An Integrated Decision Support Framework for Sustainable Road Planning in Biodiversity Hotspots

An Integrated Decision Support Framework for Sustainable Road Planning in Biodiversity Hotspots | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article An Integrated Decision Support Framework for Sustainable Road Planning in Biodiversity Hotspots Philipo Mtweve, Abraham Eustace, Qambemeda Nyanghura, Vincent Moseti, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7419361/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Global infrastructure investment will reach $ 94 trillion by 2040, posing critical threats to biodiversity in rural regions newly exposed to road access. Traditional planning approaches have failed to sufficiently incorporate stakeholder preferences and account for uncertainty, leading to decisions that often compromise ecosystems. To overcome these limitations, we introduce an integrated decision-support framework that fuses Multi-Criteria Decision Analysis (MCDA) with Bayesian Belief Networks (BBN), allowing explicit incorporation of stakeholder-derived preferences and probabilistic modeling of ecological and socioeconomic outcomes. Applying this framework to four proposed roads in Tanzania's Greater Serengeti Ecosystem, we show that conservation-oriented options (Mbulu and Eyasi) offer 15–25% better sustainability performance than conventional connectivity-maximizing routes, while serving 70% of the population with 291% higher per-capita efficiency. Thirty-year projections confirm the long-term benefits of ecosystem-sensitive designs, and cross-validation between MCDA and BBN confirms methodological robustness (r, 0.89). This study provides a scalable, evidence-based planning framework for sustainable road development. Earth and environmental sciences/Ecology/Ecosystem ecology Earth and environmental sciences/Environmental sciences/Environmental impact Bayesian networks conservation road infrastructure multi-criteria analysis sustainable development Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Global road infrastructure investment will reach $ 94 trillion by 2040, with developing countries requiring $ 2.3 trillion annually to meet sustainable road development goals¹. This expansion occurs amid accelerating biodiversity loss, with species extinction rates up to 100 times above natural background levels². The conflict between infrastructure development and biodiversity conservation is most pronounced in biodiversity hotspots, which are 36 regions that cover only 2.4% of Earth's surface yet harbor 44% of vascular plants and 35% of vertebrates, while supporting 2.3 billion people³. Infrastructure development, particularly investments into roads, is essential for rural development but also a primary driver of habitat fragmentation. Global road networks have expanded by 60% since 1990, with an additional 25 million kilometers planned by 2050⁴. Case studies for the negative impacts of roads on biodiversity and ecosystems can be found worldwide: Brazil's BR-319 highway increased deforestation by 95% within 50km⁵, Indonesia's trans-Papua highway threatens the world's largest intact tropical forest, and Africa's planned development corridors could fragment 33% of remaining intact forests⁶. Negative impacts from roads can be minimized through biodiversity-sensitive planning, but conventional approaches remain fundamentally limited. Early road infrastructure methodologies emphasized engineering feasibility and economic cost-benefit analysis, often treating environmental and social impacts as externalities to be mitigated post hoc⁷. Even where Environmental Impact Assessments (EIAs) are required, they typically occur after route selection, limiting their influence on strategic decisions. More recent approaches have incorporated Geographic Information Systems (GIS) for spatial optimization, seeking routes that minimize environmental impact while maximizing socioeconomic benefits⁸⁻¹⁰. Spatial optimization tools using Geographic Information Systems (GIS) have added environmental dimensions to route planning, but still face three persistent limitations: inadequate integration of stakeholder preferences, weak treatment of uncertainty, and limited consideration of long-term, system-level feedbacks. One significant area of uncertainty in road development planning is the unpredictability of ecological responses to road construction. For example, the introduction of a new road can lead to unforeseen changes in wildlife behavior, such as altered migration patterns or increased poaching, which can cascade into broader ecosystem impacts. This uncertainty complicates decision-making, as planners may not fully understand the long-term ecological consequences of their actions. Multi-Criteria Decision Analysis (MCDA) has emerged as a promising approach for addressing stakeholder preference integration, providing systematic frameworks for incorporating multiple, often conflicting objectives while maintaining transparency in trade-off evaluation¹¹⁻¹³. It enables explicit integration of stakeholder preferences through structured weighting processes and has been successfully applied to transportation planning, energy infrastructure, and conservation prioritization¹⁴. However, MCDA's primarily deterministic nature limits its ability to handle uncertainty and complex system dynamics. Bayesian Belief Networks (BBN) offer complementary strengths for addressing uncertainty and system complexity. BBNs model probabilistic relationships among variables, explicitly incorporating uncertainty while capturing indirect effects and feedback loops¹⁵, ¹⁶. They have proven valuable for environmental management, ecosystem service assessment, and infrastructure risk analysis¹⁷. Yet BBNs face challenges in transparent stakeholder preference incorporation and can be difficult to parameterize without extensive data. Recognition of individual method limitations has led to calls for integrated approaches that combine complementary methodologies. However, most integration attempts remain limited to sequential application rather than true methodological synthesis. The challenge lies in developing frameworks that leverage the structured preference incorporation of MCDA while capturing the uncertainty quantification and system dynamics of BBN approaches. This integration is particularly critical for road infrastructure planning in biodiversity hotspots, where decisions involve high stakes, multiple stakeholders with conflicting interests, substantial uncertainties about ecological responses, and long-term consequences that unfold over decades. Such contexts demand analytical approaches that can simultaneously address preference integration, uncertainty quantification, and temporal dynamics. East Africa represents a compelling case study for road infrastructure-biodiversity challenges. The region encompasses multiple biodiversity hotspots including the Eastern Afromontane, Coastal Forests of Eastern Africa, and Horn of Africa hotspots¹⁸, experiences rapid economic growth driving infrastructure expansion, and supports over 513 million people requiring improved connectivity. Recent regional infrastructure initiatives include the East African Railway Master Plan, Northern Corridor improvements, and multiple road development projects often negatively impacting valuable ecosystems. We use Tanzania's Greater Serengeti Ecosystem as a characteristic case study for developing and testing an integrated planning approach combining MCDA's systematic stakeholder engagement with BBN's uncertainty modeling capabilities. The GSE encompasses 30,000 km² of interconnected protected areas, including Serengeti National Park, Ngorongoro Conservation Area, and multiple game reserves¹⁹. It supports the world's largest terrestrial mammal migration (1.5 million wildebeest, zebras, and Thomson's gazelles), stores 1 million metric tons of carbon, and contains two UNESCO World Heritage Sites²⁰, ²¹. Simultaneously, the region supports 4.6 million people requiring improved market access, healthcare, and educational services²². Since 1990, four major road development options have been proposed by different government agencies and development partners to improve regional connectivity, each presenting different trade-offs between conservation and development objectives. The Northern route was proposed by the Ministry of Works and Transport (1995), the Serengeti route by tourism industry stakeholders (2008), the Eyasi route through collaborative planning between Tanzania National Parks Authority and regional development committees (2012), and the Mbulu route by conservation organizations and community-based natural resource management groups (2015). The aim of this paper is to apply the integrated transdisciplinary and interdisciplinary decision support framework to this specific question, namely which of the road options is less harmful from the perspective of biodiversity and sustainable development. This test should clearly show that the proposed methodological approach is applicable and yields helpful results. Results Stakeholder Engagement Reveals Different Preference Structures Our participatory stakeholder engagement process involved 30 regional experts and decision-makers across three structured phases: systematic literature synthesis, expert consultation, and collaborative workshops. This approach ensured representation from different sectors and stakeholder groups affected by road development while building methodological legitimacy through inclusive participation. The initial expert consultation at the Tanzania Wildlife Research Institute conference identified four criteria with 13 potential evaluation indicators across four domains: engineering feasibility, biodiversity conservation, socioeconomic development, and political interests. During the subsequent three-day stakeholder workshops, which included tourism operators, planning officials, field ecologists, road engineers, community representatives, and conservation organizations, the framework was refined to 16 criteria where three additional indicators were added (geology, aspect and Implementation feasibility). We then selected 13 indicators for spatial MCDA calculations. The selection of these 13 indicators was based on the availability of data in raster format, this selection excluded implementation feasibility, Travel time reduction and service delivery (Table 1). The Analytical Hierarchy Process (AHP) weighting revealed complex stakeholder priorities that challenged existing assumptions about the trade-offs between development and conservation, as shown in Table 1. Notably, biodiversity conservation and socioeconomic development were nearly equally valued, with weights of 36.5% and 37.5% respectively, indicating the desire for a balanced approach to environmental concerns and human needs. Within the engineering feasibility category, elevation emerged as the most critical factor at 