Global Leaf Area Index Research over the Past 75 Years: A Comprehensive Review and Bibliometric Analysis

Abstract

The leaf area index (LAI) is widely used as an important indicator and ecological parameter of vegetation structure and growth status, but the LAI lacks bibliometric analysis. To further understand the LAI’s research status and frontier dynamics, we used 75 years of data (1947–2021) from the Web of Science for scientific bibliometric analysis. The results showed that 22,276 LAI re-search papers were published from 1947 to 2021. According to the characteristics of the literature growth, LAI research can be divided into five stages: incubation, cultivation, acceleration, evolution, and outbreak periods. The research power at the different stages had different characteristics. The overall research power of the United States is number one globally, followed by China, Canada, and France. The related disciplines were widely varied, involving agriculture (the most studied field of LAI research), environmental science and ecology, remote sensing, and other fields. The development of the Google Earth engine, cloud computing platforms, and unmanned aerial vehicle technology will provide more critical support for LAI research. The results of this paper quantitatively show the development history, research hotspots, and application of LAI research and provide a reference for understanding the current situation and development trends of global LAI research.

1. Introduction

Vegetation is an important part of the terrestrial ecosystem, aquatic ecosystem, and the biosphere of the climate system, and it is also an indicator of ecological environment change [1,2]. Vegetation interacts with the atmosphere, water, and other spheres and realizes material, energy, and momentum exchange among the soil–vegetation–atmosphere through biophysical processes, such as photosynthesis and respiration [3,4], which have a significant impact on the development and phenology of the vegetation. As an important organ in the exchange of materials and energy between vegetation and the outside world, leaves play a key role in controlling many biophysical processes in the vegetation canopy. Different leaf growth conditions will cause differences in dry matter accumulation [5].

The leaf area index (LAI) is a quantitative description of vegetation’s leaves. Watson proposed the earliest definition in the field of crop science in 1947 [6]. After continuous improvement, the most widely used definition is half of the total leaf area per unit ground surface area [7]. The LAI is mainly used to find out the change rule of crop yield development, which is often used for crop growth analysis and environmental protection. It is not only an important index for describing the characteristics of vegetation structure but also a key input parameter for many ecosystems productivity, crop growth, and ecological models [4,8,9]. The LAI has been used in various fields as an important parameter for simulating terrestrial ecological processes. For example, it can be used to evaluate deforestation [10], study biological growth cycles [11], and simulate soil moisture [12], and it is often used to monitor crop growth and predict crop yield on a regional scale [13,14,15,16,17]. The LAI also provides important feedback on climate change [4,8,18]. An increase in the LAI will reduce the surface albedo and temperature in snow-free areas and increase canopy evaporation and reduce surface evaporation in tropical areas [19]. An increase in the global LAI leads to an annual increase in land evaporation of 11.4 mm. This is equivalent to more than half of the terrestrial generation that has occurred over the past three decades [20]. Remote sensing LAI data products have significantly improved the simulation of energy absorption, transmission, and interception, and the prediction of ecosystem productivity at seasonal and interannual time scales [21,22,23,24].

There are three common methods for LAI acquisition: direct measurement, indirect measurement, and remote sensing data inversion methods [8]. Direct measurement of the LAI is generally based on the destructive collection of plant leaves and indoor area measurement. This method is traditional and relatively accurate but it is destructive to the plants, time-consuming, and labor-intensive. Indirect measurement can also be used to obtain the LAI by measuring parameters that are closely related to the LAI and are easy to obtain. For example, Gregoire et al. used empirical allometric equations to directly quantify the LAI from individual plant measurements [25]. In addition, Means et al. estimated the leaf area of a single tree stem by measuring its dimensions, including the basal area and diameter [26]. Furthermore, Reinhardt et al. used a similar approach to estimate the canopy fuel characteristics in the northern Rocky Mountain forests [27]. The LAI can be measured using the LAI-2000/2200, tracing the radiation and architecture of canopies (TRAC), and other optical instruments. These measurement methods are simpler and more convenient than direct measurement and save time and labor but the disadvantage is that the measuring instruments are relatively expensive and susceptible to weather conditions. For example, Stenberg et al. evaluated the performance of the LAI-2000 plant canopy analyzer in estimating the leaf area index of Scots pine (Pinus sylvestris) stands [28]. In addition, Peper and McPherson pointed out that the ideal conditions for the indirect measurement of the LAI are cloudy or close to dawn or dusk [29]. Both the direct and indirect measurement methods are traditional. These methods are time-consuming, labor-intensive, complicated to operate, and difficult to carry out for a long period and over a large range, and they have high requirements for time and space. Thus, it is impossible to calculate the inversion and production of the LAI in large regions or globally. With the development of remote sensing technology, it is now possible to obtain vegetation information characteristics at a global scale in all-weather conditions and with continuous observations [30], and the LAI is based on remote sensing inversion, which has important application value for the study of large-scale vegetation changes. Bunnik extracted vegetation coverage and LAI from remote sensing information for the first time, which laid the foundation for LAI research based on remote sensing data [31]. Moreover, Nemani et al. used the difference between the LAI that was derived from the climate–soil–leaf area balance and the actual LAI that was derived from satellite data to estimate changes in the land cover, and they found that in India, China, and Western Europe, where the population density is high, the LAI declines dramatically [32]. Furthermore, Ganguly et al. summarized the implementation method of retrieving the vegetation LAI from Landsat surface reflectance data [33]. Additionally, Dusseux et al. estimated the LAI based on SPOT and Quickbird remote sensing images and used the temporal evolution of the LAI to identify dynamic changes in the grassland [34]. In addition, Gitelson improved the normalized difference vegetation index (NDVI), constructed a new vegetation index to estimate the LAI of wheat, and achieved good results [35]. Finally, Córcoles et al. realized the non-destructive measurement of onion canopy density using drones and established a model to analyze the relationship between canopy density and the LAI [36]. Therefore, the LAI is crucial for the in-depth study of the structural and functional properties of ecosystems and their response mechanisms to climate change. With the development of technologies such as mapping, remote sensing, and sensors, the rapid and accurate acquisition of the LAI has been greatly simplified when compared to traditional measurements. Nevertheless, the need for a fast and reliable LAI and its uses is still an important research topic in the global academic community.

Bibliometric methods originated earlier (1947) [37,38], and the concept of bibliometrics was proposed by the British intelligence scientist Pritchard in 1969 [39]. Bibliometrics is a quantitative analysis method that is based on mathematics and statistics. The basic principle is to construct the internal logic and structure of the discipline by measuring the literature and the relationship among them. This research method has been adopted by many disciplines. Bibliometric analysis can systematically evaluate the research status and internal laws of scientific literature in a certain field and predict the future evolution trend and development direction. It can also help researchers to grasp the research content of their field more directly and reduce the subjective bias of researchers through data visualization analysis so that researchers can discover the potential structure and deep mechanism of the field [40,41]. With the addition of easily accessible online databases with citation data (such as the Web of Science [WOS] and Scopus) and the development of bibliometric analysis software (such as CiteSpace [42], Bibliometrix [43], and BibExcel [44]), bibliometrics has attracted wide attention from relevant scholars. Due to its significant objective, quantitative qualities and modeling of macroscopic research, it has been applied to sociology [45], history [46], epidemiology [41], earth science [47], ecology [48], and many other fields.

