Review of Global Interest and Developments in the Research on Aquifer Recharge and Climate Change: A Bibliometric Approach

: Groundwater represents 98% of the world’s freshwater resources. These resources have been strongly impacted by the increases in temperature and in the variation in precipitation. Despite many advances, the relationship between climate change and the dynamics of aquifer recharge is still poorly understood. This study includes an analysis of 211 papers using the biblioshiny function in the bibliometric R Package. Additionally, speciﬁc papers were selected to identify limits, trends, and negative and positive impacts. The results indicated an average growth of 14.38% and a signiﬁcant increase in research from 2009. In total, 52 countries have undertaken studies in this ﬁeld, just over 26% of the total number of countries. In the papers examined, the worst projections related to reductions in recharge were identiﬁed for arid and desert areas; the highest recharges were identiﬁed in the northern regions and in areas at high altitudes, where recharge capacity is maintained or increases due to rapid snow and glacial melting resulting from temperature increases. Despite the advances achieved, more studies should be extended to analyse groundwater assessment at other latitudes to reach a complete and comprehensive understanding. This understanding should be one of the priorities for water among governments and the scientiﬁc community in order to safeguard this precious resource. This study revealed that groundwater research output increased signiﬁcantly during this period, with an average annual rate of 10.1%.


Introduction
Water and the Earth's climate are closely related. It is well-known that the hydrological cycle, particularly groundwater, is affected by climatic variations and the interactions between precipitation, temperature, and evaporation, as described by Brouyère et al. [1], Rivera et al. [2], Ranjan et al. [3], Aizebeokhai [4], Earman and Dettinger [5], Ertürk et al. [6], Stagl et al. [7], Green [8], Smerdon [9] and Berhail [10]. Each change in the climatic system induces a difference in the water system and vice versa. Due to these climatic variations, groundwater assessments have become essential, as groundwater represents approximately 98% of the world's freshwater resources. In many environments, natural groundwater discharges sustain baseflow to rivers, lakes, and wetlands during periods of low or no rainfall. When studying the impact of climate, understanding aquifer recharge is essential. Dragoni and Sukhija [11] defined groundwater recharge as a residual flux of water added to the saturated zones resulting from the evaporation, evapotranspiration, and runoff losses of precipitation, which can occur through diffuse infiltration, a preferential pathway, or surface streams and lakes. Thus, groundwater recharge is a sensitive function of climatic factors, local geology, topography, and land use. Groundwater recharge is affected by climate change in the reduction or increase of aquifer recharge volume. Natural recharge of groundwater occurs due to both diffuse rain-fed recharge and focused recharge via leakage from surface waters (ephemeral streams, wetlands, or lakes) and is highly dependent on Water 2021, 13, 3001 2 of 25 prevailing climate (precipitation and temperature variability), as well as on land cover and underlying geology. Concerning the impacts of climate change on aquifer recharge, Cuthbert et al. [12] explained that, in the next century, it will be harder for 44% of the world's aquifers to recharge, notably the shallower ones that we rely on to fill up faster. They expect that the effects of current climate changes on groundwater will manifest themselves over the next 100 years in most regions of the world.
Over the last 40 years, extensive research has been published on how climate change might impact aquifer recharge. The most sensitive regions will be mountains, semi-arid zones and arid zones where subtle shifts in seasonal weather timing and duration will result in significant changes. Authors like Orehova and Bojilova [13], Chen et al. [14], Candela et al. [15], Döll [16], Guardiola-Albert and Jackson [17], Habets et al. [18], Stigter et al. [19], Touhami et al. [20], Goodarzi et al. [21], Yagbasan [22], Moutahir et al. [23], Chavarría and Vargas [24], Lauffenburger et al. [25], Moutahir et al. [26], Fu et al. [27] and Haidu and Nistor [28] have demonstrated a diminution in aquifer recharge due to climate change effects and projected possible reductions in aquifer recharge in the future in their research. The opposite impact is aquifer overcharge, as described by Uribe et al. [29], Jyrkama and Sykes [30], Newcomer et al. [31], Serur [32] and Wöhling et al. [33], which happens in regions where snow predominates during winter. These cases of overcharge are typical in high northern and southern latitudes and at high altitudes. Despite the increasing number of studies on the impacts on groundwater, there is limited information on the impacts on aquifer recharge-dependent regions derived from the locations where studies are carried out and different countries' priorities.
Researchers have used qualitative and quantitative literature review methods, such as bibliometric analysis, to explain and correlate previous results. Bibliometric analysis is an emerging tool used in the analysis of scientific studies. Following Wang et al. [34], the process of bibliometric analysis effectively involves describing stages of development, providing a guide for the researcher through the body of knowledge and a reproducible workflow for the analysis. Pritchard et al. [35] have described how, initially, bibliometric analyses were applied to logarithmic and calculation procedures and several transmission publications. However, Broadus [36] and Persson et al. [37] have argued that this method actually further supports fair and consistent research, relying on mathematical evaluations of science, scientists or scientific activity to compile the results of a wide range of studies across different disciplines and investigate research trends in a particular field. Different authors have applied bibliometric analysis to groundwater, such as Niu et al. [38], who conducted a statistical and bibliometric analysis to determine the growth in global groundwater research over the past two decades. Similarly, Barthel and Seidl [39] carried out a bibliometric analysis to find how many collaborations have taken place in groundwater research. Wang et al. [40] applied a bibliometric analysis to climate change vulnerability research across 3004 published papers. Zhang et al. [41] investigated the temporal development of research on groundwater remediation from 1995 to 2015 using a bibliometric approach. These authors identified the potential obstacles and opportunities for researchers who currently work on groundwater contamination and remediation and related topics. Ma and Zhang [42] applied a bibliometric approach to investigate hotspots in research on submarine groundwater discharge (SGD) between 1998 and 2019. The results provided an instructive perspective on the present situation and future research direction on SGD. Furthermore, Jia et al. [43] investigated research trends and global interest in the sustainable development of groundwater over the last 40 years (from 1978 to 2017) using a bibliometric approach. This study revealed that groundwater research output increased significantly during this period, with an average annual rate of 10.1%.
