1 Introduction

Managing water, energy, food, and other ecosystem services to eradicate hunger, improve health, and build a sustainable and desirable economy is complex. In addition to the complexity within the individual resource base, the interaction and interdependence between these domains are even more complex, with different manifestations at different spatial and temporal scales (Correa-Porcel et al. 2021; Pahl-Wostl 2019). More than half of the Sustainable Development Goals (SDGs) are related to water, energy, food, and ecosystem services and have competing objectives (Papadopoulou et al. 2022). Understanding the interdependence and interactions of the water–energy–food–ecosystem is the first step to achieving sustainable development goals (Sánchez-Zarco and Ponce-Ortega 2023).

However, in many parts of the world, the sectoral approach is the dominant resource management. This method, established at all levels of government, (Terrapon-Pfaff et al. 2018) creates separate community-based structure and fosters silo thinking, addressing only one issue at a time. While sector-specific focus can aid in setting and monitoring development goals, it often overlooks the interconnectedness of water, energy, food, and ecosystem services. This disregard for interdependencies across the communities, regions, or countries can hinder broader development objectives (Pahl-Wostl 2019).

The nexus approach, which is the intersectoral approach, between water–energy–food and ecosystem services, is proposed as a critical framework for addressing the silo thinking and sectoral challenge. This framework is crucial to sustainable WEFE security at different scales, spanning from local through national to global (D’Odorico et al. 2018). In the last couple of years, many studies have been conducted to clarify the WEFE concept, indicators, and approaches (Bieber et al. 2018; Endo et al. 2017; Mahlknecht and González-Bravo 2018; Weitz et al. 2017). Recently, the effort has been extended to the use of WEFE nexus for real application on the ground and has brought multiple benefits (Abdelzaher et al. 2023; Kedir et al. 2022; Markantonis et al. 2019; Purwanto et al. 2021).

The first step in developing an integrated resource management framework is to identify important and key nexus interactions that prevail in a particular area of interest and context. Recognizing interdependencies among domains enables stakeholders to develop effective frameworks or tools for managing challenges and trade-offs. Identifying key nexus interactions is crucial for informing policymakers, researchers, and stakeholders, facilitating the development of approaches that minimize trade-offs and enhance synergies across various links.

The Tana-Beles Sub-basin exemplifies the challenges of managing natural resources with multiple stakeholders and sectoral interests. Identified as a growth area due to its potential for irrigation, hydropower, high-value agriculture, and ecotourism, the sub-basin faces significant development pressures. Large-scale projects like the Koga (7,000 ha) (Eriksson 2012), Ribb (19,925 ha) (Eriksson 2012), and Megech (7314 ha) (MEE 2010) irrigation dams alongside Tana-Beles inter-basin water transfer hydropower dam (485 MW) (Annys et al. 2019) exemplifies this. Studies report that these projects are inducing widespread wetland loss, deforestation, biodiversity decline, declining soil fertility, erosion, sedimentation, water pollution, and aquatic ecosystem degradation (Eguavoen and Tesfai 2012; Mequanent et al. 2021; Tewabe 2020). Furthermore, overpopulation, land degradation, and urbanization exacerbate these issues (Admas et al. 2022; Dile et al. 2018; Karlberg et al. 2015). The intricate relationships between these challenges necessitate a nexus approach, considering the interconnectedness of water, food, ecosysetm and energy security. A nexus approach can potentially optimize resource use, mitigate environmental impacts, and achieve sustainable development in the sub-basin. Existing research on hydrology, landscape dynamics, and water ecology within the sub-basin provides a valuable foundation for identifying key nexus interactions and formulating solutions.

The primary aim of this study is to comprehensively analyze existing nexus interactions within Tana-Beles Sub-basin. It seeks to establish a framework that guides enhanced resource management in the basin by leveraging available literature and spatial data layers. Specifically, the objectives are: 1) investigate WEFE nexus interactions present in the basin, exploring drivers, trade-offs, and synergies through a systematic review of literature; 2) examine spatial distribution of WEFE nexus interactions within the basin, emphasizing on drivers, trade-offs, and synergies; 3) conduct a stakeholder analysis to identify key stakeholders and propose optimal modes of engagement. The study will help frame the key nexus interactions in the basin and identify the key trade-offs and synergies to inform the development of sustainable natural resource management strategies in the basin.

2 Methodology

The two-fold methodology was used to identify critical WEFE nexus interactions in the Tana-Beles Sub-basin, a systematic literature review and spatial analysis by producing spatial data layers from secondary documents. The systematic review followed a three-step process: (1) identification of relevant articles from research databases, (2) screening of identified articles based on pre-established exclusion criteria, and (3) systematic review of the selected articles. For the spatial analysis, the approach involved: (1) generating spatial maps of WEFE indicators from secondary data sources and (2) overlaying these spatial maps to identify and analyze nexus interactions. Detailed descriptions of these methods are presented in the subsequent sections.

2.1 Study area

Tana-Beles sub-basin is one of the largest sub-basins of the Blue Nile and covers an area of 29,200 km2. The Sub-basin extends between 10.5° N and 12.78° N latitude and from 35.1°E to 38.2°E longitude. It comprises two topographically disconnected sub-basins: Lake Tana and Beles, both draining towards the Abay River (Blue Nile) (Fig. 1). Lake Tana sub-basin features a significant elevation range, from 1785 to 4100 m a.s.l receiving an annual precipitation of 1280 mm and maintaining an average temperature of 22 °C (Abebe et al. 2017; Goshu and Aynalem 2017). In contrast, the Beles sub-basin exhibits a warmer and sub-humid climate with average temperatures increasing southward, reaching more than 30 °C at the southernmost tip. Rainfall in Beles varies between 900 and 1400 mm annually (Weldegerima et al. 2023). This diversity in climate defines the unique landscapes and ecosystems across the Tana-Beles Sub-basin.