Bibliometric methods have been used to assess research trends in several studies (

Table 1. List of studies that used the bibliometric method which was adapted with permission from Copyright 2022, Xu et al. [49].

Reference	Fields
(Zhang et al., 2017, pp. 2010–2015) [50]	Remote sensing
(Zhang and Chen, 2020, pp. 1991–2018) [51]	Chinese Loess Plateau
(Tamiminia et al., 2020) [53]	Google Earth engine
(Li et al., 2021) [52]	Grassland remote sensing
(Zhao et al., 2022) [54]	Earth observation satellite data
(Xu et al., 2022) [49]	NDVI
This paper	LAI

). For example, bibliometric analysis methods have been used to study and analyze the research status and future trends of global remote sensing, the Chinese Loess Plateau, global grassland remote sensing, and the global NDVI [49,50,51,52]. In addition, bibliometric analysis has been used to evaluate the Google Earth engine and Earth observation satellite data applications [53,54]. Although there are many relevant studies on the LAI in the world and there is a trend of continuous growth, scientometric analysis research on the LAI has yet to be reported. When the existing literature was reviewed, it was found that there have been few studies on the long-term macro-evolution and development of the LAI during the whole historical process of research and development, and there is a lack of comparative analysis of LAI research responses and research hotspots. Therefore, quantitative analysis of the relevant literature on LAI research and mining the LAI literature information from a dynamic and time-sharing perspective could assist with the understanding of the research status and frontier dynamics in this field. It could also help researchers to understand the development process of the LAI research and provide new directions and ideas for the future. To the best of our knowledge, this is the first study on bibliometric analysis of LAI literature. Therefore, to better understand the global trends of LAI research we conducted the first bibliometric study on the LAI to provide a reference for relevant researchers to grasp the development of the field and discover valuable new research directions.

In order to achieve the research objectives, the following research questions are addressed [49]:

Q1: What are the global trends in scientific publications about the LAI?

Q2: What information is obtained from these trends?

Q3: What is the future research direction of the LAI?

The specific aims of our research were as follows:

a: To obtain bibliometric information on the 22,276 LAI studies that were extracted from the Web of Science: Science Citation Index (WOS-SCI) expanded database;

b: To use the Bibliometrix R-package and Biblioshiny to preprocess the selected data, including data importing, screening, conversion, and evaluation;

c: To determine quantitative descriptive statistics of the data, analyze the global inter-annual publications of the LAI, and use the total number of citations or h-index to determine the main countries, institutions, and authors of LAI research;

d: To use the keywords to analyze the research history and current research hotspots of the LAI.

2. Data and Methods

2.1. Literature Search Strategy

The data source of the literature retrieval directly determines the validity and accuracy of the bibliometric analysis to a large extent [55]. Web of Science is currently the largest and continuously and dynamically updated journal full-text database. Its core collection has a strict screening mechanism, according to the three laws of bibliometrics, Bradford’s law [56], Lotka’s law [57], and Zipf’s law [58], and it only includes important academic journals and international academic conference papers in various disciplines. The selection process is neutral and unbiased, with comprehensive literature, wide sources, strong authenticity, high accuracy, and reliability [59]. We observed that the abbreviation LAI has other meanings in other disciplines, which can lead to statistical errors in the data collection stage. Therefore, the subject categories of all the papers were classified and extracted, and the spelling combinations that were abbreviated as “LAI” but did not represent the LAI were screened and eliminated. For example, the LAI can also represent “left atrial isomerism”, “liver activity index”, “leukocyte adherence inhibition”, “lateral ankle instability”, and “long-acting injections” in medicine. In physics, there are also expressions related to “Langmuir adsorption isotherm”, “Lai dong model” and “DCR-LAI”. “Lai” also frequently appears as a name, place name, or part of a word. The specific combinations of LAI synonyms that could have affected this study are shown in Table S1. This step was performed to make up for the lack of references on the screening mechanism of the WOS literature retrieval system and because the Porter stem extraction algorithm of the Bibliometrix package is not intelligent enough. This may lead to some literature that contains synonyms but does not belong to the specific research category being extracted as the basic input data of the bibliometrics, thus affecting the accuracy of the research and the reliability of the results.

The WOS-SCI Expanded database was selected as the data source, repeated experiments were conducted according to the research topic, the polysemy of the word was considered, and the search rules were set to filter out the literature that was unrelated to the research topic. The search formula of the final selected advanced method was: TS = “LAI (Topic)” OR “Leaf Area Index (Topic)” NOT “Lai* LaI* LAi* lai* lAI* laI* lAi* Length of Autogamy-immaturity* Liver Activity Index* Leukocyte Adherence Inhibition* Lateral Ankle Instability * LAI-1* Lactobacillus Acidophilus* Long-acting Injectable antipsychotics* Long-Acting Injections* longyou Artificial Island* Leaf Area Infected* lefresne azofunctional Index* Local Allergic Inflammation* Location Area Identity* Left Atrial Isomerism* Langmuir Adsorption Isotherm* Eprinomectin LAI* DCR-LAI (All Fields)”. The search results returned 31,535 articles in the WOS, which was updated on October 19, 2022. All of the records were exported as “plain text files” with “full record and cited references” and each record contained the author, title, source document, abstract, and cited references.

2.2. Bibliometric Analysis Methods

Referring to the research of relevant scholars, this study proposed a six-step bibliometric method including research objective determination, research design, data collection, data analysis, data visualization, and interpretation (Figure 1). The first step was to clarify the research objectives. The main purpose of this work was to quantitatively show the development process, research hotspots, and applications of LAI research to understand the status and development trend of global LAI research and to provide a reference for the development and future direction of related disciplines. The second step was the research design phase, where the LAI was selected as the research topic, and then the research question was determined and the appropriate bibliometric method was selected. The third step was the collection of the bibliometric research data source from the WOS-SCI Expanded database, and, after determining the search criteria, the literature search returned 31,535 articles. Following recommendations [60] and adhering to internal validity, we only considered documents that met the criteria for a complete research protocol. Specifically, each document needed to contain important, essential elements such as the author’s name, title, keywords, and other information, all of which constituted the document’s directory attributes, also known as the metadata. Since the current Bibliometrix word segmentation algorithm was not intelligent enough and the extraction of some of the keywords lacked accuracy, to ensure the reliability of the conclusion, in the process of manual screening and evaluation, we eliminated the literature that did not involve the LAI by reading the literature manually. After the manual screening, 22,276 papers that were published between 1947 and 2021 and were highly related to the research topic of this paper were finally screened from the WOS-SCI Expanded database, and all the records were imported into the Biblioshiny web program and converted to Bibliometrix R data. In the fourth step, we performed a descriptive analysis of the bibliography data frame using R software [43] to perform descriptive econometric statistical analysis on all the selected documents and to create a matrix containing all the documents. The main results of the bibliometric analysis were summarized using the generic function “summary” in R. The results included the main information of the bibliographic data frame and several other tables, such as the annual scientific production, total citation per country, top manuscripts per number of citations, most productive authors, most relevant sources (journals), most productive countries, and most relevant keywords. The fifth step was to select the appropriate visualization method, and we used Biblioshiny, tidyverse (ggplot2), and VOSviewer in R software to create concept maps [49], co-citation networks, tree maps, dumbbell maps, and other graphs. Finally, we explained and discussed these findings based on the data analysis and visualization results.