In this study, we carried out a literature review using a network approach with bibloshiny bibliometrix, a package in the R environment. Aria and Cuccurullo [44] describe this package as being like an open-source tool programmed in the R language that can be used to undertake a comprehensive mapping analysis of scientific studies. This mapping analysis approach involves using specific keywords to link scientific publications as building blocks. In network terms, a hypothesis can be understood as a directed tie between two concepts or nodes. The network emerges by aggregating the hypotheses from a set of articles in the domain of the review. In this study, we aimed to review the research on climate change relating to aquifer recharge between 1980 and 2020. The following objectives were targeted: (O1) The first target was to identify the growth in research on this topic; (O2) The second target was to identify countries' contributions and the collaborations between countries, as well as trends and limitations and how these gaps have affected the interpretations of society and the international community's results; (O3) The third target was to determine the positive and negative impacts of climate variability on aquifer recharge.
In this study, we examined research contributions in English from the last 40 years. In the current section, we present the most important research findings on this topic and the goals of this study are specified. In Section 2 we explain how the objectives of this study were achieved. In Section 3, we present the results, first for objective O1, then O2 and finally O3. In Section 4, we explore these three goals as defined above. For a better understanding, this section is divided into four subsections: Section 4.1 examines the influence of the IPCC assessment report on aquifer recharge studies; Section 4.2 reflects on the impact of climate change on aquifer recharge; Section 4.3 looks at the overall techniques applied; and Section 4.4 examines the current status of research on aquifer recharge.

Sources and Selection Criteria
This review used an exhaustive compilation of papers, including articles in scientific periodicals and technical journals and chapters in edited volumes. These studies were initially identified via searches in a bibliographical database using Scopus and Google Scholar. This research was conducted using the following standardised queries: "climate change" "AND" "aquifer recharge". Documents included in the database were classified under the following categories: • Studies that demonstrated a negative or positive impact on aquifer recharge; this category included research featuring results presenting reductions or increases in the water table; • Research designed to present possible solutions for water stress, aquifer recharge, aquifer overexploitation and aquifer management; this included investigations that have developed ways to manage aquifer recharge.
The papers selected were analysed using biblioshiny, which is a statistical and graphical computer language in the R program. Figure 1 shows the workflow of biblioetrix analysis.
According to Aria and Cuccurullo [44], this open-source tool can be used for a comprehensive mapping analysis of scientific studies. The first step was data collection; in this step, only articles from SCOPUS (https://www.scopus.com/search/form.uri?display=basic# basic, accessed on 25 September 2020) were utilised. The uploaded data were converted from a CSV file to a data frame using bibliometrix. Then, this data frame was converted into a matrix, the specificities of which depended on the attribute selected. With the matrix created, data visualisation was possible in two ways: for contributions per year following a network analysis of the research production per country, or for collaborations between countries using statistical analysis.  According to Aria and Cuccurullo [44], this open-source tool can be used for a comprehensive mapping analysis of scientific studies. The first step was data collection; in this step, only articles from SCOPUS (https://www.scopus.com/search/form.uri?dis-play=basic#basic, accessed on 25 September 2020) were utilised. The uploaded data were converted from a CSV file to a data frame using bibliometrix. Then, this data frame was converted into a matrix, the specificities of which depended on the attribute selected. With the matrix created, data visualisation was possible in two ways: for contributions per year following a network analysis of the research production per country, or for collaborations between countries using statistical analysis.

Method for First Objective
To complete objective O1, all data were collected and classified following two paths: determining scientific production per year (papers published each year between 1980 and 2020) and selecting the most used keywords. To display the papers published per year, the bibliometrix function "annual scientific production" was used, and to display the most used keywords across the study period, the function "most frequent words" was used.
For better management of the information, these data were transformed into a data frame. Then, this data frame was converted into a matrix; this was possible because each manuscript's attributes were connected through the author(s), journal, keywords or publication data. These connections of different attributes were used to generate bipartite networks, which are represented as rectangular matrices (manuscripts x attributes). In these networks, the nodes were articles connected by links indicating citations and keywords shared between articles. For objective O1, the links were represented by year in order to

Method for First Objective
To complete objective O1, all data were collected and classified following two paths: determining scientific production per year (papers published each year between 1980 and 2020) and selecting the most used keywords. To display the papers published per year, the bibliometrix function "annual scientific production" was used, and to display the most used keywords across the study period, the function "most frequent words" was used. For better management of the information, these data were transformed into a data frame. Then, this data frame was converted into a matrix; this was possible because each manuscript's attributes were connected through the author(s), journal, keywords or publication data. These connections of different attributes were used to generate bipartite networks, which are represented as rectangular matrices (manuscripts x attributes). In these networks, the nodes were articles connected by links indicating citations and keywords shared between articles. For objective O1, the links were represented by year in order to present the growth in research from 1980 to 2020 and by keyword in order to demonstrate the most relevant words used during this period.