31.4%, reflecting the challenges posed by the Serengeti's topography. Soil type (25.4%) and slope (19.6%) were also significant, highlighting practical construction constraints. Furthermore, the emphasis on habitat quality (7.8%) and market connectivity (8.5%) underscored the necessity for infrastructure that supports both ecological integrity and economic opportunities, suggesting a comprehensive strategy for sustainable development. Table 1: Stakeholder-derived criteria weights and consistency measures . Domain Criteria Weight (%) Indicators Weight (%) Definition and Measurement Protocol Engineering Feasibility Terrain & Construction Difficulty 13.0 Composite measure of physical construction challenges based on topographic and geological constraints Elevation 31.4 Mean elevation above sea level (m). Higher elevations increase construction costs and technical difficulty. Measured from 30m SRTM DEM. Soil type 25.4 Soil engineering properties affecting foundation stability and construction feasibility. Classified using Harmonized World Soil Database with field validation. Scale: 1-10 (10 = excellent engineering properties). Slope 19.6 Average terrain gradient (%) affecting construction complexity and equipment requirements. Calculated from DEM using QGIS slope analysis. Higher slopes = higher construction difficulty. Geology 12.8 Bedrock type and geological stability for infrastructure foundation. Based on Tanzania Geological Survey 1:250,000 maps. Scale: 1-10 (10 = excellent foundation conditions). Aspect 10.8 Terrain orientation affecting drainage, erosion risk, and construction access. Calculated from DEM. North-facing slopes preferred for reduced erosion. Biodiversity Conservation Ecosystem Impact 36.5 Composite measure of negative impacts on ecosystem integrity, wildlife populations, and conservation effectiveness Protected areas 3.8 Distance to protected area boundaries (km). Closer proximity increases conservation conflicts and legal constraints. Measured as minimum distance to any protected area boundary. Land use/cover 4.7 Naturalness and conservation value of vegetation communities. Based on ESA WorldCover 2021 classification. Scale: 1-10 (10 = intact natural habitat, 1 = degraded/converted land). Wildlife corridors 7.6 Critical pathways for wildlife movement between protected areas, mapped by Tanzania Wildlife Authority using GPS collar data and expert knowledge. Impact measured as corridor intersection length (km). Migration routes 8.0 Seasonal pathways used specifically during wildebeest migration (May-July and November-December), distinct from year-round wildlife corridors. Based on 15-year GPS collar synthesis from Serengeti Wildlife Research Centre. Impact = route intersection length during peak migration. Habitat quality 7.8 Ecological condition and species-supporting capacity of natural areas. Assessed through: (1) vegetation density index from Landsat imagery, (2) water source proximity, (3) expert field assessments by TAWIRI researchers. Scale: 1-10 (10 = pristine habitat condition). Socioeconomic Development Development Benefits 37.5 Composite measure of positive impacts on human welfare, economic opportunities, and service access Market connectivity 8.5 Improvement in access to agricultural and livestock markets measured as: (1) reduced travel time to nearest major market (hours), (2) population within 2-hour travel time of markets. Based on existing road network analysis and market location mapping. Settlement 7.5 Number of people gaining improved road access within 10km of route alignment. Population data from Global Human Settlement Layer 2020, validated through district census data and field surveys. Service delivery 6.9 Improvement in access to healthcare, education, and government services. Measured as: (1) population within 1-hour travel time of health facilities, (2) children within 30 minutes of primary schools, (3) communities within 2 hours of district headquarters. Travel time reduction 7.3 Decrease in travel time between major population centers enabled by new road infrastructure. Calculated using network analysis comparing current vs. proposed road networks. Measured in hours saved for representative origin-destination pairs. Population served 7.2 Total population within 50km corridor of route alignment who would benefit from improved connectivity. Based on Global Human Settlement Layer with 50km buffer analysis, validated against district population statistics. Political Implementation Stakeholder Support 13.0 Feasibility of successful implementation considering political, social, and institutional factors Implementation feasibility 13.0 Composite measure including: (1) government agency support levels, (2) community acceptance based on public consultations, (3) donor/financing availability, (4) regulatory approval probability, (5) absence of major opposition groups. Assessed through stakeholder mapping and consultation feedback. Scale: 1-10 (10 = strong support, feasible implementation). Note: Consistency Ratio (overall): 0.037; Individual CR range: 0.018-0.052. Weights represent average values across all stakeholder groups derived through Analytical Hierarchy Process ( AHP ). Domain weights sum to 100.0%. Indicator weights shown are relative percentages within each domain. Individual stakeholder weights were aggregated using geometric means to maintain consistency with AHP principles. Importantly, consistency analysis yielded acceptable ratios across all participant groups (CR < 0.1), indicating logically structured preference structures rather than random responses. Individual consistency ratios ranged from 0.018-0.052, suggesting participants carefully weighted trade-offs in their decisions. Furthermore, cross-group correlation analysis revealed strong agreement among different stakeholder groups (r = 0.73-0.89), pointing to a shared understanding of regional development challenges despite different professional backgrounds. Spatial Analysis Reveals Fundamental Differences among Road Options A spatial analysis of the four proposed routes, utilizing 13 standardized datasets at a 1 km resolution, identified Eyasi Road as the most feasible and sustainable option among the four options. This route offers unique opportunities for balancing conservation and development goals. However, each route crosses distinct ecological and social landscapes and encounters various implementation challenges that significantly impact long-term sustainability outcomes. Figure 1 illustrates the spatial patterns and characteristics of the routes, displaying the Greater Serengeti Ecosystem with the four proposed road development options superimposed on the results of the MCDA suitability analysis. The color coding reflects route suitability based on integrated criteria, with green areas indicating suitable zones and grey areas marking regions deemed unsuitable for road development. The Serengeti route represents the shortest option at 287 km but crosses protected areas directly, including 198.9 km within Ngorongoro Conservation Area and Serengeti National Park boundaries. While offering apparent economic benefits through reduced travel distances, mandatory speed limits of 30-50 km/hour within protected areas actually increase total travel time, negating the distance advantage. The route serves approximately 971,000 people but requires the highest construction costs ($4.50 million per kilometer) due to extensive wildlife-friendly infrastructure requirements including animal crossing structures, specialized fencing, wildlife detection systems, and environmental monitoring protocols mandated for protected area construction. The Northern route spans 342 km connecting Musoma to Arusha via the northern ecosystem boundary, serving the largest population (2.1 million people within a 50km buffer) while intersecting critical wildlife corridors during peak migration periods. Detailed spatial analysis indicates moderate terrain challenges (average slope: 12.3%, maximum: 45%) but substantial ecological risks, with 67% of the route within 50km of protected area boundaries and 34% intersecting designated wildlife corridors used by over 900,000 migrating ungulates annually. Construction costs are high ($3.80 million per kilometer) reflecting the need for wildlife crossing infrastructure, animal detection systems, and specialized construction protocols to minimize wildlife mortality in ecologically sensitive areas. The Eyasi route follows a 398 km southern trajectory through diverse landscapes including highland forests, agricultural zones, and pastoral areas, serving 847,000 people while maintaining 15 km minimum distance from core protected areas. Terrain analysis indicates moderate construction challenges (average slope: 8.7%, soil stability index: 7.2/10) with strategic opportunities for wildlife-friendly design through careful corridor placement and crossing structures. The Mbulu route traces a 445 km southeastern boundary path, prioritizing biodiversity conservation while serving primarily rural communities (682,000 people). This route faces challenging terrain conditions (average slope: 15.2%, complex geology requiring specialized engineering solutions) but achieves the lowest construction costs ($2.45 million per kilometer) because it strategically avoids protected areas entirely, eliminating the need for expensive wildlife-friendly infrastructure. Multi-Criteria Decision Analysis Identifies Best Compromise Solutions MCDA analysis integrating stakeholder preferences with spatial datasets revealed clear performance hierarchies among route options while highlighting the complex trade-offs inherent in infrastructure planning within biodiversity hotspots. The analysis employed weighted overlay modeling using stakeholder-derived criteria weights, generating standardized scores from 0-1200 for each route across multiple sustainability dimensions. The comprehensive results of both MCDA and BBN analyses are presented in Figure 2, which demonstrates the trade-offs between biodiversity conservation and socioeconomic development across all four routes. Panel (a) shows MCDA performance scores, panel (b) illustrates the BBN network structure, and panel (c) presents BBN sustainability scores with uncertainty bounds. The Eyasi route emerges as the least harmful choice under integrated MCDA evaluation, achieving the highest combined performance with biodiversity conservation score of 1060/1200 and socioeconomic development score of 850/1200. This superior performance reflects strategic positioning that successfully balances multiple competing objectives: substantial population service (847,000 people) without critical habitat intersection, moderate terrain challenges enabling cost-effective construction while avoiding major engineering obstacles, and abundant design opportunities for wildlife-friendly infrastructure through careful corridor placement and strategically positioned crossing structures. The