3. Results and Discussion

3.1. Descriptive Bibliometric Analysis

We performed general descriptive statistical analyses using Biblioshiny in R.

Table 2. Basic information on the leaf-area-index-related literature that was identified by the bibliometric analysis.

Type	Description	Value/Number
Time span	Time of publication	1947–2021
Documents	Number of documents	22,276
Authors	Number of authors	51,324
Sources	The frequency distribution of sources as journals, books, etc.	1502
References	Number of references	479,572
Keywords plus (ID)	Number of phrases that frequently appear in the title of an article’s references	22,542
Authors’ keywords (DE)	Number of authors’ keywords	35,612
Authors’ appearances	The authors’ frequency distribution	108,514
Documents per author	Average number of authors in each document	0.434
Co-authors per document	Average number of co-authors in each document	4.87
Authors of single-authored documents	Number of single authors per articles	785
Average citations per documents	Average number of citations in each document	35.50
Annual growth rate	Average annual growth rate of documents	9.15

reflects the key information of the 22,276 LAI-related articles that were published between 1947 and 2021. In the past 75 years, an average of 297 LAI research papers have been published each year, with an average of 35.50 citations per paper. These articles involved 51,324 authors, of which 785 were single-author articles. A total of 35,612 keywords were generated, and these results demonstrated the rapid and high-quality development of LAI research.

3.2. Published Volume Analysis

3.2.1. Overall Development Trend of the Leaf Area Index

(1) Trends in the paper outputs

The number of published papers is the overall representation of the scientific community’s attention to a certain field. According to the growth and aging laws of the literature, annual statistical analysis of the number of published papers can reveal the current research status in this field and predict its research prospects and development trends [61]. The line graph in Figure 2 shows the change trend of the annual publication volume of LAI research, which directly reflects the development law of the number of LAI publications. The number of papers on LAI research increased year on year, indicating that LAI-related research has been continuously focused on and valued by scholars. Since the publication of the first publication on the LAI in 1947, the number of publications in a single year reached 1954 in 2021, with an annual growth rate of 9.15% over the 75 years. Overall, global LAI research began to increase in the middle of the 20th century. After a relatively long period of development, it advanced rapidly in the late 20th century, and the number of papers that were published increased year on year until the 21st century, when it reached a very active state, showing an exponential trend.

(2) Division of the paper output stages

The concept of a “life cycle” originates from biology and mainly describes the development process of “emergence–development–maturity/stable-decay” [62]. The evolution of disciplines also has a life cycle, and this concept has been widely used in many fields, such as politics [63], economy [64], society [65], environment [66], technology [67], and engineering [68], among others [69]. The aging mechanism and life cycle of the subject area have important research significance. Specifically, the life cycle of the discipline is a result of the objective law of scientific development and the subjective needs of information users. The knowledge and content that is studied in the discipline will inevitably experience value realization and change in the process of dissemination, resulting in regular changes in its value and utilization. Drawing on the life cycle theory and the distribution and growth trend of the LAI literature, this study divided the research and development stages of the LAI into five stages: incubation, cultivation, development, acceleration, and outbreak (Figure 2).

Incubation period (1947–1990): The curve of the number of papers over time shows that during most of the years in this stage, fewer than five papers that were related to the LAI were published, and, thus, the number of papers that were published was very small. In 1982, even though it was 35 years after the introduction of the LAI in 1947, the number of papers that were published per year was only more than 10. During this period, no papers were published in 1964, and only one paper was published in 5 years, indicating that the research on the LAI at this stage was in its infancy, so these 43 years were regarded as the incubation period of LAI research.

Cultivation period (1991–1999): In 1991, the LAI paper output increased sharply. The curve of the number of papers over time showed an obvious inflection point, and the growth curve became steeper. The number of papers increased from 47 in 1990 to 218 in 1991. The number of papers in 1991 was 4.64 times that of the previous year and the number of papers in 1991 was the sum of the number of papers in the previous 30 years. Additionally, the number of papers in the following years also showed a rapid growth trend. The 9-year average annual publication volume in this stage was about 270.56, so this stage was regarded as the cultivation period of LAI research.

Development period (2000–2005): During the 6 years of this stage, the trend of an increase in the number of published papers each year did not diminish. The annual number of papers increased steadily, ranging between 300 and 400 but it did not exceed 500, and the average annual number of papers was about 413.17. Therefore, this stage was regarded as the development period of LAI research.

Acceleration period (2006–2015): The number of papers that were related to the LAI began to show a significant increase in these 10 years. Although there were slight fluctuations during the period, the overall publication volume continued to maintain a rapid growth trend, with an average annual publication volume of 797.50 papers. Therefore, this period was regarded as the acceleration period of LAI research.

Outbreak period (2016–2021): In these 6 years, LAI research developed unprecedentedly. During this period, the number of published papers that were related to the LAI continued to increase. The number of papers that were published each year exceeded 1000, the paper output in 2021 was as high as 1954, the total number of publications was 10,278, and the average annual number was 1521, accounting for 46.14% of the 75-year cycle from 1947 to 2021. Therefore, this period was considered to be the outbreak period of LAI research.

3.2.2. Country and Institutional Distribution

(1) Country distribution

The statistics of the number of papers that were produced by different countries in each stage can more intuitively reflect the changes in the LAI research paper outputs and the country distribution of the scientific research strength in the different periods. Thus far, 123 countries have carried out LAI research, and China (4226) and the United States (4098) have published more than 4000 papers, ranking 1st and 2nd in terms of the largest number of scientific achievements, respectively, and producing far more papers than the other countries. Brazil and India, which ranked 3rd and 4th, had more than 1000 articles, producing 1178 and 1125 papers, respectively. The country with the 5th highest total number of published papers was Canada with 957 papers. The total number of published papers in these five countries accounted for 27.55, 26.72, 7.68, 7.33, and 4.30% of the total number of published papers on LAI research, respectively (Figure 3).

Among them, both China and the United States have published more than 4000 articles, far exceeding the number of papers that were published by the other countries, and they were ranked as the first echelon. Although the research on the LAI in China started relatively late, the growth in the number of published articles has been rapid. The United States started earlier and has the strongest growth, and the number of publications did not differ much from that of China. Then, Brazil and India both published more than 1000 articles and were in the second echelon, whereas Canada, Germany, France, Australia, Japan, and Italy all had more than 600 articles and were in the third echelon. Of the 12 countries with more than 500 publications, five were in Europe, three were in Asia, two were in North America, and Brazil was the only country from South America.

In addition to the number of scientific and technological achievements, the international cooperation network map can be used to measure the scientific research strength of a country (Figure 4). Among them, the United States (94), Germany (86), the United Kingdom (84), France (82), China (78), and Australia (78) had the closest cooperation, and the United States was included in a relatively high proportion of them and it was the most active country in this research. Moreover, Spain (77), the Netherlands (75), Belgium (72), Italy (71), and Canada (71) also developed rapidly in terms of international exchanges and cooperation with other countries. The other countries had less cooperation in LAI research, with fewer than 70 connections, indicating that their research had less external cooperation and exchanges and was mostly completed by local scholars. In addition, it is worth noting that economic powers, such as South Africa (54) and Russia (48), did not have a high degree of international cooperation in their LAI research.

