlation analysis. The results are shown in Table 4.
Table 4 Correlation coefficient of various news topics in the Arctic region from 2013-2019
Animal protection Shipping traffic Climate change Economic
Activities		 Military activities Regional governance Resource development Melting ice caps
Animal protection 1.0000 0.6217 0.7749 0.3240 0.3986 0.7586 0.4437 0.9701
Shipping traffic 0.6217 1.0000 0.1113 0.3716 0.1289 0.2053 0.4591 0.6684
Climate change 0.7749 0.1113 1.0000 0.0423 0.2843 0.9241 0.4914 0.7970
Economic activities 0.3240 0.3716 0.0423 1.0000 0.8159 0.1507 0.3534 0.2063
Military activities 0.3986 0.1289 0.2843 0.8159 1.0000 0.4624 0.3336 0.2617
Regional governance 0.7586 0.2053 0.9241 0.1507 0.4624 1.0000 0.5001 0.7947
Resource development 0.4437 0.4591 0.4914 0.3534 0.3336 0.5001 1.0000 0.5138
Melting ice caps 0.9701 0.6684 0.7970 0.2063 0.2617 0.7947 0.5138 1.0000
There is a strong positive correlation between climate change and regional governance and between ice cap melting and animal protection. In addition, shipping traffic has a strong positive correlation with ice cap melting, and economic activity is strongly correlated with military activity. These eight categories are positively correlated with each other and influence each other but are dominated by climate change and melting ice caps. This indicates that melting ice caps due to climate change is the central driver to affect international relations in the Arctic. A series of environmental problems brought about by climate warming have posed new challenges to developing Arctic countries. While the melting ice caps have also brought new opportunities for Arctic resource development, channel openings, and economic development, both opportunities and challenges have moved the Arctic region to a climate change-driven geopolitical era.

4 Evolution of spatiotemporal patterns of national (regional) interaction networks in the Arctic region based on news topics

The number of news items reflects the importance of a certain topic, and the increase or decrease of the category activity of the news items is often related to changes in observed data. To demonstrate the feasibility of using news data instead of real monitoring data to study inter-country (regional) interaction networks, this study obtained monitoring data for the minimum sea ice extent in the Arctic from September 2013 to September 2019 from the National Snow and Ice Data Center (NSIDC) and performed correlation analysis with the news heat data for ice cap melting from 2013 to 2019 (Figure 1).
Figure 1 Correlation coefficients between the number of melting ice cap news pieces and sea ice area in the Arctic from 2013-2019
News for Arctic ice cap melting showed an extremely strong negative correlation with the sea ice minimum extent monitoring data with a correlation coefficient of -0.97 (p<0.001), which is statistically significant. From Figure 1, the trend of the sea ice extent is inversely proportional to the news articles of ice cap melting. This indicates that it is feasible to use news data instead of real monitoring data to study the interaction network between countries (regions) as a strong correlation is necessary. This study examines the characteristics of national (regional) interaction networks in both temporal and spatial dimensions based on the analysis of news topic data in the Arctic region and explores the evolutionary patterns of geopolitical relations in the Arctic region.