Tana-Beles Sub-basin is home to a diverse array of socio-economic activities and environmental challenges. With a population exceeding four million individuals, the region relies predominantly on agrarian pursuits with subsistence farming and livestock husbandry serving as the primary sources of livelihood (Assefa et al. 2023). Agriculture in the area is characterized by smallholder farming practices, often conducted on marginal lands, leading to issues such as soil erosion, deforestation, and land degradation (Assefa et al. 2019). The sub-basin grapples with water scarcity and inefficient water management practices, exacerbated by climate variability and recurrent drought events (Yimam et al. 2021). This contributes to persistent poverty and food insecurity in the region (Taye et al. 2021). Initiatives aimed at mitigating these challenges involve promoting sustainable agricultural practices, enhancing water resource management, and bolstering resilience to climate change impacts through community-based adaptation strategies (Ayalew et al. 2022).

The Tana-Beles sub-basin exemplifies the complex interdependencies within the WEFE framework. Energy availability and supply are dominated by reliance on traditional biomass fuels like firewood and dung, contributing to deforestation (Karlberg et al. 2015). Regarding food yield and supply, agricultural productivity is constrained by climate change, land degradation, and limited access to modern inputs and techniques (Assefa et al. 2020a, b; Bizimana et al. 2023). Interventions aimed at enhancing food security include promoting climate-resilient agricultural practices and improving market linkages for smallholder farmers (Stein et al. 2014). Ecosystem services such as water provisioning and soil fertility maintenance underpin agricultural productivity but are threatened by land degradation and habitat fragmentation (Aligas et al. 2023). Integrated approaches that promote sustainable resource management and enhance resilience to environmental stressors are essential for addressing these challenges and improving livelihood opportunities for local communities.

2.2 Systematic literature review approaches

2.2.1 Literature sourcing

Since WEFE nexus research is an emerging field, studies on land and water resources in the Tana- Beles Sub-basin have predominantly focused on these elements in isolation, with few explicitly employing a WEFE nexus framework to analyze their interconnectedness. Consequently, we employed a proxy approach, collecting literature that explored interactions between individual or composite WEFE indicators via keyword searches in Scopus, capturing the interplay of various WEFE elements. The search employed two types of search terms: 1) terms based on the names of the principal sub-watersheds within the Tana-Beles Sub-basin, such as “Ethiopia” AND “Lake Tana’’ OR “Beles” OR “Gumera” OR “Megech” OR “Dirma” OR” Gelda” OR “Gilgel Abay” OR “Enfranz” OR “Rib” OR “Gemero” OR “Beles” and 2) terms related to the specific WEFE components relevant to the sub-basin, such as “Nexus” OR “WEFE” OR “WFE” OR “Food” OR “Water” OR “Energy” OR “Ecosystem” OR “Environment” OR “water volume” OR “water quality” OR “hydrology” OR “water balance” OR “climate change” OR “Livelihoods” OR “Agriculture” OR “aquatic ecosystem” OR “groundwater” OR “human impact” OR “siltation” OR “hydrochemical” OR “physicochemical” OR “phytoplankton” OR “zooplankton” OR “evapotranspiration” OR “river inflow” OR “fauna” OR “flora” OR “watershed” OR “catchment” OR “basin” OR “water resource” OR “irrigation”. This search strategy yielded 509 articles (Fig. 2).

Fig. 1
figure 1

Location of Tana-Beles Sub-basin

Full size image
Fig. 2
figure 2

Article search and filtering process employed in the systematic review process

Full size image

2.2.2 Literature screening criteria

Articles retrieved from the search process were screened for inclusion based on predefined criteria, assessing their relevance to Tana-Beles Sub-basin nexus interactions considering spatial (study area), conceptual scope, and multiple nexus indicators (Fig. 2). Spatial scope relevance ensured studies were conducted within the Tana-Beles Sub-basin or provided relevant sub-basin insights. Conceptual scope relevance was assessed to see if studies offered pertinent information on WEFE elements within the sub-basin. Studies focusing on phenomena beyond WEFE were excluded. Articles deemed relevant underwent a final evaluation step, assessing consideration of multiple (at least two) nexus interactions. This involved examining evidence of relationships between nexus elements. Articles solely focused on a single WEFE element or lacking information on multiple nexus interactions were excluded. The screening process was conducted manually by reviewing article abstracts. This resulted in a selection of 102 articles (Fig. 2) relevant for further analysis. These articles then proceeded to the data extraction and systematic review stages.

The selected articles were passed to database creation steps, which consisted of two steps: bibliometric data extraction from Scopus and manual study characterization using the Nexus framework. Metadata extraction utilized Scopus’ automated metadata characterization feature, exporting the saved study list to a CSV file. Study characterization employed a predefined data registration template, assigning unique IDs based on Scopus metadata. Nexus-relevant information was recorded in designated columns which contained WEFE elements, specific indicators, and methods.

2.2.3 Systematic review analysis approach

Bibliometric analysis was conducted using the bibliometrix R package (Aria and Cuccurullo 2017) on metadata extracted from Scopus. The analysis included: (1) an overall publication overview, providing basic information such as period and temporal trend, (2) a conceptual structure map of articles; generated using the top 300 keywords (Aria and Cuccurullo 2017). While producing the conceptual structure, keywords were manually normalized to identify synonyms and exclude place names and common technical terms. The conceptual structure graph categorizes studies into research themes based on the keyword co-occurrence network. We utilized the Walktrap algorithm to generate the keyword conceptual structure graph. Notably, Walktrap offers significant advantages for bibliometric analysis through its ability to precisely identify and group interconnected keywords and automatically determine optimal cluster numbers (Yang and Leskovec 2012).

2.2.4 Analysis of key drivers

The drivers represent recent or ongoing changes, both anthropogenic (human-induced) and natural, that are hypothesized to influence WEFE within the region. We employed a content analysis approach to identify these drivers manually from the objectives and result parts of the selected studies. Then, we analyzed the co-occurrence of these drivers with specific WEFE components. This involved counting studies mentioning drivers, diagnosing their influence on WEFE components, identifying critical nexus interactions, linking drivers across multiple components, and assessing their causal effects on the entire nexus.

