We conducted a bibliometric collection on 20 May 2021 using topic search (TS) based on the Science Citation Index Expanded (SCI-E) database in “Web of Science Core Collection.” We selected SCI-E database because it provides comprehensive coverage of the most important and influential research outputs across the world.

We selected the eight main deserts and four so-called sandy lands in China as our search terms (See details Figure 1). These twelve deserts are ordered as followings: the Taklimakan Desert, the Gurbantunggut Desert, the Badain Jaran Desert, the Tengger Desert, the Kumtag Desert, the Qaidam Desert, the Kubuqi Desert, the Ulan Buh Desert, the Mu Us Sandy Land, the Hunshandake Sandy Land, the Horqin Sandy Land and the Hulunbuir Sandy Land. Hence, we retrieved the following outputs in detail: TS = [Taklimakan or Gurbantunggut or “Badain Jaran” or Tengger or Kumtag or (Qaidam and Desert) or (Hobq or Kubuqi) or “Ulan Buh” or (“Mu Us” and Desert) or (Hunshandake or Otindag) or Horqin or ((Hulunbuir or “Hulun Buir”) and (Desert or “Sandy land”))]. Because deserts are not distributed in all parts of some target areas, such as in the Qaidam Basin and the Hulunbuir grassland, we had to constraint our search strings to “desert” or “sandy land.” The document types in our research included articles, books, reviews and letters, but we did not include meeting abstracts, book reviews, and corrections.

The earliest study in the SCI-E database was from 1986, and thus we searched all publications between 1986 and 2020. A total of 2,600 documents were identified, and the three most productive research regions were the Tengger Desert (597 documents), the Taklimakan Desert (590 documents), and the Horqin Sandy Land (499 documents) (Figure 2A). However, less publications were founded for the Kubuqi Desert, the Ulan Buh Desert, the Hulunbuir Sandy Land, and the Kumtag Desert. The annual outputs for the Taklimakan Desert had significant increase after 2006, and the annual outputs for the Tengger Desert and the Horqin Sandy Land increased quickly from 2012 (Figure 2B). The most common publication type was research article, which accounted for 98.5% of the total outputs, and the remaining publications types included reviews (1.33%) and letters (0.1%).

To evaluate the research on deserts in China since 1986, we conducted our bibliometric analysis in the following five components: temporal evolution of annual outputs, main research areas, temporal evolution of keywords, marked journals, dominant institutions, and most productive authors.

Outputs and citations are the two main metrics that represent publication quantity and quality, respectively. Average number citations per output was used to show the average numbers for one paper cited by others in the local database (Zhang and Chen, 2020). To identify the most productive countries, we classified outputs by the nationality of their corresponding authors. International cooperation rate (ICR) for publications was measured as the proportion of papers whose authors come from at least two countries and was used to describe the cooperation patterns of one country with other countries (Bouabid et al., 2016). In addition, a concept map based on correspondence analysis in bibliometrics was used to show the clusters of research topics (Cuccurullo et al., 2016). Marked journals was assigned by the number of publications they carried on the subject based on Bradford’s law for bibliometric analysis (Gong et al., 2019).

We used the bibliometrix R package (Aria and Cuccurullo, 2017) to perform data mining and analysis. This package provides powerful and comprehensive tools for bibliometric analysis and has been widely used (Rodríguez-Soler et al., 2020; Niknejad et al., 2021). We used the ggplot2 R package (Wickham, 2016) to visualize the results.

