Progress and Hotspot Analysis of Bibliometric-Based Research on Agricultural Irrigation Patterns on Non-Point Pollution

Abstract

With the constant advancement of irrigation technology and the continuous expansion of irrigation areas, non-point source pollution (NPS) caused by agricultural activities has posed a persistent threat to ecosystems and biological safety. Against this backdrop, it is imperative to lay scientific foundations for green, sustainable, and high-quality agricultural development through a thorough review of the relevant research progress. In this study, bibliometric methods are adopted to comprehensively analyze and visualize the current state and key literature on agricultural irrigation and NPS pollution from 2010 to July 2024. The focus of this study is specifically on summarizing the research hotspots and development trends of different irrigation methods and the mechanisms behind their impacts on NPS pollution. The results indicate that publications from the United States and China account for 63.8% of the total, but the fragmentation of research efforts remains, suggesting a necessity to strengthen international and regional collaboration. There are three institutions with the highest publication output, namely Northwest A&F University, Hohai University, and the Chinese Academy of Sciences. The subjects identified as the key areas of research on irrigation-related NPS pollution (IRR-NPS) include precision irrigation, rapid water pollution response, spatiotemporal management, interdisciplinary integration, wastewater treatment, and crop models. Regarding future research, it is necessary to focus attention on real-time precision irrigation, standardized crop models, data accuracy, spatiotemporal pollution coordination, pollution purification technology development, interdisciplinary integrated governance, and the innovative applications of soil improvement technologies. In addition to offering theoretical support and practical guidance for the management of agricultural NPS pollution, this study also provides management and technical support for policymakers, which is beneficial for advancing agricultural irrigation technology and environmental preservation.

1. Introduction

Agricultural irrigation plays a crucial role in modern agricultural production, but the underestimation of its impact on NPS pollution should be avoided as much as possible. Since the 1970s, the use of chemical fertilizers and pesticides worldwide has been gradually popularized, which makes agricultural NPS pollution an increasingly prominent issue [1]. To address this global challenge, a series of management measures have been implemented in many countries across the globe, such as optimizing fertilization, improving irrigation technologies, and enhancing cooperation in agricultural NPS pollution monitoring [2,3]. For nearly half a century, the application of various efficient water-saving irrigation technologies, such as drip irrigation and sprinkler irrigation, has been gradually promoted due to technological advancements and increased environmental awareness, contributing significantly to the reduction of agricultural NPS pollution [4,5].

Notably, there remain multiple challenges facing the mitigation process, although modern irrigation technologies and management practices have alleviated agricultural NPS pollution to some extent in recent years. According to research, the use of improper irrigation methods is a major contributor to nitrogen fertilizer loss and water pollution. For instance, conventional flood irrigation causes 20–70% of applied nitrogen to leach away [6], which results in the waste of resources and the severe pollution of water bodies [7]. In comparison to the surface pollution caused by improper fertilization, the amount of pesticide pollution is less, mainly dominated by organochlorine and organophosphorus. However, it usually lasts longer and is likely to cause water pollution and eutrophication of water bodies [8]. Characterized by dispersed sources and diverse pollutants, agricultural NPS pollution makes the control process complex and costly [9]. Additionally, various pollutants, such as nitrogen and phosphorus, can cause significant deterioration in water quality after entering water bodies through runoff [10]. Currently, there are still no systematic monitoring and evaluation methods in place to accurately quantify and compare the effectiveness of different irrigation practices in reducing pollution [11]. Moreover, the difficulty of its monitoring and management increases due to the complex spatiotemporal variability of NPS pollution [12]. In the course of agricultural production, it often proves challenging to coordinate irrigation management, environmental protection, and resource utilization, which makes it difficult to control NPS pollution [13]. This study aims to reveal the effects and mechanisms of different irrigation methods on pollution reduction through a systematic review of the scientific issues related to the impact of agricultural irrigation on NPS pollution. It is expected to provide a scientific reference for the development of effective agricultural NPS pollution prevention and control strategies. Specifically, it contributes to optimizing irrigation technologies and management practices, enhancing the sustainability of agricultural production [14], reducing fertilizer loss, improving efficiency in nitrogen use, mitigating the negative environmental impacts of agriculture [15], and promoting the formulation of more effective agricultural NPS pollution control policies. Thus, sustainable agricultural development and water environment preservation can be promoted [16,17].

Focusing on the main source of pollution is based on surface source pollution generated in the process of agricultural irrigation. Therefore, irrigation-related NPS pollution (IRR-NPS) is analyzed in this paper. To be specific, bibliometric analysis is conducted using the CiteSpace 6.1.R6 software tool to systematically review the literature on the impact of agricultural irrigation on NPS pollution. First, the research literature on agricultural irrigation and NPS pollution from 1976 to the present is collected. Then, the data are processed and visualized with CiteSpace. Furthermore, the research hotspots and development trends of different irrigation methods are analyzed to reveal their impact mechanisms on NPS pollution. Finally, the results of the literature analysis are discussed to propose sensible recommendations on irrigation management and pollution control measures, providing theoretical support and practical guidance for the future practice of agricultural NPS pollution management [18,19].

2. Materials and Methods

2.1. Data Collection and Preparation

Prior to 2010, scholarly works revolved mainly around conducting isolated investigations into non-point source pollution or irrigation practices [20,21]. Moreover, throughout the 1980s and 1990s, widespread industrial activities contributed largely to non-point source pollution. However, the focus of non-point source pollution shifted toward agricultural pollution with the steadily increasing consciousness of sustainable environmental development. Meanwhile, research on irrigation methods was centered on enhancing the efficiency of water resource utilization [22]. Beyond 2010, the progressive evolution of agricultural irrigation technology prompted a shift in focus toward the non-point pollution caused by agricultural irrigation, with industrial pollution increasingly recognized as point-source pollution [23]. The selected search period includes studies from January 2010 to July 2024 in the Web of Science (WOS) Core Collection database. CiteSpace enables more than just traditional systematic reviews; considering that the contents cited in different studies overlap, the challenges of future research areas are underrepresented, and the data are outdated. It supports the tracking of emerging trends and the identification of key developments by providing a valuable, timely, repeatable, and flexible method for the visual analysis of literature, enabling [24].

To include as many relevant documents as possible through fuzzy queries, the keywords were set as follows: #1 (TS=irrigation mode or TS=irrigation method) and (TS=non-point source pollution); #2 (TS=irrigation NEAR/5 mode or TS=irrigation NEAR/5 method) and (TS=non-point source pollution); #3 (TS=“irrigation mod*” or TS=“irrigation meth*”) and (TS=non-point source pollution); and #4 (TS=water-saving irrigation) and (TS=non-point source pollution). All four formulas were linked with the logical operator “OR”, and the terms related to “metal” were excluded to remove the references to heavy metal NPS pollution. Allowing for the limitations on language and document type, the document types were set to “articles, reviews”, and the language was set to “English”. In total, 1627 documents were obtained, with 200 foreign authors from WOS (abbreviated as WOS-OA). To reduce errors while improving accuracy, all documents were imported into Endnote for step-by-step sorting, including the removal of duplicates and irrelevant records. Finally, 913 documents were retained as the foundational data for this study, with the date of data collection falling on 6 July 2024.