Method for Second Objective
Objective O2 took into account the evolution of research per country, collaborations between countries, trends, and limitations. These goals were using the following methods:

•
Articles published per country and collaborations between countries were determined by employing a network matrix plotted using descriptive statistical analysis of countries' scientific production and a collaboration word map. As was explained for the first objective, these connections of different attributes were used to generate bipartite networks that could be represented as rectangular matrices. For this item, the network was based on scientific publications that had references to other scientific publications, generating a further network. These networks were analysed to capture meaningful properties of the underlying research system, particularly in order to determine the influence of bibliometric units, such as publications by country and collaborations undertaken between countries; • The trends and the limitations in this review were identified by reading each study selected for the third objective; thus, the second and third objectives were closely linked.

Method for Third Objective
It was possible to complete objective O3 by reading the abstracts of the papers selected from the database. Only abstracts with positive or negative results for aquifer recharge were selected. These positive or negative results had to refer to increases or decreases in groundwater recharge. From the abstracts selected, 122 studies were considered.

Results
The study of groundwater recharge from a climatic point of view began in the 1980s, when the scientific community saw that the climate change was accelerating more and more rapidly. In total, 211 studies were gathered, of which 160 belonged to the Scopus search source and 51 to the Google Scholar search. After selecting the papers, we begun to complete objectives O1, O2 and O3, as described in Section 1. With regard to objective O1, it was found that research in this field showed an annual growth rate of 14.38% from 1988 to 2020 ( Figure 2). From this total growth, 79.62% of the comprehensive studies were carried out between 2011 and 2020, and of this percentage, 30.81% occurred between 2019 and 2020. The most relevant words used in this data frame were "climate change", "aquifer" and "recharge" or "recharging". In all the papers referenced in this review, the term "climate change" was the essential keyword used by the authors, followed by "aquifer recharge". Figure 2 presents the growth in total publications in the Scopus collection and Google Scholar from 1980 to 2020, and the most used keywords in the data frame. With regard to objective O2, the results showed that 52 countries (Scopus data) have contributed to creating a database of documents and information related to the effects of climate variability on the vulnerability of groundwater recharge. The geographic distribution was arranged to show the highest and the lowest numbers of studies among developed countries, developing countries and underdeveloped countries; underdeveloped countries are often the ones most affected by climate change. During the 1980s and 1990s, Australia and the United Kingdom established themselves as pioneers in this type of research. Nevertheless, climate change studies on aquifer recharge extended to developing countries after the 2000s and to underdeveloped countries after 2011. The motivation that has driven the studies in underdeveloped countries is the need to study water quality and With regard to objective O2, the results showed that 52 countries (Scopus data) have contributed to creating a database of documents and information related to the effects of climate variability on the vulnerability of groundwater recharge. The geographic distribution was arranged to show the highest and the lowest numbers of studies among developed countries, developing countries and underdeveloped countries; underdeveloped countries are often the ones most affected by climate change. During the 1980s and 1990s, Australia and the United Kingdom established themselves as pioneers in this type of research. Nevertheless, climate change studies on aquifer recharge extended to developing countries after the 2000s and to underdeveloped countries after 2011. The motivation that has driven the studies in underdeveloped countries is the need to study water quality and find new sources of supply. Figure 3 presents the contributions of different countries since 1980. This figure illustrates the countries that have been conducting studies on climate change and groundwater per decade, the legend represents the frequency of study by countries. For example, research was only carried out in Australia during the 1980s, while by the 2010s different countries on all continents have addressed the issue of climate change and groundwater.
Water 2021, 13, x FOR PEER REVIEW 7 Country collaboration during the 1980s was noted only for Australia. This rese focused on a water drainage model beyond the root zone used to predict the effec varying rainfall and evaporation rates on the discharge and recharge of aquifers an consequent levels of water tables, considering soil salinity. During the 1990s, two stu were carried out. The first was by Green and Marsh [45], who undertook research i United Kingdom and maritime Western Europe. These researchers examined impo questions concerning the resilience of current water management strategies in light o temperature increment. They showed that water management in Europe, as elsewhe underpinned by long-term stability in rainfall, runoff, and aquifer recharge patt However, this first study did not have an impact on the public and political spheres to the uncertainty associated with predictions based on global circulation models second study was carried out by Kilsby et al. [46]; these authors described the applic of an upscaled physically-based hydrological modelling system to the Arkansas Red R basin in the United States. This model was designed to be used as a tool to model the phase of the hydrological cycle and the impacts of climate variability and land-use cha Over the last 20 years, some of the collaborations that took place between coun Country collaboration during the 1980s was noted only for Australia. This research focused on a water drainage model beyond the root zone used to predict the effects of varying rainfall and evaporation rates on the discharge and recharge of aquifers and the consequent levels of water tables, considering soil salinity. During the 1990s, two studies were carried out. The first was by Green and Marsh [45], who undertook research in the United Kingdom and maritime Western Europe. These researchers examined important questions concerning the resilience of current water management strategies in light of the temperature increment. They showed that water management in Europe, as elsewhere, is underpinned by long-term stability in rainfall, runoff, and aquifer recharge patterns. However, this first study did not have an impact on the public and political spheres due to the uncertainty associated with predictions based on global circulation models. The second study was carried out by Kilsby et al. [46]; these authors described the application of an upscaled physically-based hydrological modelling system to the Arkansas Red River basin in the United States. This model was designed to be used as a tool to model the land phase of the hydrological cycle and the impacts of climate variability and land-use change.