After the first publication in SCI-E database in 1986, the annual output of research on deserts in China increased exponentially (Figure 3) with an average annual growth rate of 9.45%. Specifically, we divided the annual output into three rising periods: 1) ready-to-grow period: steady with slight rising from 1986 to 2000 with 58 publications totally; 2) steady growth period: publications were risen to 563 during 2001–2010, with an average annual growth rate of 87%; 3) explosive growth period: there were 1,957 publications from 2011 to 2020, with nearly 200 papers published per year. The remarkable growth occurred after 2012, during which time 71.8% of the total outputs were published. This dramatic increase followed the global trend of increasing academic publications in recent years (Mu et al., 2016; Zia, 2021). At the same time, it coincided with the remarkable growth of scientific research in China in the past 10 years. For example, Zhang and Chen (2020) found that almost 90% of the outputs about the Chinese Loess Plateau were published between 2006 and 2018. Dong et al. (2020) concluded that more than 40% of all publications about the Qinghai-Tibetan Plateau were published between 2013 and 2018. Therefore, the most likely explanation for this explosive increase in research output on deserts in China is due to the rapid economic development and increased support (e.g., more funding) from the Chinese government for scientific research (Dong et al., 2020). At the same time, it indicated that more attention have been paid to the environmental and ecological research on desert areas in Northern China. From this trend, it can be predicted that the annual publications of the research on deserts in China will continuous to increases in the next few years.

In contrast to the continuous growth in annual output, the average citations per output showed a fluctuant trend (Figure 3). The similar increasing trend for the average citations per output to the annual outputs were found during 1985–2003. This might be explained that researchers started to pay more attention to scientific research on deserts in this period, leading to a synchronous growth between outputs and citations under the circumstance that a relative small number of publications exists. The remarkable decrease in citations occurred after 2004, whereas annual production of publications increased rapidly during this period. One reason for the declining citation rate despite the increasing publication rate is that researchers had too abundant references to cite. Due to lags in publishing time, later outputs usually had lower citations while older outputs gradually lose their influence as they are replaced by more recent publications (Zhang and Chen, 2020). Therefore, what should be noted is that our analysis showed only the current citations of newly published papers and could not show their peak number of citations. Altogether, because of the continuing growth of outputs in desert research in China, it’s not surprising that the average number of citations per output decreased gradually.

Analysis of research areas can help to reveal the core subjects responsible for increasing productivity in the research field (Zhang and Chen, 2020). By analyzing the contents using Clarivate Analytics in Web of Science, each publication can be assigned to at least one research area. Based on the classification of the Web of Science research areas, the total number of research areas increased from 9 in 2000 to 32 in 2020. The 10 most productive areas were environmental sciences and ecology (accounting for 35.5%), geology (28.8%), agriculture (13.2%), physical geography (12.5%), soil science (12.5%), water resources (10.7%), meteorology and atmospheric science (10.6%), plant science (7.15%), engineering (4.31%), and geochemistry geophysics (3.85%) (Figure 4). This reveals that the research on deserts on China were mainly focused on earth sciences, although they covered a wide range of other research areas such as paleontology, energy fuels, business economics, nutrition dietetics, remote sensing, and archaeology. Some of the research of China’s deserts in the area of energy fuels concerned resource exploitation and petrochemical industry (Wang et al., 2008); business economics area involved the trade of livestock and Chinese herbal medicine (Song et al., 2021); remote sensing area regraded the use of geographic information system, remote sensing and machine learning method to evaluate and regulate the environment in the desert lands in Northern China (Xu et al., 2011; Filonchyk et al., 2018).

Due to limitations of water and nutrients availability and the scare of vegetation, desert lands are considered as one of the most ecologically fragile regions in the world. Hence, environment and ecology is usually an important focus for scientific research in this field (Wang et al., 1999; Li et al., 2018). It was confirmed by our results that the research area of environmental sciences and ecology accounting for 35.5% of the total publications for the research on deserts in China. Ecological restoration and environmental protection have always been important tasks in the regulation of desert areas in Northwestern China (Wang, 2009; Li et al., 2017). Meanwhile, ecological engineering construction in Northwestern China in the past 30 years has also driven a substantial amount of scientific research (Li et al., 2009; Chen et al., 2019). For example, based on the protective system of the Shapotou section of the Baotou-Lanzhou Railway, researchers analyzed the soil water carrying capacity of vegetation and the model for sand fixation by revegetation in aeolian desert areas where precipitation levels are less than 200 mm (Li et al., 2014). In addition, the research of the engineering practices for ecological restoration in the Horqin Sandy Land identified and parameterized various ecological models and techniques for the ecological restoration of desertified grassland (Zuo et al., 2012; Miao et al., 2015). Vice verse, research work in these areas could provide a scientific basis for decision-making by the government to formulate restoration policies, and technical guidance for the ecological restoration industry (Wang et al., 1999).