route's integrated score of 960/1200 represents a 104% performance advantage over the Northern route (470/1200) and substantially outperforms other connectivity-focused alternatives. The Mbulu route demonstrates exceptional biodiversity protection performance (1200/1200), representing perfect scores across all conservation criteria including protected area avoidance, wildlife corridor preservation, habitat fragmentation minimization, and endangered species protection. This outstanding environmental performance positions the route as the gold standard for ecosystem-sensitive infrastructure development. Its socioeconomic performance (750/1200) reflects moderate development benefits due to serving a smaller total population (682,000 people), but the route achieves superior cost-effectiveness with the lowest construction costs by strategically avoiding protected areas. Comprehensive sensitivity analysis across alternative weighting schemes confirms the robustness of these results while revealing decision contexts where different routes might be preferred based on varying stakeholder priorities. When biodiversity weights increase to 70% (development: 30%), Mbulu and Eyasi routes strengthen their comparative advantages with scores increasing by 12% and 8% respectively, while Northern route performance declines substantially with an 18% score decrease. Bayesian Belief Network Analysis Reveals System Dynamics and Uncertainty BBN modeling addresses fundamental limitations of deterministic MCDA by incorporating uncertainty, causal relationships, and system dynamics often missing from static evaluation approaches. Our participatory BBN development process engaged the same stakeholder groups in network structure definition and parameter estimation, ensuring methodological consistency with MCDA while adding crucial probabilistic dimensions that capture the inherent unpredictability of infrastructure outcomes in complex socio-ecological systems. The final BBN architecture includes 35 nodes representing infrastructure development drivers, ecological impact pathways, and socioeconomic outcome mechanisms connected through 100 carefully validated causal relationships, as shown in the network structure component of Figure 2. Network structure emerged through systematic stakeholder workshops where participants mapped causal pathways based on their professional experience, supplemented by literature evidence from 47 analogous developments in similar ecosystems. BBN analysis identified the Mbulu route as optimal under uncertainty conditions, achieving an overall sustainability score of 0.52±0.05 (on a 0-1 probability scale). This score represents the probability of achieving positive outcomes across integrated biodiversity and livelihood metrics, with higher values indicating greater likelihood of sustainable development success. The Mbulu route's superiority reflects its exceptional biodiversity protection probability (0.58, indicating 58% chance of maintaining ecosystem functionality) combined with substantial positive livelihood impacts (0.47, representing moderate but reliable socioeconomic benefits). The Eyasi route ranked close second place with sustainability score 0.47±0.04, demonstrating balanced performance across outcome domains while maintaining low uncertainty. The Northern route ranked third (0.41±0.03) despite high positive livelihood impacts (0.50), severely limited by poor biodiversity outcomes (0.31). The Serengeti route performed worst (0.40±0.06) with both poor biodiversity (0.37) and moderate livelihood outcomes (0.42), plus high uncertainty reflecting unpredictable protected area consequences. Long-term Temporal Projections Reveal Sustainability Trajectories Thirty-year sustainability projections incorporating cumulative environmental impacts through annual decay functions revealed stark differences in route resilience and long-term viability. Our temporal modeling employed route-specific decay rates calibrated to regional deforestation trends: conservation-oriented routes (Mbulu, Eyasi) experience slower ecosystem degradation (4.6-4.9% annually) while connectivity-focused routes (Northern, Serengeti) face accelerated decline (6.2-6.8% annually) due to induced development pressures. The critical transition analysis revealing permanent sustainability disadvantages for connectivity-focused infrastructure strategies is presented in Figure 3. Panel (a) shows thirty-year trajectories demonstrating initial convergence (0-10 years), critical transition (15-25 years), and permanent divergence (25-30 years), with conservation routes outperforming connectivity routes. Panel (b) shows Northern route components with economic benefits declining while environmental costs compound exponentially after year 15. Panel (c) illustrates performance comparison showing the Northern route crossing critical thresholds with permanent sustainability disadvantages. Conservation-oriented routes maintain substantially higher sustainability throughout projection periods, with critical divergence emerging during years 10-15 when cumulative impacts begin overwhelming initial benefits for connectivity-focused alternatives. The Mbulu route sustains performance longest, maintaining 0.32 by year 20 and 0.12 by year 30 (62% and 23% of initial performance). The Northern route drops to 0.22 by year 20 and 0.08 by year 30, reflecting compound environmental degradation. Critical transition analysis reveals precise thresholds where route performances converge initially but diverge permanently after year 15, creating irreversible sustainability disadvantages for connectivity-focused strategies. During the initial decade, the Northern route maintains 79% of the Mbulu route's performance as immediate economic benefits partially compensate for environmental costs. However, years 15-25 demonstrate accelerating divergence driven by exponential compounding of ecological impacts habitat fragmentation, wildlife population declines, and ecosystem service losses intensifying 40% faster than linear economic benefit decline. Cross-method validation and integration insights Cross-validation between MCDA and BBN approaches provides critical validation while revealing complementary insights that support decision-making. The relationship between the two analytical approaches and their efficiency implications are illustrated in Figure 4, which shows strong correlation (r=0.89) between MCDA and BBN approaches, with conservation-oriented routes demonstrating superior efficiency per capita served. Correlation analysis demonstrates strong overall agreement (r = 0.89, p < 0.01) between method rankings, confirming that both approaches identify two alternating road options but with very minor differences despite fundamentally different analytical foundations. However, detailed comparison reveals important nuances in method emphases that prove valuable rather than problematic. MCDA favors the Eyasi route's balanced approach to immediate trade-off optimization, while BBN identifies Mbulu route's superior uncertainty resilience and long-term sustainability. The integration reveals critical efficiency insights: Mbulu achieves 0.762 sustainability points per 1,000 people served versus Northern's 0.195 points per 1,000 people, representing 291% higher efficiency. These findings challenge conventional assumptions that maximizing population connectivity yields optimal outcomes, demonstrating instead that targeted sustainable infrastructure provides superior long-term value per capita served. Discussion Challenging conventional infrastructure paradigms Our results fundamentally challenge prevailing assumptions in road infrastructure planning that prioritize immediate connectivity benefits over integrated sustainability outcomes. The systematic underperformance of the Northern and Serengeti routes selected through traditional economic optimization criteria and currently under implementation demonstrates how narrow decision frameworks can yield globally suboptimal solutions with severe long-term consequences. Importantly, our analysis reveals that routes crossing protected areas (Serengeti and Northern) require the highest construction costs ($4.50 and $3.80 million per kilometer respectively) due to mandatory wildlife-friendly infrastructure including animal crossing structures, detection systems, specialized fencing, and environmental monitoring protocols. Despite these expensive mitigation measures, these routes still generate poor sustainability outcomes, questioning the economic logic of protected area development even when environmental safeguards are implemented. The superior performance of conservation-oriented routes reveals a critical insight contradicting conventional development wisdom: sustainable road infrastructure can achieve substantial connectivity benefits while providing better long-term outcomes at lower costs. The Mbulu route exemplifies this paradox, achieving the highest sustainability scores while requiring the lowest construction costs ($2.45 million per kilometer) by strategically avoiding protected areas entirely. This challenges fundamental assumptions about cost-effectiveness in infrastructure planning, suggesting that "conservation-by-design" represents the most economically rational approach. The successful MCDA-BBN integration addresses a fundamental sustainability science challenge: combining stakeholder preference incorporation with uncertainty quantification in complex decision contexts. Our approach demonstrates that methodological integration leverages complementary strengths while addressing individual limitations, providing more robust decision support than single-method approaches. The participatory stakeholder engagement process reveals sophisticated understanding of infrastructure trade-offs among local experts that conventional consultation processes often fail to capture. The strong consensus across different stakeholder groups (r = 0.73-0.89) demonstrates that meaningful agreement on complex trade-offs is achievable when supported by structured analytical frameworks. This contradicts common assumptions that stakeholder conflicts are irreconcilable and must be resolved through political rather than technical processes. BBN modeling proves particularly valuable for capturing indirect effects and feedback loops that deterministic approaches systematically miss. The probabilistic framework revealed how infrastructure decisions propagate through complex socio-ecological systems, generating cascading consequences that unfold over years or decades. Cross-validation provides crucial robustness testing while revealing complementary insights about different decision aspects. The framework demonstrates that conservation and development can be synergistic when supported by rigorous analytical tools and participatory planning processes. Our identification of Eyasi and Mbulu routes as superior alternatives provides immediate policy