2.2.5 WEFE nexus analysis

WEFE indicators directly measured, estimated, or quantified within the selected studies were identified. These indicators represent primary data collected by the researchers or secondary data compiled during the research process. They were extracted manually and categorized into two groups (WEFE 1 and WEFE 2) based on their roles within the conceptual frameworks presented in each study. Indicators assigned to the WEFE 1 category represented independent variables, hypothesized to influence other variables. Conversely, WEFE 2 category indicators functioned as dependent variables, potentially influenced by other factors. Following this, Co-occurrence analysis assessed how often two indicators were studied together, identifying nexus interactions like trade-offs, and positive, and negative synergies.

These interactions were drawn and visualized using a chord diagram and Sankey diagram, showing co-occurrence frequency alongside nexus type. Chord and Sankey diagram are useful to illustrate the interconnections between indicators across multiple components of a system. Higher frequencies suggest established relationships, while lower ones highlight areas needing further research.

2.3 Spatial explicit nexus analysis

The systematic review identified key WEFE nexus interactions in the Tana-Beles sub-basin. However, due to the absence of a comprehensive nexus-based study for the entire sub-basin, the review could not pinpoint the locations of WEFE interactions across the sub-basin. To achieve this, we first built a spatial database of major WEFE indicators using Woreda (local name for district) boundaries within the basin as spatial units. The specific indicators included food indicators (per capita crop production, proportion of cultivated arable land, crop water consumption), water indicators (green water availability), energy indicators (wood biomass energy consumption), and ecosystem indicators (proportion of essential life-supporting area, biomass yield).

2.3.1 Estimation of major WEFE indicators

The spatial analysis started with preparing a spatial database of major WEFE indicators using secondary data sources. Among the data sources, coefficients and constant variables related to food (crop water footprint) and energy (biomass energy consumption) components were derived from scientific studies rooted in primary data collection within the Tana-Beles sub-basin (Table 1). However, crop water productivity coefficients specific to the study area were not readily available for certain crops. In such cases, we adopted estimates from analogous regions that align with the biophysical characteristics of the Tana-Beles Sub-basin. Additionally, land cover data of the study area was derived from the dynamic world global land cover data (Brown et al. 2022) for the years 2022 and 2023 through a median reducer approach to harmonize and reduce the complexity of multi-temporal image sources.

Table 1 Data sources used to produce spatially explicit nexus interactions in the Tana-Beles Sub-basin
Full size table

2.4 Food subsystem

The food system’s performance was assessed using three indicators: per capita crop production, food water footprint, and proportion of cultivated arable land.

Per capita crop production (PCP) was estimated using a three-year (2019–2021) average of main-season crop yield and cultivated area for major grain crops in the study woredas. Data were obtained from the Ethiopian Statistical Service (CSA 2018, 2020, 2022). Disaggregation from zonal to woreda level was necessary to capture spatial variations in grain production within the sub-basin. This involved calculating the relative share of cultivated area for each crop from zonal data and multiplying it by the cropland extent of each woreda (sourced from the dynamic world land cover class). Finally, woreda-level crop production was estimated by multiplying crop area data with crop yield data.

To make a more realistic estimate, we deducted 40% of the gross crop production to consider production costs, based on data from the literature (Elias et al. 2017; Meskel and Gemechu 2017). We then divided the resulting value by the prevailing population data of each woreda to estimate per capita crop production.

$$ {\text{PP}}_{ij} = \frac{{\left( {Y_{ij } . A_{ij} } \right)0.4}}{{{\text{Population of Wored}}a_{i} }}{ } $$

Were,

Pij = Per capita crop production in woreda i (in kg/person/year).

Yij = Crop yield of crop j in woreda ij.

Aij = Cultivated area of crop j in woreda i.

Crop water consumption (CWC) CWC was estimated for each woreda by coupling crop water productivity (CWP) data from relevant literature (Cai and Rosegrant 2003; Cook et al. 2006) with woreda-level crop production (derived from per capita estimates). CWP are mathematical models that describe the relationship between the amount of water used by a crop (typically measured as evapotranspiration) and its resulting yield (Allen et al. 1998). Data was prioritized based on contextual relevance (climate, topography, soil, farming system) to inform crop-specific water consumption estimates (Table 1). These estimates were then scaled by annual production data to derive total water footprints for each woreda’s crop production system.

Proportion of cultivated arable land (PCAL) Potentially arable land within each woreda was estimated using slope, land cover, and digital elevation models (DEMs). Areas exceeding a 30% slope gradient were excluded due to cultivation limitations. Land cover classes incompatible with farming, such as water bodies and settlements, were further eliminated. While water availability, soil characteristics, and temperature were initially factored in, no inherent limitations for crop production were identified throughout the sub-basin; hence, these variables were excluded for further analysis. Consequently, uncultivated arable land to the year 2022 was assessed for each woreda based on slope and land cover classifications. This information enabled the calculation of the cultivated land to the available arable land ratio for each woreda, serving as an indicator for the food subsystem.

2.5 Water subsystem

Green water (GW) availability, defined as the soil moisture readily accessible for plant uptake, was employed as the primary indicator for basin-wide water resource assessment (Egan 2011). This selection of green water is due to the critical role green water plays in the food production system compared to blue water (surface and groundwater runoff excluding environmental flow requirements) and greywater (domestic and industrial wastewater) in the Tana-Beles basin. Spatially explicit green water availability for each woreda (administrative unit) was derived from a Soil and Water Assessment Tool (SWAT) model output for the Tana-Beles basin, as established by WLRC (2019). This model output provides hydrological attributes for 1600 hydrologic response units (HRUs). The woreda-level estimate was calculated by averaging the green water availability (originally in millimeters) across all HRUs within each woreda. Conversion to million cubic meters (MCM) was subsequently performed using the respective woreda area to obtain the total annual green water availability.