Figure 5 shows the top 10 countries in terms of the total and average citations for LAI research papers from 1947 to 2021. For the total citation frequency, the United States had the highest total citation number of all the countries (254,936). The total citation of its articles was about 2.5 times that of the second most citations, and the United States was followed by China (93,957), Canada (48,047), France, (44,742), Australia (36,059), Germany (29,282), Spain (28,711), the United Kingdom (26,830), Italy (22,064), and the Netherlands (18,905). In terms of the average article citations, the United States had the highest average number of citations (62.21), which was followed by France (56.07), Canada (50.21), the Netherlands (49.88), Australia (47.20), Spain (47.14), the United Kingdom (46.66), Germany (36.60), Italy (32.35), and China (22.23). The average number of citations in the United States, France, and Canada was higher than 50, indicating that the scientific research results that were produced by these three countries had great influence and a high number of reads and citation value. China ranked 2nd in terms of the total number of citations, indicating that China has academic influence in this field but due to a large number of papers and because many are of low quality, the average number of citations ranked last among the 10 countries. Thus, there was still a certain gap when compared to the advanced level of the foreign countries, which needs to be further studied. Articles with high citations are generally leading or innovative, and the research content and conclusions are often worthy of the attention of scholars in related fields. Therefore, the United States shows a leading position in LAI-related research. The production of scientific research results is inseparable from the joint efforts of scientists around the world. Under the trend of globalization, cooperation between countries for scientific research has played a positive role. Thus, countries that participate in research in LAI-related fields also generally have foreign cooperation.

(2) Institutional distribution

There were 7325 institutions worldwide that conducted LAI research.

Table 3. Top 20 institutions according to the total number of citations for leaf-area-index-related research.

Institution	Country	TC	TA
University of Arizona	USA	12,903	64
University of Wisconsin	USA	11,780	101
GSFC	USA	9167	90
Boston University	USA	8074	58
University of Nebraska	USA	7653	121
Oregon State University	USA	7579	82
University of California, Berkeley	USA	7560	69
University of Maryland	USA	7162	91
University of Montana	USA	7037	44
NCAR	USA	6128	24
CCRS	CANADA	6041	49
University of Toronto	CANADA	5489	92
University of British Columbia	CANADA	5423	65
Beijing Normal University	CHINA	5302	180
IGSNRR, CAS	CHINA	5197	150
Universitat de València	SPAIN	4777	70
University of New Hampshire	USA	4614	33
China Agricultural University	CHINA	4449	152
University of Florida	USA	4304	116
Nanjing Agricultural University	CHINA	4157	106

Abbreviations: TA, total number of articles; TC, total number of citations; GSFC, Goddard Space Flight Center; NCAR, National Center for Atmospheric Research; CCRS, Canada Centre for Remote Sensing; IGSNRR, CAS, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences.

shows the top 20 institutions in the world in terms of the total citations for LAI research and the influence of each institution, which was evaluated according to the number of citations for papers that were published by the research institution. The top 10 institutions with the highest citation frequency were all located in the United States, which corresponds with the United States having the most total and average citations for LAI research papers, reflecting the strong research and scientific influence of the United States in this research field. In particular, the University of Arizona (12,903), with the highest total citations, and the second-ranked University of Wisconsin (11,780) had the most representative research work and led the development of LAI research. The institutions that were ranked 11th–13th were the Canada Centre for Remote Sensing (6041), University of Toronto (5489), and University of British Columbia (5489) in Canada, which echoes the top ranking of Canadian LAI research papers in terms of average citations, indicating that the scientific research results that were published by these research institutions in LAI research had a high degree of international recognition. There were four institutions from China, they were Beijing Normal University (5302), the Institute of Geographic Sciences and Natural Resources Research (5197), China Agricultural University (4449), and Nanjing Agricultural University (4157); because of their large number of citations, they had an important impact on LAI research. Many research institutions from China have published many papers on LAI research and have developed rapidly but the average citation frequency was not high, and the quality of the research and development has much room for improvement. Notably, although the number of papers that were published by the University of Valencia (4777) in Spain was not very large, the citation frequency was relatively high. Therefore, the scientific research achievements and outstanding contributions of this institution toward LAI-related research are worthy of recognition.

3.2.3. Disciplines and Periodical Distribution

(1) Disciplines Distribution

The subject classification system of the WOS database is more refined and scientific than many of the other subject classification systems, and it covers 252 subjects. Each journal and book that is included in the WOS core collection belongs to at least one subject category and can be assigned to up to six subject categories. The subject of an article can be used to roughly understand its research direction [52]. In this study, a total of 22,276 publications on LAI research involved 46 independent subject categories and 35 pairs of interdisciplinary subject categories, accounting for 32.12% of the total categories. The research areas that were covered by the literature increased from 1 in 1947 to 52 in 2021 (Figure 6).

Table 4. Top 10 disciplines in terms of literature numbers from 1947 to 2021.

Rank	WOS Subject Category	Number of Articles	Proportion /%
1	Agriculture	7541	33.85
2	Environmental sciences and ecology	6378	28.63
3	Remote sensing	3912	17.56
4	Imaging science and photographic technology	3402	15.27
5	Plant sciences	3003	13.48
6	Forestry	2654	11.91
7	Geology	2544	11.42
8	Meteorology and atmospheric sciences	2038	9.15
9	Water resources	1457	6.54
10	Engineering	1229	5.52

lists the statistics of the top 10 disciplines. The top 10 research fields included 20,917 LAI-related research papers, accounting for 93.90% of the total number of papers. The research disciplines of the LAI were widely distributed, and the discipline with the highest number of publications was agriculture, which published 7541 papers from 1947 to 2021, accounting for 33.85% of the total number of papers. The 2nd highest was environmental science and ecology, with a total of 6378 papers, accounting for 28.63% of the total number of papers. The fields of the number of papers that were ranked 3rd to 5th were relatively similar, and they were remote sensing (3917), imaging science and photographic technology (3407), and plant sciences (3000), accounting for 17.56, 15.27, and 13.48% of the total published volume, respectively. Thus, the research content of the LAI was mainly based on agricultural disciplines, focusing on ecology, remote sensing, and botany in the natural sciences. There were relatively few studies in the field of social sciences, but it is developing towards a multidisciplinary structure and diversification.

Figure 7 shows the changes in the focus areas of LAI research. Before 1980, the main research fields were agriculture and botany. By 1990, research on LAI in engineering, meteorology, and atmospheric science and forest science gradually developed. Then, forestry, ecology, and natural science, and remote sensing showed explosive growth and became the leading fields in terms of LAI literature output. Since 2000, the relevant disciplines have maintained a rapid growth momentum year on year. The convening of the United Nations Millennium Summit and the establishment of the eight Millennium Development Goals have resulted in researchers paying more and more attention to environmental and ecological changes [71], which explains the publications in this field. According to the total number of citations in each research field, imaging science and photography technology, plant science, forestry, and meteorology and atmospheric science had more citations. Some remote sensing indicators related to the LAI, such as the NDVI [72,73,74], red-edge chlorophyll vegetation index [75,76], harvest index [77,78], enhanced vegetation index [79,80], and soil adjusted vegetation index [81,82], were also widely cited.

