Forest resource management, refugee integration, and food security in rural Zambia: balancing sustainability and equity

Ruesink, Brigitte; Gronau, Steven Article — Published Version Forest resource management, refugee integration, and food security in rural Zambia: balancing sustainability and equity Food Security Provided in Cooperation with: Springer Nature Suggested Citation: Ruesink, Brigitte; Gronau, Steven (2025) : Forest resource management, refugee integration, and food security in rural Zambia: balancing sustainability and equity, Food Security, ISSN 1876-4525, Springer Netherlands, Dordrecht, Vol. 17, Iss. 5, pp. 1273-1299, https://doi.org/10.1007/s12571-025-01547-3 This Version is available at: https://hdl.handle.net/10419/330218 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Vol.:(0123456789) Food Security (2025) 17:1273–1299 https://doi.org/10.1007/s12571-025-01547-3 RESEARCH Forest resource management, refugee integration, andfood security inrural Zambia: balancing sustainability andequity BrigitteRuesink1,2 · StevenGronau1 Received: 8 March 2024 / Accepted: 24 April 2025 / Published online: 18 June 2025 © The Author(s) 2025 Abstract Africa’s rising refugee numberslead to integration increasinglybeing replaced by repatriation. Investigating the long-term effects of refugees on host areas is crucial for sustainable integration, as the population increase puts pressure on limited natural resources. While existing literature addresses the environmental impacts of refugees, behavioral models rarely focus on this issue. This study uses an Agent-Based Model to simulate interactions between refugees, hosts, and forest resources. The objective isto (1) quantify the impact of refugee settlements and host communities on forest resources, (2) assess the effects of varying refugee settlement sizes on sustainable forest utilization and food security, and (3) evaluate how labor cooperation influences deforestation. The modelapplies a 2018 dataset from a refugee hosting community in rural Zambia, including 277 households, and comprehensive supplemental secondary data.Results show that forest reduction is driven by the need for firewood and land for refugee settlements, significantly reducing theforest area. Revealed deforestation threatens sustainableforest ecosystems and impacts food security by diminishing access to wild fruits and edible insects, crucial to local diets. Cooperation between refugees and host communities in slash-and-burn farming temporarily boost food production,but accelerates forest reduction. This leads to long-term resource depletion and competition. Highlighteddynamics showthat, if unmanaged, refugee influxes can exacerbate food insecurity in rural refugee settings. Agroforestry and policy interventions focusing on sustainable land use, property rights, and alternative energy sources are essential to balance refugee needs with forest preservation and food security in host communities. Keywords Agent-based modeling· Refugees· Deforestation· Food security· Zambia· Sustainable development 1 Background By the end of 2023, the number of forcibly displaced individuals worldwide reached 117.3 million, including 43.4 million refugees, reflecting the growing challenges of forced migration driven by conflict, economic instability, and climate change (UNHCR, 2024). The ongoing Ukraine War significantly contributed to this increase. Understanding the impact of these movements requires not only an analysis of the countries of origin but also the host countries, many of which are already vulnerable (Barman, 2020; Khaled, 2021; Schneiderheinze & Lücke, 2020). Of the 32.5 million refugees in mid-2022, 7.1 million were hosted by the least developed countries, including 783,000 in Southern Africa, where fragile conditions exacerbate integration challenges such as food security (UNHCR, 2022a, 2023a). Zambia plays a significant role in the refugee context, hosting around 71,000 refugees, or about one-tenth of the total number of refugees in Sub-Saharan Africa (UNHCR, 2023b). The country faces high levels of poverty and food insecurity and depends on external aid to support citizens and refugees (WFP, 2023). In recent years, the COVID19 pandemic and climate change have further worsened thesituation (WFP, 2023). While a long migration history connects the Democratic Republic of the Congo (DRC) and Zambia (Clark, 2002; Larmer, 2016; Mkandawire & Daka, 2018), more recent conflicts led to a significant influx of refugees from the DRC to Zambia in 2017 (UNHCR, 2018; Vlassenroot & Verweijen, 2017). * Brigitte Ruesink [email protected] 1 Institute forEnvironmental Economics andWorld Trade, Leibniz University Hannover, Königsworther Platz 1, 30167Hannover, Germany 2 Institut für Umweltökonomik und Welthandel, Leibniz University Hannover, Königsworther Platz 1, 30167Hannover, Germany 1274 B.Ruesink, S.Gronau In response, the Mantapala refugee settlement was implemented in early 2018 in Northern Zambia, near the DRC border. The settlement, hosting over18,000 refugees, brought increased attention and aid to the area, benefiting both refugees and the local host community (UNHCR, 2018, 2021b, c, 2022b; WFP, 2023; WFP & UNHCR, 2021). With 66% of refugees not intending to return to the DRC (UNHCR, 2021b), integration into the host community is essential. The settlement was established within the Comprehensive Refugee Response Framework (CRRF), with support from national and international organizations to provide agricultural assistance and services to refugee and host households, fostering integration into daily life and economic activities like farming (Mwansa, 2022; UNHCR, 2019; WFP, 2023; WFP & UNHCR, 2021). This study focuses on the socioenvironmental dynamics of the Mantapala region, a remote forest area. Food security has become a significant concern in Mantapala, where a 2020 joint needs assessment by the World Food Programme (WFP) and the United Nations High Commissioner for Refugees (UNHCR) revealed that 44% of households exhibit poor or borderline food consumption (UNHCR & WFP, 2020). Although the report finds Zambia’s food availability at the national level sufficient to meet market demands, this did not translate to all communities. Smallholder farmers, particularly refugee smallholders, faced resource limitations that hindered their ability to cultivate available land, contributing to localized food scarcity. Most households relied on market purchases and inkind food assistance to meet their needs, with some refugee households engaging in labor-for-food exchanges (UNHCR & WFP, 2020). Zambia’s dependence on agriculture contributes to high deforestation rates, averaging 0.245% annually from 2000 until 2014 (Republic of Zambia, 2016), with even higher rates of 0.91% from 2015 to 2016 recorded in the Nchelenge district, where Mantapala is located (Global Forest Watch, 2023). Given the population growth rate of 3.6% in therespective Luapula Province (Zambia Statistics Agency, 2022) and the dependency on forest resources, substantial pressure on the environment is exacerbated by the influx of refugees (Weber etal., 2023). Slash-and-burn agriculture, a common practice in Nchelenge, further contributes to deforestation when performed unsustainably, with ash from burned trees used as fertilizer to enrich nutrient-depleted soils (Kaluba etal., 2021). The predominant Miombo woodlands, which provide firewood, charcoal, wild fruits, and edible insects (important forlocal food systems), are particularly affected (Brockerhoff etal., 2017; Ickowitz etal., 2021; Syampungani etal., 2009). The degradation of these forests directly undermines food security by reducing the availability of forest-derived foods that supplement diets and sustain livelihoods in Mantapala (Gronau etal., 2019). Refugee settlements can bring