2.6 Energy subsystem

Wood biomass, constituting over 85% of the total energy supply, is the primary household energy source in the Tana-Beles Sub-basin due to limited access to electricity (Guta 2014). Therefore, wood biomass energy consumption (WBEC) per household was chosen as an indicator to quantify the energy subsystem within the Water–Energy–Food Nexus (WEFE) framework.

Average wood biomass energy consumption per household ranges from 2.4 to 2 tons of dry wood annually (Guta 2014; Woldeamlak Bewket 2005). An estimate of 1.94 tons per household was identified as the most representative value for the sub-basin and subsequently multiplied by the total number of households to calculate woreda-level WBEC. Besides, the vegetation area required to support the average wood biomass energy consumption of the respective Wordedas was calculated using the biomass yield productivity index of each woreda based on data from WBISP 2004; and MEFCC 2017.

2.7 Ecosystem subsystem

For the ecosystem subsystem, two indicators were used to measure its status: biomass yield and essential life-supporting areas (ELSA). The biomass yield of the sub-basin was calculated based on the biomass yield estimates for different agroecological regions in Ethiopia provided by WBISP (2004), specifically for shrubland and forest biomes as they constitute the main sources of fuel wood in the study area (Woldeamlak Bewket 2005). Then, the total biomass yield of each woreda was estimated by multiplying the coefficient values by the area coverage of shrublands and forests. This gives the mean annual biomass yield of the study woreda in tons.

ELSA concept (Ervin et al. 2022) was employed to measure the ecosystem subsystem’s interaction with other components within the sub-basin. ELSA was defined as land areas outside crop production and settlements providing crucial ecosystem services, encompassing relatively intact forests and woodlands, shrublands, and swamp areas (Ervin et al. 2022). Spatial extents of these land cover classes were quantified for each woreda using the dynamic world land cover map. Subsequently, the relative share of ELSA within each woreda’s total land area was computed and used as an indicator of ELSA abundance.

2.7.1 Spatial explicit WEFE nexus interactions

WEFE indicator estimations produced average values for each indicator within each woreda (district) of the Tana-Beles Sub-basin. These values were stored in a vector format spatial database as an attribute table linked to the Woreda shapefile, allowing for spatial analysis. Individual indicator maps were then generated based on these estimated values. Next, we employed a spatially explicit approach to identify nexus interactions between indicators. This involved establishing threshold values to determine the type of interaction present in a specific location. These thresholds were based on the predicted effect of one indicator on another when a change occurs, as informed by relevant literature (Table 2). An attribute selection criterion was then established in the spatial database. This criterion compared specific WEFE indicators within each woreda and categorized the type of nexus interaction based on the predefined threshold values. The results were stored in a separate attribute field named “Nexus Type” (Trade-off, Synergy, or Balanced). Finally, nexus interaction maps were produced using ArcGIS Pro’s bivariate color symbology functions. This function visually represents the quantitative relationship between two indicators in a specific location using a graduating color scheme.

Table 2 Criteria used to determine nexus interaction in the Tana-Beles Sub-basin
Full size table

2.7.2 Correlation analysis of the WEFE indicators

In addition to identifying relationship types using the previously described method, a correlation matrix approach was employed to assess the strength and statistical significance of relationships among WEFE indicators across the entire sub-basin (Friendly 2002). This analysis involved constructing a data table where each row represented the average value of an indicator for a specific woreda (23 rows corresponding to the administrative divisions within the Tana-Beles basin). Columns represented the individual WEFE indicators (7 total), corresponding to WEFE indicators considered in the study. The corrplot package in R statistical software (Wei and Simko 2021) was then used to generate a correlation matrix visualization. This analysis enabled the identification of the types of nexus interactions among WEFE components at the sub-basin scale.

2.8 Stakeholder identification

A stakeholder mapping exercise was conducted to identify key actors involved in WEFE management activities within the Tana-Beles sub-basin. This exercise utilized findings from the systematic review results and a stakeholder consultation workshop held on May 16–17, 2022, in Bahir Dar City, Ethiopia, with 35 participants representing diverse stakeholder groups. Results from both the literature review and consultation were integrated into a stakeholder matrix. The rows in the matrix list stakeholders, while columns capture characteristics relevant to their WEFE involvement. These characteristics include stakeholder category (key, primary, or secondary), role in WEFE component management, scale of operation (local, basin, district/zone, regional, national, international), implementation capacity (low, medium, high), level of interest (high, medium, low), and influence/power (high, medium, low). Additionally, the matrix documents stakeholders’ primary interests, objectives, and contributions to successful nexus implementation. The relevant information was extracted from the reviewed documents manually which were subsequently integrated with the result of the stakeholder consultation workshop; where the participants provided information for the stakeholders they represent. The combined data set was then used to generate a power-interest relationship graph, visualizing the key stakeholders’ roles in WEFE component management and operation.

3 Results and discussion

3.1 A systematic review of nexus interactions

Our systematic literature review identified 102 peer-reviewed articles published between 1991 and 2021 that investigated WEFE nexus interactions within the Tana-Beles Sub-basin. The following sections detail the findings extracted through this review process.

3.1.1 Characterization of nexus literature

Figure 3 shows the spatial distribution of studies in different parts of the Tana-Beles Sub-basin. The reviewed studies show a varied focus on different sub-basins within the Tana-Beles basin. Among them, 26 studies encompassed the entire Tana-Beles basin as part of their broader analysis of the Abay basin, including all sub-basins within Tana-Beles. An additional 14 studies focused on the entire Lake Tana Sub-basin, with their spatial scope encompassing all watersheds within the basin. The remaining studies concentrated on individual watersheds within the Tana-Beles Sub-basin. Among watershed-specific studies, Gumera (10 studies), Gilgel Abay (9 studies), and Gelda (8 studies) received the most attention. Watersheds west and south of Lake Tana were studied less frequently. Notably, Gemero and Garno, large watersheds northeast of the lake, lacked any available research despite their size (Fig. 4).

Fig. 3
figure 3

Spatial distribution of WEFE-based studies in Tana-Beles Sub-basin by major watersheds.)