Over the last 20 years, some of the collaborations that took place between countries, specifically the studies in which the impact of climate change on the recharge of aquifers was taken into account, aimed to develop good techniques for water resource management. Among these collaborations, the cooperation between Mexico and the United States [47] involved studying the Upper Rio Grande basin, where aquifer recharge was analysed through the biophysical, hydrologic, agronomic and economic impacts due to climate change, water supply limits and continued population growth. One collaboration between Canada, Germany and Italy [48] focused on adjusting the scale of climate models to that of basin models due to the numerous studies that have recognized the incongruities in the match between the spatial scales of general circulation models and hydrological models. Another example of collaboration is that between Mexico and Spain. Molina-Navarro et al. [49] undertook a study of the semi-arid Guadalupe River basin where aquifer recharge is expected to decrease by up to 74%, with a consequent reduction of groundwater flow. This cooperation between the authors was due to the fact that water resource management has become a challenging task in this region as it has been aggravated by vulnerability to climate change. At another latitude, in a collaboration between South Africa and Ethiopia, Ebrahim et al. [50] carried out a review of water resource management, taking into account the effects of high inter-annual climate variability and growing water demand on aquifer recharge. Figure 4 presents the seven countries with the highest levels of collaboration.
On the other hand, the trends identified with regard to objective (O2) have varied over the decades: (T1) During the 1980s, there was a trend in studies about the consequences of the increase in greenhouse gases on the agricultural sector, food production and energy; (T2) The resilience of current water management strategies in the face of very unusual climate conditions was a goal within the scientific community; (T3) Temperature could increase by up to 2 • C by the end of the 21st century, and precipitation could vary drastically at a global level, causing radical consequences for groundwater. All this based on the fact that the effects of current climate variability are an anthropogenic result; (T4) The review determined that the most significant change will be decreasing water-table levels in Mediterranean areas and semi-arid and arid regions; (T5) One of the most important trends was managing aquifer recharge as a measure to reduce hydrological and climate change vulnerabilities.
Despite all the advances shown in Figure 3, this research topic has presented severe limitations since its inception: (L1) During the study period, the first challenge for researchers was to answer the following questions: how would a doubling of CO 2 in the atmosphere impact the climate, and how could this climate affect water resources? (L2) One of the significant limitations of these scenarios was the overestimation of the radioactive forcing associated with the doubling of atmospheric CO 2 and, therefore, the underestimation of the equilibrium climate sensitivity to doubling CO 2 . These limitations were reflected in the inexactness of the parameterisation, the lower resolution used to represent the study areas and the lack of studies and information about the relationship between climate and aquifer recharge; (L3) The limitations in these years were evident in the fact that the spatial resolution of the models was not always compatible with the hydrologic models; as such, the global climates models could not be directly used at local or regional scales. The global climate models did not account for small-scale, regional and local processes due to the large grid sizes incorporated; (L4) The most hindering factor for climate scenarios was their interpretation (and misinterpretation) by society. It was argued that these would only achieve public and political credibility when they could demonstrate coherence over time with the hydrological consequences captured by climate programs; (L5) Despite the creation of a model, the variation in the water table was not explored more extensively. One significant gap was the lack of data concerning water-table fluctuation, which limited the comprehension of natural aquifer recharge. This limit is standard in developing and underdeveloped countries; (L6) Concerning the program used to respond to goals O1 and O3, the most significant limitation involved the visualisation and interpretation of the results shown in Figure 3. This figure shows the countries where research on aquifer recharge and climate change has already been undertaken; nevertheless, the program shows all countries, which may confuse the reader. The program does not show the specific aquifer or the exact region studied, but only shows all countries; (L7) One of the most significant practical limits was the lack of interest and commitment on the part of governments and rulers with regard to investing resources in groundwater research; in such regions, the quality and quantity of water is not a national priority. Additionally, examinations of the land surface component of climate modelling typically neglected groundwater. Smerdon [9], following the research carried out by Green et al. [51], argued that the lack of groundwater observations not only makes it difficult to inform management decisions but also limits the scientific understanding and evaluation of climate and hydrologic simulation models; (L8) Concerning this research, one important limitation is that this study only examined English publications, thus excluding publications about this topic in other languages. However, we took into account several studies undertaken in different geographic zones.
The positive and negative impacts described with regard to objective O3 indicate whether climate change is causing a reduction or increase in groundwater recharge. These impacts are measured through two climate parameters (precipitation and temperature) and one physical parameter (evaporation or evapotranspiration on the soil surface). These three variables are fundamental in determining an aquifer's capacity to be supplied after a period of rain or during the most optimal seasons for recharging (see Section 1, second paragraph). With increasing temperatures, evaporation increases, which results in soil drying. The most sensitive regions with negative impacts are mountains and arid zones, where subtle shifts in the timing and duration of seasonal weather influence recharge significantly. According to Tuinhof et al. [52], in arid and desert zones, climate variability has a substantial impact on rainfall and drought. The opposite situation is aquifer overcharge (which can be considered a positive effect in terms of the volume of water that could reach the saturated zone); this happens in regions where snow predominates during winter. These overcharges have been measured in high northern and southern latitudes, as well as in areas at high altitudes, where the temperature is projected to increase. Such temperature increases melt the snow and glacial ice in a shorter period time, adding additional water to aquifers. Through this review, it was possible to determine that climate change is not the only challenge facing groundwater. In addition to climate variability, the overexploitation of hydrogeological resources affects the capacity of aquifer recharge through water consumption either by populations or in agricultural or industrial activities. All these effects are discussed in Section 4. Table 1 shows the most significant studies concerning to increase in recharge and decrease recharge for North America and the Caribbean. Table 2 presents studies found for South America, Table 3 for Africa, Table 4 for Asia, Table 5 for Middle East, Table 6 for Europe and Table 7 for Australia and Oceania.