both benefits and challenges to host communities. On the positive side, the presence of refugees often attracts international aid, improving infrastructure, such as schools and health clinics (Bilgili etal., 2019; UNHCR & WFP, 2020; Zhou etal., 2022). These improvements strengthen the host community’s resilience and enhance food security by increasing access to essential services (UNHCR & WFP, 2020). Additionally, the interaction between refugees and hosts can stimulate local economic activities and build trust (Alix-Garcia etal., 2019; d’Errico etal., 2022; Fajth etal., 2019; Kim etal., 2022). However, the benefits caused by refugee arrivals are not always sustainable. The sudden increase in population caused by refugee influxes puts significant pressure on natural resources, including water, wetlands, and forests (Bernard etal., 2020; Bildirici etal., 2023; Jaafar etal., 2020). Langer etal. (2015) identify the initial phase of construction and the disaggregation of the settlement as critical periods for deforestation. Once settled, refugees’ energy demands, particularly for firewood, exacerbate vegetation loss near settlements (Kranz etal., 2015). Agricultural expansion also plays a role, with Langer etal. (2015) associating deforestation with land conversion, while Maystadt etal. (2020) observe shifts from forests to cropland without a clear correlation to refugee numbers. Other studies, such as those by Makunga and Misana (2017), highlight settlement growth, wood extraction, and agricultural activities as primary deforestation drivers. Notably, the literature shows mixed results, with some studies reporting positive effects on vegetation near refugee settlements (Kranz etal., 2015; Salemi, 2021), while others link population density to forest loss (Phiri etal., 2019; Richardson etal., 2021). An important aspect of these socio-environmental dynamics in refugee-hosting contexts is the issue of land rights, which play a central role in resource management and social cohesion. In Zambia, insecure property rightscan lead migrants to clear allocated land entirely, even when they do not intend to cultivate all of it, exacerbating tensions with the host community (Unruh etal., 2005). When hosts perceive their property rights as insecure or their access to a resource as restricted due to the presence of refugees, they may be less motivated to manage their land sustainably (Unruh etal., 2005; Vallino, 2014). Even when land use rights are clearly defined, they may not effectively prevent deforestation or resource conflicts. Ferrer Velasco etal. (2023) report that only 16% of Zambian stakeholders perceive land use rights and protected areas as practical tools against deforestation, reflecting a broader skepticism about their practical impact. The CRRF, adopted by 193 UN members, emphasizes the need for integrated refugee management approaches that address immediate needs and long-term impacts on host communities, particularly food security (UNHCR, 1275 Forest resource management, refugee integration, and food security in rural Zambia: balancing sustainability and equity 2001–2023). Countries, such as Zambia, have implemented the CRRF to promote sustainable refugee-host relationships, focusing on durable solutions that benefit both groups (Carciotto & Ferraro, 2020; Dare & Abebe, 2018). However, integration in SSA often faces challenges, including resource competition for essentials like land and firewood, particularly during the early phases of refugee arrivals (Barman, 2020; Langer etal., 2015). Such competition can exacerbate tensions and create negative perceptions of refugees, complicating integration efforts (Gronau & Ruesink, 2021; Skinner & Gottfried, 2017; Whitaker & Giersch, 2015). The growing focus on refugee integration highlights the need to better understand the role of host communities in managing environmental and resource challenges, which remains underexplored (Phillimore, 2020). The continuously rising number of refugees worldwide (UNHCR, 2022a) and ongoing climate change (IPCC, 2023) emphasize the urgency of developing long-term solutions for refugee situations. Since refugees often remain in host countries for extended periods, typically 10 to 15 years (Devictor & Do, 2017), repatriation alone is insufficient. Instead, cooperative and effective resource management by refugees and hosts is critical for fostering sustainable resource use and improving food security for both groups (UNHCR & WFP, 2020). Initiatives such as reforestation projects and improved cookstove programs aim to address these conflicts while fostering cooperation (Tabi Eckebil etal., 2022). However, their success requires a holistic approach that balances the needs of refugees and hosts, preventing well-intentioned aid from unintentionally harming local populations (Khaled, 2021). Agent-Based Models (ABMs) have been widely applied in refugee research to simulate migration dynamics (Collins & Frydenlund, 2016; Entwisle etal., 2020; Groen etal., 2019; Hébert etal., 2018; Suleimenova & Groen, 2020; Vanhille Campos etal., 2019) and post-arrival dynamics in host communities, particularly in Europe (Araya etal., 2021; Boshuijzen-van Burken etal., 2020; Paolillo & Jager, 2020), as well as to study disease spread within refugee settlements (Crooks & Hailegiorgis, 2014; Gilman etal., 2020). In deforestation research, ABMs have been used to explore the impacts of landuse decisions (Müller-Hansen etal., 2019; Zhou etal., 2022), livelihood strategies (Kiruki etal., 2019), and policy interventions on forest resources (Richardson etal., 2021; von Essen & Lambin, 2023). Weber etal. (2023) initiated an important exploration of the impact of refugee settlements on forest stocks. However,their study highlights the need for a more detailed examination of these dynamics. Our study bridges this gap by expanding on Weber etal.’s work by varying settlement sizes and incorporating labor cooperation between the refugee and host communities. This approach allows us to assess the sustainability of settlement capacities for forest resource preservation and examine the socio-environmental effects of economic interactions between these groups. In this study, an ABM is developed to analyze the effect of refugees on their host environment. It is based on a dataset collected in the refugee-hosting community of Mantapala in 2018 and supplementary secondary data. The secondary data includes desk research and interview outcomes with the local forestry department to specify behavioral rules, forest composition, and tree regrowth. Mantapala is an ideal setting to explore the interplay between refugee populations, natural resource management, and food security. It can illustrate how refugee influxes can worsen vulnerabilities or drive development through successful integration. This study adds to current research by investigating the causes of deforestation in the refugee context through a bottom-up approach, where we model individual behaviors and interactions to simulate the complex system of deforestation dynamics. It provides a base model that can be extended by countermeasures to simulate their effectiveness before implementation. For example, the promotion of sustainable agricultural practices (agroforestry), development of alternative energy sources (provision of improved cooking devices or implementation of solar energy projects), reforestation initiatives, or strengthening land use and forest management policies (Deißler etal., 2024; Ferrer Velasco etal., 2023; Kinyili, 2022; Luzi etal., 2019; Njenga etal., 2021). The model visualizes how different population sizes, slash-and-burn agriculture, and labor cooperation between refugees and hosts affect the development of forest cover in rural Zambia. By isolating firewood demand as a primary deforestation driver, the study offers nuanced insights into sustainable resource management strategies that address forest depletion while enhancing food security in refugeehosting communities. This research addresses three key questions: (1) To what extent do refugee settlements and host communities affect forest resources in sub-Saharan African countries? (2) How do varying refugee settlement sizes impact forest resource depletion and food security in rural host communities? (3) To what degree does cooperation between refugees and host communities in agricultural activities influence the development of forest covers? The paper is structured as follows: Chapter two describes the data and ABM. In chapter three, the model results are presented and discussed in chapter four, including limitations and future directions. The paper concludes in chapter five with a summary of the new researchinsights and future works. 