4.1 Characteristics of the temporal evolution of the interaction network

With melting ice caps and surges in socio-economic and military activities in the Arctic region, the competition for its dominance has become increasingly intense. The network of national (regional) interactions in the Arctic region has developed toward greater participation and closer interactions between countries but differs in detail depending on the intrinsic characteristics of the various interaction themes.
The network of national (regional) interactions on the theme of regional governance reflects its character as the most immediate topical issue in the Arctic. As shown in Figure 2, these eight countries (regions) in the Arctic have been actively involved in the Arctic regional governance from 2013-2019 with high interactions between all countries. The closeness of interactions has been increasing year-over-year and reached very close levels by 2019. 2018 was briefly characterized by an extremely close Russia-Canada interaction and more balanced interactions between other countries (regions), which shows the characteristics of the dominance of major powers. However, given the rise of the intensification of climate change and resource competition in the Arctic, countries (regions) in the Arctic Circle will not give up their right to participate in Arctic affairs because of their national power, but they will gradually expand the breadth and depth of their participation based on their characteristics.
Figure 2 Changes in the national interaction network of regional governance in the Arctic from 2013-2019
Most of the national (regional) interaction networks have gone through a process from sparse to dense, including military activities, shipping and transportation, climate change, ice cap melting, and animal protection. The national (regional) participation and sparse interaction were at the beginning and active participation and densification interactions were later. However, the duration of this change process varies between themes. For topics of immediate interest to each country (such as military activities and shipping traffic) that are normally related to Arctic encirclement movements, control of shipping routes, and resource development, the interaction network diagrams have a transition from sparse to dense from 2014-2015, respectively (Figures 3 and 4). From 2013-2019, military interactions in the Arctic gradually intensified with initial military interactions occurring primarily between Russia and Canada followed by closer military interactions between Russia and Norway. By 2019, Russia had close military interactions with most countries (regions) in the Arctic Circle. The literature shows that Russian military activities in the Arctic are markedly confrontational (Sun and Ma, 2017), which suggests a future pattern of one-to-many military interactions in the Arctic that is dominated by Russia.
Figure 3 Changes in the national interaction network of military activity in the Arctic from 2013-2019
Figure 4 Changes in the national interaction network of channel traffic in the Arctic from 2013-2019
Figure 4 shows temporal changes in the interaction network of shipping traffic theme countries (regions). Unlike military activities, the shipping traffic network tended to be dense in 2015, while 2016 is the year of closest interactions. Although the Arctic countries still maintain some interaction with each other after this, the closeness of interaction between countries (regions) starts to decrease to ordinary levels. The sea ice extent has been decreasing yearly since 2014 and dropped to a seven-year low by 2016. This provides natural conditions for Arctic navigation, and Arctic countries have seized this opportunity to develop their shipping economies. After 2016, the Arctic Sea ice extent rebounded and the boom in shipping development receded, causing its interactive closeness to decline. However, shipping traffic has always been one of the competed resources for the Arctic in the context of global warming, and countries will remain close in shipping traffic for some time (Liu et al., 2015). When the next plunge in sea ice extent occurs, the degree of close interaction will likely increase as dramatically as it did in 2016.
For themes not directly related to Arctic interest contests, such as climate change, ice cap melting, and animal conservation, the interaction network map spans a longer period from sparse to dense, and the interaction closeness is lower than previous themes. As shown in Figures 5 and 6, Arctic countries did not interact more closely on the topic of climate change until 2018, but the ice cap melt interaction network only reached a denser structure in 2019. In both categories, Greenland has become a major participant, actively interacting with other countries in the Arctic region on climate change and melting ice caps. This is related to its unique geography, where 84% of Greenland’s territory is covered by snow and ice. The ice cap covers an area of 180 km², and its ice cap has an average thickness of 2300 m, while climate warming and ice cap melting directly impact it (Guo and Wang, 2017). Figure 7 shows the temporal changes in the interaction network of animal protection thematic countries (regions). Although it reached a more balanced network structure for national interactions in 2016, its overall level of close interaction is low, and Arctic animal protection still faces serious challenges.
Figure 5 Changes in the national interaction network for climate change in the Arctic from 2013-2019
Figure 6 Changes in the national interaction network of shrunken ice cap theme from 2013-2019
Figure 7 Changes in the national interaction network for animal protection in the Arctic from 2013-2019.
Temporal variations in the interaction network for resource development and economic activity has distinctive features (Figure 8). The closeness of their interactions increased from 2013-2015 but suddenly decreased in 2016, and then again increased year-over-year. 2016 was the year with the smallest Arctic Sea ice extent in the last seven years. As Arctic ice cap melting intensifies, environmental threats have slowed the pace of resource exploitation. Economic activity has been affected by this with weakening interactions. The difficulty of resource exploitation and underdeveloped economic activities in the Arctic are the main reasons why these two themes interact much less closely than the others (Figure 9).
Figure 8 Changes in the national interaction network for resources in the Arctic from 2013-2019
Figure 9 Changes in the national interaction network for economic activity in the Arctic from 2013-2019
Although the closeness of interactions between the different themes for the eight Arctic states varies over time, it is generally seen to be strongly related to Arctic climate change. Ice cap melting caused by climate change is an important driving factor that affects the interactions of these eight countries (regions) in terms of shipping and transportation, resource development, economic activities, animal protection, etc. Thus, the Arctic has entered an era of climate change-driven geopolitics. The area of sea ice in the Arctic has gradually decreased in recent years, which has facilitated the construction of Arctic waterways and resource development. Countries with Arctic interests have been involved in these activities. At the same time, a series of environmental risks brought by climate warming has also touched the nerves of Arctic countries, making the interactions between them co-exist with cooperation and competition and with opportunities and challenges. However, the current competition between Arctic states (regions) is based primarily on the perspective of Arctic development. This has manifested in a limited policy and action game within the framework of governance, and the Arctic geopolitical relations will develop in the direction of win-win cooperation.