guidance for Tanzania while offering a replicable template for a more sustainable infrastructure planning in biodiversity hotspots globally. Successful implementation requires addressing technical capacity constraints; as integrated MCDA-BBN analysis requires specialized skills currently limited in many planning agencies. However, our experience demonstrates that participatory approaches can generate valuable results with modest technical resources when supported by appropriate facilitation. Future research should expand comparative analysis across additional biodiversity hotspots to test framework generalizability while identifying context-specific adaptations. Integration with emerging technologies including artificial intelligence and automated satellite monitoring could enhance evaluation frameworks while reducing implementation costs. Methods Study area and infrastructure development context The Greater Serengeti Ecosystem encompasses approximately 30,000 km² of interconnected protected areas in northern Tanzania, representing one of Africa's most significant conservation landscapes²³. The ecosystem includes Serengeti National Park (14,750 km²), Ngorongoro Conservation Area (8,292 km²), three game reserves (Maswa, Ikorongo-Grumeti, Kijereshi totaling 4,500 km²), Wildlife Management Areas (Makao, Ikona covering 2,200 km²), and Lolindo Game Controlled Area (1,000 km²) ²⁴. This diverse protected area network hosts the world's largest terrestrial mammal migration, with approximately 1.5 million wildebeest, 200,000 zebras, and 300,000 Thomson's gazelles participating in annual movements covering over 1,800 km²⁵. The ecosystem supports 70+ large mammal species, 500+ bird species, and provides critical habitat for endangered species including African wild dogs, cheetahs, and black rhinoceros. It sequesters approximately 1 million metric tons of carbon annually and contains two UNESCO World Heritage Sites plus one UNESCO Biosphere Reserve²⁶, ²⁷. Human populations within and around the GSE total approximately 4.6 million people with annual growth rates of 2.8%, creating increasing pressure for infrastructure development²⁸. Economic activities include pastoralism (affecting 65% of households), small-scale agriculture (35% of land use), tourism (contributing $1.8 billion annually to national economy), and natural resource extraction, with limited road access constraining market participation and service delivery²⁹, ³⁰. Four road development options have been proposed since 1990 to improve regional connectivity, each representing different institutional priorities and development philosophies: Northern route (Ministry of Works and Transport, 1995, emphasizing economic connectivity), Serengeti route (tourism industry stakeholders, 2008, prioritizing tourism access), Eyasi route (collaborative planning between Tanzania National Parks Authority and regional development committees, 2012, seeking balanced development), and Mbulu route (conservation organizations and community-based natural resource management groups, 2015, prioritizing biodiversity protection). Stakeholder Engagement and Data Collection We employed systematic three-phase data collection designed to ensure methodological rigor while building legitimacy through inclusive participation³¹, ³². The approach integrated established participatory research protocols with structured analytical frameworks to balance stakeholder ownership with technical precision³³. Phase 1: Literature Review and Expert Identification involved systematic review of 47 peer-reviewed publications on Serengeti ecology and road impacts using structured search protocols across multiple databases (Web of Science, PubMed, Google Scholar, African Journals Online). Search terms combined ecosystem identifiers with infrastructure and impact keywords, applying quality criteria for peer-reviewed publications and authoritative gray literature from 1990-2023³⁴. This phase identified key research gaps, established baseline knowledge, and informed expert selection criteria. Phase 2: Expert Consultation occurred at Tanzania Wildlife Research Institute annual conference (December 2022) with 14 specialists selected based on publication record (minimum 5 Serengeti-focused publications), field experience (minimum 10 years' regional work), and disciplinary diversity (ecology, economics, engineering, policy). Structured interviews employed standardized protocols covering impact pathway identification, parameter estimation feasibility, and stakeholder mapping for workshop selection. Individual consultation sessions (15 minutes each) were followed by 2-hour group synthesis to identify consensus areas and knowledge gaps. Phase 3: Collaborative Workshop (January 15-17, 2024, Dar es salaam) included 16 participants representing six stakeholder categories selected using power-interest matrix analysis³⁵. Participants included tourism operators (2, representing >$50M annual revenue), planning officers (4, from regional and district levels), field ecologists (2, PhD-level with >5 years GSE research), road engineers (4, professionally certified with rural experience), community representatives (2, elected leaders from affected areas), and conservation NGOs (2, with technical expertise and regional presence). Workshop design employed structured facilitation with daily rotating mixed-sector groups to prevent dominance effects and ensure cross-pollination of perspectives. Daily sessions followed systematic protocols: Day 1 focused on criteria development through individual brainstorming (30 minutes), small group synthesis (90 minutes), and plenary integration (120 minutes). Day 2 addressed weight estimation using Analytical Hierarchy Process training (2 hours) followed by individual implementation with consistency checking (5 hours). Day 3 covered Bayesian network development through collaborative structure mapping (3 hours) and parameter elicitation using probability wheels and structured scenarios (4 hours). All sessions were recorded with participant consent, with multiple independent note-takers ensuring accurate documentation. Multi-Criteria Decision Analysis Implementation MCDA implementation integrated established Analytical Hierarchy Process protocols with spatial analysis using standardized geographic datasets³⁶⁻³⁸. The approach followed systematic protocols for criteria standardization, weight determination, and spatial evaluation to ensure methodological rigor and reproducibility. Criteria Development and Standardization employed participatory identification of 16 potential indicators across four domains (engineering feasibility, biodiversity conservation, socioeconomic development, political implementation), refined to 13 spatially-analyzable indicators based on data availability and measurement feasibility. Each indicator received operational definitions, measurement protocols, and standardization procedures to ensure consistent evaluation across route alternatives. Standardization employed Min-Max normalization to 0-100 scales with higher values indicating better performance, accounting for indicator directionality (cost vs. benefit measures). Weight Determination used Analytical Hierarchy Process with structured pairwise comparisons at domain and indicator levels³⁹. Individual participants completed comparison matrices using established 9-point scales, with consistency checking using Consistency Ratio thresholds (CR < 0.1). Group weight aggregation employed geometric mean approaches to maintain AHP consistency properties while preserving individual preferences. Final weights underwent sensitivity analysis across alternative aggregation methods and stakeholder subgroups to test robustness. Spatial Analysis used QGIS 3.28 with 13 standardized 1km datasets covering terrain (elevation, slope, aspect from 30m SRTM DEM), geology (Tanzania Geological Survey 1:250,000 maps), soils (Harmonized World Soil Database v2.0), land use (ESA WorldCover 2021), protected areas (Tanzania National Parks Authority boundaries), wildlife corridors and migration routes (Tanzania Wildlife Authority GPS collar data), settlements (Global Human Settlement Layer 2020), and infrastructure (OpenStreetMap with ground-truthing). All datasets underwent comprehensive preprocessing including coordinate standardization (EPSG:32736), resolution harmonization, extent standardization, and quality validation through field verification at 50 stratified random points. Route-specific analysis evaluated performance within 1km buffer zones using zonal statistics, accounting for construction impact areas and immediate effects. Final suitability scores derived through weighted linear combination: S = Σ (wi × si), where S = overall score, wi = criterion weight, si = standardized score. Comprehensive sensitivity analysis tested alternative weighting schemes, normalization approaches, and buffer distances to assess result stability. Bayesian Belief Network Development and Analysis BBN development employed participatory modeling approaches combining stakeholder knowledge with literature-based calibration to create robust probabilistic frameworks for uncertainty analysis⁴⁰,⁴¹. The methodology integrated established BBN development protocols with participatory research methods to ensure technical rigor while maintaining stakeholder ownership. Network Structure Development used systematic group processes ensuring stakeholder ownership while maintaining technical feasibility. Participatory workshops employed structured techniques including individual variable identification (10-15 variables per participant), small group pathway mapping using sticky-note visualization on large-format charts, systematic validation of causal relationships using three criteria (logical consistency, empirical support, expert consensus), and complexity management through iterative simplification while preserving essential dynamics. The consensus network comprised 35 nodes organized in three hierarchical levels: infrastructure drivers (4 nodes: route characteristics, construction timeline, traffic volume, investment level), intermediate impacts (8 nodes: habitat degradation, wildlife mortality, corridor disruption, migration interference, edge effects, poaching access, pollution impact, agricultural expansion), and outcome variables (5 nodes: market access, healthcare access, education access, employment generation, tourism revenue). Each node received operational definitions with 2-3 discrete states and clear measurement protocols. The final architecture included 100 causal relationships, each validated using established criteria and documented with supporting rationale. Parameter Estimation involved structured expert elicitation for conditional probability tables using established protocols for environmental decision-making⁴². Individual elicitation employed systematic four-step approaches: node state definition with operational criteria, scenario presentation