2 Materials andmethods 2.1 Data The primary data for this study originated from a census survey conducted in April 2018 as part of the "Food Security in rural Zambia" (FoSeZa) project. The survey covered 277 1276 B.Ruesink, S.Gronau host households across eight villages within a 10 km range of the Mantapala refugee settlement in northernZambia(Luapula Province, Nchelenge District). A structured questionnaire collected information on household demographics, agricultural practices, land use, forest resource utilization, and interactions with refugees, supplemented with geographic village coordinates for spatial context. Additional data sources were utilized to supplement the dataset. Data from Gronau and Ruesink (2021), ILO (2011), UNHCR (2020), and WFP and UNHCR (2021) informed household behavior patterns. Tree composition and regrowth behavior were modeled using data from Montfort etal. (2021). Interviews with the local forestry departmentof the region provided further insights into forest management practices. A descriptive analysis of the collected dataset established baseline conditions of host community dependence on forest resources, labor needs, and refugee interactions. Key variables analyzed included firewood consumption, land size, cultivation, and settlement contact areas. All this information provide insights into pre-settlement conditions and resource dependencies. The analysis includedthe calculation of key variables'mean values, standard deviations, and ranges. 2.2 Agent‑based Modeldesign andimplementation 2.2.1 Core processes andpurpose This study employed an Agent-Based Model (ABM) to capture the complex effects of refugee and host communities on shared forest resources. ABM simulates household behavior and aggregate effects on resource sustainability, enabling the exploration of emergent phenomena critical for understanding socio-environmental dynamics and policy decisions (Gilbert, 2019; Heckbert etal., 2010; Squazzoni, 2012; Wilensky & Rand, 2015). Implemented in NetLogo 6.3.0 (Wilensky, 1999), the ABM followed the ODD (Overview, Design concepts, Details) protocol (Grimm etal., 2006, 2010, 2020) for transparency and reproducibility. The entire protocol is provided in the supplementary information. The purpose of this ABM was to explore how population dynamics, including general growth and refugee inflows, affect forest resource sustainability in rural Zambia over three decades. The model investigates two core dynamics: competition versus cooperation and environmental degradation versus sustainable use. Refugee influxes increase resource demand, potentially causing conflicts over land and firewood, while cooperative practices can mitigate competition and promote sustainability (UNHCR & WFP, 2020). The model simulated these dynamics through firewood collection, slash-and-burn agriculture, and labor cooperation, providing insights into refugee-host impacts on food security and forest sustainability. The ABM addressed these processes through three key questions: (1) How do refugee settlements and host communities affect forest resources? (2) How do varying refugee settlement sizes impact forest resource depletion and food security? (3) How does labor cooperation influence deforestation? Recognizing that high population density can lead to unsustainable forest use (Phiri etal., 2019; Richardson etal., 2021), the model simulated these dynamics based on micro-processes, providing insights into refugee-host impacts on food security and forest sustainability. 2.2.2 Key components andsetup The ABM developed for this study captured the complex socio-environmental dynamics between refugee and host communities in Mantapala, Zambia. As illustrated in Fig.1, the model was structured around three primary components: refugee and host communities, natural resources, and institutional settings, collectively representing resource use, cooperation, and environmental sustainability dynamics. Both communities were modeled as agents, each representing a household with distinct characteristics, behaviors, and resource needs. For a detailed list of all assumptions made in the model setup and their justification, see Table3 in the Appendix. Host households were parameterized using primary survey data on demographics, land use practices, and resource dependencies, supplemented by literaturebased assumptions (Gronau & Ruesink, 2021; ILO, 2011; Weber etal., 2023). Refugee settlement areas were designed to reflect deforestation from housing and resource use, with a baseline reference size for housing of 8,000 hectares accommodating 25,000 refugees (UNHCR, 2020; WFP & UNHCR, 2021). Scenarios adjusted these zones to explore varying population sizes. Refugee households were assigned firewood demands based on the host dataset (Clark, 2002; Larmer, 2016; Vlassenroot & Verweijen, 2017) and were placed within cleared settlement zones. Stochasticity accounted for household formation and distribution variability within the villages and behavioral dynamics. Natural resources were further considered, particularly forests, which both groups rely on for survival and livelihood. Forests were modeled as two-dimensional grids of 13,000 hectares divided into 200 × 260 plots, each 0.25 hectares, representing the study area (Gronau etal., 2018). Plots were parameterized with tree age and density using data from Montfort etal. (2021) and Weber etal. (2023). Baseline conditions were established by simulating 15 years of pre-settlement slash-and-burn agriculture, reflecting traditional land-use practices in Mantapala (Kaluba etal., 2021). Randomness in tree density across plots ensured spatial variability and annual updates for tree growth and harvesting simulated dynamic forest interactions. The final aspect of the institutional setting mediates resource access and influences cooperation or conflict 1277 Forest resource management, refugee integration, and food security in rural Zambia: balancing sustainability and equity between refugees and host communities. Property rights assigned host households with ownership to specific plots, restricting access to privately owned or settlement areas. Host households continued claiming near plots as their own until they reached the number of owned plots indicated in the survey. They could hire refugee labor, reflecting cooperative interactions. Stochasticity in cooperation probabilities ensured variability in agent decisions, capturing the unpredictability of real-world behaviors. 