vulnerability to climate change. At another latitude, in a collaboration between South Af rica and Ethiopia, Ebrahim et al. [50] carried out a review of water resource managemen taking into account the effects of high inter-annual climate variability and growing wate demand on aquifer recharge. Figure 4 presents the seven countries with the highest level of collaboration. On the other hand, the trends identified with regard to objective (O2) have varied over the decades:

Influence of the Intergovernmental Panel on Climate Change (IPCC) Assessment Reports on Aquifer Recharge Studies
The creation of the first global climate model (GCM) in the 1960s and 1970s by Sawyer [145] and Broecker [146] boosted the establishment and improvement of new institutions for the investigation of climate change focused on the increase in greenhouse gases. From this time, the scientific community began to analyse what was happening with the climate, including the effects of greenhouse gases on the climate, global warming and the effects of carbon dioxide in the atmosphere. At this time, the National Aeronautics and Space Administration (NASA) undertook its first work on climate modelling with Hansen et al.'s [147] research. This research deployed methods created in the 1970s, using a simple energy-balance model to project future warming. Seven years later, the same team presented a new global climate model dividing the world into discrete grid cells of 8 degrees latitude by 10 degrees longitude. This new GCM included aerosols and various greenhouse gases in addition to CO 2 , as well as essential cloud dynamics. Despite all the water problems at this time, these first studies were focused on measuring the impact of climate change on the energy sector, agricultural production and the demand for water for human consumption.
In this decade, the most crucial international program in the evaluation of scientific knowledge related to climate change was created. The creation of the Intergovernmental Panel on Climate Change (IPCC) in 1988 opened the doors for studies and essential information about the effects of climate change on hydrogeological resources. The first report was published in 1991 and integrated four scenarios (A, B, C and D) according to carbon dioxide concentrations. In this first report, scenario A presented a case with few or no measures to curb emissions, while the other three scenarios were based on increasing control levels. In 1992, the United Nations held the Conference on Environment and Development in Rio de Janeiro where the IPCC presented six more scenarios (IS92a-f), considering future greenhouse gas and aerosol precursor emissions based on assumptions concerning population and economic growth, land use, technological changes, energy availability and fuel mix during the period between 1990 and 2100. The Second Assessment Report was presented in 1995 and the third IPCC assessment report in 2001, showing four new scenarios (A1, A2, B1 and B2) that were used to simulate the future of groundwater according to the different scenarios for the weather. These scenarios were classified following the next criteria. Scenario A1 considered the rapid economic growth, high population growth by 2050, and efficient technologies. Scenarios A2 describes a very heterogeneous world based on self-sufficiency and conservation, continuous population growth, and slow technological evolution. Scenario B1 considers the rapid population growth by 2050, but with an economy based on services and information. Introduction of clean technologies with practical resources, green economy, natural resource management and climate sustainability. Scenario B2 is based on slow population growth and economic, social, and environmental sustainability. In 2007, the IPCC presented a fourth report, moving from the previous scenarios to Representative Concentration Pathways (RCPs). In 2014, the penultimate report was published, resulting in an increase in studies from all parts of the globe. Figure 5 illustrates the yearly scientific production for the four decades from 1980 to 2020. Figure 5 shows the evolution of the growth in research carried out on this topic, but it is clear that there is a relationship between the scientific production and the IPCC assessment reports. Evidence for this can be observed between 2007 and 2014, when the fourth and the fifth reports were published. As well as these advances, from the mid-1980s the scientific community began undertaking studies on the possible effects of climate change on groundwater resources based on the scenario involving an increase in greenhouse gases. In agreement with Smerdon [9] and Green et al. [51], between approximately 2000 and 2009 the complexity of modelling increased to represent coupled processes and the growing realization that multiple GCMs were needed due to their varying predictions. Unquestionably, this trend of increasing complexity has continued to the present day. During this time, other studies were devoted to testing methods and models based on climate data with the objective of investigating the impacts on water resources, like Changnon et al.'s [148] research. The IPCC was one of the first to develop possible climate change scenarios for that decade, but different researchers also started to improve and develop climate scenarios in order to undertake projects and studies, such as Vaccaro [149], who used global climate models including the Goddard Institute for Space Studies model, the Geophysical Fluid Dynamics Laboratory model, and the Oregon State University model. Three of these models were based on average monthly changes, and the maximum water deficit effect was also used. This first aquifer recharge study highlighted one of the essential conditions required to fill aquifers: the threshold of accumulated precipitation in the soil moisture reservoir. The most notable impact of the IPCC reports was a remarkable growth in papers after the Fourth Assessment Report.  