covering all parent node combinations, numerical estimation using calibrated probability wheels with confidence intervals, and consistency checking through probability rules and logical constraints. Group consensus building used equal-weight linear pooling with outlier identification (>2 standard deviations), structured discussion of reasoning behind divergent estimates, and final consensus with documentation of persistent disagreements. Probabilistic Inference and Analysis used R statistical software with specialized packages: bnlearn for network structure learning and validation, gRain for probabilistic inference and belief updating, and Rgraphviz for network visualization. Uncertainty propagation employed Monte Carlo analysis with 1,000 iterations incorporating parameter uncertainty (normal distributions around expert estimates), structural uncertainty (alternative network specifications), and scenario analysis (systematic infrastructure development variations). Temporal Analysis and Cross-validation Temporal modeling employed differential decay functions calibrated to regional empirical data to project 30-year sustainability trajectories⁴³, ⁴⁴. Cross-validation used multiple approaches to assess framework robustness and methodological reliability. Decay Rate Calibration used regional deforestation studies and infrastructure impact meta-analyses to establish route-specific decay parameters. Calibration employed 15-year Landsat analysis of forest loss around existing Serengeti roads with spatial buffers (1km, 5km, 10km), wildlife population monitoring data from Serengeti Research Institute (1990-2020), tourism revenue trends from Tanzania National Parks Authority, and ecosystem service assessment studies. Statistical analysis used exponential decay functions St = S0 × e^(-αt) fitted to empirical data with route-specific parameter estimation. This was done on the assumption that the current past 20-year trend (2000 – 2020) of changes will remain the same. Conservation-oriented routes (as defined by stakeholders) received decay rates of 4.6-4.9% annually, while connectivity-focused routes showed 6.2-6.8% rates based on induced development pressure patterns. Uncertainty Analysis employed Monte Carlo simulation with 1,000 iterations using Latin Hypercube Sampling for efficient parameter space coverage. Analysis incorporated parameter uncertainty (±10-20% variation around mean estimates), model uncertainty (ensemble modeling of alternative decay functions), scenario uncertainty (multiple development intensity pathways), and stochastic variation (historical environmental fluctuations). Convergence testing verified simulation adequacy while sensitivity analysis identified critical parameters requiring additional research attention. Cross-validation employed multiple approaches including correlation analysis (rank correlations, score correlations, uncertainty overlap assessments), bootstrap validation (n=500) for robustness testing across stakeholder subsamples and parameter variations, external validation through independent expert reviews, and comparison with outcomes from analogous infrastructure developments. Statistical significance testing used non-parametric methods (Kruskal-Wallis for multiple groups, Mann-Whitney U for pairwise comparisons) with Bonferroni correction. Effect sizes reported Cohen's d for continuous variables and Cliff's delta for ordinal comparisons. Integration Analysis compared MCDA and BBN rankings using correlation coefficients, assessed efficiency metrics (sustainability per capita served), and evaluated complementary insights from different analytical approaches. Results demonstrated strong overall agreement (r = 0.89) while revealing valuable methodological complementarity for different decision aspects. Declarations Data Availability All research data supporting the conclusions are available in the supplementary materials. Comprehensive results tables including route performance metrics and economic analysis are provided in Supplementary Tables S1-S4. Detailed methodology protocols, additional results tables, and extended sensitivity analyses are provided in Supplementary Materials S1-S6. Acknowledgements We thank Tanzania Commission for Science and Tanzania Wildlife Research Institute for research permits, and data access. We acknowledge the 30 workshop participants for sharing their expertise and time, and the communities of the Greater Serengeti Ecosystem for their insights and local knowledge. Funding provided by the Federal Ministry of Research, Technology and Space of Germany under the LANUSYNCON Project. Author Contributions P.J.M., A.E., and L.B.-F. conceived the study and developed the methodological framework. P.J.M., Q.M.N., and V.M. designed and conducted stakeholder engagement activities. P.J.M., T.K., and C.B. performed spatial data analysis and MCDA implementation. P.J.M. and A.E. developed BBN models and conducted uncertainty analysis. P.J.M., and Q.M.N. coordinated field activities and stakeholder relations. P.I. and L.B.-F. supervised methodology development and provided conceptual guidance. P.J.M. led manuscript writing with substantial contributions from all authors. All authors reviewed and approved the final manuscript. Competing Interests The authors declare no competing financial or non-financial interests related to this research. References Global Infrastructure Hub. Global Infrastructure Outlook, Infrastructure Investment Needs 50 Countries, 7 Sectors to 2040 (2017). Ceballos, G. et al. Accelerated modern human-induced species losses: entering the sixth mass extinction. Sci. Adv. 1, e1400253 (2015). Myers, N. et al. Biodiversity hotspots for conservation priorities. Nature 403, 853-858 (2000). Laurance, W.F. et al. A global strategy for road building. Nature 513, 229-232 (2014). Fearnside, P.M. & Graça, P.M.L. BR-319: Brazil's Manaus-Porto Velho highway and the potential impact of linking the arc of deforestation to central Amazonia. Environ. Manage. 38, 705-716 (2006). Laurance, W.F. et al. Reducing the global environmental impacts of rapid infrastructure expansion. Curr. Biol. 25, R259-R262 (2015). Flyvbjerg, B. et al. How common and how large are cost overruns in transport infrastructure projects? Transp. Rev. 23, 71-88 (2003). Geneletti, D. Reasons and options for integrating ecosystem services in strategic environmental assessment of spatial planning. Int. J. Biodivers. Sci. Ecosyst. Serv. Manag. 7, 143-149 (2011). Kaneko, S. et al. Method of designing forest road construction scenario with GIS-based calibration using sustainable forestry model. Results Eng. 24, 103078 (2024). Çalışkan, E. Geographic Information System(GIS) and Roadeng usage to determine the environmentally sensitive forest road route in mountainous terrain. Eur. J. Interdiscip. Stud. 2, 18-25 (2016). Malczewski, J. & Rinner, C. Multicriteria Decision Analysis in Geographic Information Science (Springer, 2015). Yannis, G. et al. State-of-the-art review on multi-criteria decision-making in the transport sector. J. Traffic Transp. Eng. Engl. Ed. 7, 413-431 (2020). Kiker, G.A. et al. Application of multicriteria decision analysis in environmental decision making. Integr. Environ. Assess. Manag. 1, 95-108 (2005). Mendoza, G.A. & Martins, H. Multi-criteria decision analysis in natural resource management: a critical review of methods and new modelling paradigms. For. Ecol. Manag. 230, 1-22 (2006). McCann, R.K. et al. Bayesian belief networks: applications in ecology and natural resource management. Can. J. For. Res. 36, 3053-3062 (2006). Uusitalo, L. Advantages and challenges of Bayesian networks in environmental modelling. Ecol. Model. 203, 312-318 (2007). Aguilera, P.A. et al. Bayesian networks in environmental modelling. Environ. Model. Softw. 26, 1376-1388 (2011). Brooks, T.M. et al. Global biodiversity conservation priorities. Science 313, 58-61 (2006). Olson, D.M. & Dinerstein, E. The Global 200: priority ecoregions for global conservation. Ann. Mo. Bot. Gard. 89, 199-224 (2002). Carpenter, S.R. et al. Science for managing ecosystem services: beyond the Millennium Ecosystem Assessment. Proc. Natl. Acad. Sci. 106, 1305-1312 (2009). Folke, C. et al. Resilience thinking: integrating resilience, adaptability and transformability. Ecol. Soc. 15, 20 (2010). African Development Bank. East Africa Regional Integration Strategy 2018-2022 (2018). Sinclair, A.R.E. et al. Serengeti IV: Sustaining Biodiversity in a Coupled Human-Natural System (University of Chicago Press, 2015). Walelign, S.Z. et al. Roads and livelihood activity choices in the Greater Serengeti Ecosystem, Tanzania. PLOS ONE 14, e0213089 (2019). Thirgood, S. et al. Can parks protect migratory ungulates? The case of the Serengeti wildebeest. Anim. Conserv. 7, 113-120 (2004). Dobson, A.P. et al. Road will ruin Serengeti. Nature 467, 272-273 (2010). Holdo, R.M. et al. Predicted impact of barriers to migration on the Serengeti wildebeest population. PLoS ONE 6, e16370 (2011). Msoffe, F.U. et al. A new approach to the classification of land use in Africa: the case of the Greater Serengeti-Mara ecosystem. Afr. J. Ecol. 49, 98-111 (2011). Homewood, K. et al. Staying Maasai? Livelihoods, Conservation and Development in East African Rangelands (Springer, 2009). Nelson, F. et al. The evolution and reform of Tanzanian wildlife management. Conserv. Soc. 5, 232-261 (2007). Reed, M.S. Stakeholder participation for environmental management: a literature review. Biol. Conserv. 141, 2417-2431 (2008). Reed, J. et al. A theory of participation: what makes stakeholder and public engagement in environmental management work? Restor. Ecol. 25, S7-S17 (2017). Cornwall, A. & Jewkes, R. What is participatory research? Soc. Sci. Med. 41, 1667-1676 (1995). Petticrew, M. & Roberts, H. Systematic Reviews in the Social Sciences: A Practical Guide (Blackwell Publishing, 2006). Bryson, J.M. What to do when stakeholders matter: stakeholder identification and analysis techniques. Public Manag. Rev. 6, 21-53 (2004). Saaty, T.L. The Analytic Hierarchy Process (McGraw-Hill, 1980). Saaty, T.L. Decision making with the analytic hierarchy process. Int. J. Serv. Sci. 1, 83-98 (2008). Ishizaka, A. & Labib, A. Review of the main developments in the analytic hierarchy process. Expert Syst. Appl. 38, 14336-14345 (2011). Beier, P. & Noss, R.F. Do habitat corridors provide connectivity? Conserv. Biol. 12, 1241-1252 (1998). Castelletti, A. & Soncini-Sessa, R. Bayesian networks and participatory modelling in water resource management. Environ. Model. Softw. 22, 1075-1088 (2007). Voinov, A. & Bousquet, F. Modelling with stakeholders. Environ. Model. Softw. 25, 1268-1281 (2010). O'Hagan, A. et al. Uncertain Judgements: Eliciting Experts' Probabilities (John Wiley & Sons, 2006). Forman, R.T.T. et al. Road Ecology: Science and Solutions (Island Press, 2003). Jaeger, J.A.G. et al. Predicting when animal populations are at risk from roads: an interactive model of road avoidance behavior. Ecol. Model. 