2.2.3 Construction andagent dynamics The model simulated long-term socio-environmental dynamics over 30 years using discrete annual steps (Weber etal., 2023). An important design concept of the model was the reliance on predefined, rule-based interactions between households and forest resources. Figure2 illustrates the annual decision-making processes central to the ABM’s depiction of household behavior. Please refer to the full ODD protocol for a detailed explanation of the sub-processes. The model updated environmental conditions at the start of each simulation year, including tree growth (age and density) and recovery from previous slash-and-burn activities, following ecological parameters based on secondary data (Montfort etal., 2021; Weber etal., 2023). All households then assessed their annual firewood demands based on resource stocks and individual characteristics. Refugee and host households followed the steps depicted in the first part of Fig.2 to meet their firewood demand. They sensed the availability of mature trees (aged over 15 years) and identified accessible patches (mature trees available and the patch does not belong to another household) with the highest tree density. If sufficient trees were available, households interacted with the plot by cutting Fig. 1 Conceptual foundation illustrating interactions between host and refugee communities and their influence on environmental dynamics. The figure depicts key factors such as resource competition, labor cooperation, and settlement size and their key impact on forest resources 1278 B.Ruesink, S.Gronau Fig. 2 Annual decision-making processes of host and refugee households in firewood collection and agricultural activities. The figure illustrates key decisions made in the ABM regarding firewood collection frequency, agricultural land use, and labor cooperation. Arrows represent decision pathways. Diamonds indicate decision points, while rectangles describe actions 1279 Forest resource management, refugee integration, and food security in rural Zambia: balancing sustainability and equity the demanded number of trees and resetting their demand to zero. All mature trees were harvested if resources were insufficient, and the unmet demand was reduced accordingly. Multiple households could access and interact with the same patch annually without spatial limitation. Property rights restricted firewood collection to owned plots or common forest areas, consistent across all simulations. If firewood demands remained unmet after 104 collection attempts (twice weekly), this unmet demand was recorded, tracking resource insecurity and highlighting households struggling to meet basic needs. After all firewood collection activities were completed, host households performed the steps indicated in the second part of Fig.2. Each host household decided whether to hire refugee labor to overcome agricultural constraints based on a probability factor. This probability was influenced by factors such as the household’s previous interactions with the refugee settlement (whether positive or negative), the household head’s age and education, proximity to the refugee settlement, and the household’s well-being compared to the previous year (Gengo etal., 2018; Gronau & Ruesink, 2021). Households adapted their slash-and-burn behavior after hiring refugee labor, expanding cultivation by the cultivated area per working-age household member (ILO, 2011). However, labor cooperation only increased desired cultivation and not total land ownership, limiting its feasibility if land resources were insufficient. Finally, the host households interacted with plots by slash-and-burn agriculture, clearing the desired number of plots (if accessible) and taking home a predefined amount of firewood. Plots were available for slash-and-burn agriculture if they belonged to the household (i.e., owned) and were not in the cultivation or fallow phase. The model incorporated mediated interactions through resource competition, which intensified as resources became scarcer. To conclude the yearly cycle, population growth was simulated at a rate of 3.6%, based on data from Luapula Province (Zambia Statistics Agency, 2022). 2.2.4 Scenarios The model evaluated multiple scenarios to analyze the socio-environmental dynamics of refugee-host interactions and their impact on forest resources. Table1 provides an overview of these scenarios defined by variations in key factors, including the presence of a refugee settlement, the number of refugees, and the inclusion or exclusion of slashand-burn agriculture or labor cooperation. For example, the “Medium—S&B” scenario simulated a refugee settlement of 5,000 refugees, utilizing a refugee settlement area of 1,600 hectares, with slash-and-burn agriculture but no labor cooperation. This adaptability allowed for a bottom-up perspective, exploring “what-if” scenarios and assessing potential outcomes under different configurations. The insights gained provide a robust basis for future policy decisions to balance resource sustainability and refugee-host dynamics. 2.3 Model outputs andvalidation The primary model output was the total remaining tree numbers reflecting deforestation levels, which is essential for understanding the environmental consequences of resource use. Outputs also included unmet firewood demands, offering insights into conditions that foster sustainable practices or exacerbate competition. A key emergent phenomenon is the “tragedy of the commons” (Hardin, 1968), where rational short-term actions by households lead to overuse and eventual depletion of resources. This outcome arose as households, following predefined rules, pursued their goals without coordinated resource management, leading to resource scarcity and exclusion from use. By simulating Table 1 Overview of the simulated model scenarios The table summarizes key characteristics of the scenarios, including the presence of a refugee settlement, the number of refugees, the use of slash-and-burn agriculture, and the option for labor cooperation Scenario Refugee settlement Number of refugees Slash-and-burn agriculture Labor cooperation Base—S&B No 0Yes No Base—No S&B No 0No No Small—S&B Yes 2,500 Yes No Medium—S&B Yes 5,000 Yes No Large—S&B Yes 10,000 Yes No Small—No S&B Yes 2,500 No No Medium—No S&B Yes 5,000 No No Large—No S&B Yes 10,000 No No Small—Labor Cooperation Yes 2,500 Yes Yes Medium—Labor Cooperation Yes 5,000 Yes Yes Large—Labor Cooperation Yes 10,000 Yes Yes 1280 B.Ruesink, S.Gronau forests as common-pool resources (Ostrom, 2019), the model highlighted the risk of overuse and its implications for food security. Unlike models that extrapolate from past data, this ABM used structural knowledge to project future scenarios that differ significantly from past observations. The model’s reliability was ensured by running multiple simulations with different random seeds, verifying consistency despite inherent randomness. To ensure the robustness of the model, validation against empirical data focused on three key patterns: (1) the positive effects of slash-and-burn agriculture on soil composition and tree density, (2) the general and localized deforestation rate, and (3) increased deforestation around refugee settlements. These validation steps aligned with the KIA protocol (Troost etal., 2023) to ensure robustness and reliability. During model construction, key parameters and structural assumptions were systematically varied, including refugee settlement size (from no settlement to 10,000 residents), the inclusion of slashand-burn agriculture, and labor cooperation. These tests revealed scenario-specific outcomes consistent with theoretical and empirical expectations, demonstrating the model’s capacity to differentiate distinct outcomes under varied configurations. Sensitivity analyses of parameters such as tree density, minimum tree age, and the area for labor cooperation were all grounded in literature and empirical data. Tree composition was varied using data from Chidumayo (2019) and ultimately parameterized with Montford etal. (2021), ensuring that the representation of forest dynamics accurately reflected regional ecological contexts. These sensitivity analyses further supported the model’s robustness and suitability. 