Figure 5 shows the evolution of the growth in research carried out on this topic, but it is clear that there is a relationship between the scientific production and the IPCC assessment reports. Evidence for this can be observed between 2007 and 2014, when the fourth and the fifth reports were published. As well as these advances, from the mid-1980s the scientific community began undertaking studies on the possible effects of climate change on groundwater resources based on the scenario involving an increase in greenhouse gases. In agreement with Smerdon [9] and Green et al. [51], between approximately 2000 and 2009 the complexity of modelling increased to represent coupled processes and the growing realization that multiple GCMs were needed due to their varying predictions. Unquestionably, this trend of increasing complexity has continued to the present day. During this time, other studies were devoted to testing methods and models based on climate data with the objective of investigating the impacts on water resources, like Changnon et al.'s [148] research. The IPCC was one of the first to develop possible climate change scenarios for that decade, but different researchers also started to improve and develop climate scenarios in order to undertake projects and studies, such as Vaccaro [149], who used global climate models including the Goddard Institute for Space Studies model, the Geophysical Fluid Dynamics Laboratory model, and the Oregon State University model. Three of these models were based on average monthly changes, and the maximum water deficit effect was also used. This first aquifer recharge study highlighted one of the essential conditions required to fill aquifers: the threshold of accumulated precipi-

Impact of Climate Change on Aquifer Recharge
In recent years, groundwater has taken up a critical position in the concerns of the scientific community. However, what are the advances that have been made so far? The first point to note is that this resource is widely recognized as a precious public resource, above all in arid and semi-arid areas. This is case on the European continent, specifically in the Mediterranean regions. In these regions, groundwater is an important resource due to the climate conditions and high water consumption. As well as these factors, Haidu and Nistor [30] and Santoni et al. [114] showed that projections and scenarios present the worst future yet: a decrease in precipitation, variability in recharge potential and changes in inter-annual climate variability. This situation can be clearly seen in Portugal, where Oliveira et al. [118] estimated a recharge decrease of 45% in 2007. In another case in Spain, described by Aguilera and Murillo [117], the percentage of recharge obtained from precipitation decreased progressively throughout the 20th century, which may indicate a significant decrease in available underground resources.
The work of other authors, such as Borji et al. [109], is in agreement with these observations. Borji et al. argue reduced precipitation as a consequence of vegetation cover destruction and land-use change can result in a reduction in aquifer recharge. This reduction in precipitation affects the recharge of the aquifer when rainfall decreases, and the courses of rivers can change, which can affect the water table. Other effects that have been observed include higher evapotranspiration, marine intrusion into coastal aquifers and variations in water tables. Marine intrusion into coastal aquifers is a negative impact that is already affecting various places. According to Bloetscher et al. [56], increases in the groundwater level resulting from sea level rise can create significant challenges for recharge of island and coastal aquifers. One example of the impact of climate change and sea level rise is Thailand. For this region, Srisuk and Nettasana [104] argued that reductions in the groundwater head caused by lower rainfall could aggravate the impacts of sea level increases, unleashing effects such as reductions in groundwater recharge rates, flow and discharge and possible increases in aquifer temperatures, which may increase the levels of bacteria, pollution from pesticides, nutrients, and heavy metal contamination, to name just some secondary effects. Similarly, increased flooding could increase the flushing of urban and agricultural waste into groundwater systems, especially into unconfined aquifers, and further deteriorate groundwater quality. All of these consequences are occurring not only in Thailand but in all vulnerable coastline and island (Oceania and the Caribbean) areas.
On the other hand, Gremaud and Goldscheider [150] observed that in regions where snow predominates during the winter, over-recharge of aquifers could occur as a result of increasing temperatures and thus cause faster melting of snowy and glacial areas. Particular regions in the Alps, Himalayas and near the Andes will undergo increases in recharge. These events will also have other consequences, such as increased water tables, more moisture in soils, and more significant river flows. Besides melting ice, other events can imply an increase in recharge. For example, in Niger, a change in land use has changed the aquifer recharge amount. For this case, Favreau et al. [81] have shown that the water table in southwest Niger increased continuously from 1963 to 2007, resulting in an increase of four metres despite a 23% deficit in monsoonal rainfall between 1970 and 1998. An increase in aquifer recharge has also been identified in Chile; this research, carried out by Uribe et al. [29], indicated that recharge reached up to 6% of the annual precipitation.
In the context of these aquifer effects, different countries are affected by the same transboundary aquifers, which can sometimes be a limitation for decision making concerning water management. Some transboundary aquifers are heavily affected by climate change, as is the case for the six riparian states Ghana, Burkina Faso, Benin, Cote d'Ivoire, Mali and Togo. McCartney et al. [88] found that groundwater recharge can be expected to decrease for this region, which would have serious consequences for the rural poor, food security and economic growth in the riparian countries. However, there are important transboundary aquifers for which there are no in-depth studies about climate change impacts; for example, the Guarani aquifer. Research on the impacts of climate change on the water resources of the Guarani aquifer is limited. Studies so far have shown an incremental increase in precipitation, raising the risk of flood and probably affecting the aquifer water level. The opposite phenomenon has been observed in the groundwater resources of the Mekong. According to Cooley et al. [77] and Bui et al. [103], groundwater resources in the delta will decline during the 21st century. However, according to Jayakumar and Lee [151], the literature presents limited information on the groundwater resources of the Mekong basin, and detailed studies have not been conducted up to now.