(2) Periodical distribution

Academic journals are important tools for academic communities to engage in scientific research activities. The statistics and analysis of the source publications of LAI research papers can determine the main core journals with strong influence in this field. In terms of the subject categories of the journals, it is helpful for researchers to select key journals according to their research directions for literature review and submission. Therefore, this paper used the h-index and Journal Citation Reports (JCR) impact factor to characterize the level of the journals. The h-index combines the rankings of the total number of publications and the total number of citations, which can be more effective at characterizing the quality and level of the journals [83]. The JCR impact factor is an indicator that is recognized by the academic community to characterize the academic level of journals [84]. The literature on LAI research from 1947 to 2021 came from 1436 journals. In this study, the top 10 journals were counted with regard to the number of publications, and they were ranked from high to low according to the h-index. The results are shown in

Table 5. Top ten journals ranked by the number of local citations in leaf-area-index-related research.

Sources	H Index	N. LC	ND	IF	Country of Origin
Remote Sensing of Environment	238	79,324	992	13.85	USA
Global Change Biology	217	15,267	216	13.21	UK
IEEE Transactions on Geoscience and Remote Sensing	216	15,515	213	8.125	USA
Forest Ecology and Management	152	12,920	383	4.384	The Netherlands
International Journal of Remote Sensing	151	21,247	550	3.531	UK
Agricultural and Forest Meteorology	144	35,725	843	6.424	The Netherlands
Field Crops Research	127	14,438	494	6.145	The Netherlands
Agronomy Journal	113	14,532	350	2.65	USA
Journal of Geophysical Research: Atmospheres	106	13,428	224	5.217	USA
Remote Sensing	81	13,804	1039	5.349	Switzerland

Abbreviations: N. LC, number of the total local citation; IF, impact factor in 2021.

. The average impact factor for the 10 journals was 6.89, with a maximum of 13.85. According to Bradford’s law, the distribution of the LAI research papers in relevant journal information sources was not balanced, and the source journals were highly dispersed. The ten most influential journals were in the core area (Zone 1). These journals published relatively more high-level literature on LAI research and had a strong influence in related fields. Among the top five journals with the most published papers, the Remote Sensing of Environment’s h-Index, number of scientific publications, and impact factor (IF) all ranked 1st, which was much higher than those of the other journals, indicating that the articles that were included in this journal had a high degree of attention in the LAI-related research field. Researchers pay close attention to original and rigorous research results of universal significance. For the h-index, Global Change Biology ranked 2nd, and IEEE Transactions on Geoscience and Remote Sensing ranked 3rd. These two journals had a relatively high publication volume, and their IF was also at the forefront. Therefore, with increasingly severe global climate change and ecological environment changes, the vigorous development of various remote sensing platforms, and the continuous expansion of application fields, the attention of scholars and authoritative professional journals to the LAI is rapidly increasing (Figure 8).

3.3. Author Analysis

The analysis of the main influential authors helps in understanding the main researchers in the field and the direction of their attention, which could help researchers to conduct academic exchanges and cooperation according to their own scientific research needs. We ranked 51,324 authors, with 22,276 papers from high to low according to the h-index [55]. A total of 785 independent authors published 951 single-authored papers, and the average co-authorship per paper was 4.87. In addition, the international collaboration authorships amounted to 30.22%, and each author contributed an average of 0.434 papers. These results also show that LAI researchers usually form research teams and academic groups to conduct collaborative research in the field. The relevant information of the top authors was selected (

Table 6. Top ten most influential authors ranked by their h-index.

Author	H Index	G Index	M Index	TC	NP	PY_start	Country
Chen J.M.	59	118	1.844	14,645	179	1991	Chian
Baret F.	56	105	1.75	11,892	105	1991	France
Myneni R.B.	56	100	1.806	14,679	100	1992	USA
Running S.W.	47	63	1.306	13,004	63	1987	USA
Black T.A.	46	82	1.438	7619	82	1991	Canada
Gower S.T.	45	54	1.364	9450	54	1990	USA
Weiss M.	40	59	1.667	7538	59	1999	France
Knyazikhin Y.	39	60	1.5	7077	60	1997	USA
Coops N.C.	36	64	1.636	4186	68	2001	Canada
Ciais P.	35	61	1.842	8375	61	2004	France

Abbreviations: TC: Web of Science Core Collection times cited count; NP: number of scientific productions; PY_start: first year published.

): Chen J.M. (59), Baret F. (56), Myneni R.B. (56), Running S.W. (47), Black T.A. (46), Gower S.T. (45), Weiss M. (40), Knyazikhin Y. (39), Coops N.C. (36), and Ciais P. (35). In the WOS database, Chen J.M. had the highest total citations in the field of LAI research, which was up to 14,645 times, and his h-, g-, and m-indexes also ranked 1st. The h-indexes of Baret F. from France and Myneni R.B. from the United States were both 56, and their total citations were 11,892 and 14,679, respectively. Of the 10 authors in the table, 4 were from the United States, 3 were from France, 2 were from Canada, and 1 was from China.

3.4. Paper Influence Analysis

The number of citations of the papers represents the influence of the papers, reflects the acceptance of their academic achievements by their peers, and is an important indicator of the academic influence when evaluated by other researchers. The highly cited literature can be used to gain insights into influential authors and important research topics in a certain field within a certain period. After automatic analysis and calculation using the Bibliometrix platform, the local citation score (LCS, number of citations within the field) and the global citation score (GCS, total number of citations in the WOS) of the top 10 papers were obtained (

Table 7. Top ten papers according to their local citation score.

Title	Senior Author	Journal	Year	LC	GC
Overview of the radiometric and biophysical performance of the MODIS vegetation indices [85]	Huete A.	Remote Sensing of Environment	2002	794	5136
Global products of vegetation leaf area and fraction absorbed PAR from year one of MODIS data [86]	Myneni R. B	Remote Sensing of Environment	2002	711	1359
Review of methods for in situ leaf area index determination: Part I. Theories, sensors and hemispherical photography [87]	Jonckheere I.	Agricultural and Forest Meteorology	2004	708	1021
Hyperspectral vegetation indices and novel algorithms for predicting green LAI of crop canopies: Modeling and validation in the context of precision agriculture [88]	Haboudane D.	Remote Sensing of Environment	2004	631	1448
Defining leaf area index for non-flat leaves [7]	Chen J.M	Plant, Cell & Environment	1992	620	863
Derivation of Leaf-Area Index from Quality of Light on the Forest Floor [89]	Jordan C. F	Ecology	1969	610	1415
Ground-based measurements of leaf area index: a review of methods, instruments and current controversies [90]	Breda N.J. J	Journal of Experimental Botany	2003	596	895
Direct and Indirect Estimation of Leaf Area Index, fAPAR, and Net Primary Production of Terrestrial Ecosystems [91]	Gower S. T	Remote Sensing of Environment	1999	546	837
Estimating Corn Leaf Chlorophyll Concentration from Leaf and Canopy Reflectance [92]	Daughtry C.S. T	Remote Sensing of Environment	2000	494	1335
Leaf area index of boreal forests: Theory, techniques, and measurements [5]	Chen J.M	Journal of Geophysical Research	1997	492	673

and

Table 8. Top ten papers according to the global citation score.