Frequent keywords analysis can provide vital information about research trends and reveal new research directions (Xie et al., 2008; Sun et al., 2012; Zhang and Chen, 2020). Besides our search terms (e.g., Taklimakan Desert, Tengger Desert, Horqin Sandy Land, Badain Jaran Desert etc.), the most frequent keywords were related to “sand” and “soil,” including “sandy land,” “sand dunes,” “soil properties,” and “soil organic carbon,” with the frequencies of 623, 424, 222 and 206, respectively (Figure 5). At the same time, “water” was another frequent keyword and was associated with “soil moisture,” “soil water,” and “semi arid,” with the frequencies of 481, 473 and 290, respectively. These results indicated that construction projects aimed to stabilize sand dunes and improve soil and water conditions are still a major focus of research on China’s deserts (Su and Zhao, 2003; Li et al., 2007; Zou et al., 2010; Zhang and Zhao, 2015). In addition to these terms, “biological soil crusts” was also a frequent keyword with a frequency of 275. As a key biotic component of desert ecosystems, biological soil crusts can help maintain soil stability, influence soil hydrological processes, fix carbon and nitrogen, and promote the establishment of vascular plants and microorganism communities (Zhang et al., 2007; Li et al., 2010; Colica et al., 2014). The role of biological soil crusts in desert lands has been recognized and highly valued in recent years (Li et al., 2018), and even artificial crusts have been developed and applied to stabilize sandy dunes and in ecological restoration of arid and semi-arid ecosystems (Bu et al., 2017; Zhao et al., 2021).

Compared to the period of 1986–2010, the frequent keywords showed several changes during 2011–2020 (Table 1). The rank of “sand dunes” fell to sixth in period of 2011–2020 from the third during 1986–2010, despite that the frequency increased from 117 to 307. We also found the “dust storm” disappeared from the 15 most frequent keywords list in the period of 2011–2020. The above findings likely reflect that the mitigation of environmental issues resulted by Chinese government’s great efforts such as the Three-North Shelter Forest Program (Wang et al., 1999; Bao et al., 2018; Chen et al., 2019). However, with the global warming intensification, we found that the keyword “climate change” ranked into the top 10 with the frequency up to 242 between 2011 and 2020 (Table 1). The widespread focus on climate change in this research field was confirmed by the concept map of topics shown in Figure 6. The total research outputs can be clustered into three main topics, with the largest cluster focusing on “desert” and “desertification” and the other two clusters focusing on “water” and “climate,” respectively. These results indicated that more attention has been paid to the potential impacts of climate change on desert areas in Northern China (Shi et al., 2007; Dong et al., 2014; Zeng et al., 2016). For example, the climate transition from warm-dry to warm-wet in Northwestern China has been hotly debated in recent years (Yang et al., 2021). Altogether, climate change, combined with desertification, will remain an important research topic relevant to deserts in China.

In addition, it should be noted that the “Tibetan Plateau” became one of the 15 most frequent keywords in the period of 2011–2020 with more than two hundreds frequencies. As the “third pole of the world,” the Tibetan Plateau distributes alpine meadow and other kinds of grasslands, but why here became to a hot area relating to desert research? We explained this by that the land degradation and even desertification has been occurring in the Tibetan Plateau especially the region near the Qaidam Basin because of the grassland conversion to croplands, overgrazing, global warming and so on (Cao et al., 2004; Huang et al., 2020).