4.2 Spatial distribution patterns of interaction networks

Differences in comprehensive national power, geographical location advantages, and interests of Arctic countries have led to a unique pattern of national (regional) interaction networks. The results of counting the number of news articles in different sentiment categories in the interactive network of the Arctic countries (regions) in 2019 are shown in Table 5. The proportion of “positive” and “very positive” news in the interactions between two countries to the total number of news other than “neutral” is calculated to indicate the proportion of positive sentiment in the interactions. The total number of news items in each sentiment category for interactions between the two countries is calculated as an indication of their closeness. Finally, the color gradient is used to indicate the percent of positive sentiment, and the line width is used to indicate the interaction closeness. The spatial distribution of the interaction network in the Arctic countries (regions) in 2019 is obtained (Figure 10).
Table 5 Sentiment analysis of national interactive networks in the Arctic in 2019
Interaction Very negative Nega-
tive		 Neu-
tral		 Posi-
tive		 Very positive Interaction Very negative Nega-
tive		 Neu-
tral		 Posi-
tive		 Very positive
Canada-Russia 1 125 2351 21 0 Finland-Sweden 0 10 893 10 0
Norway-Russia 0 54 2107 20 0 Canada-Sweden 0 18 857 1 0
Denmark-Greenland 0 10 1848 7 0 Greenland-Norway 0 4 818 19 1
Canada-Greenland 1 27 1568 18 0 Iceland-Russia 0 5 815 12 0
Greenland-Russia 0 41 1485 11 0 Denmark-Sweden 0 10 754 12 0
Denmark-Russia 0 9 1435 8 0 Greenland-Iceland 1 25 686 20 0
Canada-Norway 0 12 1394 17 0 Canada-Finland 0 45 647 8 2
Finland-Russia 0 72 1301 4 0 Canada-Iceland 1 8 632 16 0
Norway-Sweden 0 18 1136 14 0 Denmark-Finland 0 4 637 8 1
Russia-Sweden 0 29 1046 3 0 Denmark-Iceland 0 3 624 12 0
Denmark-Norway 0 7 957 19 0 Finland-Iceland 0 3 523 14 0
Canada-Denmark 0 5 953 10 0 Iceland-Sweden 0 7 487 8 0
Iceland-Norway 0 7 916 28 0 Greenland-Sweden 2 15 412 0 0
Finland-Norway 1 23 903 16 0 Finland-Greenland 0 8 380 1 0

Note: The values in the table are the number of news items in different sentiment categories in the national interactive network; sentiment scores of [-100, -10) are “very negative,” [-10, -5) are “negative,” [-5, 5) are “neutral,” [” and [10,100] are “very positive.” The table is arranged in descending order of the total number of news items.