Full size image
Fig. 4
figure 4

Temporal evolution of WEFE-based studies in the Tana-Beles Sub-basin by data analysis methods. The brightness of the bars represents the number of studies published each year. Darker colors indicate a higher number of studies, while lighter colors indicate fewer studies

Full size image

The temporal trends of the studies reveal a sharp increase in research output following 2009, coinciding with the beginning of substantial investments in the nation’s education sector. The evolution of studies over time also indicates a notable rise in the complexity and diversity of methodological approaches since the second half of the 2000s. Early studies (before the 2000s) were predominantly reliant on confirmatory (Abebe et al. 2001; Green and Mengestou 1991) and inferential analysis (Nagelkerke and Sibbing 1998), gradually transitioning towards numerical (Kebede et al. 2006) and physical-based (Chebud and Melesse 2009) modeling, and subsequently towards geospatial-based analysis (Eggen et al. 2016; Engineering 2020). Despite this advancement, nexus-based analysis has yet to emerge as a prevalent approach, with only a few studies adopting this methodology in recent years.

3.1.2 Conceptual structures of the reviewed studies

A thematic map of articles produced through the similarity of author-nominated keywords is presented in Fig. 5. Each cluster in the graph represents articles with approximate similarity in research thematic areas. The size of the clusters represents the relative importance of the articles in the research agenda. The x-axis represents the degree of centrality for each theme; which quantifies a theme’s connectedness with other themes across the entire network. Conversely, the y-axis depicts the density of a theme, representing the interconnectedness between keywords within a particular cluster, reflecting the theme’s level of development. Consequently, the visualization trends from left to right along the x-axis, with centrality increasing, and density (degree of development) increases from bottom to top along the y-axis (Aria and Cuccurullo 2017; Cobo et al. 2011; Forliano et al. 2021).

Fig. 5
figure 5

A keyword-based conceptual structure graph of the selected articles (developed based on Aria and Cuccurullo 2017)

Full size image

Analysis of the thematic map reveals distinct research areas. The Motor themes, focused on water balance and hydrological modeling, occupy the upper-right quadrant, indicating well-developed and influential research. Basic research themes, emphasizing land and water resource utilization, are positioned in the lower-right quadrant, holding significant influence despite lower internal coherence. Niche themes, such as those related to household socioeconomics, are in the upper-left quadrant, exhibiting strong internal connectivity but limited influence on other clusters. The lower-left quadrant depicts emerging or declining areas with limited relevance to the motor team’s current focus. These areas lack development and are characterized by keywords related to aquatic ecosystem functions (Fig. 6).

Fig. 6
figure 6

Chord diagram depicting the interlinkages between the four WEFE components. The flows/arcs quantify the interlinkages within the nexus. This thickness of the arcs is proportional to the frequency of cases in the literature linking those components (indicators).

Full size image

Analysis of the conceptual structure results through the WEFE nexus framework reveals motor research aligns with the water component, while environmental studies fall under basic research. Food and energy themes have limited development. This suggests a focus on characterizing individual WEFE indicators, particularly water, with less emphasis on integrated analysis across the entire WEFE system.

3.1.3 Major WEFE nexus interactions and associated drivers

The complex nexus links between the four WEFE subsystems obtained from the systematic review are presented using the chord diagram (Fig. 6). In the chord diagram, each subsystem takes a share of the circle proportional to the number of cases that the subsystem is indicated in the database, and the arcs drawn between the subsystems show the interconnections between the subsystems. The arcs have the same color as the originating subsystem and end up in the same or another subsystem. In the case of linking the same subsystem, the arc is linking the two indicators within a subsystem. The size of the arc is proportional to the size of the nexus links. The result shows that a large arc is observed between the water and ecosystem components indicating that large case studies are conducted linking the two components, followed by the interlinks between water and food subsystems(Fig. 6). The limited nexus interactions discussed in the papers are between energy versus food, and energy versus ecosystem.

The evolving phenomena driving research on major WEFE components is further illustrated in Fig. 7. The y-axis represents key study drivers, while the x-axis depicts WEFE elements (water, ecosystem, food, energy) considered by the respective studies. The frequency of colored boxes at intersections indicates the number of studies with a specific driver influential on a WEFE component. Landscape alteration emerged as the primary study driver, studied frequently against all WEFE components. This alteration includes deforestation (ecosystem), land degradation (food), increased material transport to water bodies (water), and disruptions to biomass energy (energy). Notably, seven out of the eight drivers, excluding climate change, stem from anthropogenic activities, highlighting the profound impact of human-induced environmental changes on WEFE systems.

Fig. 7
figure 7

Distribution of studies across evolving phenomena of study drivers and major Water-Energy-Food-Ecosystem (WEFE) components.

Full size image

To examine which issues are discussed against another nexus system, we represented the WEFE indicators in a Sankey diagram shown in Fig. 8 which represents the specific WEFE system indicators interlinked with another indicator.

Among the indicators represented in Fig. 8, the ecosystem component exhibits the strongest interlinkages, both within its subsystem and with the other subsystems. These connections include climate change’s impact on water availability and storage (water indicators), as well as the influence of water quality on aquatic ecosystems and fish production. Additionally, food production demonstrates significant interlinkages with other subsystems, particularly cropland expansion’s effects on various water indicators (groundwater recharge, lake water quantity, and quality) and ecosystem indicators (land degradation, soil erosion, wetland area loss, and degradation). The pursuit of increased food production to enhance livelihoods in the sub-basin often entails tradeoffs concerning water and ecosystem services.

Fig. 8
figure 8

Nexus links between indicators across (and within) the WEFE subsystem and their nexus interactions (tradeoffs and synergies). Numbers in white boxes represent the number of studies that considered each specific indicator pair. Indicators in the WEFE 1 category are considered causal factors, influencing the indicator they connect to in the WEFE 2 category. Conversely, WEFE 2 indicators are considered effect variables, reflecting the outcomes or consequences of the connected WEFE 1 indicator. This classification is based on the conceptual framework of the reviewed studies.