Another impact on aquifer recharge is the impact of human actions, as described by Díaz-Cruz and Barceló [152], Shah [99], Akbarpour and Niksokhan [153], Borji et al. [109], Urrutia et al. [76], Running et al. [74] and Chitsazan et al. [112]. This anthropogenic impact can affect the recharge of aquifers as a result of modifications in land use. For example, in the Sahel, a semi-arid region between the Sahara to the north and the Sudanese savannah to the south, Seguis et al. [80] found that the anthropogenic effect has resulted in a transition from a wet period to a dry period. This was caused by land-use modification at the same time as the occurrence of superficial currents, causing an increase in runoff of 30 to to70%, producing more significant infiltration and a continuous increase in the water table. In other work, paleoclimatic studies have demonstrated vigorous aquifer recharge, and more significant rainfall has been simulated and identified as a consequence of past climate changes events (see Kulongoski et al. [55]). In general, different factors can cause groundwater degradation, such as population growth and overexploitation of groundwater, especially for irrigation and specifically in areas of high evapotranspiration. Other factors that affect hydrogeological recharge include the presence of cultivated areas, modifications of cultivation patterns, changes in the industrial water consumption pattern, the absorption of water per capita and devastation of natural resources. In addition to these modifying patterns, anthropogenic effects can result in substantial modifications to the water table and groundwater quality. The study of the impact of human beings is thus necessary for the interpretation of recharge scenarios, especially during the current century in which, according to Wöhling et al. [33] and Khalaj et al. [111], anthropogenic factors will play an essential role in the control of the hydrological regime.

Overall Techniques Applied
Over 40 years of studies and research development, aquifer recharge has been measured in different fields and with different methods and tools. Many of these tools are primarily comprised of mathematical and statistical models that have evolved over the years in order to be able to adjust more and more effectively to reality and places of study. Therefore, positive behaviour can be observed in the development of tools for the study and analysis of this field. Parts of the studies analysed in this review were carried out using hydrological or hydrogeological models, empirical statistic models and evapotranspiration models, such as the work by Uribe et al. [29] and Monterroso-Rivas et al. [64]. Other used ecohydrological models that link climatic variables to groundwater level, such as Chen et al. [63], Schuol et al. [84] and Wilcox and Thurow [154]. Methods from other fields have also been applied for this topic, such as chemistry methods. McIntosh and Walter [155], Ma and Edmunds [156], Adiaffi et al. [157], Schachtschneider and February [85], Van Hengstum et al. [158], McCartney et al. [143], Alidina et al. [159], Szucs et al. [160], González-Acevedo et al. [66], Niinikoski et al. [161], Lauffenburger et al. [25], Santoni et al. [114], Bahir et al. [91], Libera et al. [162], Bam and Bansah [93] and Hamdi et al. [163] have applied techniques such as isotopic analyses and hydrochemistry. A specific and interesting field often drawn on in groundwater studies in recent years is that of geographic information systems; in this domain, researchers such as Mango et al. [86], Baruffi et al. [123], Fiseha et al. [127], Braca et al. [132], Kolusu et al. [92], Zhang et al. [83], Fathi et al. [164] and Ajjur and Mogheir [165] have used interface tools between geographic information systems (GISs) and remote sensing. According to Green et al. [51], satellite remote sensing (RS) undoubtedly represents the most powerful method for detection and monitoring of environmental and climate change at a truly global scale.
In addition to the methods mentioned above, methods like time series and setting the minimum and maximum temperature and precipitation using a global climate model, as described by Sulis et al. [48], have been carried out for public water management. Moreover, analyses of time series of the fluctuations of the water table, empirical models of precipitation and recharge, non-parametric studies based on the Mann-Kendall test and programming languages have been employed by Obergfell et al. [166], Pulido-Velazquez et al. [71] and Malekinezhad and Banadkooki [110]. In this same domain, Perrin et al. [101] and Baruffi et al. [120] improved quantitative methodologies by adapting them to regional contexts, rendering them capable of assessing water resources at the watershed scale, as well as the impact of management measures. In the field of mathematics, Drumheller et al. [167], Masud et al. [140], Magallanes-Quintanar et al. [73], Haj-Amor et al. [168] and Mwetulundila and Atangana [169] have applied algorithms and mathematical models to study the flow of water within a fractured aquifer with permeable and impermeable rock. Other research has taken into account the evolution of water table fluctuations as an indicator of climate change consequences; for example, Tedd et al. [122], Callegary et al. [58] and Veiga et al. [126]. Studies taking into account extreme events, such as the consequences of negative human impacts, have been carried out by Trásy et al. [170], Hirata and Conicelli [78], McCartney et al. [88] and Trásy et al. [170]. Finally, in the last forty years one of the most important and commonly used programs for modelling groundwater flows was developed: modflow. In the 1990s, this software was mainly used for analysing groundwater flow and contaminant transport models under different conditions, as described by Hariharan and Shankar [171]. However, in the last two decades it has been used to evaluate the impact of climate change on aquifer recharge, as described by Woldeamlak et al. [172], Mas-Pla et al. [121], Molina et al. [125], Regnery et al. [173], Toure et al. [90], Malekinezhad and Banadkooki [110], Al-Maktoumi et al. [174], Chunn et al. [175], Ostad-Ali-Askari et al. [176], Liu et al. [177] and Xiang et al. [178]. Among the more commonly used techniques and tools, geographic information systems, modflow and hydrological models were identified here.

Current Status of the Evaluation of Aquifer Recharge and Water Resources Management
Among the alternatives to climate change effects, Díaz-Cruz and Barceló [152], Regnery et al. [173], Palma et al. [65], Fabbri et al. [129], Rupérez-Moreno et al. [179], Drumheller et al. [167], De Giglio et al. [180], Salameh et al. [181] and Mwetulundila and Atangana [169] have suggested that artificial recharge is a viable option to secure water sources for the future. For this option, direct and indirect techniques can be used, similar to those described by De Giglio et al. [180] and Ward and Pulido-Velazquez [47], for which aquifer recharge depends on the hydrological characteristics of the specific area, such as climate, geology, exploitation and the use of the land in the region. Some researchers, such as Gonzalez-Serrano et al. [182] and Faramarzi et al. [108], have suggested that wastewater treatment and artificial irrigation are appropriate options for regions with seasonal stress. Another solution is to adopt more efficient irrigation technologies to reduce valuable return flows and limit aquifer recharge. However, achieving real water savings requires designing institutional, technical and accounting measures that accurately track and financially reward water depletion reduction. It is important to note that, despite the scientific community's interest in finding viable ways to manage aquifer recharge from an artificial point of view, this does not provide security for the water balance.