Figure 10 Spatial distribution of national interactive networks in the Arctic in 2019
As seen in Figure 10, the degree of all eight nodes of the graph is 7. There are relatively close interactions between any two of these eight countries (regions) in the Arctic region, which shows there are frequent interactions and that all countries are actively involved in Arctic affairs. In terms of the sentiment orientation of the interactions, there are few very negative and very positive news articles in 2019, and the proportion of neutral news ranges from 92.17% to 99.09%, which accounts for the vast majority. This indicates that international relations in the Arctic region were generally more stable in 2019. However, there were still 15 groups of interactions with less than 50% positive sentiment. Among these, the negative sentiment interactions between Russia and Canada are the most obvious, which keeps a tense situation in the Arctic region.
Russia and Canada are the two of the eight Arctic states with the strongest combined power and have an important influence on the Arctic region in terms of their emotional orientation when interacting with other countries. Russia ranks first in 2019 by participating in 18.56% of the total interactions for all eight Arctic countries. In Arctic affairs, Russia has adhered to the principle of scientific research as a forerunner, military as the backing, cooperation as the means, and pre-emption as a basis for Arctic affairs (Luo and Li, 2020). Figure 10 shows that Russia’s interactions with all countries (regions) are dominated by negative sentiment, except for interactions with Iceland, which are dominated by positive sentiment. This has much to do with the ongoing strategic squeeze on Russia by the US-led NATO countries in recent years, which has seen Russia become increasingly active in Arctic affairs and display a more tough stance to find new strategic pivots. Canada’s involvement in Arctic affairs is second only to Russia at 14.8 percent. Its emotional orientation when interacting with Arctic countries is highly factional, as evidenced by the predominance of positive emotions when interacting with Denmark, Norway, and Iceland, which are also NATO members, while there is a predominance of negative emotions with other countries (regions).
In the interaction network of countries (regions) in the Arctic region, Iceland’s interactions with all countries are almost always dominated by positive sentiment, even though the percentage of positive sentiment in interactions with Greenland is less than 50% (43.48%). Iceland is a small and sparsely populated country and is the smallest among the eight Arctic countries. Due to its limited combined national power, it is the least involved in Arctic affairs. At the same time, its economic development is highly dependent on the natural resources of the Arctic region. This leads to the predominance of positive emotions in its interactions with countries.
Greenland has attracted significant attention from various countries due to its important geographical location, rich mineral resources, and political demands to achieve independence. Among them, Canada, Russia, and Denmark are the three countries with the heaviest share in Greenland’s interaction network and a high share of negative sentiment. Greenland is located at the intersection of the Northeast, Northwest, and Central routes of the Arctic, which is an important strategic position is (Jacobsen, 2020). Against the backdrop of competing forces for control in the Arctic shipping lanes, Canada interacts particularly closely with Greenland because of its northern Arctic archipelago in the important Northwest Passage (Shen et al., 2021). Greenland is also an important arena for US-Russian strategic rivalries, with the famous Thule Air Base located on the northwest coast of Greenland and close interactions between Russia and Greenland at a time when US-Russian relations were tense. The reason for the very close interaction between Denmark and Greenland is that Greenland is not yet fully independent from Denmark. Denmark has strengthened its military and economic control over Greenland to safeguard its Arctic interests.
The spatial distribution of the interaction network of Arctic countries (regions) generally shows patterns dominated by Russia and Canada. Other Arctic countries (regions) also play a role in important issues and themes specific to their countries based on their characteristics. With the intensification of climate change and resource competition in the Arctic, these countries (regions) will not give up their right to participate in Arctic affairs because of their national power, but will gradually expand the breadth and depth of their participation based on their characteristics.
As an official observer state of the Arctic Council, China is also actively participating in Arctic affairs, which is not only a necessity of national interests but also a legitimate right given to China by international law and is an international responsibility given to China by the times. Under the interactive network of international relations in the Arctic led by Russia and Canada, China should take the construction of the “Silk Road on Ice” as a grip to steadily promote economic and trade cooperation with countries (regions) in the Arctic Circle; actively participate in the opening of shipping lanes, ecological protection, and resource conservation and utilization in the Arctic region; and make positive contributions to the social and economic prosperity and sustainable development.