Full size image

Within the water subsystem, the dynamics of lake water volume and its downstream consequences are the most frequently studied nexus interactions. Lake water volume indicators exhibit the highest level of interlinkage with irrigation farming and cropland expansions, highlighting the tradeoffs between the food and water systems. Furthermore, water storage frequently interacts with various ecosystem indicators (climate change, land cover change, and water hyacinth infestation), characterized by tradeoff nexus types that suggest ongoing water storage degradation in response to ecosystem disturbances.

3.1.4 Conceptual framework of critical nexus types in Tana-Beles Sub-basin

By synthesizing the information from the frequency of studies by major WEFE components (Fig. 6), study driver frequencies (Fig. 7) and WEFE indicator inter-linkage analysis (Fig. 8), we produced a nexus-based framework that highlights critical nexus interactions in the Tana-Beles Sub-basin, as depicted in Fig. 9.

Figure 9 shows the major food subsystem in the sub-basin which includes crop farming, livestock rearing, and fishing. Crop production is the main driver of ecosystem deterioration and declining water storage volumes in the sub-basin (Asmamaw et al. 2021). It is also the subject of the most scientific attention, with a focus on its impact on other WEFE subsystems. Cropland expansion directly alters ecosystem subsystems through deforestation, wetland degradation, soil erosion, and soil quality deterioration (Dessie et al. 2014; Eggen et al. 2016; Erkossa et al. 2014; Hassen and Assen 2018). These impacts trigger cascading effects throughout the terrestrial and aquatic ecosystems of the sub-basin. Among these impacts, deforestation is one of the well-documented issues in the literature (Abebe et al. 2021; Enku et al. 2014; Gashaw et al. 2020; Ligdi et al. 2010), which aggravates soil erosion and nutrient leaching effects of the food production system on the ecosystem (Moges et al. 2017). The other tradeoff of the food system is that the recent increase in agricultural fertilizer application has caused increased nutrient losses to streams and lakes in the sub-basin (Dersseh et al. 2019; Kaba et al. 2014). One of the empirical pieces of this impacts is water pollution and eutrophication impacts in Lake Tana (Dersseh et al. 2022, 2020a) which have caused the rapid expansion of water hyacinth in the lake (Dersseh et al. 2019; Enyew et al. 2020; Nega et al. 2021) affecting 300–1500 hectares of the lake’s surface area (Ayalew 2014; Dersseh et al. 2020b; Worqlul et al. 2020). Water hyacinth is a major focus of scientific research and conservation efforts due to its potential impacts on other lake ecosystem components and fishery production (Degaga 2018; Gezie et al. 2018). Most scientific studies have focused on the factors affecting water hyacinth expansion, with less attention on its impact on the lake ecosystem.

The interaction of food and water components is another WEFE nexus issue of significant scientific attention. The two subsystems directly interact through irrigation water use and the water balance effects of cropland expansion (Abera et al. 2021; Asmamaw et al. 2021; Assefa et al. 2020a, b). The sub-basin has two types of irrigation schemes: large-scale (Alemayehu et al. 2010; Gebreyes et al. 2020) and small-scale (Abera et al. 2019; Taye et al. 2021; Worqlul et al. 2021). Large-scale schemes use water from dams (Alemayehu et al. 2010), while small-scale schemes tap water directly from rivers (Taye et al. 2021). Both types of irrigation have been reported to reduce the availability of water in the downstream areas, including Lake Tana, by diverting water away from the natural flow (Abera et al. 2021; Alemayehu et al. 2010).

Another significant interaction within the WEFE nexus is the impact of climate change on the other subsystems. Climate change is becoming increasingly evident in the Tana-Beles Sub-basin, causing rainfall irregularities and extreme temperature variations. These changes are impacting the food production system by reducing crop production and productivity, as well as water availability (Ayele et al. 2016; Setegn et al. 2011; Tigabu et al. 2021).

The energy subsystem is the least studied of the WEFE nexus interactions in the Tana-Beles Sub-basin. The available evidence suggests that it interacts with the water and food subsystems through the construction of hydropower dams. On the water subsystem, hydropower dams regulate river flow, which alters the natural flow of draining rivers affecting the availability of surface water in the downstream areas including Lake Tana (Annys et al. 2019). On the food subsystem, hydropower dams inundate grazing land, which reduces livestock production and productivity for small-scale farmers in the sub-basin.

3.2 Spatial explicit nexus interactions

3.2.1 Spatial distribution of nexus indicators

This section presents the spatial distribution of key WEFE indicators (i.e., food, water, energy, and ecosystem).

Food Per capita crop production in the Tana-Beles sub-basin averages 375 kg/year, but varies significantly. Northern areas produce 400–500 kg/year, while southern areas fall below 250 kg/year (Annex I). This likely reflects differences in livelihoods and food consumption patterns. Cultivated land also varies spatially, exceeding 90% in the north and dropping to less than 5% in the south. Crop water consumption follows a similar pattern, reaching over 400 MCM/year in the north and less than 50 MCM/year in the south.

Water Green water availability is estimated at 19,751 MCM/year, with higher volumes in the south (over 1500 MCM/year) compared to the north (less than 100 MCM/year) (Annex I).

Energy Wood biomass dominates household energy, accounting for 1.5 billion tons annually. Spatial variations exist, with Mecha and Fogera exceeding 200,000 tons/year, while most areas consume 50,000–100,000 tons/year (Annex I).

Ecosystem both ecosystem indicators (ELSA and biomass yield) are higher in the south than in the north. Northern ELSA is less than 15% and biomass yield is below 51,500 tons/year, while the south boasts ELSA exceeding 75% and biomass yield reaching 450,000 tons/year (Annex I). This suggests a healthier ecosystem in the south.