2.2. Research Methodology

From the perspective of bibliometrics, important noun phrases were extracted from titles, abstracts, and keywords to perform co-word, co-occurrence, burst, and emerging trend analyses of the literature. This approach can be adopted to explore international collaboration networks, core authors, and development trends in the research field.

CiteSpace refers to data visualization software developed by Professor Chaomei Chen of the School of Information Science and Technology at Drexel University using the Java platform [15,25]. Because of its high performance in knowledge graph visualization and bibliometric analysis, it ranks as the most widely used bibliometric analysis software at present [26]. CiteSpace has been widely applied in the research on subsurface drip irrigation [27], groundwater salinization [28], and crop water footprint [29], with significant results produced. In this study, CiteSpace 6.1.R6 software is applied to conduct statistical and visual analyses of relevant literature for irrigation improvement and pollution reduction. The analysis focuses on a range of indicators such as publication number, authors, institutions, keywords, and research hotspots (Figure 1).

As shown in the knowledge graph, a larger node area suggests a higher frequency of occurrence or citation. The measure of node importance is Betweenness Centrality (BC), whose value usually ranges from 0 to 1. When BC values exceed 0.1, they are considered above the median [30]. The clustering tools in CiteSpace include topic recognition, keyword extraction, and multidimensional clustering. Regarding clustering methods, the Log-Likelihood Ratio algorithm (LLR) [31] is used in this paper to perform cluster analysis on the cited literature. The clustering effect is evaluated against the Modularity (Q value) and Silhouette (S value) indices. When the Q value exceeds 0.5, it indicates significant cluster structure; when the S value exceeds 0.7, it indicates reasonable clustering results [32].

2.3. Formatting of Mathematical Components

Betweenness Centrality (BC) is a critical parameter used to indicate influence. A node with high BC is considered to exert significant influence and control within the network, acting as a stronger pivot for promoting collaboration in the field. This is conducive to identifying key figures, research frontiers, and connections between research areas. The formula for this parameter is expressed as follows:

$${BC}_i={\sum }_{\alpha \ne \beta \ne \gamma }{\displaystyle \frac{{\eta }_{\alpha \gamma }^{\beta }}{{\delta }_{\alpha \gamma }}}$$

(1)

where ${BC}_i$ represents the Betweenness Centrality; ${\delta }_{\alpha \gamma }$ means the number of shortest paths from node $\alpha$ to node $\gamma$; ${\eta }_{\alpha \gamma }^{\beta }$ denotes the number of shortest paths passing through node $\beta$ out of the total ${\delta }_{\alpha \gamma }$ shortest paths from node $\alpha$ to node $\gamma$.

Connection strength ($\mathrm{C}\mathrm{o}\mathrm{s}$) is an important parameter used to measure the closeness of collaboration, and its value is directly proportional to the degree of cooperation. The calculation formula for this parameter is expressed as follows:

$$\mathit{Cos}(x_{ij}, s_i, s_j)={\displaystyle \frac{X_{ij}}{\sqrt{S_i{S}_j}}}$$

(2)

where $\mathit{Cos}(x_{ij}, s_i, s_j)$ represents the connection strength; $X_{ij}$ denotes the co-occurrence frequency of $i$ and $j$; $s_i$ stands for the frequency of $i$’s occurrence; $s_j$ refers to the frequency of $j$’s occurrence.

In cluster analysis, the Q-value (Modularity) and S-value (Silhouette) are two commonly used evaluation metrics. The Q-value is used to assess the community structure within a network by comparing the actual number of connections to the expected number, thus evaluating the closeness between communities. The calculation formula is as follows:

$$Q=\sum _k({\displaystyle \frac{e_{kk}}{m}}-{\left({\displaystyle \frac{a_k}{m}}\right)}^2)$$

(3)

where $e_{kk}$ is the number of connections within community $k$, $a_k$ is the total number of connections in community $k$, and $m$ is the total number of connections in the network. When the Q-value is close to 1, it indicates significant modularity in the network, with dense connections within each community and sparse connections between communities. Conversely, when the Q-value is close to 0, it suggests a random distribution of nodes within the network, lacking a clear community structure.

The S-value is used to evaluate the cohesion and separation of clusters. For a given data point $k$, the S-value is defined as follows:

$$S\left(k\right)={\displaystyle \frac{b\left(k\right)-a\left(k\right)}{max(a\left(k\right),b\left(k\right))}}$$

(4)

where $a\left(k\right)$ represents the average distance from point $k$ to other points within the same cluster, and $b\left(k\right)$ represents the average distance from point $k$ to the nearest other cluster. The S-value ranges from −1 to 1, with values closer to 1 indicating better clustering quality. Values close to −1 indicate that point $k$ is closer to other clusters rather than its own cluster. By averaging the S-values of all points, an overall S-value for the clustering can be obtained, which is used to assess the rationality of the clustering structure.

3. Results and Discussion

3.1. Disciplinary Co-Occurrence Analysis

To demonstrate the thematic distribution of journals in the IRR-NPS research field from a macro perspective, the LRR algorithm of CiteSpace software was applied to cluster the CR, DOI, and other fields in the literature database. Then, a discipline contribution analysis chart was created through the dual-map overlay feature of CiteSpace (Figure 2). In the Web of Science (WOS) database, there are over 10,000 journals classified by their respective disciplines, each of which is represented by different colors. These disciplines are distributed in different positions in the source area and the reference area. More specifically, citing journals are positioned on the left side of the image, while cited journals are positioned on the right side [33]. Labels represent the disciplines covered by the journals, and the colored links from left to right indicate citation paths, with thicker lines suggesting more citations. There are two main citation paths showing that Veterinary/Animal/Natural Science journals frequently cite research from Plant/Ecology/Zoology journals (cited 12,670 times) and Environmental/Toxicology/Nutrition journals (cited 9331 times). The f value on the colored links indicates the citation frequency of the journal, while the z value, indicating the size of each node, is the normalized f value, showing a positive correlation with it. Obviously, the z and f values of the citation lines for Plant/Ecology/Zoology journals are greater in comparison to Environmental/Toxicology/Nutrition journals. It indicates the greater the number of node connections, higher co-occurrence, and stronger dominance of the former.

3.2. Analysis of Annual Publication Number and Countries of Origin

As a significant indicator of activity and development trends in a particular field [34], the annual publication number is useful for analyzing research trends, understanding distribution patterns, evaluating index status, co-authorship conditions, and identifying research hotspots [35]. According to the number of articles and reviews on IRR-NPS published from 2010 to July 2024 (Figure 3), there is a relationship between the collected data and the annual publication number of relevant countries. Overall, the publication number has increased steadily over time, despite a slight decline in 2013, 2015, 2017, and 2020. The largest number of publications was recorded in 2022, indicating the increased attention paid to the issues with improved irrigation and NPS pollution reduction. The scope of IRR-NPS research is expanding.