5 Conclusions

This paper studies the evolution of geo-relations in the Arctic region based on a news database. It uses the LDA and labeled LDA models to mine popular topics in the Arctic region, constructs a national (regional) interaction network model, reveals the spatial and temporal evolution characteristics of the national (regional) interaction network in the Arctic region through model analysis, and provides suggestions for China’s active participation in Arctic affairs. The main findings are as follows.
(1) The top issues of interest in the Arctic include military activities, climate change, shipping and transportation, regional governance, melting ice caps, animal conservation, resource development, and economic activities. These eight categories are positively related to each other, but climate change and ice melting are dominant. With global warming, melting ice caps have become a major driving factor that affects changes in Arctic geopolitical relations. Thus, the Arctic has entered a climate change-driven geopolitical era.
(2) There is a strong correlation between news on ice cap melting and changes in sea ice monitoring data. Thus, it is feasible to study the interaction network between countries (regions) using news data instead of real detection data.
(3) As ice caps melt, socio-economic and military activity in the Arctic has proliferated and the struggle for its dominance has become increasingly fierce. In the last seven years, the network of national (regional) interactions in the Arctic region has evolved toward greater participation and closer interactions between countries, but with different details depending on the intrinsic characteristics of different interaction themes. There have been high levels of interactions in regional governance themes, with Canada, Russia, Norway, Sweden, Denmark, Finland, Iceland, and Greenland all showing high levels of enthusiasm for regional governance in the Arctic. National (regional) interaction networks on the themes of military activities, shipping traffic, climate change, melting ice caps, and animal protection have undergone a sparse to dense process, but the network structure has tended to be dense at different rates depending on their closeness to countries’ immediate interests. Thematic networks of resource development and economic activities have been influenced by natural environmental changes, such as climate change and melting ice caps, and have undergone an interaction closeness process of increase, decrease, and gradual increase.
(4) The interaction network of countries (regions) in the Arctic shows a spatial distribution pattern dominated by Russia and Canada. Russia and Canada, as the comprehensive powers in the Arctic region, have been actively involved in Arctic affairs and interact closely with neighboring countries on topical issues. Particularly important topics include regional governance, shipping traffic, and military activities. Other countries (regions) also play a role in important issues and themes that they are interested in according to their characteristics. With the intensification of climate change and resource competition in the Arctic, countries (regions) in the Arctic Circle will not give up their right to participate in Arctic affairs because of their national power but will gradually expand the breadth and depth of their participation.

References

Publishing order | Descend order by publishing year | Descend order by cited within
[1]
Chen X Q, Yuan L H, Shen S et al., 2019. Analysis of the geo-relationships between China and its neighboring countries. Acta Geographica Sinica, 74(8): 1534-1547. (in Chinese)

DOI

[2]
Cheng C X, Shi P J, Song C Q et al., 2018. Geographic big-data: A new opportunity for geography complexity study. Acta Geographica Sinica, 73(8): 1397-1406. (in Chinese)

DOI

[3]
Du D B, Qin D H, Ma Y H et al., 2020. The age of cryopolitics is coming. Bulletin of Chinese Academy of Sciences, 35(4): 514-522. (in Chinese)

[4]
Guan X G, Li Z F, 2021. The economic effect of the “Polar Silk Road” on China and the Arctic countries along the route. Journal of Dalian University of Technology (Social Sciences), 42(1): 45-57. (in Chinese)

[5]
Guo P Q, Wang Junjie, 2017. The geopolitical influence of Greenland independence. Contemporary International Relations, (8): 58-64. (in Chinese)

[6]
He Q S, 2010. Climate change and Arctic geopolitical game. Foreign Affairs Review (Journal of Foreign Affairs College), 27(5): 113-122. (in Chinese)

[7]
Huang J X, Zhang T Y, Wang L et al., 2020. Spatial distribution patterns and the accessibility evaluation of oil and gas resources in Russia. Acta Geographica Sinica, 75(9): 2009-2024. (in Chinese)

DOI

[8]
Jacobsen M, 2020. Greenland’s Arctic advantage: articulations, acts and appearances of sovereignty games. Cooperation and Conflict, 55(2): 170-192.