3.2.2 Correlations among WEFE indicators in the Tana-Beles Sub-basin

A correlation analysis was employed to examine the nexus types among the indicators across the sub-basin (Fig. 10). Accordingly; cultivated land and crop production positively correlated with water consumption (synergy), while negatively with green water (tradeoff). This implies that locations with high crop water consumption tend to have more cultivated land and crop production. The tradeoff between crop water consumption and green water availability suggests that green water availability and food production have an inverse relationship. In other words, the highest crop production amounts and most intensively cultivated areas are found in places where green water availability is relatively lower. The other interaction of the food subsystem is with the ecosystem component, characterized by a negative correlation (tradeoff) between both indicators of the food component (proportion of cultivated land and per capita crop production) and two of the ecosystem components (biomass yield and ELSA). This indicates that as the proportion of cultivated land increases and per capita crop production gets higher, the availability of land to accommodate essential ecosystem services reduces (Fig. 9).

Fig. 9
figure 9

A conceptual framework of major nexus interactions in the Tana-Beles Sub-basin was identified from the literature.

Full size image
Fig. 10
figure 10

Correlation graph of Water-Energy-Food-Ecosystem indicators in the Tana-Beles Sub-basin (CL: cultivated land, PCG: per capita grain production, CWC: crop water consumption, WBEC: Wood biomass energy consumption, GW: Green water, BY: biomass yield, ELSA: Essential life supporting areas).

Full size image

Within the energy subsystem, WBEC is positively correlated with crop water consumption and cultivated land. This means that places with high WBEC tend to also have a high intensity of cultivated land and high crop water consumption, and vice versa. In contrast, WBEC is not significantly correlated with two of the ecosystem indicators. The correlation between green water availability and two ecosystem components (biomass yield and ELSA) is one of the highest positive correlations observed among the indicators considered. This suggests that areas with high biomass production rates and abundant ELSA are also characterized by less cultivated land and more abundant water availability for plant growth.

3.2.3 Spatial nexus tradeoffs and synergies within the sub-basin

The analysis of WEFE indicators across the study sub-basin reveals complex spatial interactions, as shown in Fig. 11. Each map illustrates the relationship between two indicators, with shading indicating four categories: both high, both low, high–low, and low–high. For example, the crop water consumption-green water availability map identifies areas with high consumption and low availability (high–low) and vice versa (low–high). Both high and both low represent areas where both indicators fall within the respective categories. The remaining maps (Food–Ecosystem, Energy–Food, Energy–Ecosystem) follow the same legend scheme. Labels denote the dominant nexus type (Trade-off—T, Synergy—S, Balanced—B) within each administrative division, based on predefined thresholds (Table 2). These spatial patterns highlight distinct interactions between WEFE indicators across the sub-basin.

Fig. 11
figure 11

Tradeoff- and synergies among different WEFE indicators in different parts of the Tana- Beles sub-basin (produced using ArcGIS Pro's bivariate color symbology).

Full size image

When looking at the interaction of the water and food indicators, we found a sub-basin-wide crop water consumption to green water (CPC-GWA) ratio of 21%, indicating the presence of sufficient water sources for crop production at sub-basin scale, considering the recommended ~ 30% environmental flow reservation from the total available green water (Egan 2011; Rouillard et al. 2022). However, this ratio varies significantly across different parts of the sub-basin. In the Northern regions, the CPC-GWA ratio exceeds 50%, hinting at potential tradeoffs between food and water security. Conversely, the CPC-GWA ratio in the middle of the sub-basin declines to less than 40%, indicating a more balanced state. The southern parts exhibit high synergetic potential with a CPC-GWA ratio of 25% or lower.

In terms of food and ecosystem relationships, the sub-basin is characterized by distinct interactions among the Northern and Southern parts. In the Northern parts, the food production system has occupied nearly all the available spaces for crop production without leaving enough space for ELSA. Hence the nexus in this part of the sub-basin is characterized by a tradeoff between the two subsystems. On the other hand, in the southern parts, a significant area is left uncultivated which leaves abundant space for a proper proportion of ELSA. The nexus in this part of the sub-basin is then turned to synergies, which implies the presence of a potential opportunity to expand croplands without posing a significant threat to the ELSA.

The distinct spatial patterns of crop water consumption and green water availability are also reflected in the nexus of food and energy subsystems. The proportion of cultivated land and WBEC is generally higher in the northern parts than in the south. This suggests that there is limited opportunity to meet the prevailing biomass energy shortages in the north by expanding forests and shrublands without impacting food production. In contrast, the southern part of the sub-basin has extensive land areas that can accommodate both food and biomass energy production, presenting potential synergies among the two WEFE subsystems (Fig. 9).

The relationship between WBEC and biomass yield is inverted between the northern and southern parts of the sub-basin. In the north, WBEC is high, but biomass yield is low, suggesting a tradeoff among the subsystems, where wood biomass utilization rates exceed the ecosystem’s sustainable biomass production capacity. In the south, biomass yield is higher than WBEC, indicating potential synergies between the two subsystems (Fig. 9).

3.3 Stakeholder analysis

Effective WEFE nexus management in the Tana-Beles sub-basin necessitates stakeholder analysis and engagement. This identifies key actors, their interests, and decision-making influence. This information is crucial for crafting sustainable basin-wide policies.

The stakeholder analysis out of 102 studies that were reviewed identified six categories of stakeholders: communities, local government, international organizations, public institutions, federal government, and universities and research institutes. The studies mention community and local governments accounting for 22.5% and 19.6%, respectively as the main stakeholders in the sub-basin that benefits, impacts, and/or works in the WEF nexus. Organizations that include investors, NGOs, ministries, and some government agencies account for 12.7% of the mentioned stakeholders in the articles. Institutions that include religious (Monasteries of Orthodox Church) entities are mentioned in 11.7% of the articles.

Figure 12 presents a power-interest matrix derived from the findings of reviewed studies and stakeholder consultation workshops. Stakeholders are plotted based on their level of interest (X-axis) and implementation power (Y-axis). This visualization allows for targeted stakeholder engagement strategies. Stakeholders positioned in the top right quadrant (high interest, high power) require regular consultation and direct involvement in WEFE intervention strategies. Conversely, those in the low right quadrant (high interest, low power) warrant advanced notice of emerging WEFE issues. Keeping them informed fosters their cooperation in interventions. Stakeholders in the top left quadrant (high power, low interest) should be consulted strategically to secure their support for large-scale WEFE interventions. Finally, stakeholders in the bottom left quadrant (low power, low interest) may not necessitate direct involvement but should be kept informed when necessary.