3 Results 3.1 Descriptive results The descriptive analysis highlights the rural host community’s heavy reliance on forest resources for food and energy. Table2 summarizes the key characteristics of the surveyed households, illustrating their dependency on forest resources, the challenges related to land use, and their contact with the nearby refugee settlement. As shown in Table2, 85% of the 277 households reported collecting firewood, and 59% consumed wild fruits the week preceding the survey. These findings emphasize the essential role that forest resources play in the host community's daily life and food security. A notable disparity exists between the average land ownership and actively cultivated area. While households own an average of 6.74 hectares of land, only 2.05 hectares are under cultivation. Labor shortages emerged as a significant constraint, with 34% of households identifying this as the main reason for leaving land uncultivated. However, other factors also contribute to this issue: 17% of households reported waiting for land to recover from slash-and-burn practices, while 16% cited financial limitations as a barrier to cultivation. The survey also revealed natural resource competition between the host community and the nearby refugee settlement, with 46% of host households reporting such an experience. Furthermore, 36% of the households hold a forest use license, which grants regulated access to forest resources. 3.2 Impact ofrefugee settlements andhost communities onforest resources Figure3 presents the predicted deforestation rates of the base models, which include only the host community before the implementation of a refugee settlement in later scenarios. These predictions are compared to annual deforestation rates reported by literature: a national rate for Zambia of 0.245% (2000–2014) (Global Forest Watch, 2023) and a localized rate for the Nchelenge district of 0.91% (2015–2016) (Republic of Zambia, 2016). These observed rates reflect conditions before the arrival of refugees in 2017. In the base Table 2 Sample characteristics of the host community in the Mantapala study area SD Standard deviations; Min Minimum value; Max Maximum value The table provides summary statistics, including mean, standard deviation (SD), minimum (Min), and maximum (Max) values, for relevant variables of the survey Variable Mean SD Min Max Distance to settlement (km) 4.80 1.60 2.71 8.79 Household size (members) 6.04 2.36 1 13 Respondent age (years) 42.60 14.58 17 88 Respondent education (years) 6.66 2.98 0 16 Gender (1 = female) 0.20 0.40 0 1 Land size (owned, hectares) 6.74 5.63 0 25 Land size (cultivated, hectares) 2.05 1.68 0 10 Uncultivated land due to:  Labor shortage (1 = yes) 0.34 0.47 0 1  Wait for recovery S&B (1 = yes) 0.17 0.38 0 1  Financial limitation (1 = yes) 0.16 0.37 0 1  License for forest use (1 = yes) 0.36 0.48 0 1  Firewood collection (1 = yes) 0.85 0.36 0 1  Consumed wild fruits in last 7 days (1 = yes) 0.59 0.49 0 1 Contact with refugees:  Real-life experience (1 = yes) 0.93 0.25 0 1  Competition for natural resources (1 = yes) 0.46 0.50 0 1  Wage competition (1 = yes) 0.10 0.30 0 1  Employment opportunities (1 = yes) 0.52 0.50 0 1 1287 Forest resource management, refugee integration, and food security in rural Zambia: balancing sustainability and equity 5 Conclusion Refugees have numerous positive effects on their host communities, such as improved access to schooling, health facilities, infrastructure, and stimulating economic activities. However, a large increase in population can lead to greater competition for natural resources. This study utilized an Agent-Based Model (ABM) to explore the impact of refugee influxes on forest resources and food security in rural Zambia. Our research aimed to address (1) how refugee settlements and host communities affect forest resources, (2) the impacts of varying settlement sizes on sustainable utilization of the forest and food security, and (3) the role of labor cooperation between refugees and hosts in mitigating forest loss. The ABM provided a nuanced understanding of the complex interactions between population growth, resource depletion, and food security, offering a valuable tool for future policy-oriented researchand implementations. The model showed several key findings directly answering the research questions: (1) Population pressure was identified as the critical driver of forest resource depletion, with larger settlements unable to sustain tree regeneration under high firewood demand. While the initial settlement construction had a modest impact on forest resources in smaller settlements, cumulative population pressure led to depletion across all scenarios. (2) No sustainable settlement size was identified, as even the smallest modeled scenario resulted in unsustainable outcomes due to cumulative pressures. (3) Labor cooperation between refugees and hosts presented mixed outcomes on food security. While it facilitated expanded agricultural activities, this benefit was countered by the accompanying decrease in forest resources. Beyond addressing these research questions, the study revealed further insights into resource use and environmental dynamics. While slash-and-burn agriculture temporarily improves soil fertility and forest cover, its benefits are quickly outweighed by the unsustainable firewood demand driven by population pressure. Firewood collection, identified as the primary driver of forest loss, led to unsustainable forest use even without agricultural expansion. This reduction in forest resources negatively impacts two key dimensions of food security: availability and stability. Declines in wild fruits and edible insects diminish food availability, while the long-term decline in forest-based food resources undermines food stability, increasing vulnerability for both refugee and host communities. This increased competition for diminishing resources may also exacerbate tensions between hosts and refugees, hindering integration efforts. Although property rights enforcement alleviates firewood competition for hosts, its success in fostering integration is contingent on adequate energy and resource support for both groups. These results emphasize the importance of timely interventions during the establishment of refugee settlements, particularly those exceeding 10,000 residents, which require immediate support to prevent severe deforestation. Smaller settlements, such as those with 2,500 refugees, leave more room to take action. Measures such as promoting agroforestry, introducing improved cooking stoves, and initiating reforestation efforts are means to reduce firewood dependency and mitigating resource conflicts. These measures are essential for environmental sustainability, enhancing food security, and fostering positive host-refugee interactions. The insights provided by this study contribute to the growing body of knowledge on the environmental tradeoffs of refugee settlement planning, offering a nuanced perspective on the dynamics of forest resource depletion and the need for integrated solutions. While the current model focuses on rural Zambia, its findings can inform similar settings within SSA's Comprehensive Refugee Response Framework (CRRF). Future research should focus on refining the model by incorporating policy interventions to explore sustainable outcomes. Addressing the refugee situation effectively from the onset can mitigate forest loss, reduce resource conflicts, increase food security, and ultimately support the successful integration of refugees into their host communities. Such integration helps refugees find a new home and enables impoverished rural host communities to benefit from the presence of refugees in sustainable and mutually beneficial ways. 1288 B.Ruesink, S.Gronau Appendix