DOI

[9]
Li W B, Sun L, Zhang D K, 2008. Text classification based on Labeled-LDA model. Chinese Journal of Computers, 4: 620-627. (in Chinese)

[10]
Li Z F, 2015. Arctic countries network and China’s Arctic strategy. Northeast Asia Forum, 24(2): 31-44. (in Chinese)

[11]
Liu D H, Ma Y R, Wang C J et al., 2015. Developments of Arctic passage resources under global climate change. China Population Resources and Environment, 25(Suppl.1): 6-9. (in Chinese)

[12]
Luo Y J, Li F, 2020. The great powers’ Arctic game and China’s Arctic energy security: Probing into the advancement path of the “Ice Silk Road” framework. Journal of International Security Studies, 38(2): 91-115. (in Chinese)

[13]
Ma M Q, Yuan W, Ge Q S et al., 2019. Big data analysis of social development situation in several regions along the “Belt and Road”. Progress in Geography, 38(7): 1009-1020. (in Chinese)

[14]
Ma X M, 2019. China’s Arctic policy on the basis of international law: Identification, goals, principles and positions. Marine Policy, 100: 265-276.

DOI

[15]
Michael B, 2017. Crises and international cooperation: An Arctic case study. International Relations, 31(4): 375-402.

DOI

[16]
Shen X Y, Zhang Y, Chen C S et al., 2021. Long-term spatial and temporal variations of sea ice in the northwest passage of the Canadian Arctic Archipelago. Chinese Journal of Polar Research, 33(1): 71-87. (in Chinese)

[17]
Su Y, Lan Z Y, Lin Y R et al., 2016. Tracking disaster response and relief following the 2015 Nepal Earthquake. 1: 495-499.

[18]
Sun Q J, Ma J G, 2017. Deterrence strategy of Russia in geopolitical game in the Arctic region. Journal of SJTU (Philosophy and Social Sciences), 25(1): 14-21. (in Chinese)

[19]
Wang P, Gao C, Chen X M, 2015. Research on LDA model based on text clustering. Information Science, 33(1): 63-68. (in Chinese)

DOI

[20]
Xiao Y, 2018. Greenland: The anchor of Danish Arctic strategy’s transformation? Pacific Journal, 26(6): 78-86. (in Chinese)

[21]
Xie X G, Cheng X B, 2019. The construction of China-Russia “Ice Silk Road” under the influence of Russian Arctic Policy’s adjustment. Journal of Liaoning University (Philosophy and Social Sciences), 47(1): 184-192. (in Chinese)

[22]
Xiong C R, Wang L M, Liang M L, 2018. The construction of the “Polar Silk Road” and its geo-strategic significance on China. Academic Exploration, 229(12): 31-37. (in Chinese)

[23]
Xu G M, 2021. Arctic order in a changing world: Generation mechanism and change logic. Russian, East European & Central Asian Studies, 1: 106-124, 157-158. (in Chinese)

[24]
Ye B H, Cheng Y, Wang L et al., 2019. A systematic review on geo-relations research in the Arctic. Progress in Geography, 38(4): 489-505. (in Chinese)

DOI

[25]
Ye B H, Cheng Y, Wang L et al., 2021. The geo-economics relationship in the Arctic Region. World Regional Studies, 30(2): 234-244. (in Chinese)

DOI

[26]
Zhu B L, 2016. Interpretation of Canada’s Arctic strategy from the perspective of medium sized countries. Forum of World Economics and Politics, 4: 141-155. (in Chinese)

Options
Outlines

/

〈 〉