Fig. 12
figure 12

Power-Interest relationships of stakeholders in the Tana-Beles Sub-basin.

Full size image

4 Discussion

A review of WEFE research in the Tana-Beles sub-basin revealed a gap in understanding WEFE interactions. While sectoral studies addressed individual indicators, a framework for comprehensive nexus analysis was lacking. To address this, we combined a systematic review with spatial analysis. The review provided a critical foundation, while the spatial analysis identified specific locations and magnitudes of synergies and trade-offs.

Both approaches converged on the challenges in the northern Tana-Beles sub-basins food system (crop farming, livestock rearing, fishing). This region struggles with marginal food production, limited arable land, and competition for space with essential ecosystems. This pressure creates land scarcity for expansion and biomass production for energy, especially concerning the area’s rapid population growth and low crop productivity.

The result from the review literature implies the profound impact of food production on the water cycle primarily through hydrological alterations and stream drying. However, spatial analysis reveals abundant green water resources across the sub-basin, suggesting that expanding crop production would have a relatively modest impact on water availability. This is supported by the observation that crop water consumption remains below 40% of green water resources in most regions. These seemingly contradictory findings suggest a disconnect between water sources and availability. The presence of sufficient green water volume implies that the mean annual rainfall has the potential to sustain food production. However, the absence of robust water storage infrastructure hampers the optimal utilization of these resources to make water available at the required time and location. Farmers rely on readily accessible water sources instead of investing in sustainable storage mechanisms. Addressing this infrastructure deficiency is crucial for maximizing water availability and ensuring the resilience and long-term sustainability of the water–food nexus in the studied sub-basin.

The energy-ecosystem nexus remains least understood, with no empirical evidence revealed about their interactions in literature. Spatial analysis, however, reveals critical energy-ecosystem nexus issues in the sub-basin. Energy deficits in the northern escarpments stem from excessive biomass utilization exceeding the ecosystem’s production potential. The limited vegetation cover is insufficient to meet growing fuelwood demands, posing severe land degradation threats.

Additionally, spatial analysis highlights the limited potential to address biomass energy deficits due to food system constraints on afforestation and reforestation efforts. This indicates that one strategic development intervention will be devising methods to meet the mounting demands for food and energy within the constraints of limited available land. Notably, this challenge is most pronounced in the northern part of the basin, characterized by a pronounced tradeoff between food and the other WEFE components. Conversely, the southern region of the sub-basin presents a distinct scenario where potential synergies between the food-ecosystem and food-energy components exist. However, the realization of these synergies has been impeded by social, economic, and political constraints, and there is a gap in scientific understanding of the mechanisms required to harness these potential synergies for addressing WEFE issues within the sub-basin. This knowledge deficit underscores the need for further research and policy development to unlock the latent opportunities for sustainable resource management in the southern part of the sub-basin.

The key stakeholders in the Tana-Beles sub-basin include farmer and fishery communities, local and regional governments, organizations and institutions, the federal government, and researchers. These stakeholders have a variety of interests in the basin, including water for irrigation and fishing, hydropower generation, water supply, research and education, and water resource management and planning.

The result highlighted that different stakeholders like farmers and fishery communities benefit from the basin by using its water for irrigation of their farmlands and fishing to sustain their livelihoods. Whereas local government and regional governments like the Bureau of Agriculture, Bureau of Energy and Mines, Bureau of Environmental Protection, Land Administration and Use, the Abbay Basin Authority, and the Organization for Rehabilitation and Development in Amhara benefit from the basin by using it for water resource management and planning and water supply interventions. The Ethiopian Electric Power Corporation benefits from the basin by using its water for hydropower generation. Organizations and institutions like the Amhara Regional Agriculture Research Institute, and Bahir Dar University, benefit from the basin by using it for research and education to study the impacts of climate change and other environmental issues. The level of coordination among the stakeholders in the Tana-Belese sub-basin is challenging. This is because the stakeholders have different interests and priorities, and they often tend to plan and work in a silo format.

5 Conclusion and future recommendation

This study investigated the WEFE nexus within the Tana-Beles Sub-basin, Ethiopia. We employed a two-pronged approach: a systematic review of existing literature on WEFE indicators and a complementary spatial analysis. This combined methodology allowed for a comprehensive assessment of WEFE interactions within the basin.

A review of WEFE literature identified knowledge gaps and strengths regarding WEFE interactions. Water assessments often focused heavily on measuring the physical availability of water, neglecting the economic aspects that might limit access. Food system studies tend to portray conflicts with other WEFE components, highlighting the need for solutions that enhance food production with minimal environmental impact. Energy research is the least developed, with a focus on hydropower’s trade-offs with ecosystems, leaving broader impacts and benefits unclear. While research prioritizes established ecosystem threats like climate change, land use, and pollution, exploration of novel ecosystem-based solutions, synergistic interactions within ecosystems, and their carrying capacity for WEFE demands remain limited. Despite progress in individual WEFE component analysis (water, food, energy), a nexus approach remains underexplored due to siloed land–water research methods. A growing body of WEFE literature and advanced assessment frameworks offer an opportunity for seamless integration, fostering holistic ecosystem understanding.

Spatial analysis with empirical data revealed previously overlooked WEFE interactions across the Tana-Beles basin. While trade-offs dominated the northern Tana sub-basin (land–water stress, food insecurity), the southern Beles sub-basin (warm, low-population lowlands) exhibited potential synergies for production gains with minimal WEFE disruption. This spatial disparity underscores the need for spatially explicit WEFE assessments. Our approach, despite using a limited set of indicators and approximate thresholds, effectively highlighted these critical nexus issues. Future research can refine this methodology by incorporating broader indicators and advanced spatial techniques, particularly valuable in the absence of robust WEFE models and data management systems.