Table 1. Number of published articles per country.

Rank	Number of Papers	Percentage %	Centrality	Burst	Country
1	451	54.34	0.64		CHINA
2	91	10.96	0.33	2.3	USA
3	50	6.02	0.2	1.44	SPAIN
4	49	5.90	0.07	3.22	INDIA
5	29	3.49	0.12	4.9	AUSTRALIA
6	27	3.25	0.13		ITALY
7	26	3.13	0.04	2	IRAN
8	22	2.65	0.16	2.87	FRANCE
9	18	2.17	0.13	2.72	ENGLAND
10	16	1.93	0.09	3.81	GERMANY
11	16	1.93	0.03	2.96	CANADA
12	16	1.93	0.06		EGYPT
13	15	1.81	0.12	1.7	TURKEY
14	15	1.81	0.01		ISRAEL
15	14	1.69	0.05		PAKISTAN
16	14	1.69	0.13	3.62	NETHERLANDS
17	11	1.33	0.05		SAUDI ARABIA
18	13	1.57	0.05	2.92	PORTUGAL
19	10	1.20	0.14		BRAZIL
20	10	1.20	0.05	1.43	SOUTH AFRICA

lists the number of publications by country over the past 14 years. The United States initially led research on IRR-NPS. Then, the field experienced a period of slow growth from 2012 to 2017. Notably, China has implemented policies on water resource management since 2012 [36], which has caused a rise in the number of published articles. For this reason, China surpassed the United States. Prior to 2013, the amount of relevant literature was relatively limited. In the years following 2013, the research entered an initial development stage. Despite a slight fluctuation in the number of published articles across China, it generally showed an increasing trend. In contrast, the United States was significantly outnumbered by China in publications, although the number fluctuated less.

3.3. Analysis of Publishing Institution and Authors

The analysis of publishing institutions and authors plays a crucial role in understanding the structural development and changes within a research field [37]. As revealed by analyzing the articles in the IRR-NPS research field (

Table 2. Top 28 institutions with the largest number of publications in IRR-NPS research.

Rank	Frequency	Centrality	Burst	Year	Institution
1	41	0.06	5	2017	Northwest A&F Univ
2	37	0.08	1.72	2010	Chinese Acad Sci
3	37	0.04	2.46	2012	Hohai Univ
4	32	0.06		2012	Chinese Acad Agr Sci
5	28	0.06	3.95	2014	China Agr Univ
6	22	0.04	1.95	2010	China Inst Water Resources & Hydropower Res
7	16	0.01		2019	North China Univ Water Resources & Elect Power
8	15	0.04		2016	Wuhan Univ
9	15	0.01	2.74	2017	Shihezi Univ
10	12	0.02	0.71	2014	Beijing Normal Univ
11	11	0.01	2.39	2021	Kunming Univ Sci & Technol
12	11	0	1.24	2010	CSIC
13	10	0.02	2.95	2020	Xian Univ Technol
14	10	0.01	2.28	2021	Yangzhou Univ
15	10	0		2022	Minist Agr & Rural Affairs
16	9	0.01		2019	Univ Chinese Acad Sci
17	8	0		2022	Gansu Agr Univ
18	8	0	3.28	2019	Lanzhou Univ
19	7	0		2018	Guangxi Univ
20	6	0.01		2022	Northeast Agr Univ
21	6	0	3	2020	Univ Tehran
22	5	0.02		2021	Yangtze Univ
23	5	0.01		2019	Inner Mongolia Agr Univ
24	5	0.01	1.1	2011	Agr Res Org
25	5	0		2020	Univ Girona
26	5	0	0.77	2017	Beijing Acad Agr & Forestry Sci
27	5	0	2.21	2010	INRA
28	5	0	3.04	2010	Univ Florida

), there are twenty-eight publishing institutions with five or more articles published annually. Of these, twenty-four are in China, two in Spain, one in Iran, and one in the United States. China is predominant in this field, with Northwest A&F University as the highest-producing institution, followed by the Chinese Academy of Sciences, Hohai University, and the Chinese Academy of Agricultural Sciences. On average, they have over 30 publications annually. To a large extent, this is attributed to China’s increasing national efforts invested in pollution control over recent years.

The co-authorship visualization (Figure 4) involves 479 nodes and 225 links, with a density of 0.002. This reflects the level of collaboration and knowledge exchange among authors in the research field. In this context, a density of 0.002 suggests relatively sparse connections between nodes. That is to say, the actual established connections are outnumbered by the total potential connections in the network. This implies a pressing need to promote cross-institutional and cross-disciplinary collaboration, establish tighter cooperation networks, and address the challenges in the field collectively for the ultimate improvement in research efficiency and depth.

3.4. Keyword Analysis

3.4.1. Exploration of Hot Topics in Review Articles

For a comprehensive exploration of the research hotspots and directions in the IRR-NPS field, the keywords obtained from 24 review articles within the literature database were manually analyzed in this paper (

Table 3. Statistical analysis of keyword information in review articles.

Rank	Label	Count	Keywords
1	energy use	29	management × 3, performance × 2, model × 2, wheat × 2, deficit irrigation × 2, use efficiency × 2, performance measure, drip irrigation, design, network, irrigation system management, irrigation modernization, decision support system, input cost, environmental sustainability, agricultural water management, crop, low discharge, integrative management, on demand, optimization, agriculture
2	berry crop	18	abiotic stress × 3, plant abiotic stress, humic acid, fruit yield, foliar application, arbuscular mycorrhizal fungi, fruit specy, gene expression, heavy metal, growth promoting rhizobacteria, brassica napus, chlorophyll fluorescence, hydrogen peroxide, drought stress, fruit quality, drought tolerance
3	sewage treatment plant effluent	13	domestic sewage, posttreatment, anaerobic treatment, waste water treatment, activated sludge, scale uasb, treatment system, municipal wastewater, aerobic treatment, post treatment, expanded granular sludge, polishing pond, uasb reactor
4	nutrient use efficiency	15	growth × 4, perennial fruit, fertilizer use efficiency, rhizosphere hybridization, basin irrigation, long term, potassium fertilization, nutrient-microbe synergy, nagpur mandarin, microbial consortium, citrus reticulata blanco, co2 enrichment
5	intensive cotton	13	plant × 2, irrigation × 2, increases stand establishment, economic benefit, lint yield, challenges and countermeasure, light and simplified cultivation, bt cotton, intensive farming technology, cultivation, accumulation
6	collaborative catchment-scale management	11	framework, suspended sediment, food industry, wireless sensor network, surface, land use, eutrophication, pesticide, agricultural activity, pollution, time
7	cover	12	climate change × 3, wind, quality, agricultural irrigation reservoir, suppressant monolayer, temperature, dispersion behavior, stored water, shade cloth cover, spreading rate
8	actual evapotranspiration	9	drought × 2, nitrogen fertilizer rate, olea europaea, fulvic acid, actual evapotranspiration, corn grain yield, humic substance, component

Note: The numbers following the ‘×’ symbol under the ‘Keywords’ section of the table indicate the frequency of occurrence for each keyword.