Model outcomes inlabor cooperation scenarios Fig. 6 Predicted changes in tree numbers over 30 years for six scenarios, varying refugee settlement sizes (Small = 2,500/Medium = 5,000/ Large = 10,000 refugees) and inclusion or exclusion of labor cooperation. Solid lines represent scenarios without labor cooperation, while dashed lines represent scenarios with labor cooperation. Blue lines refer to a settlement size of 2,500 refugees, orange lines to a settlement size of 5,000 refugees, and green lines to a settlement size of 10,000 refugees 1289 Forest resource management, refugee integration, and food security in rural Zambia: balancing sustainability and equity Fig. 7 Households unable to fulfill their annual firewood demand over 30 years, comparing six model scenarios with varying refugee settlement sizes (Small = 2,500/Medium = 5,000/Large = 10,000 refugees) and the option for labor cooperation between refugees and hosts. a refers to the total number of households unable to meet firewood demand across all six model scenarios. Solid lines represent scenarios without labor cooperation, while dashed lines represent scenarios with labor cooperation. Blue lines refer to a settlement size of 2,500 refugees, orange lines to a settlement size of 5,000 refugees, and green lines to a settlement size of 10,000 refugees. b refers to the share of refugee and host households within their population unable to meet firewood needs in scenarios with a settlement size of 2,500 refugees, with and without labor cooperation. Solid lines represent scenarios without labor cooperation, while dashed lines represent scenarios with labor cooperation. Blue lines refer to refugee households, and orange lines to host households 1290 B.Ruesink, S.Gronau Model assumptions andtheir justification Table 3 Elaboration and justification for the Agent-Based Model (ABM) assumptions Assumption Source Justification/Elaboration Model setup Setup host households Survey data These variables are based on the households’ individual answers to the questionnaire. We directly import them into our model, and the households can access them to determine their individual demands/behaviors. All data refers to the last year before the survey Initiated variables: HHID: Household Identification Number village: The household’s village land-owned: The household’s owned land in hectare land-cultivated: The household’s cultivated land area in hectares trees-home-use: The number of trees the household collected for home use area-slash-burn: The area the household cultivated with slash-and-burn agriculture in hectare fallow-time-before-cultivation: The number of years the household left their land fallow after slash-and-burn activities before cultivating it trees-slash-burn-home-use: The number of trees the household took home after slash-and-burn activities for home use usage-time-cleared-land: The number of years the household used cleared land for cultivation purposes fallow-time-before-slash-burn: The number of years the household left the land fallow before performing slash-andburn activities again Setup refugee households (WFP & UNHCR, 2021) The number of refugees is divided by the household size of 5 (WFP & UNHCR) to represent households instead of individuals. This assumption is made as households use firewood together and to adjust the data to match the host household dataset Setup refugees as a one-time influx (UNHCR, 2021b, c) The model implements the refugees’ arrival as a one-time event instead of a constant influx. This assumption is based on the setting in Mantapala. The refugees arrived in a transition center and moved to Mantapala a few months after the initial influx to Zambia. The main occurrence (and significant change for the host community) happened in early 2018 Setup refugee settlement (UNHCR, 2020; WFP & UNHCR, 2021) The refugee settlement area is cleared of trees (and no new trees are regrowing during the model run). Based on the literature, the reference size is 8,000 hectares for 25,000 refugees. The size adjusts linearly to the number of refugees in the model run Setup trees (Montfort etal., 2021; Weber etal., 2023) Based on Montfort etal., a random, normally distributed number of trees with a mean value of 81 and a standard deviation of 6.32 is created on each plot (0.25 hectare) 1291 Forest resource management, refugee integration, and food security in rural Zambia: balancing sustainability and equity Table 3 (continued) Assumption Source Justification/Elaboration Setup tree composition (Kaluba etal., 2021) As the host community already performed slash-and-burn agriculture before the refugees arrived (Kaluba etal.), it is simulated that they practiced S&B for 15 years in the setup. This way, a more sophisticated tree composition is created, which better represents the common agricultural practice  Setup villages and farms Survey data The host households move to the GPS coordinates associated with their village. Afterward, they continue claiming free plots around them until they reach the number of plots owned indicated by the survey Structural processes and rules Accessible plots Survey data Host households can access their own plots for firewood collection and forest plots. Refugees can only access the forest plots. Plots are only accessible with at least one tree above the minimum age present Agricultural practice Survey data Host households use farmland in different ways after it was affected by slash-and-burn agriculture (For each period, the duration was indicated by the household in the survey): First, land is fallow Second, land is cultivated Third, land is fallow before it can be used for S&B again Annual deforestation rates of 0.245% for Zambia and 0.91% for Nchelenge (Republic of Zambia, 2016) (Global Forest Watch, 2023) This rate was chosen based on observed deforestation rates in Zambia from 2000 until 2014 and for Nchelenge from 2015 to 2016. These were the available rates closest to the time of data collection in 2018 Cut trees Survey data The households cut trees based on their defined firewood demand for the year. Households move to the nearest accessible plot with the highest number of available trees (above minimum age). Hosts prefer their own over forest plots. The households start to cut down the oldest trees as they provide the most energy. Two cases are possible for the collection of trees: 1) Enough trees to fulfill demand: The household cuts down the demanded number of trees and sets its remaining demand for the year to zero 2) There are not enough trees: Households cut down all available trees on the plot and reduce their yearly demand by this amount Cut trees limit (Adeyonu, 2014; Bardasi & Wodon, 2006; Vianello, 2016) Households cut trees until they fulfilled their demand for the year or until they collected twice a week for a year (104 times). We define the system as unsustainable if they collect more than twice weekly as the households have less time for other activities, like education and farming 1292 B.Ruesink, S.Gronau Table 3 (continued) Assumption Source Justification/Elaboration Decision-making process of agents The decision-making process is simplified to focus on key behaviors relevant to the model’s objectives, such as resource collection and land use Define firewood demand Survey data The households first define their firewood demand for the year. Therefore, the stock (if there is any) is used to fulfill the demand. If the stock cannot fulfill the demand, the remainder is the amount they must collect for the year Firewood demand is constant over time Survey data The model bases the annual firewood demand on the survey data without accounting for potential changes in consumption due to climate change or deforestation. As environmental conditions evolve, households may need to adjust their usage or