Title	Senior Author	Journal	Year	LC	GC
Overview of the radiometric and biophysical performance of the MODIS vegetation indices [85]	Huete A.	Remote Sensing of Environment	2002	794	5136
Estimates of global terrestrial isoprene emissions using MEGAN (Model of Emissions of Gases and Aerosols from Nature) [91]	Guenther A.	Atmospheric Chemistry and Physics	2006	49	2684
FLUXNET: A New Tool to Study the Temporal and Spatial Variability of Ecosystem-Scale Carbon Dioxide, Water Vapor, and Energy Flux Densities [93]	Baldocchi D.	Bulletin of the American Meteorological Society	2001	0	2594
A global analysis of root distributions for terrestrial biomes [94]	Jackson R. B	Oecologia	1996	94	1897
On the relation between NDVI, fractional vegetation cover, and leaf area index [95]	Carlson T. N	Remote Sensing of Environment	1997	334	1754
Sentinel-2: ESA’s Optical High-Resolution Mission for GMES Operational Services [96]	Drusch M.	Remote Sensing of Environment	2012	163	1693
Hyperspectral vegetation indices and novel algorithms for predicting green LAI of crop canopies: Modeling and validation in the context of precision agriculture [88]	Haboudane D.	Remote Sensing of Environment	2004	631	1448
Derivation of Leaf-Area Index from Quality of Light on the Forest Floor [89]	Jordan C. F	Ecology	1969	610	1415
Global products of vegetation leaf area and fraction absorbed PAR from year one of MODIS data [86]	Myneni R. B	Remote Sensing of Environment	2002	711	1359
Estimating Corn Leaf Chlorophyll Concentration from Leaf and Canopy Reflectance [92]	Daughtry C.S. T	Remote Sensing of Environment	2000	494	1335

Abbreviations: LC: local citations; GC: global citations.

).

According to the LCS and GCS scores, the most influential paper was an overview of the radiation and biophysical properties of the moderate-resolution imaging spectroradiometer (MODIS) vegetation indices. The results of this paper showed that there was a good correspondence between the reflectance and vegetation index values of the canopy top that were measured in the air and the MODIS sensor values in the four concentrated test sites of the semi-arid grass/shrub, savanna, and tropical forest biomes. By comparing the time profiles of the MODIS-NDVI, NOAA-14, and 1 km AVHRR-NDVI, it was found that the MODIS-based index had higher fidelity. This paper also presented the dynamic range of the MODIS VIs and evaluated their sensitivity in discriminating vegetation differences in sparse and dense vegetation regions [85]. The 2nd most influential paper according to the LCS (ranked 9th for the GCS) was “Global products of vegetation leaf area and fraction absorbed photosynthetically active radiation from year one of MODIS data”. This study developed and implemented a collaborative algorithm based on the retrieval of the fraction of photosynthetically active radiation that was absorbed by the vegetation green LAI and surface reflectance from the canopy reflectance data that were measured using MODIS and a multi-angle imaging spectroradiometer (MISR), and the performance of the algorithm was evaluated, and the product results were validated using field data [86]. The 3rd most influential paper according to the LCS was not in the top 10 according to the GCS and it was a review paper on the methods for in situ LAI determination. It focused on the current measurement methods for the plant leaf area, reviewed different types of leaf area measurement methods, and covered the research progress of digital image processing technology in measuring plant leaf area. On this basis, the advantages and disadvantages of the leaf area measurement methods and future development trends were discussed [87]. The 4th most influential paper according to the LCS (ranked 7th for the GCS) proposed a new hyperspectral vegetation index algorithm for predicting the green LAI of the crop canopy. The focus of this study was to reduce the variability of the LAI estimates due to changes in the leaf chlorophyll concentration so that the LAI of the crop canopy could be determined simply and accurately for agricultural management [88]. The 5th most influential paper according to the LCS did not appear in the top ten list according to the GCS. It was a research article that defined the LAI of non-flat leaves. This study investigated the projection coefficients of several objects including spheres, cylinders, half cylinders, triangles, and square rods to eliminate the leaf area index of non-flat leaves by mathematical derivation and numerical calculations This paper suggested that the LAI of the non-flat leaves should be defined as half of the total cross-sectional area per unit ground surface area, which has an important influence and far-reaching significance on the related research on leaf area [7]. The 6th most influential paper according to the LCS (ranked 8th for the GCS) determined and measured the LAI of forests using the ratio of 800 to 675 μm light on the forest floor based on the principle that leaves absorb more red than infrared light, and it concluded that the more leaves that exist in the canopy, the greater the ratio of light [89]. The 7th and 8th most influential papers according to the LCS also did not appear in the top ten list according to the GCS. The 7th most influential paper according to the LCS was a review article. It was a systematic overview and summary of the direct and indirect methods of ground measurements of the LAI, the required instruments, their advantages and disadvantages, and the accuracy of the results. This paper comprehensively discussed the two main causes of the differences, namely the aggregation and contribution of stems and branches, and proposed some recent theoretical or technical solutions. It pointed out that the accuracy, sampling strategy, and spatial validity of LAI measurements must be evaluated to ensure the quality of all LAI-dependent canopy ecophysiological and biophysical processes for measurement and modeling purposes [90]. The 8th most influential paper according to the LCS used direct and indirect methods to estimate the LAI, fraction of absorbed photosynthetically active radiation, and net primary production of terrestrial ecosystems. The study summarized the carbon allocation patterns of the major terrestrial biomes, discussed emerging allocation patterns that could be incorporated into the global net primary production model, and provided light use efficiency coefficients for the major biomes, noting the need to address inconsistencies in the radiation, dry matter, and carbon allocation units [91]. The 9th most influential paper according to the LCS (ranked 10th for the GCS) estimated the chlorophyll concentration of maize leaves by the reflectance of the leaves and canopy, selected the wavelength that was sensitive to the chlorophyll concentration, used the radiative transfer model to simulate the canopy reflectance, and proposed a strategy for detecting crop nitrogen status using remote sensing data [92]. Furthermore, the tenth most influential paper according to the LCS did not appear in the top ten list according to the GCS. This study supported the International Boreal Ecosystem–Atmosphere Study through the LAI values that were obtained by several research teams and by using different methods for a broad range of boreal forest types. These methods included destructive sampling and optical instruments: the TRAC instrument, the LAI-2000 plant canopy analyzer, hemispheric photography, and the sunfleck rangefinder. The results showed that these instruments underestimated the LAI of the leafy boreal forests, and they demonstrated that optical techniques combined with limited direct leaf sampling methods can be used to quickly obtain accurate LAI measurements [5].