). Review articles highly comprehensive in nature, due to various analyses and findings synthesized from different scholars and teams within the IRR-NPS field. These review articles provide more thorough, in-depth, and clear methods of presenting research hotspots and directions. Additionally, they are highly innovative and closely aligned with the academic frontier and existing issues in IRR-NPS research, for which their keywords can reflect the latest research progress and trends. High-frequency keywords reflect high attention in the field, while the evolution over time suggests a shift in research hotspots. According to the analysis of keywords obtained from these reviews, aside from the category related to “energy use,” significant attention is also drawn to berry crop cultivation and wastewater treatment in the process of crop cultivation. This suggests the extensive application of IRR-NPS technologies in crop cultivation, with relatively limited research on actual evapotranspiration. Additionally, there are other significant research directions, including spatial-temporal coordination management and the control of other factors [37]. Therefore, it is expected that technological innovations in IRR-NPS can prompt the extensive discussions on how to reduce energy wastage, improve the capacity of wastewater treatment, and enhance watershed management [38,39].

3.4.2. Analysis of Keyword Co-Occurrence

By revealing the frequency of keywords in a research field, keyword co-occurrence mapping is conducive to identifying the research hotspots and core content [40]. Each node represents a keyword, node color indicates the year of occurrence, and node size suggests the frequency of occurrence. If a keyword appears more frequently, its node becomes larger. The lines between nodes represent the strength of co-occurrence (Figure 5). Additionally,

Table 4. Top 20 most frequently cited keywords used in the literature pertaining to IRR-NPS research.

Rank	Number of Papers	Centrality	Burst	Year	Keywords
1	111	0.19		2011	Yield
2	92	0.12		2010	Growth
3	89	0.1		2012	Irrigation
4	82	0.13		2012	Management
5	72	0.15		2011	Soil
6	71	0.2	1.98	2010	Water
7	64	0.17	4.29	2011	System
8	61	0.13		2011	water use efficiency
9	58	0.08		2014	drip irrigation
10	57	0.08		2014	Quality
11	56	0.1		2014	climate change
12	48	0.07		2014	Impact
13	47	0.09	5.41	2012	Performance
14	47	0.06		2014	Model
15	47	0.03		2018	use efficiency
16	38	0.09		2018	grain yield
17	36	0.09		2011	deficit irrigation
18	31	0.04		2013	Productivity
19	29	0.04		2018	winter wheat
20	27	0.06		2015	Nitrogen
21	26	0.04		2012	Crop
22	26	0.02		2016	Evapotranspiration
23	23	0.04	1.83	2016	Plant
24	21	0.03		2016	Maize
25	21	0.05		2014	Fertilizer
26	20	0.05	2.11	2011	Area
27	20	0.05		2010	Stress
28	20	0.06		2011	Efficiency

lists the keywords with an occurrence frequency of ≥20. The co-occurrence map shows 465 nodes, 1495 links, and a density of 0.0139. There are 28 keywords with a higher frequency than 20. As a concise summary of the research hotspots in academic papers, keywords represent the core and essence of these papers, indicating the research hotspots in the field.

As shown in Figure 5, there are nine highly central keywords: water (0.2), yield (0.19), system (0.17), soil (0.15), management (0.13), water use efficiency (0.13), growth (0.12), irrigation (0.1), and climate change (0.1). Water is identified as the carrier of agricultural NPS pollution. In terms of agricultural production, yield and water quality are significantly affected by the accumulation of soil salinity, water productivity, and nitrogen leaching [41]. Through the exploration of irrigation technologies, researchers develop various irrigation system models to lower irrigation costs and reduce agricultural NPS pollution [42]. In the context of growing environmental awareness, proper water-saving irrigation system management can reduce secondary soil salinization [43] and decrease nitrogen application without affecting or even increasing yield. Thus, water use efficiency is improved, and agricultural NPS pollution is reduced [44,45]. Additionally, in agricultural irrigation, climate change not only causes salinization and a reduction in water content, affecting irrigation efficiency. With the same yield maintained, there may be an increase in NPS pollution emissions [46,47].

3.4.3. Keyword Clustering Relating to IRR-NPS Research Literature According to Time of Citation

Through the keyword clustering map, a timeline graph was created to visualize the evolution of keywords over time. It clearly indicates the evolution of keywords and thematic changes within each cluster. The keywords in different clusters were arranged in chronological order to illustrate the time distribution of each cluster’s keywords, which allows the evolution of research hotspots to be better understood [48]. As shown in Figure 6, the Q value of the keyword clustering is 0.5643, and the S value is 0.8366. There are 13 labels produced by this graph, generated through keyword selection and the log-likelihood ratio clustering algorithm. These labels were selected for their strong representativeness (size > 10) and satisfactory clustering tightness (silhouette > 0.7). There are 863 links among 447 independent nodes. The size of each node represents the number of occurrences. A larger node indicates a greater significance, which is conducive to understanding the influence between clusters [49,50].

Between 2010 and 2014, the keyword timeline view focused on keywords related to traditional agricultural management and irrigation technology, such as management, water use efficiency, evapotranspiration, and transpiration. Irrigation technologies and water quality models advanced with the improvement in understanding the growth mechanisms of different crops and the exploration of water and fertilizer management practices. From 2014 to 2024, sensor controllers continued to advance, which was accompanied by a shift to more specific technologies such as drip irrigation, climate change impacts, quality, efficiency, and nitrogen and fertilizer use. It is suggested that the research conducted in the years after 2014 is more focused on how to improve efficiency and address the challenges posed by climate change.

For a better understanding of the research hotspots and future directions in this field, the keywords with similar themes and clear evolution paths were categorized into four topics. The first one focuses on irrigation water resource management, including clusters #4 Irrigation Management, #5 Irrigation Modernization, #6 Irrigation Volume, #7 Water Use Efficiency, and #13 Residential Irrigation. The primary years of research carried out on these topics are 2018, 2017, 2013, and 2014. The second one revolves around the management of agricultural production practices, involving clusters #0 Organic Fertilizer, #8 Grain Yield, #12 Agricultural Practices, and #14 Multivariate Statistical Analysis, with key years being 2016, 2019, 2016, and 2012. The third one is centered on examining various environmental impacts, including clusters #1 Climate Change, #2 Optimization, #3 Microbial Desalination Cells, and #11 Treated Wastewater, with the primary years of research including 2018, 2017, 2017, and 2015. The fourth one explores the patterns of plant growth, focusing on clusters #9 Removal and #10 Rainfall Simulator, with significant research activity performed in 2020 and 2019.