The number of journals covering desert research on China has also increased significantly, and more journals were involving in this field. Before 2000, publications were founded in only 45 journals, and the total research outputs have been published in 482 journals now. Based on Bradford’s law of bibliometrics (Pope, 1975), the 18 most productive journals were assigned to the core class (Table 2). These journals published 880 (33.8%) of the total 2,600 papers. The five most productive journals were Journal of Arid Land (109 papers), Chinese Science Bulletin (74 papers), Catena (64 papers), Environmental Earth Sciences (59 papers), and Journal of Arid Environment (56 papers). As the most productive journal, Journal of Arid Land has exhibited the fastest increase since its establishment in 2009 (Figure 7). Meanwhile, Chinese Science Bulletin, as a classical journal in this research field, showed almost no increase after changed the journal title changed to Science Bulletin in 2015. Overall, the publications in these marked journals displayed a similarly dramatic increasing trend with the annual outputs showed in Figure 3.

The top three journals with the highest citations were Journal of Geophysical Research: Atmospheres, Journal of Arid Environments, and Palaeogeography Palaeoclimatology Palaeoecology. Although Journal of Arid Land had the largest number of publications, it did not have the highest number of citations. The most likely reason is that papers in Journal of Arid Land were mainly published only recently (after 2009), and it takes time for newly published papers to be cited. In summary, the scope of the most productive journals covering desert research on China are about earth science in arid land, involving environmental and ecological management, land degradation and restoration process, ecosystem health and sustainable development, and environmental change.

All of the outputs were distributed in 43 countries, and most studies were conducted by Chinese authors (2,198 outputs; 84.6%) (Table 3). The other four most productive countries were Japan (4.73%), the United States (3.19%), Germany (1.39%) and the United Kingdom (1.00%). As the deserts are located in China, it’s not uncommon that the international cooperation rate for Chinese authors was relatively low (ICR = 0.263; Table 3), ranking ninth among the top 10 most productive countries. Despite Chinese authors accounted for the most of outputs, the average number of citations per output for Chinese authors was low (18.02), ranking only eighth among the top 10 countries with the highest number of citations (Table 4). The average number of citations to a certain extent could represent the overall research quality and the global influence for a specific field (Mu et al., 2016; Zhang and Chen, 2020). Hence, these results highlight that Chinese researchers should improve the average number of citations of their outputs to expand their global influence.

According to our analysis, 1,653 institutions worldwide have engaged in research on Chinese deserts, and the top 20 most productive institutions contributed 2,595 papers (including collaborative publications). The Chinese Academy of Sciences (CAS) accounted for the largest number of papers (1,477 papers; 56.8%) (Table 5). Specifically, Northwest Institute of Eco-Environment and Resources (NIEER, CAS), University of Chinese Academy of Sciences (UCAS), Lanzhou University, and Xinjiang Institute of Ecology Geography (XIEG, CAS) had significantly more publications than other institutions. Most of the dominant institutions were located in the Northwestern region of China.

The analysis of authors revealed the researchers who had more outputs and made important contributions. There were 5,373 authors in total, of whom 3,283 (61.1%) had published only one paper. The 10 most productive authors were all from China and contributed a total of 601 papers (23.1%) (Table 6). Li X R, Zhao X Y, and Zhang Y M were the top three most productive authors, and Li X R, Wang X P, and Yang X P were the authors with the highest number of citations. Therefore, Li X R was the author with both the highest number of publications and citations. In addition, we found that the 10 most highly cited papers in this research work were all published before 2010 (Table 7). The citation numbers is usually an important indicator to evaluate the quality of publications (Zhang and Chen, 2020). However, the continuous increase of annul publications in the recent 10 years has resulted in the decline of citations per publication even for the high-quality papers. Therefore, we should take a more rational view to recognize and assess the citation numbers of publications today.