seek alternative energy sources. However, we want to create a “sticking to the status quo” analysis to evaluate forest and food security changes without adjustments Hire refugees Survey data (Gronau & Ruesink, 2021; ILO, 2011) Households hire refugees for cultivation purposes with a probability of the household’s state variable “probability-laborcooperation”. Hiring a refugee means that the desired cultivation area increases by the number of hectares per member in the working age of the household. The assumption is that the host households can now cultivate more of their owned land because of the additional labor Households collect firewood a maximum of two times per week With the collection of two times per week, we already see issues with firewood provision. We acknowledge that, in reality, the households would increase the collection times but don’t see the need to implement this in the model as we have already proven an unsustainable behavior at this point. During the model construction, a higher number was tested, which led to more households being able to fulfill their demand and fewer trees in the later model stages compared to the presented model outcomes Population growth rate of 3.6% (Zambia Statistics Agency, 2022) The population (refugees and hosts) grows at a rate of 3.6% annually. This is based on the Zambia Statistics Agency's average annual population growth rates from Luapula Province between 2010 and 2022 1293 Forest resource management, refugee integration, and food security in rural Zambia: balancing sustainability and equity Table 3 (continued) Assumption Source Justification/Elaboration Model schedule 1. Model setup 2. Annual processes: a. Plots recover from S&B, and trees grow b. All people collect firewood in a random order c. The hosts hire refugees d. The hosts perform slash-and-burn agriculture (if desired) e. Population growth We start the yearly processes by letting the environment grow. Afterward, we implement human activities. We start with the firewood collection as it is a continuous activity. Afterwards, the hosts hire refugees and perform their agricultural activities. During the model creation, this order was changed without significant changes in the results For the yearly firewood collection, we assume a random order of collection. An order could invalidate our results as it cannot be proven with our data, nor was it visible in the literature that firewood collection happens in a specific order of households. Different orders (first hosts, then refugees, and vice versa) were tested during model construction, leading to similar trends in the presented model version Slash-and-burn practice Survey data (1) The household moves to a random plot that is accessible for slash-and-burn agriculture (it belongs to the household and is currently not in one of the slash-and-burn phases) (2) The plot is cleared from trees, and the household takes a specific number of trees home (indicated by survey data). These trees are added to the stock, which is used first to fulfill demand (3) The plots set their status to slash-and-burn (tree growth adjusts accordingly) The spatial Scale of the model is 13,000 hectares (Gronau etal., 2018) The Mantapala study region is 13,000 hectares in size. In our model, each plot represents 0.25 hectares, allowing a precise plot distribution to the host households based on the survey data Temporal Scale of 30 years (Weber etal., 2023) The model runs for 30 years to see the refugees’ effect on deforestation after their initial arrival 1294 B.Ruesink, S.Gronau Table 3 (continued) Assumption Source Justification/Elaboration Tree growth and recovery on farms Survey data (Montfort etal., 2021) Interview with the local forest department Each year the farm's trees grow older, and new trees grow The regrowth of new trees depends on the status of the plot cultivation. The tree density is based on Montfort etal. The duration of cultivation and fallow periods are imported from the survey data (1) Plot was not cultivated by slash-and-burn agriculture, or this practice happened at least 36 years ago: 64 new trees/ha with a maximum of 324 trees/ha (2) The Plot is currently cultivated: No trees are growing on this plot, which is currently cultivated with crops (3) The plot was cultivated by slash-and-burn agriculture and is currently not in cultivation status. 135 new trees/ha with a maximum tree density (maximum number of trees/ha) dependent on time passed since S&B. 4 thresholds: a. 1year: 660 trees/ha b. 8years: 592 trees/ha c. 17 years: 408 trees/ha d. 29 years: 512 trees/ha Tree growth in forest (Montfort etal., 2021) Interview with the local forest department Each year, the trees in the forest grow older, and new trees grow. 64 new trees grow per hectare per year (estimation of forest official), with a maximum of 324 trees per year (Montfort etal.) Parameter values Contact probability set to an individual value for households (Gronau & Ruesink, 2021) Based on Gronau and Ruesink, we define a contact probability with their results for the bivariate probit model. Four variables are defined as significant for the contact: 1) Age54: Is the responder at least 54? Yes: -0.0964 2) EducPri: Did the respondent finish primary education? Yes: + 0.1360 3) LongDist: Is the household in a village far from the camp? Yes: -0.1139 4) BetterLife: Does the household have better life satisfaction than the year before? Yes: + 0.080 Hectare per member of working age is set to individual value for households Survey data (ILO, 2011) To define how productive the household is in cultivating land, we divide the hectare of land the household cultivated in a year by the number of household members above the age of 14 (ILO, 2011). This value is used for the “hire refugees” process Minimum age of trees Interview with the forestry department Households cannot cut a tree with an age below 15. We assume that households do not collect small trees as they provide less energy. Based on the interview, we identified an age of 15 as a reasonable threshold 1295 Forest resource management, refugee integration, and food security in rural Zambia: balancing sustainability and equity Table 3 (continued) Assumption Source Justification/Elaboration The number of cuttable trees on each plot is set to an individual number Interview with the forestry department Trees cannot be cut if they are under 15 years old. This is the estimated age at which the trees are big enough to be considered by the households for collection. The number of cuttable trees on a plot is the number of trees present above the minimum age of 15 Number of refugees in the model: 2,500/5,000/10,000 (Weber etal., 2023) Given that Weber etal. found no sustainable forest resource management for a settlement size of approximately 18,000 refugees, we decided to test smaller settlement sizes The initial number of trees is 324 trees/ha (Montfort etal., 2021) The number of trees indicated by Montfort etal The probability of labor cooperation is 0.5: Both labor cooperation and no cooperation have the same probability (The final decision for cooperation is still depending on the availability of land) Survey data (Gengo etal., 2018) Option1: HH experienced no impact from the refugee settlement Option 2: HH experienced positive contact, but HH is employed at camp. It is assumed that the positive contact is based on this employment, as Gengo etal. indicate positive effects of employment opportunities for hosts Probability labor cooperation is 1: Host household open to labor cooperation Survey data (Gronau & Ruesink, 2021) The host household experienced positive contact with refugees, and no member is employed at the settlement. Therefore, the household is open to cooperating in land cultivation, based on the connection between contact and opinion found by