Topic 1, “Irrigation Water Resource Management”, represents a core topic where numerous research nodes are involved. These nodes are closely interconnected, underscoring the importance of irrigation water management in IRR-NPS research and the sustained focus on this topic. According to the timeline of keyword occurrences in Figure 6, early studies focused mainly on improving water use efficiency and deficit irrigation. Research indicates a positive correlation between water stress and water use efficiency, particularly throughout the growth cycle of potatoes. During growth, moderate water stress can improve water use efficiency by up to 78% [51]. Over time, the focus of research conducted on irrigation water management has gradually shifted to modernization, water conservation, and the rationalization of irrigation practices. In this process, various modern water-saving irrigation measures have been popularized, such as Internet of Things (IoT) technologies and smart irrigation controllers. In comparison to traditional irrigation management methods, these modern irrigation technologies, such as IoT-based systems, have been demonstrated as effective in reducing water consumption by up to 34% [52,53]. This is vitally important for sustainable agricultural development, as it can enhance water use efficiency while reducing the water waste caused by poor management and improper allocation. In terms of residential irrigation, it is essential to develop and apply smart irrigation controllers, given the complexity of irrigation conditions. According to the measures and data from studies, the use of smart controllers can reduce water use by 15 to 40 percent, which means 17 percent water savings through the use of smart irrigation controllers. Therefore, water resource efficiency can be improved by optimizing plant selection and landscape layout through the development of smart controllers for complex conditions in residential irrigation collectively. Contributory to household and community water savings, this effectively reduces the risk of surface source pollution [54]. Overall, there is a gradual evolution of irrigation water resource management towards greater precision and intelligence, which is conducive to improving the efficiency of agricultural water use and significant to environmental protection and sustainable agricultural development. In Topic 2, the significant influencing factors in the management of agricultural production practices include organic fertilizers, grain yield, agricultural practices, and multivariate statistical analysis. According to research, the use of organic fertilizers can improve both soil quality and crop nutrient uptake, thus reducing the dependency on chemical fertilizers. For example, in comparison to the use of conventional urea alone, the combined use of controlled-release nitrogen fertilizers and organic fertilizers for rice cultivation significantly enhances root growth and activity. In addition to increasing grain yield and improving resource use efficiency, this approach also mitigates nitrogen loss and NPS pollution [55]. Grain yield is an indicator that directly reflects the efficiency of crop water and nutrient use. It is possible to boost rice yields while improving water and nitrogen use efficiency by optimizing irrigation and fertilization strategies, such as implementing alternate wetting and drying irrigation combined with nitrogen fertilizer application. Thus, resource waste is effectively reduced [56]. Various agricultural practices, including tillage, fertilization, and irrigation techniques, exert a significant effect on nitrogen loss from farmland. Reasonable agricultural management, such as the appropriate application of nitrogen and the combined use of organic and chemical fertilizers, can be relied on to reduce nitrogen loss while lowering the risk of groundwater and surface water pollution. According to the research conducted in the North China Plain, the medium nitrogen application (MN) combined with organic fertilizers is effective in restricting nitrate nitrogen residues and losses in the soil [57]. Revealing the hadrochemical characteristics and sources of contaminants in groundwater, multivariate statistical analysis facilitates the assessment of the impact that agricultural practices have on water quality. For instance, the multivariate statistical analysis conducted for the studies of groundwater in mining areas has shown that natural processes are the major influencing factors in groundwater chemistry, and that mining activities have a limited impact on shallow groundwater quality. This provides a scientific reference for groundwater resource management and NPS pollution prevention [58]. These agricultural management strategies and analytical tools play an essential role in enhancing resource use efficiency, mitigating nutrient loss, and evaluating the environmental impacts of agricultural activities. Topic 3, “Environmental Impact”, has consistently been crucial to evaluating the sustainability of IRR-NPS research. Climate change has a significant impact on global agricultural production. In comparison to Equal Scarcity (ES) methods, the use of Yield Stress (YS) irrigation methods leads to a higher efficiency in water resource allocation. By prioritizing the crops more sensitive to water scarcity, this approach improves the yields of multiple crops, generates more net profits for farmers, and enhances water retention for downstream ecosystems [59,60]. Additionally, water–nitrogen coupling management strategies can be optimized to ensure that appropriate levels of irrigation and nitrogen application allow cotton’s water use efficiency (WUE), nitrogen use efficiency (NUE), yield, and economic benefits to exceed 90% of their maximum values. This provides scientific evidence for effective local water and fertilizer integrated management [61,62]. Providing innovative solutions to saline water and wastewater treatment, microbial desalination cells (MDCs) can help promote the reuse of water in agricultural irrigation. For instance, the desalination rate of saline water can reach 1.2 g L⁻¹ d⁻¹ following MDC treatment, with the rate of organic compound removal approaching 70%. This represents an effective pathway to the reuse of saline and wastewater [63,64]. Research has also indicated that wastewater reuse can be achieved through reverse osmosis (RO) processes. Under optimal operating conditions, RO can achieve a 99.4% salt rejection rate and a flux of 31.6 L/m²h, demonstrating the suitability of treated wastewater for agricultural irrigation. Thus, the reliance on fresh water resources can be reduced [65,66]. By improving irrigation efficiency, optimizing resource management, and promoting wastewater reuse, these strategies contribute effective solutions to NPS pollution in agricultural irrigation and the sustainable development of agriculture. Topic 4, “Plant Growth Pattern Studies”, focuses primarily on removal technologies and rainfall simulators. Removal technology involves a system where microbial fuel cells (MFC) are combined with anion exchange membranes (AEM) to effectively remove boron from irrigation water, thus mitigating its adverse effects on plant growth. In the pretreatment mode, this system leads to a boron removal rate of 40–50%, while in the post-treatment mode, the removal rate can reach 80–90%, with boron concentrations reduced to the irrigation water standard of 2–5 mg L⁻¹. Apart from improving water quality, this technology also provides an environmentally friendly solution to agricultural irrigation by generating current and reducing conductivity [67,68]. Through the simulation of various rainfall conditions, rainfall simulation technology optimizes crop root distribution and growth. In the experiments conducted by Ningxia University, the ridge-furrow rainfall harvesting (RFRH) system, combined with plastic and biodegradable film mulching, boosted grain yields by 30% and 25% on average over two years. This system improves root distribution in the upper soil layer, which enhances the efficiency of water and nutrient absorption, thus reducing water waste in agricultural irrigation and mitigating the risk of NPS pollution [69,70]. With these technologies applied, irrigation efficiency is improved, which curtails the excessive use of fertilizers and pesticides. Thus, the water pollution caused by agricultural activities is effectively controlled.

3.4.4. Analysis of Keyword Bursts

A keyword burst chart was used to better analyze research trends (Figure 7). Since the emergence of new topics represents the research frontier [71], the default value for the minimum duration of burst years decreased from 2 to 1, resulting in 20 burst words. “Year” indicates the time the keyword first appeared, “Begin” and “End” represent the start and end times of the burst, and “Strength” indicates the intensity of the burst. A higher burst intensity means that the frequency of a keyword increased more significantly during that period, reflecting the current development trends and hotspots in IRR-NPS research [72]. The length of the segments in the chart is proportional to the duration of the burst. The light blue part indicates the period prior to the appearance of the keyword, the dark blue part represents the time the keyword burst began, and the red part shows the duration of the burst.