185, 329-348 (2005). Additional Declarations There is NO Competing Interest. Supplementary Files SupplementaryMaterialsDecisionPaper.docx Supplementary Materials: Balancing Development and Conservation: A Decision Framework for Sustainable Road Planning in Biodiversity Hotspots rs.pdf Reporting Summary Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7419361","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":517241491,"identity":"8222c7e0-cd6a-4930-b377-d2899842b0ad","order_by":0,"name":"Philipo Mtweve","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA9klEQVRIiWNgGAWjYDACCcbGAwkIrg1RWhogWtjA3DSYuAEeLQwMBxgQWg4T1mIu3dxw4OEOGwbd+T2Gj3kqztsbHD/A+OEHwx97XFos5xxsOJB4Jo3B7BiPsTHPmduJG84kMEv2MBgw49JicCMRqKXtMFAL7zbJmW23EwyAXpMGSrARo2X7z5n/ztkbnH/A/BsowUOULQwfGw4wbriRwAayRQKnX2Ykgv3CY3Ys/7PEh2PJiTNvPGyz7DEwxhlk5hLpDx/+3GEjZ3b4WOKHhBo7e77zyYdv/KiQwxliYLMYGxiQXQ7k4otJmJZRMApGwSgYBbgBAIwAW+rE5/rtAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0009-0008-9004-0730","institution":"Bonn University","correspondingAuthor":true,"prefix":"","firstName":"Philipo","middleName":"","lastName":"Mtweve","suffix":""},{"id":517241492,"identity":"1bffb741-d7a7-4680-a74a-8f90df15f9ac","order_by":1,"name":"Abraham Eustace","email":"","orcid":"","institution":"Tanzania Wildlife Management Authority (TAWA)","correspondingAuthor":false,"prefix":"","firstName":"Abraham","middleName":"","lastName":"Eustace","suffix":""},{"id":517241493,"identity":"a0169bf5-bd8a-4eae-ad0e-994e250b0d0f","order_by":2,"name":"Qambemeda Nyanghura","email":"","orcid":"","institution":"Sokoine University of Agriculture (SUA)","correspondingAuthor":false,"prefix":"","firstName":"Qambemeda","middleName":"","lastName":"Nyanghura","suffix":""},{"id":517241494,"identity":"cae5c04b-3f17-4682-85b8-5c6400e11bc7","order_by":3,"name":"Vincent Moseti","email":"","orcid":"","institution":"University of Bonn","correspondingAuthor":false,"prefix":"","firstName":"Vincent","middleName":"","lastName":"Moseti","suffix":""},{"id":517241495,"identity":"81ce3f1f-afe8-4806-b4a1-d23c61bb78af","order_by":4,"name":"Tanja Kramm","email":"","orcid":"","institution":"University of Cologne","correspondingAuthor":false,"prefix":"","firstName":"Tanja","middleName":"","lastName":"Kramm","suffix":""},{"id":517241496,"identity":"0d11af56-9251-47b5-a462-a9eb1242e6df","order_by":5,"name":"Pierre Ibisch","email":"","orcid":"https://orcid.org/0000-0001-9820-9272","institution":"Eberswalde University for Sustainable Development","correspondingAuthor":false,"prefix":"","firstName":"Pierre","middleName":"","lastName":"Ibisch","suffix":""},{"id":517241497,"identity":"94936edb-bc0e-46fe-86ec-5cdbdfcbe23b","order_by":6,"name":"Lisa Biber-Freudenberger","email":"","orcid":"","institution":"University of Bonn","correspondingAuthor":false,"prefix":"","firstName":"Lisa","middleName":"","lastName":"Biber-Freudenberger","suffix":""}],"badges":[],"createdAt":"2025-08-20 16:25:31","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7419361/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7419361/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":91827914,"identity":"df76e4da-49db-479c-89f7-950f99a3e784","added_by":"auto","created_at":"2025-09-22 08:52:43","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":714711,"visible":true,"origin":"","legend":"","description":"","filename":"AnIntegratedDecisionSupportFrameworkforSustainableRoadPlanninginBiodiversityHotspots.docx","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/d5ebd41c0a108213add790c4.docx"},{"id":91830139,"identity":"5cbc4823-cd29-4d20-9c4c-81ef2d3f0fb5","added_by":"auto","created_at":"2025-09-22 09:00:43","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":8074,"visible":true,"origin":"","legend":"","description":"","filename":"NCOMMS2565888.json","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/c546982f42faeecaf7d9a29c.json"},{"id":91826965,"identity":"dad96995-6a02-4f10-84cf-b06fd19ba404","added_by":"auto","created_at":"2025-09-22 08:44:43","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":279105,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterialsDecisionPaper.docx","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/ab9f92fdcbd8122839c59b25.docx"},{"id":91826958,"identity":"ced17b0c-067c-4e84-840d-65637f948523","added_by":"auto","created_at":"2025-09-22 08:44:43","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":110158,"visible":true,"origin":"","legend":"","description":"","filename":"NCOMMS25658880enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/d7a51b3a6c13904a0ff7299b.xml"},{"id":91826961,"identity":"4412a7ff-7a4d-4610-b3d5-af1ab0088f73","added_by":"auto","created_at":"2025-09-22 08:44:43","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":305567,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/09eb48807da423e5d6928a90.png"},{"id":91827947,"identity":"28941cd2-b3d8-49f4-bd77-4ba15f5052bf","added_by":"auto","created_at":"2025-09-22 08:52:43","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":138208,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/d9e6b4034a4ecb11fcc03e7d.png"},{"id":91830140,"identity":"3bc3abb4-935d-4efa-9b6a-03aad1e01221","added_by":"auto","created_at":"2025-09-22 09:00:43","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":179695,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/6315114eb432572e2869428f.png"},{"id":91827976,"identity":"5abd8ff9-d349-44b8-af94-ac2aaa307ed9","added_by":"auto","created_at":"2025-09-22 08:52:43","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":33495,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/00fbf0246cfef57ab4f15a9d.png"},{"id":91826959,"identity":"fb9bbf2c-3239-4516-bbf9-68c5d9fd654f","added_by":"auto","created_at":"2025-09-22 08:44:43","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":55877,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/fc6ce38d1a8028f10b8559eb.png"},{"id":91826966,"identity":"978428bb-f286-4b1d-8c28-26ea4737b618","added_by":"auto","created_at":"2025-09-22 08:44:43","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":29494,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/65ec86673ad3f3a259b8f7a3.png"},{"id":91826963,"identity":"28eef701-f68b-4084-a74e-d136978669e2","added_by":"auto","created_at":"2025-09-22 08:44:43","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":37985,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/860d7493e3f4f32f0ccbf968.png"},{"id":91827927,"identity":"1eea363e-09bb-468b-bc6e-372962b8555b","added_by":"auto","created_at":"2025-09-22 08:52:43","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":19182,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/aa0463b1d001739343370cfa.png"},{"id":91826971,"identity":"4cd215e4-0cab-4ea6-9009-07327444595a","added_by":"auto","created_at":"2025-09-22 08:44:43","extension":"xml","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":105814,"visible":true,"origin":"","legend":"","description":"","filename":"NCOMMS25658880structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/5d33014095fdf9712552efa0.xml"},{"id":91826968,"identity":"b911735c-8024-47f4-b398-a56d93a74f87","added_by":"auto","created_at":"2025-09-22 08:44:43","extension":"html","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":118059,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/c2a9aba4a5cfc28e15bd9134.html"},{"id":91826951,"identity":"741dfce5-1c9b-4a39-bf63-0570ee62daa7","added_by":"auto","created_at":"2025-09-22 08:44:43","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":305567,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eGreater Serengeti Ecosystem showing four proposed road development options with MCDA suitability analysis. Colors indicate route suitability based on integrated criteria (green = Suitable, grey = Not Suitable). The Eyasi route (blue) emerges as optimal under integrated analysis, balancing connectivity needs with biodiversity protection.\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/75ca30ee0fed8b3a89419aec.png"},{"id":91826952,"identity":"47e9b0dc-960c-47fc-ba31-bcd4b293b798","added_by":"auto","created_at":"2025-09-22 08:44:43","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":138208,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eMulti-criteria analysis and Bayesian network assessment. a, MCDA performance scores showing biodiversity conservation versus socioeconomic development trade-offs. Eyasi emerges as optimal with balanced high performance, while Mbulu achieves maximum biodiversity protection. b, BBN network structure showing causal relationships between infrastructure development and outcomes through 17 nodes and 34 causal pathways. c, BBN sustainability scores with uncertainty bounds from Monte Carlo analysis. Mbulu leads overall sustainability (0.52±0.05) followed by Eyasi (0.47±0.04). Probability distributions from 1,000 Monte Carlo iterations show statistical significance of performance differences (p\u0026lt;0.01) between conservation-oriented routes and connectivity-focused alternatives.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/17485ae465c7ef1970976881.png"},{"id":91826957,"identity":"9a488689-e837-43c5-a717-c5ea9552095a","added_by":"auto","created_at":"2025-09-22 08:44:43","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":179695,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eA, thirty-year trajectories showing initial convergence (0-10 years), critical transition (15-25 years), and permanent divergence (25-30 years) with conservation routes (solid) outperforming connectivity routes (dashed). b, Northern route components showing economic benefits declining while environmental costs compound exponentially after year 15. c, Performance comparison illustrating Northern route crossing critical thresholds (0.2, 0.1) with permanent sustainability disadvantages.\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/9a6b9d71c08c5e068dd54f2c.png"},{"id":91826953,"identity":"84db88df-dadf-410f-a381-75db34af6bd1","added_by":"auto","created_at":"2025-09-22 08:44:43","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":33495,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eCross-method validation showing strong correlation (r=0.89) between MCDA and BBN approaches, with conservation-oriented routes demonstrating superior efficiency per capita served.\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/c63b8736badd9984086ea0c0.png"},{"id":107489780,"identity":"45c78be5-2797-4e60-9463-1d907d94ad5f","added_by":"auto","created_at":"2026-04-22 02:48:57","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1446158,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/a92f7f74-f950-4347-a40c-e7cb3f8a727e.pdf"},{"id":91826955,"identity":"e782c4db-a6c3-4386-9fc3-8a61deb8413f","added_by":"auto","created_at":"2025-09-22 08:44:43","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":279105,"visible":true,"origin":"","legend":"Supplementary Materials: Balancing Development and Conservation: A Decision Framework for Sustainable Road Planning in Biodiversity Hotspots","description":"","filename":"SupplementaryMaterialsDecisionPaper.docx","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/da08143b7a7c73e390ecd070.docx"},{"id":91826970,"identity":"a9229212-4f06-4c1e-b9db-7346e66f5bc2","added_by":"auto","created_at":"2025-09-22 08:44:43","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":3254405,"visible":true,"origin":"","legend":"Reporting Summary","description":"","filename":"rs.