None of the four papers ranked 3rd to 6th in the GCS list appeared in the LCS top ten list. The third-ranked paper on the GCS list described the global network of micrometeorological flux measurement sites, “FLUXNET”. It currently has more than 140 stations running on a long-term and continuous basis to measure the exchange of carbon dioxide, water vapor, and energy between the biosphere and the atmosphere [93]. The development and application of this new tool provide new avenues and options for LAI studies. The fourth-ranked paper on the GCS list analyzed the rooting patterns of global terrestrial biomes, compared the distribution of the various plant functional groups, and discussed the advantages and possible disadvantages of the study in the context of root biomass and root function [94]. The fifth-ranked paper on the GCS list discussed the relationship between the NDVI, vegetation cover, and LAI using correlations between simple radiative transfer models that included vegetation, soil, and atmospheric components. The study mentioned that depending on how the LAI is defined, the LAI and partial vegetation cover may not be completely independent quantities, so care should be taken when using the LAI and partial vegetation cover independently in a model because LAI may have a partially sensitive dependence on the NDVI [95]. The sixth-ranked paper on the GCS list was the introduction of Sentinel-2 satellite remote sensing data. The availability of Sentinel-2 data provides a more refined 10 m spatial resolution and 10- or 5-day revisit cycle time series images, which is an unprecedented opportunity for LAI research. Sentinel-2 data have quickly become the main data source for LAI research [96]. In general, the top ten most influential articles in the LCS were related research on the vegetation index, including LAI review articles, measurement algorithms, estimation models of the LAI, and remote sensing methods to determine the LAI. From the list of the top ten most influential articles in the GCS, the articles that ranked 3rd, 4th, and 6th in the list introduced new approaches and new tools that were related to LAI research, respectively. In addition, the other articles in the GCS list focused on LAI-related topics.

3.5. Keyword Analysis

Keywords can usually reflect the research theme of the literature and the core content of the research results. The statistics and analysis of keywords can be used to accurately grasp the research hotspots and future research directions in a certain research field, and they also provide further development directions for future scientific research. A total of 35,612 author keywords were detected in 22,276 papers on the LAI that were published between 1947 and 2021. Figure 9 is a tree diagram of the top 20 keywords. To more accurately display the themes and research contents of the papers that received the most attention in the different stages of LAI research, this study analyzed and interpreted the research topics by combining the temporal changes in the high-frequency words and highly cited papers in each stage (

Table 9. High-frequency words in the different stages and the characteristics of their distribution.

Stage	High Frequency Words (Partial)	Characteristics of Distribution
Incubation period (1947–1990)	Leaf area, leaf area index, growth, yield, harvest index, canopy, biomass, evaporation, grain yield, and soil	The high-frequency hot topic words are more dispersed and have not formed obvious rules. Harvest index, biomass, evaporation and so on have developed earlier and have strong research continuity;
Cultivation period (1991–1999)	Yield, photosynthesis, carbon dioxide, seedling, water stress, climate change, hemispherical photography, net primary production, vegetation canopy, and forest ecosystem	Compared with the previous period, the research topics have both continuations (such as yield, vegetation canopy) and new growth (such as photosynthesis, water stress, climate change and net primary production, etc.), and climate change has attracted attention;
Acceleration period (2000–2005)	Temperature, nitrogen, water balance, variability, classification, spectral reflectance, long term, precipitation, nutrition, and deforestation	The high-frequency hot topics are relatively concentrated, and the popularity growth, change, and continuity are more significant. Classification, spectral reflectance, and precipitation nutrition suddenly become new hot topics at the end of this stage.
Evolution period (2006–2015)	Growth, retrieval, remote sensing data, time series, regression, remote estimation, global product, identification, mechanism, and MODIS LAI	The high-frequency hot topics began to form a heat stability and growth law. The research heat of remote sensing data, remote estimation, global product, MODIS LAI, and other remote sensing methods and products continued to grow. Time series, regression, identification and other research topics and methods maintained a high degree of attention
Outbreak period (2016–2021)	Climate change, nitrogen use efficiency, leaf chlorophyll content, machine learning, induced chlorophyll fluorescence, UAV, Sentinel-2, vegetation mapping, health, and Gaussian processes	The research hotspots are more abundant, and the research on different high-frequency topics is more in-depth, lasting, and stable. The overall heat of climate change, machine learning, UAV and Sentinel-2 as research hotspots continues to rise. Themes such as use efficiency, leaf chlorophyll content, induced chlorophyll fluorescence, vegetation mapping and gaussian nitrogen processes remained steady and increased.

). The high-frequency words at each stage and their temporal distribution characteristics are shown in Figure 9.

Figure 10 is a dumbbell plot of the author keywords over time, where the x-axis represents the year, and the y-axis represents the keywords. The top ten keywords by frequency were the LAI, remote sensing, LAI, evapotranspiration, yield, photosynthesis, Modis, biomass, NDVI, and climate change. Of these keywords, “LAI” appeared the most frequently. “Remote Sensing” is one of the most important research fields for the modern LAI, and the study of the LAI based on remote sensing methods has become a hot topic. There were 1421 papers from 1989 to 2021, and “MODIS” had the 7th highest frequency and was the most widely used sensor in LAI-related research [97,98,99,100], with 596 papers from 1997 to 2021. “Evapotranspiration” [101,102,103] includes surface water evaporation and water transpiration in plants. It is an important part of maintaining the land surface water and surface energy balances. The LAI is also an important input parameter of the soil water evapotranspiration model. When the LAI is derived from remote sensing methods it is an excellent proxy for calculating transpiration by upscaling the field measurements. It is very important to include the field data in a global scale model for accurately determining transpiration. Additionally, “yield” [104,105,106] is the focus of agriculture, forestry, botany, and other disciplines. The LAI is an important quantitative index that reflects the photosynthetic capacity and growth and development status of the crop groups. A higher LAI is conducive to photosynthesis, the accumulation of organic products, and the formation of yield. Through the acquisition of the LAI, information such as crop growth, yield, disease, insect pests, and energy exchange can be well monitored. Then, “photosynthesis” [107,108] is the sum of a series of complex metabolic reactions, the basis for the survival of the biological world, and an important medium for the Earth’s carbon and oxygen cycle. The leaf area determines the photosynthetic physiological activity of the leaves to a certain extent and can reflect the utilization level of the light energy that is absorbed by the plant leaves. Moreover, the “biomass” [109,110,111] can reflect crop growth as an important indicator for monitoring and estimating crop yield, and it is closely related to the LAI. Biomass is an important part of the global carbon cycle. Monitoring the vegetation biomass with the LAI can provide basic data for the study of energy balance and energy flow in terrestrial ecosystems. In addition, “NDVI” is highly correlated with the LAI, and the NDVI data are a powerful tool for the large-scale estimation of the LAI and seasonal and phenological trends [95,112,113]. With the development of remote sensing technology, it is possible to obtain vegetation information characteristics on a global scale and under all weather conditions and continuous observation. When the LAI is based on remote sensing inversion, it has important application value for the study of large-scale vegetation changes. In addition, “climate change” [4,114,115,116] has had a profound impact on vegetation, a key component of the biosphere, which plays an important role in regulating the Earth’s climate and providing ecosystem services and is also very sensitive to global changes. If the leaf area is based on remote sensing methods, the index data are an important parameter for the systematic analysis of global trends and their drivers.