3.5. Analysis of Reference Co-Citation

A co-citation relationship is formed when multiple documents are cited by the same paper (Figure 8). Citation counts are considered a crucial indicator of quality and academic attention for a document, and highly cited papers often exert significant academic influence. Co-citation relationships can be analyzed to reveal the knowledge structure and system of a specific research field [73]. A majority of co-citation nodes appear after 2015 (except for #5 Irrigation Methods), as irrigation methods are essential for precision irrigation and IRR-NPS research. The following perspectives are considered to improve the accuracy of the analysis: chronological order, citation frequency, and time span.

Chronologically, the earliest cited paper among the top 30 co-cited articles is the 2015 publication by Berbel, J, “Literature review on the rebound effect of water-saving measures and analysis of a Spanish case study”. His comprehensive explanation of the “irrigation rebound effect” was unprecedented. Through empirical research, Berbel proposed that water-saving irrigation technology can mitigate the rebound effect caused by non-irrigation water loss rather than the use of water pricing to offset the rebound effect [74]. Next, as demonstrated through two field experiments conducted in Junfeng Pan’s 2017 paper, “Water productivity and nitrogen use efficiency of rice under different water management and fertilizer-N inputs in South China,” alternate wetting and drying irrigation can save water by 24.1% and 71.4% without affecting nitrogen inputs while maintaining yield. It was also verified that there was no significant interaction between water and nitrogen ratios in biomass production and yield [75]. Subsequently, in Barzegari M’s 2017 paper, “Irrigation and nitrogen managements affect nitrogen leaching and root yield of sugar beet,” ordinary furrow irrigation (OFI) was compared with variable alternate furrow irrigation (VAFI) and fixed alternate furrow irrigation (FAFI) in sugar beet. According to the experimental results, VAFI and FAFI reduced nitrate leaching by 46% and 52%, respectively. In comparison to OFI, the resistance to water stress was enhanced, nitrogen fertilizer use was reduced, and NPS pollution risk was mitigated [76]. As indicated in Agbna GHD’s 2017 publication, “Effects of deficit irrigation and biochar addition on the growth, yield, and quality of tomato”, under deficit irrigation, biochar soil amendments led to a significant increase in soil organic matter and total nitrogen. Meanwhile, the levels of nitrate and ammonium nitrogen in the soil are reduced, and the risk of soil acidification and groundwater pollution are mitigated. This boosts yield while curtailing pollutant emissions [77].

In terms of citation frequency, it represents the academic frontier of IRR-NPS research to some extent. Produced by the R Core Team of the R Foundation for Statistical Computing, cited eight times, the top-ranked work provides statistical analysis support and multidisciplinary guidance for IRR-NPS research from 2018 to 2024. This involves soil dry-wet cycle physical and chemical property analyses [78], with appropriate water-saving irrigation reduction models identified for different seasons [79]. Additionally, the best application period is determined for deficit irrigation [80]. Secondly, as indicated in Grafton RQ’s 2018 publication, “The paradox of irrigation efficiency”, irrigation efficiency is not limited to a measure of water savings, given the regional water quantity unevenness. It requires extensive accounting, measurement, and uncertainty evaluation to track water resources for higher precision in irrigation [81]. According to Ishfaq M’s 2020 publication, “Alternate wetting and drying: A water-saving and eco-friendly rice production system”, relative to traditional flooding, alternate wetting and drying irrigation improves rice quality as this practice not only reduces water consumption (25–70%), CH4 emissions (11–95%), arsenic (13–90%), and mercury (5–90%), but also increases yield (10–20%) and zinc concentrations [82]. Alternate wetting and drying irrigation is more economically feasible than other irrigation techniques such as drip and sprinkler irrigation.

Based on an analysis of the number of clustered nodes, their durations, and their sudden contributions, the research can be divided into two phases. The first phase (2010–2018) includes clusters #4 Irrigation Efficiency, #6 Soil and Populations, and #12 Water and Nitrogen Use Efficiency. Among them, #4 Irrigation Efficiency is of the most significance, with 18 nodes, the longest duration, the most abrupt contributions, and the knowledge base established for IRR-NPS research. As demonstrated in a comparative study conducted by Borrego-Marin in 2018 using long-term data collected from the Guadalquivir River Basin, with the improvement of irrigation efficiency in modern irrigation systems, diffuse nitrogen pollution can be reduced by 5484 tons annually by reducing nitrogen fertilizer application and optimizing irrigation management, while fertilizer application can also be reduced by 14,777 tons of CO₂ equivalent [83]. In cluster #6 Soil and Populations, research on partial root-zone drying irrigation has revealed that soil fertility and water use efficiency can be improved by reducing nitrogen and phosphorus losses and enhancing root growth, while plant adaptability and the sustainable development of communities can also be promoted [84]. For cluster #12 Water and Nitrogen Use Efficiency, it is indicated by studies that improving water use efficiency from 40% to 50% can lead to a reduction of approximately 14.4% in the total nitrogen (TN) load in agricultural land. Through field monitoring and model simulations, evidence has been collected to suggest that nitrogen loss can be mitigated by optimizing water and nitrogen use efficiency, which significantly reduces NPS pollution in semi-arid agricultural regions [85]. The second phase (2018–2024) mainly involves clusters #0 Quality, #1 Alternate Wetting and Drying Irrigation, #2 Agriculture, and #3 Irrigation Policy. There are plenty of nodes in these clusters. Accounting for 57% of the top 11 clusters, these represent the forefront of knowledge. As the most prominent cluster, #0 Quality focuses on quality, yield, water and nitrogen use efficiency, and deficit irrigation. Regarding rice production, the combination of deficit irrigation with controlled-release urea can reduce irrigation water usage by 38.1%, increase rainfall utilization to 44.4%, maintain a high rice yield of 8776 kg/ha, and reduce total nitrogen losses by 48.3% (runoff) and 33.6% (leaching). Thus, the efficiency of water and nitrogen use is enhanced [82,86]. Cluster #1 Alternate Wetting and Drying Irrigation focuses on analyzing how this technique significantly reduces nitrogen losses in rice paddies. Distinct from traditional continuous flooding, alternate wetting and drying irrigation can mitigate total nitrogen (TN) losses by 31.6% to 38.9%. Notably, TN runoff losses are reduced by 25.0% to 35.3% when nitrogen management is optimized, which is highly effective in mitigating NPS pollution [87]. Regarding cluster #2 Agriculture, research shows that dual benefits can be created by optimizing irrigation strategies for reducing IRR-NPS pollution. On the one hand, it can improve water use efficiency by 10%, alleviate the total nitrogen (TN) load in agricultural land by approximately 14.4%, and significantly reduce nitrogen leaching into groundwater or surface runoff [88]. On the other hand, it reduces excessive water consumption while curbing nitrogen and phosphorus emissions in semi-arid regions (e.g., regulated deficit irrigation RC-CI mode), which mitigates the risk of water eutrophication while improving rice yield by 5.8% [89]. Finally, as highlighted in cluster #3 Irrigation Policy, it is crucial to promote the application of water-saving technologies and improve irrigation efficiency during policy implementation, while avoiding the irrigation efficiency paradox or Jevon’s paradox. In this way, the reduction of IRR-NPS can be facilitated [90,91,92]. Newer and smaller clusters, such as #5 Hydraulic Retention Time, #8 Stable Isotope, #10 Rice, and #11 Surface Soil Moisture, represent the emerging research hotspots as well. As shown through various HRT condition experiments in cluster #5 Hydraulic Retention Time, longer HRTs can promote pollutant adsorption or decomposition by the soil, which reduces deep water percolation and surface runoff [93]. As revealed by cluster #8 Stable Isotope research, stable isotopes play a crucial role in irrigation water quality monitoring and isotope tracing, as reflected primarily in the measurement of ion concentrations in closed aquifers (CT), the analysis of water-rock interactions, and the investigation into the impact of mineral dissolution on groundwater chemistry [94]. Cluster #10 Rice has highlighted the significant water consumption and potential for non-point source pollution in rice cultivation, alongside a massive potential for genetic improvement. According to the studies conducted in the Poyang Lake region, a 10.4% water-saving effect can be achieved in rice cultivation, with drainage water pollution reduced by 20.4% through integrated water and fertilizer management. The discovery of additive genetic variance (sigma(2)A) is significant for regulating the genetic expression of desired traits in these studies [95]. Cluster #11 Surface Soil Moisture is essential for the optimization of IRR-NPS. According to the recent research focusing on this area, irrigation events can be precisely detected by combining Sentinel-1 SAR data with the Optirrig model, which improves the efficiency of water resource management and mitigates environmental impacts. Additionally, the use of the ORCHIDEE land surface model in simulations is beneficial in understanding the limitations of these models under different climatic and landscape conditions. This is of significance to improving irrigation strategies and mitigating the negative impact of agricultural activities on water quality [96].