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7419361/v1/9d51ffde17f902edf40cfcd9.pdf"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"An Integrated Decision Support Framework for Sustainable Road Planning in Biodiversity Hotspots","fulltext":[{"header":"Introduction","content":"\u003cp\u003eGlobal road infrastructure investment will reach \u003cspan\u003e$\u003c/span\u003e94 trillion by 2040, with developing countries requiring \u003cspan\u003e$\u003c/span\u003e2.3 trillion annually to meet sustainable road development goals\u0026sup1;. This expansion occurs amid accelerating biodiversity loss, with species extinction rates up to 100 times above natural background levels\u0026sup2;. The conflict between infrastructure development and biodiversity conservation is most pronounced in biodiversity hotspots, which are 36 regions that cover only 2.4% of Earth's surface yet harbor 44% of vascular plants and 35% of vertebrates, while supporting 2.3\u0026nbsp;billion people\u0026sup3;.\u003c/p\u003e\u003cp\u003eInfrastructure development, particularly investments into roads, is essential for rural development but also a primary driver of habitat fragmentation. Global road networks have expanded by 60% since 1990, with an additional 25\u0026nbsp;million kilometers planned by 2050⁴. Case studies for the negative impacts of roads on biodiversity and ecosystems can be found worldwide: Brazil's BR-319 highway increased deforestation by 95% within 50km⁵, Indonesia's trans-Papua highway threatens the world's largest intact tropical forest, and Africa's planned development corridors could fragment 33% of remaining intact forests⁶.\u003c/p\u003e\u003cp\u003eNegative impacts from roads can be minimized through biodiversity-sensitive planning, but conventional approaches remain fundamentally limited. Early road infrastructure methodologies emphasized engineering feasibility and economic cost-benefit analysis, often treating environmental and social impacts as externalities to be mitigated post hoc⁷. Even where Environmental Impact Assessments (EIAs) are required, they typically occur after route selection, limiting their influence on strategic decisions.\u003c/p\u003e\u003cp\u003eMore recent approaches have incorporated Geographic Information Systems (GIS) for spatial optimization, seeking routes that minimize environmental impact while maximizing socioeconomic benefits⁸⁻\u0026sup1;⁰. Spatial optimization tools using Geographic Information Systems (GIS) have added environmental dimensions to route planning, but still face three persistent limitations: inadequate integration of stakeholder preferences, weak treatment of uncertainty, and limited consideration of long-term, system-level feedbacks. One significant area of uncertainty in road development planning is the unpredictability of ecological responses to road construction. For example, the introduction of a new road can lead to unforeseen changes in wildlife behavior, such as altered migration patterns or increased poaching, which can cascade into broader ecosystem impacts. This uncertainty complicates decision-making, as planners may not fully understand the long-term ecological consequences of their actions.\u003c/p\u003e\u003cp\u003eMulti-Criteria Decision Analysis (MCDA) has emerged as a promising approach for addressing stakeholder preference integration, providing systematic frameworks for incorporating multiple, often conflicting objectives while maintaining transparency in trade-off evaluation\u0026sup1;\u0026sup1;⁻\u0026sup1;\u0026sup3;. It enables explicit integration of stakeholder preferences through structured weighting processes and has been successfully applied to transportation planning, energy infrastructure, and conservation prioritization\u0026sup1;⁴. However, MCDA's primarily deterministic nature limits its ability to handle uncertainty and complex system dynamics.\u003c/p\u003e\u003cp\u003eBayesian Belief Networks (BBN) offer complementary strengths for addressing uncertainty and system complexity. BBNs model probabilistic relationships among variables, explicitly incorporating uncertainty while capturing indirect effects and feedback loops\u0026sup1;⁵, \u0026sup1;⁶. They have proven valuable for environmental management, ecosystem service assessment, and infrastructure risk analysis\u0026sup1;⁷. Yet BBNs face challenges in transparent stakeholder preference incorporation and can be difficult to parameterize without extensive data.\u003c/p\u003e\u003cp\u003eRecognition of individual method limitations has led to calls for integrated approaches that combine complementary methodologies. However, most integration attempts remain limited to sequential application rather than true methodological synthesis. The challenge lies in developing frameworks that leverage the structured preference incorporation of MCDA while capturing the uncertainty quantification and system dynamics of BBN approaches.\u003c/p\u003e\u003cp\u003eThis integration is particularly critical for road infrastructure planning in biodiversity hotspots, where decisions involve high stakes, multiple stakeholders with conflicting interests, substantial uncertainties about ecological responses, and long-term consequences that unfold over decades. Such contexts demand analytical approaches that can simultaneously address preference integration, uncertainty quantification, and temporal dynamics.\u003c/p\u003e\u003cp\u003eEast Africa represents a compelling case study for road infrastructure-biodiversity challenges. The region encompasses multiple biodiversity hotspots including the Eastern Afromontane, Coastal Forests of Eastern Africa, and Horn of Africa hotspots\u0026sup1;⁸, experiences rapid economic growth driving infrastructure expansion, and supports over 513\u0026nbsp;million people requiring improved connectivity. Recent regional infrastructure initiatives include the East African Railway Master Plan, Northern Corridor improvements, and multiple road development projects often negatively impacting valuable ecosystems.\u003c/p\u003e\u003cp\u003eWe use Tanzania's Greater Serengeti Ecosystem as a characteristic case study for developing and testing an integrated planning approach combining MCDA's systematic stakeholder engagement with BBN's uncertainty modeling capabilities. The GSE encompasses 30,000 km\u0026sup2; of interconnected protected areas, including Serengeti National Park, Ngorongoro Conservation Area, and multiple game reserves\u0026sup1;⁹. It supports the world's largest terrestrial mammal migration (1.5\u0026nbsp;million wildebeest, zebras, and Thomson's gazelles), stores 1\u0026nbsp;million metric tons of carbon, and contains two UNESCO World Heritage Sites\u0026sup2;⁰, \u0026sup2;\u0026sup1;. Simultaneously, the region supports 4.6\u0026nbsp;million people requiring improved market access, healthcare, and educational services\u0026sup2;\u0026sup2;.\u003c/p\u003e\u003cp\u003eSince 1990, four major road development options have been proposed by different government agencies and development partners to improve regional connectivity, each presenting different trade-offs between conservation and development objectives. The Northern route was proposed by the Ministry of Works and Transport (1995), the Serengeti route by tourism industry stakeholders (2008), the Eyasi route through collaborative planning between Tanzania National Parks Authority and regional development committees (2012), and the Mbulu route by conservation organizations and community-based natural resource management groups (2015).\u003c/p\u003e\u003cp\u003eThe aim of this paper is to apply the integrated transdisciplinary and interdisciplinary decision support framework to this specific question, namely which of the road options is less harmful from the perspective of biodiversity and sustainable development. This test should clearly show that the proposed methodological approach is applicable and yields helpful results.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eStakeholder Engagement Reveals Different Preference Structures\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOur participatory stakeholder engagement process involved 30 regional experts and decision-makers across three structured phases: systematic literature synthesis, expert consultation, and collaborative workshops. This approach ensured representation from different sectors and stakeholder groups affected by road development while building methodological legitimacy through inclusive participation.\u003c/p\u003e\n\u003cp\u003eThe initial expert consultation at the Tanzania Wildlife Research Institute conference identified four criteria with 13 potential evaluation indicators across four domains: engineering feasibility, biodiversity conservation, socioeconomic development, and political interests. During the subsequent three-day stakeholder workshops, which included tourism operators, planning officials, field ecologists, road engineers, community representatives, and conservation organizations, the framework was refined to 16 criteria where three additional indicators were added (geology, aspect and Implementation feasibility). We then selected 13 indicators for spatial MCDA calculations. The selection of these 13 indicators was based on the availability of data in raster format, this selection excluded implementation feasibility, Travel time reduction and service delivery (Table 1).\u003c/p\u003e\n\u003cp\u003eThe Analytical Hierarchy Process (AHP) weighting revealed complex stakeholder priorities that challenged existing assumptions about the trade-offs between development and conservation, as shown in Table 1. Notably, biodiversity conservation and socioeconomic development were nearly equally valued, with weights of 36.5% and 37.5% respectively, indicating the desire for a balanced approach to environmental concerns and human needs. Within the engineering feasibility category, elevation emerged as the most critical factor at 31.4%, reflecting the challenges posed by the Serengeti\u0026apos;s topography. Soil type (25.4%) and slope (19.6%) were also significant, highlighting practical construction constraints. Furthermore, the emphasis on habitat quality (7.8%) and market connectivity (8.5%) underscored the necessity for infrastructure that supports both ecological integrity and economic opportunities, suggesting a comprehensive strategy for sustainable development.