The “leaf thickness” [117,118] and “specific leaf” [119,120,121] are the keywords that had the longest duration of attention. They are relatively stable leaf physiological traits of vegetation and are closely related to the LAI, photosynthesis, leaf nitrogen content, and leaf development. Additionally, “transpiration” [122,123,124] is one of the most important complex physiological processes of vegetation, and “earliness” [125,126] is an important research topic in agriculture, forestry, and botany. The “harvest index” [127,128,129] is an important component of crop yield. Furthermore, the “humid tropics” [130,131] has become the focus of LAI research due to its unique climate and geographical location. In Figure 10, the farther the red dots are on the right and the bigger the blue dots, the closer the occurrence time of the corresponding keyword is to the present and the greater the number of published papers, respectively, which can reflect the research trend. In terms of sensors that are related to LAI research, unmanned aerial vehicle (“UAV”) and “Sentinel-2” and their “radiative transfer model” have been the hotspots in LAI research in recent years. In the past five years, the number of keywords using “UAV” in LAI research papers has shown a growth trend that started from nothing to rapid growth. Thanks to the increasing maturity of UAV-related technologies in recent years and their rapid popularization in the civil field, there were 173 articles in 2015–2021. The UAV remote sensing method can not only meet the high-resolution requirements but can also obtain large-scale real-time, non-destructive, and reliable vegetation information and greatly reduce the workload in the field. The research on vegetation that has been based on UAV is fruitful and has resulted in a lot of progress in related research fields. Because of its unique advantages, it is mainly used for remote sensing applications of medium and small scale and high precision, such as vegetation index extraction of the regional high precision LAI [132,133,134,135,136] and vegetation coverage [137,138,139], crop yield estimation at a field scale [140,141,142], fine classification and feature recognition [143,144,145], and high-throughput crop phenotype research [142,146,147]. The Sentinel-2 data have three bands in the “red edge” region, which can calculate the intensity of the chlorophyll reflection peak more accurately and improve the accuracy of atmospheric correction. It is an ideal data source for accurate and efficient vegetation monitoring [148]. It also provides the conditions for low-cost, process-oriented, and normalized monitoring of the medium and high spatial resolution LAI in specific geographical areas [149,150,151]. Moreover, it is widely used in crop classification, area extraction [152,153], biomass estimation [154,155], crop chlorophyll [156,157] and moisture information remote sensing monitoring [158,159,160] and other research areas. The “random forest” [161,162,163] and “machine learning” [164,165,166] methods have been the most prevalent keywords in recent years, with 173 and 180 papers, respectively. Then, “precision agriculture” [167,168,169] has become an important form of modern agricultural production to rationally utilize agricultural resources, increase crop yields, reduce production costs, and improve the ecological environment, and it has received more and more attention. The research on “climate change” and “the driving mechanism of carbon sink change” requires the LAI as a key input for dynamic process models, such as climate models and carbon cycle models. Thus, the research and application of the LAI should be carried out on a regional and global scale as global remote sensing LAI product data are increasingly used [116,170,171,172].

4. Conclusions and Prospects

4.1. Conclusions

This study focused on the development trend of global LAI research. Based on a data set of 22,276 documents on LAI research from 1947 to 2021, the bibliometric analysis comprehensively reveals the trend of LAI research, the distribution of research power, and the evolution of research topics in the past 75 years. Descriptive statistics, publication volume analysis, and discipline distribution analysis are essential means of evaluating the status quo and hotspots of literature research in terms of bibliometrics. Additionally, influence and author analyses are essential references for evaluating hotspots in the field of literature research and their academic influence. Then, keywords are an essential part of academic journal papers. As a highly refined summary of the research content of the article, they can represent the hot issues of the current research in this field. These findings can be used to assist researchers in fully understanding the research progress and hotspots of the LAI, and they have particular guiding significance for the current research on and development of the LAI.

(1)

In terms of the overall development trend, global LAI research has shown a sharp increase since the early 1990s, and it is still very active. The number of publications increased exponentially, from one article in 1947 to 1954 in 2021. For literature growth over time, the 75-year time window was divided into five stages: the incubation, cultivation, acceleration, evolution, and outbreak periods.

(2)

In terms of the main force of the research, the United States ranked 1st in the paper output in each period, and the University of Arizona and the University of Wisconsin formed the main force in leading and promoting LAI research. For international scientific research cooperation, large-, medium-, and small-scale cooperation networks were formed among different countries. The United States occupied a dominant position in the scientific research cooperation network and established close cooperative relationships with many countries. China, the United States, Brazil, and India were the major countries of LAI research. The most influential journals included Remote Sensing of Environment, Global Change Biology, and IEEE Transactions on Geoscience and Remote Sensing. Chen J.M., Baret F., and Myneni R.B. were the foremost researchers.

(3)

In terms of research disciplines and application fields, LAI research involved many disciplines. Agriculture, forestry, vegetation, environmental science and ecology, and remote sensing science are the first significant fields to receive attention. Due to in-depth research and interdisciplinary integration, LAI has gradually expanded to imaging science, photography technology, meteorologies, and atmospheric sciences, such as science, hydrology, and engineering. With the rapid development of satellite remote sensing technology and high-precision sensors, the research fields and research methods involving the LAI are also increasing yearly. Due to scientific research cooperation, geography and landforms, and other factors, different countries have commonalities in selecting research topics, regional differences, and distinctive features.

(4)

The LAI research themes showed a substantial stage and ecological development dependence for the evolution of high-frequency words and high-cited papers. Relevant research started from early research on fundamental issues such as “growth”, “yield” and “canopy” and gradually changed to the complex physiological characteristics and essential physiological processes of vegetation, the material cycle and energy flow of ecosystems, global climate change and its ecological effects, and other macro issues. There has also been a focus on global climate change, biodiversity, and other issues in the 21st century. These trends reflect that in the context of social, ecological, and technological development and global change, the scientific issues of LAI research tend to be in-depth and complex, and scientific research tends to cover greater depth and become more integrated.

However, this study still has some limitations due to limited tools and data filtering strategies. The stem extraction algorithm and screening mechanism of the WOS document retrieval system need more reference to professional knowledge, which may lead to the extraction of some documents that contain synonyms but do not belong to specific research categories as primary input data. Flaws in this algorithm may affect the accuracy and reliability of the results of bibliometric studies. In addition, the current Bibliometrix word segmentation algorithm needs to be more intelligent, and some keywords have uncertainty and a lack of accuracy in terms of effective extraction and characterization. Therefore, subsequent bibliometric research should enhance the semantic understanding of the citation data to improve the make the extraction of the bibliometric keywords more automated and accurate. The quantitative evaluation of publications using bibliometric indicators is objective, transparent, and repeatable, but it may also ignore the focus of qualitative evaluation. An accurate assessment of the quality of scientific work should be obtained by analyzing the relevant indicators of publications and combining bibliometric indicators with peer review. Since the advantages of bibliometric indicators correspond to the disadvantages of peer review, it would allow for a fairer and more accurate assessment of the scientific research. The detailed indicators of the bibliometric methods and the standardization of the results also need to be further explored.

4.2. Prospects

With the addition of new data sources and the rapid development of cloud computing platforms, new research methods, such as machine learning, deep learning, random forest, and the Google Earth engine, have also brought new development opportunities and directions for LAI research. Multidisciplinary integration will still be the main trend of LAI research in the future. The LAI will be more widely used in agroforestry (including global vegetation change, extensive agricultural monitoring, phenology, yield estimation, and plant transpiration) and environmental science and ecology (including climate change, precipitation, and drought monitoring). With the improvement of people’s living standards and increasing health concerns, the LAI will play a more important role in the fields of public health and human activity research in the future and affect a wider range of disciplines. The development of more and more advanced sensor and imaging technology and multi-source data fusion reconstruction technology will be able to produce LAI products that can provide a higher spatial and temporal resolution and longer time series. With the continuous development of UAV technology, there is still much room for improvement in the research and application of UAV in terms of digitization, informatization, and the automatic extraction of the vegetation index, and this is expected to provide more important support for the research and development of the LAI.