4. Discussion and Recommendations

Data show that surface water extraction accounts for about 86% of agricultural irrigation water, and 30–50% of the Earth’s surface has been contaminated by IRR-NPS pollution [97,98,99]. In China, agricultural nitrogen and phosphorus contribute 81% and 93% to water pollution, respectively. Water pollution has a severe impact on human health, for which there is a pressing need to control the NPS pollution caused by agricultural irrigation [100,101,102]. As shown in this study, the subjects that have gradually become the forefront research directions in IRR-NPS include precision irrigation, rapid water pollution response, spatiotemporal coordinated management, interdisciplinary integration, wastewater treatment, and crop models. To propose more precise, convenient, and practical implementation suggestions (Figure 9), the following five discussions are based on research from 2010–2024. Betweenness Centrality (BC) is a key parameter used to represent influence, and a node with high BC is considered to have significant influence and control within the network, playing a more important part in enhancing collaboration in the field. This is conducive to identifying the key figures, research frontiers, and the connections between research areas. The formula is expressed as follows:

• Select Practical, Low-Cost Precision Irrigation Strategies:

In recent years, the difficulty of implementing and promoting lightly polluted crop cultivation has been reduced by optimizing irrigation methods with low costs and environmental adaptability. This is a conclusion drawn by analyzing crop growth mechanisms, climate, soil properties, economic costs, and water-fertilizer management [103]. First, the main crops studied in IRR-NPS research include rice, wheat, corn, and cotton. Rice is exemplary in this regard. Given its uniqueness in water resource requirements, the NPS in rice cultivation increases largely due to the disappearance of wetlands and the conversion of drylands to rice fields, with temperature identified as a significant influencing factor [104,105,106]. Moreover, unlike traditional flooding, alternate wetting and drying irrigation is effective in reducing NPS, with heavy metals lowered by 13–90% [82,86]. In the research on drought climate adaptability of rice, Super Rice 9108 was developed through genetic engineering. Compared to conventional shallow water irrigation (FSI), rainwater-controlled irrigation (RC-CI), and straw-covered dry farming (DPS), the lodging index was reduced by 24% and 16%, respectively. However, the average income of RC-CI rose by 5.8%, while that of DPS shrank by 4.4%. According to research, RC-CI is the best irrigation model under severe water deficit conditions [89]. In eastern and southern Africa, three perspectives were taken to conduct an applicability analysis of rainwater harvesting and controlled irrigation: rainwater collection, residual statistics, and economic benefits. There were two strategies proposed: planting short-lived grains in non-drought rainy seasons and using perennial crops (such as cassava) during drought seasons to ensure yield [107]. Among them, shallow wet irrigation, alternate wetting and drying irrigation, and rainwater harvesting irrigation were investigated for applicability, with shallow wet irrigation determined to be highly applicable, affordable, and effective in pollution control. Adopting appropriate water-saving irrigation technology is also regarded as a promising management method for sustainable production [87,108].

As water-saving irrigation technology advances constantly, research has been conducted to reveal that precision irrigation can reduce NPS pollution by 15.8–44.1% [48]. Precision irrigation has become a key research focus in the IRR-NPS. With the development of cloud platforms, sensors, and other technologies, precision irrigation strategies are influenced mainly by water distribution and integrated management efficiency [84,109,110]. The continuous progress in data collection technology has resulted in the shift from various methods of water distribution management (such as water resource allocation schemes [111,112] and adaptive assessments [113,114]) to irrigation information classification [115,116], artificial intelligence irrigation (ACO/LIDM) [117,118], and dynamic irrigation (NSAE-ANFIS) [119,120]. With the application of multiple sensors and intelligent algorithms, it is practical to accurately reverse total nitrogen (TN) in real time, with a maximum inversion frequency of a minute scale (5 min) [121]. Meanwhile, ammonia emissions and soil salinity become controllable [122,123,124]. By evaluating irrigation adaptability under different environments [125,126], economically adaptive crop models can be selected (such as Optirrig/FTSMC) [96,127,128].

2.

Accelerate Innovative Research on Multi-Segment NPS Pollution Control Methods:

It is argued in most studies that management techniques can be divided into source control, process pollution reduction, and terminal purification, depending on the generation and development of IRR-NPS pollution [129,130,131]. In this study, the focus is on wastewater treatment and spatially coordinated governance. This is mainly because, besides the wide span and recent content of the reviewed literature, the sources of NPS pollution encompass not only agricultural activities but also livestock breeding, microplastics, rural waste, and straw leachate pollution [132,133,134,135]. Also, this study covers the development process and forefront directions of NPS pollution caused by agricultural irrigation, with different emphases.