Another way to achieve energy effectiveness and climate change mitigation, according to Alqahtani and Sale [183], is to reduce energy consumption related to pumping groundwater. In order to achieve sustainable management of the recharge of aquifers, Bam and Bansah [93] and Hejazian et al. [139] have stated that it is essential to identify the sources of water that replenish the aquifer in watersheds dominated by irrigation farming and to determine the most vulnerable areas that may be impacted by human activities and climate change. Alternatively, Stuyfzand and van der Schans [131], have presented a trend among studies conducted in the current decade that seek to quantify groundwater as part of the technique known as managed aquifer recharge (MAR). Authors like Tuinhof et al. [52], Díaz-Cruz and Barceló [184], Ward and Pulido-Velazquez [47], Bloetscher et al. [56], Tapsuwan et al. [141], Ward and Dillon [185], Kendon et al. [124], Ferrant et al. [186], Gnadlinger [187], Perales-Momparler et al. [188], Fisher-Jeffes et al. [82], Salameh et al. [181], Cruz-Ayala and Megdal [75] and Zhang et al. [189] agree the MAR is one of the promising groundwater engineering approaches to water security. This trend is due to the increases in population, economic development and rising demand for water from the agricultural and industrial sectors, for which supply and demand is a significant problem; this can be found, for example, in Africa. In this continent, climatic variability results in unreliable and uncertain water availability and contributes to water insecurity, particularly in arid and semi-arid areas where water storage infrastructure is limited.
According to Ebrahim et al. [50], MAR, which consists of the purposeful recharge and storage of surface runoff and treated wastewater in aquifers, serves various purposes, one of which is to provide a means to mitigate the adverse impacts of climate variability. Despite an unclear scope for this technology in Africa, the prevalence and range of MAR experiences in Africa have not been extensively examined. However, MAR includes recharge planning and the control of surface runoff to mitigate the adverse impact of climate variability. Among the risks, Díaz-Cruz and Barceló [184] mention the possible contamination of aquifers from organic micro-pollutants, and salinization of the water could be one of the secondary effects that the artificial recharge of aquifers could cause.
Despite advances in this field, there are many regions yet to be studied, including geographic areas such as Central America, Central Asia, the South Pacific and several African countries. In the case of Central America, this region needs to improve aquifer recharge from the point of view of hydroclimatic variations. This region is strongly affected by droughts and the effects of El Niño and La Niña phenomena, which are becoming more oppressive and intense as a consequence of global warming. According to França et al. [190], some of the most extreme hydrological events were associated with La Niñainduced changes in precipitation and river flow in 1989, 1999, 2009 and 2012. African countries must ensure the sustainability of aquifers since much of the population living on this continent depends on groundwater to live. Asian countries should increase the number of studies undertaken and apply methods to investigate the recharge and management of aquifers; this is urgent due to the high density of the population, social problems, droughts and water scarcity. The small islands of Oceania and the Caribbean, destined to suffer the worst consequences of rising sea levels, must consider recharge and develop research on aquifers because aquifers in this part of the planet are the largest sources of freshwater for the inhabitants. In this context, Holdin et al. [68] carried out research on 43 small islands distributed worldwide, of which 44% were in a state of water stress. Recharge was projected to increase by as much as 117% on 12 islands situated in the western Pacific Ocean and Indian Ocean and projected to decrease by up to 58% on the remaining 31 islands.
To conclude this section, climate change is a strong threat to aquifer recharges. Climate variability is causing more intense and frequent cyclonic storms, such as hurricanes, cyclones and typhoons, with more extreme events expected in regions already affected by tropical cyclones, including Central America and the Caribbean, East Africa, most of Asia, Australia and the Pacific islands.

Conclusions
This bibliometric study analysed a total of 211 studies, from which research trends, contributions of and collaborations between countries and the impacts of climate change on aquifer recharge were identified. The trends showed that arid and semi-arid regions are the most vulnerable to climate change, the effects of which are reflected in the decreases in piezometric levels and in recharge due to variability in precipitation. Undoubtedly, the creation of the IPCC in 1988 and the reports published about climate change during the last 30 years have promoted the study of groundwater, the last two decades being essential in the development of this subject.
This bibliometric review shows that the impact of climate change on the recharge of aquifers is reflected in decreases in the water table due to temperature increases, rainfall variability and evapotranspiration increases. The second effect identified is the overrecharge of aquifers. This over-recharge is an effect of melting glaciers and precipitation increases at high altitudes and high latitudes. Another effect of climate change identified is saline intrusion due to sea level rises, which pollutes freshwater and increases the piezometric level. This research presents the most important studies published in English; however, at the same time, this is a limitation, as studies published in other languages were not taken into account. Thanks to the studies conducted over the past decades, it has become possible to integrate these effects into the water management of some regions and create policies to help protect groundwater.
Author Contributions: All authors have worked, read and agreed to the published version of the manuscript. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.