Gronau and Ruesink The probability of labor cooperation is 0: The host household is not open to labor cooperation Survey data (Gronau & Ruesink, 2021) Host households experience negative contact with refugees and are therefore not open to cooperating for land cultivation. Based on the connection between contact and opinion found by Gronau and Ruesink Tree demand is set to individual value for host households Survey data The households add up the number of trees they collected for home use and the number of trees they took home from slashand-burn activities in the year before the survey Tree demand is set to individual value for refugee households Survey data (Clark, 2002; Larmer, 2016; Vlassenroot & Verweijen, 2017) The tree demand of refugee households is based on the assumption that they behave similarly to the host households. We justify the assumption using the same environment in which the two groups interact. Additionally, the Democratic Republic of Congo and Zambia are connected by a long migration history, supporting the assumption of similar livelihood strategies. They set their demand to a random-normal number with a mean value of 203 and a standard deviation of 409 based on the survey data (Minimum of zero) Each assumption is described in detail, including its basis in empirical data, literature references, and its relevance to the dynamics modeled in the study 1296 B.Ruesink, S.Gronau Supplementary Information The online version contains supplementary material available at https:// doi. org/ 10. 1007/ s1257102501547-3. Acknowledgements The dataset originates from the “FoSeZa – Food Security in rural Zambia” project, funded by the German Federal Ministry of Food and Agriculture (BMEL) [2813 FSNU11]. We want to thank the people living in Mantapala for their continuous support. The authors also thank the Zambian Ministry of Agriculture, the Ministry of Fisheries and Livestock and the Zambia Agriculture Research Institute (ZARI) for field work assistance at the study site. Author contributions Conceptualization: Brigitte Ruesink, Steven Gronau; Methodology: Brigitte Ruesink; Formal analysis and investigation: Brigitte Ruesink; Writing – original draft preparation: Brigitte Ruesink; Writing – review and editing: Brigitte Ruesink, Steven Gronau; Funding acquisition: Steven Gronau; Supervision: Steven Gronau. Funding Open Access funding enabled and organized by Projekt DEAL. This research is part of the Leibniz Young Investigator Grant by the Leibniz University Hannover, Germany. Grant number LYIG-08–2019-11. Data availability The data and model code to recreate the results are available in the supplementary information. Further data presented in this study is available on request from the authors. Declarations Ethics approval Ethical clearance for field research design was granted by the University of Zambia (UNZA). Consent to publish Consent for publication is not applicable, as the manuscript does not contain data that could be used to identify survey participants. Competing interests The authors have no relevant financial or nonfinancial interests to disclose. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. References Adams,M. (2022). Green charcoal and livelihoods in Zambia. Caritas Scotland. https:// www. carit as. eu/ greenc har c oalandlivel ihoodsinzambia/. Accessed 07.11.2023. Adeyonu,A.G. (2014). Gender time allocation and farming households’ poverty in rural Nigeria. World Journal of Agricultural Sciences, 2(5), 123–136. http:// eprin ts. lmu. edu. ng/ 585/. Accessed 05.06.2022. Agblorti, S. K. M., & Grant, M. R. (2019). Conceptualising obstacles to local integration of refugees in Ghana. Refugee Survey Quarterly, 38(2), 195–213. https:// doi. org/ 10. 1093/ rsq/ hdz005 Alix-Garcia, J., Walker, S., & Bartlett, A. (2019). Assessing the direct and spillover effects of shocks to refugee remittances. World Development, 121, 63–74. https:// doi. org/ 10. 1016/j. world dev. 2019. 04. 015 Araya, F., Faust, K. M., & Kaminsky, J. A. (2021). Agent-based model of hosting communities’ perceptions of water and wastewater infrastructure during the German refugee crisis. Journal of Management in Engineering, 37(4), 4021035. https:// doi. org/ 10. 1061/ (ASCE) ME. 19435479. 00009 38 Bardasi,E., & Wodon, Q. (2006). Measuring time poverty and analyzing its determinants: Concepts and application to Guinea. https:// mpra. ub. unimuenc hen. de/ id/ eprint/ 11082. Accessed 09.10.2023. Barman,B.C. (2020). Impact of refugees on host developing countries. In S. K. Das & N. Chowdhary (Eds.), Refugee crises and third-world economies: Policies and perspectives (Vol. 6, pp.103–111). Emerald Publishing Limited. https:// doi. org/ 10. 1108/ 978-18398219052020 1011 Bernard,B., Aron,M., Loy,T., Muhamud,N.W., & Benard,S. (2020). The impact of refugee settlements on land use changes and vegetation degradation in West Nile Sub-region, Uganda. Geocarto International, 1–19. https:// doi. org/ 10. 1080/ 10106 049. 2019. 17040 73 Bildirici,M.E., Castanho,R.A., Couto,G., & Yılmaz Genç,S. (2023). Refugees, traditional energy consumption, environmental pollution, and deforestation: Fourier BARDL method. 2211–467X, 48, 101109. https:// doi. org/ 10. 1016/j. esr. 2023. 101109 Bilgili, Ö., Loschmann, C., Fransen, S., & Siegel, M. (2019). Is the education of local children influenced by living near a refugee camp? Evidence from host communities in Rwanda. International Migration, 57(4), 291–309. https:// doi. org/ 10. 1111/ imig. 12541 Boshuijzen-van Burken, C., Gore, R., Dignum, F., Royakkers, L., Wozny, P., & Shults, F. L. (2020). Agent-based modelling of values: The case of value sensitive design for refugee logistics. Journal of Artificial Societies and Social Simulation, 23(4), Article 6. https:// doi. org/ 10. 18564/ jasss. 4411 Brockerhoff, E. G., Barbaro, L., Castagneyrol, B., Forrester, D. I., Gardiner, B., González-Olabarria, J. R., Lyver, P. O., Meurisse, N., Oxbrough, A., Taki, H., Thompson, I. D., van der Plas, F., & Jactel, H. (2017). Forest biodiversity, ecosystem functioning and the provision of ecosystem services. Biodiversity and Conservation, 26(13), 3005–3035. https:// doi. org/ 10. 1007/ s105310171453-2 Carciotto, S., & Ferraro, F. (2020). Building blocks and challenges for the implementation of the global compact on refugees in Africa. Journal on Migration and Human Security, 8(1), 83–95. https:// doi. org/ 10. 1177/ 23315 02420 908247 Chidumayo, E. N. (2019). Is charcoal production in Brachystegia-Julbernardia woodlands of Zambia sustainable? Biomass and Bioenergy, 125, 1–7. https:// doi. org/ 10. 1016/j. biomb ioe. 2019. 04. 010 Clark, J. F. (2002). The African stakes of the Congo war. Palgrave Macmillan. Collins,A.J., & Frydenlund,E. (2016). Agent-based modeling and strategic group formation: A refugee case study. In 2016 Winter Simulation Conference (WSC). IEEE. https:// doi. or g/ 10. 1109/ wsc. 2016. 78221 84 Crooks, A. T., & Hailegiorgis, A. B. (2014). An agent-based modeling approach applied to the spread of cholera. Environmental Modelling & Software, 62, 164–177. https:// doi. org/ 10. 1016/j. envso ft. 2014. 08. 027 d’Errico, M., Mariani, R. D., Pietrelli, R., & Rosati, F. C. (2022). Refugee-host proximity and market creation in Uganda. The Journal of Development Studies, 58(2), 213–233. https:// doi. org/ 10. 1080/ 00220 388. 2021. 19617 49 Dare, O., & Abebe, A. M. (2018). Regional solutions and the global compact on refugees: The experience from Africa. International Journal of Refugee Law, 30(4), 704–706. https:// doi. org/ 10. 1093/ ijrl/ eey063 Deißler,L., Mausch,K., Grote,U., Karanja,A., & McMullin,S. (2024). Aspirations, ambitions and the adoption of diverse fruit trees – a case study of the livelihood effects in Kenya. 2666–7193, 16, 100544. https:// doi. org/ 10. 1016/j. tfp. 2024. 100544