The wastewater treatment in IRR-NPS requires pollution reduction during wastewater discharge. Based on the pattern of pollution diffusion, there are three stages involved in the treatment process: field reduction grass ditches, pond wetlands, and main ecological ditches. In field reduction grass ditches, the preliminary purification of surface and groundwater is enabled through various processes such as improved fertilization techniques, including slow-release fertilizers [136], integrated water and fertilizer technology [137], and organic fertilizer replacement technology [138]. The improvement of soil moisture requires return flow irrigation [139], rainwater harvesting irrigation [140], soil measurement techniques [141], and soil erosion control [142]. In pond wetlands, further purification takes place through plant absorption, sediment adsorption, and microbial degradation. As revealed by experiments, plant filter strips, bio-sand filters, and artificial wetlands are effective in reducing pollutant levels [143,144]. In main ecological ditches, terminal purification is conducted by directing water to discharge zones. According to research, aquatic biota sedimentation remediation (symbiotic algae) contributes to improving water quality and reducing NPS pollution [145], which indicates an innovative pathway to future watershed quality management.

In recent years, there have been some challenges presented to the collection of NPS pollution control and information due to the insufficiency of satellite image resolution, diverse data collection methods, and data lag. As remote sensing, GIS, and wireless sensor technology advance, a gradually rising popularity has been gained by their application in the settings like IRR-NPS pollution monitoring and spatiotemporal analysis. Through spatiotemporal coordinated management simulation and pollution tracking research, pollution management efforts can be effectively controlled [146,147,148]. In some cases, the existing three-stage purification interception methods are suboptimal, which is attributed to the impact of rainfall, temperature, and other factors. Therefore, it is recommendable to improve wastewater treatment technologies into multiple stages and promote technological innovation. Regarding spatiotemporal coordinated management, it is necessary to take into account the regional differences in NPS pollution characteristics and formulate targeted management plans. It is possible to improve data collection convenience, pollution controllability, and the accuracy of management plans by integrating multi-segment NPS pollution control methods with spatiotemporal coordinated management in the future. This is conducive to further promoting the purification and prevention of NPS pollution.

3.

Programming Languages Aid in Integrating Remote Sensing and Sensor Technology in IRR-NPS Strategies

IRR-NPS research indicates limited author collaboration and relatively sparse interdisciplinary connections. Since 2018, there has been a steady upward trend shown by the research conducted on programming languages, remote sensing, and sensors. Through effective integration and innovation, it is feasible to enhance the feasibility of precision irrigation, devise intelligent irrigation decision-making schemes, and develop crop model applicability. The YOLOv7 algorithm and computer vision can be leveraged to facilitate the supervised classification of coffee fruit maturity, which makes it possible to take informed decisions on irrigation, fertilization, and other timely field management measures [149]. By integrating drones, IoT, machine learning (ML), and communication technology (ICT), intelligent irrigation decision-making schemes can be developed [150]. Inversion technology combines Sentinel-1 soil moisture data with the Optirrig crop model to develop irrigation monitoring methods, with two filters used to enhance the correlation between irrigation events and rainfall [127]. For wide-area irrigation facilities and inspection status monitoring, a comprehensive service platform based on RFID, GPS, and GIS can be adopted to achieve the optimal control and utilization of agricultural irrigation nodes [151]. In terms of future interdisciplinary research, a concerted effort should be invested in the integration of advanced remote sensing technology, precise sensor data, sophisticated crop models, and cutting-edge artificial intelligence algorithms. Furthermore, this integration should be supported by fostering collaboration among a diverse group of professionals, including agricultural scientists, hydrologists, environmental scientists, data analysts, legal scholars, and policy analysts. This is purposed to promote innovation and overcome the challenges of IRR-NPS.

4.

Increase Innovative Research on Soil Improvement Measures

According to IRR-NPS research, good soil is useful for filtering pollution, improving nutrient retention, enhancing water infiltration, reducing soil erosion, and improving biodiversity. The focus of existing soil improvement research is on saline-alkali soil management and soil fertility regulation. As science and technology constantly advance, greenhouse gas emissions and soil salinization have gradually become the major concerns for sustainable agricultural development and NPS pollution prevention. Therefore, it is particularly important to carry out innovative research on soil improvement measures. In this study, a review is conducted on physical and chemical improvements alongside the engineering soil improvement technologies primarily involving water-saving irrigation techniques. In terms of soil physical improvement, the research focuses on mulching agronomy and biochar. Compared to no mulching, both plastic mulching and straw mulching are beneficial for rice cultivation as they can reduce CH₄ emissions, significantly lower the risk of global warming, improve nitrogen use efficiency, and reduce fertilizer use [152,153]. Rice straw biochar and rice husk biochar have significant effects on soil fertility improvement and pH value regulation [154]. Regarding chemical improvement measures, studies have revealed that gypsum improves saline-alkali soil by lowering pH, reducing conductivity, and increasing porosity [155]. Represented by mineral-source potassium humate, some chemically synthesized additives play a part in maintaining soil fertility and health [156]. Additionally, there is a lack of research focusing on biological improvements in IRR-NPS. Therefore, it is imperative to further innovative research on soil improvement technologies and increase research investment. In the future, more comprehensive and effective improvement results can be achieved through a comprehensive application of various soil improvement methods.

5.

Research on measures to increase collaborative use of datasets and software.

Firstly, the literature database used in this study is derived solely from the Web of Science core dataset, and there is no analysis of the literature collections from other databases such as CNKI and Scopus. Consequently, there are certain limitations. To enhance its completeness, it is necessary to expand the literature through future research for further analysis. Secondly, since CiteSpace involves macro analysis and retains core knowledge, parameter adjustment or integration with other visualization software is required to ensure the comprehensiveness and accuracy of the results [157,158]. Additionally, the use of more in-depth methods like meta-analysis is still requisite if a detailed analysis of a single cluster is considered necessary [159,160,161].

5. Conclusions

In this paper, a systematic review of IRR-NPS studies from 2010 to 2024 is conducted to demonstrate the research trends and technological innovations in the field. Despite a theoretical framework constructed for optimizing irrigation patterns, many challenges remain in practical applications. To ensure research results translate into actionable solutions for agriculture, key future directions are suggested. Precision irrigation technology has proven effective in improving water resource utilization, reducing fertilizer loss, and mitigating environmental pollution. However, high costs and complexity limit its adoption, especially for smallholders. Future research should focus on reducing costs and simplifying processes for wider use. Significant regional differences in soil, climate, and crops require adaptable irrigation strategies. Regional pilot projects can help assess technology effectiveness in specific environments, ensuring local needs are met. Government policy and financial incentives are also critical to overcoming economic barriers to technology adoption, such as subsidies, low-interest loans, and technical support for farmers. Interdisciplinary collaboration is vital for addressing non-point source pollution by integrating water management, soil science, environmental engineering, and smart agriculture. Strengthened collaboration can develop more integrated management solutions. Long-term data monitoring is essential for assessing the environmental impacts of irrigation patterns and providing evidence for future policy and technology optimization.

Overall, this study provides theoretical support for IRR-NPS control and practical guidance for technology application and policy formulation. The results are expected to contribute to sustainable agricultural development and environmental protection by reducing technology costs, promoting regionally adaptive strategies, and enhancing interdisciplinary cooperation. Future research should focus on precision irrigation, crop models, data accuracy, spatiotemporal pollution management, and integrated interdisciplinary solutions.