\u003c/p\u003e\n\u003cp\u003eTable 1: \u003cstrong\u003eStakeholder-derived criteria weights and consistency measures\u003c/strong\u003e.\u0026nbsp;\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"596\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eDomain\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eCriteria\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eWeight (%)\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eIndicators\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eWeight (%)\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eDefinition and Measurement Protocol\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eEngineering Feasibility\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTerrain \u0026amp; Construction Difficulty\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e13.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eComposite measure of physical construction challenges based on topographic and geological constraints\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eElevation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e31.4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMean elevation above sea level (m). Higher elevations increase construction costs and technical difficulty. Measured from 30m SRTM DEM.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSoil type\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e25.4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSoil engineering properties affecting foundation stability and construction feasibility. Classified using Harmonized World Soil Database with field validation. Scale: 1-10 (10 = excellent engineering properties).\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSlope\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e19.6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAverage terrain gradient (%) affecting construction complexity and equipment requirements. Calculated from DEM using QGIS slope analysis. Higher slopes = higher construction difficulty.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eGeology\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e12.8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBedrock type and geological stability for infrastructure foundation. Based on Tanzania Geological Survey 1:250,000 maps. Scale: 1-10 (10 = excellent foundation conditions).\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAspect\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e10.8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTerrain orientation affecting drainage, erosion risk, and construction access. Calculated from DEM. North-facing slopes preferred for reduced erosion.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eBiodiversity Conservation\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eEcosystem Impact\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e36.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eComposite measure of negative impacts on ecosystem integrity, wildlife populations, and conservation effectiveness\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eProtected areas\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3.8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDistance to protected area boundaries (km). Closer proximity increases conservation conflicts and legal constraints.\u0026nbsp;Measured as minimum distance to any protected area boundary.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLand use/cover\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e4.7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNaturalness and conservation value of vegetation communities. Based on ESA WorldCover 2021 classification.\u0026nbsp;Scale: 1-10 (10 = intact natural habitat, 1 = degraded/converted land).\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eWildlife corridors\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e7.6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCritical pathways for wildlife movement between protected areas, mapped by Tanzania Wildlife Authority using GPS collar data and expert knowledge. Impact measured as corridor intersection length (km).\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMigration routes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e8.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSeasonal pathways used specifically during wildebeest migration (May-July and November-December), distinct from year-round wildlife corridors. Based on 15-year GPS collar synthesis from Serengeti Wildlife Research Centre. Impact = route intersection length during peak migration.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHabitat quality\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e7.8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eEcological condition and species-supporting capacity of natural areas. Assessed through: (1) vegetation density index from Landsat imagery, (2) water source proximity, (3) expert field assessments by TAWIRI researchers. Scale: 1-10 (10 = pristine habitat condition).\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eSocioeconomic Development\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDevelopment Benefits\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e37.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eComposite measure of positive impacts on human welfare, economic opportunities, and service access\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMarket connectivity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e8.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eImprovement in access to agricultural and livestock markets measured as: (1) reduced travel time to nearest major market (hours), (2) population within 2-hour travel time of markets. Based on existing road network analysis and market location mapping.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSettlement\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e7.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNumber of people gaining improved road access within 10km of route alignment. Population data from Global Human Settlement Layer 2020, validated through district census data and field surveys.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eService delivery\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e6.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eImprovement in access to healthcare, education, and government services. Measured as: (1) population within 1-hour travel time of health facilities, (2) children within 30 minutes of primary schools, (3) communities within 2 hours of district headquarters.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTravel time reduction\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e7.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDecrease in travel time between major population centers enabled by new road infrastructure. Calculated using network analysis comparing current vs. proposed road networks. Measured in hours saved for representative origin-destination pairs.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePopulation served\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e7.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTotal population within 50km corridor of route alignment who would benefit from improved connectivity. Based on Global Human Settlement Layer with 50km buffer analysis, validated against district population statistics.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003ePolitical Implementation\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStakeholder Support\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e13.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eFeasibility of successful implementation considering political, social, and institutional factors\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eImplementation feasibility\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e13.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eComposite measure including: (1) government agency support levels, (2) community acceptance based on public consultations, (3) donor/financing availability, (4) regulatory approval probability, (5) absence of major opposition groups. Assessed through stakeholder mapping and consultation feedback. Scale: 1-10 (10 = strong support, feasible implementation).\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003eNote: Consistency Ratio (overall): 0.037; Individual CR range: 0.018-0.052. Weights represent average values across all stakeholder groups derived through Analytical Hierarchy Process (\u003c/em\u003e\u003cem\u003eAHP\u003c/em\u003e\u003cem\u003e). Domain weights sum to 100.0%. Indicator weights shown are relative percentages within each domain. Individual stakeholder weights were aggregated using geometric means to maintain consistency with AHP principles.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eImportantly, consistency analysis yielded acceptable ratios across all participant groups (CR \u0026lt; 0.1), indicating logically structured preference structures rather than random responses. Individual consistency ratios ranged from 0.018-0.052, suggesting participants carefully weighted trade-offs in their decisions. Furthermore, cross-group correlation analysis revealed strong agreement among different stakeholder groups (r = 0.73-0.89), pointing to a shared understanding of regional development challenges despite different professional backgrounds.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSpatial Analysis Reveals Fundamental Differences among Road Options\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA spatial analysis of the four proposed routes, utilizing 13 standardized datasets at a 1 km resolution, identified Eyasi Road as the most feasible and sustainable option among the four options. This route offers unique opportunities for balancing conservation and development goals. However, each route crosses distinct ecological and social landscapes and encounters various implementation challenges that significantly impact long-term sustainability outcomes.\u003c/p\u003e\n\u003cp\u003eFigure 1 illustrates the spatial patterns and characteristics of the routes, displaying the Greater Serengeti Ecosystem with the four proposed road development options superimposed on the results of the MCDA suitability analysis. The color coding reflects route suitability based on integrated criteria, with green areas indicating suitable zones and grey areas marking regions deemed unsuitable for road development.\u003c/p\u003e\n\u003cp\u003eThe Serengeti route represents the shortest option at 287 km but crosses protected areas directly, including 198.9 km within Ngorongoro Conservation Area and Serengeti National Park boundaries. While offering apparent economic benefits through reduced travel distances, mandatory speed limits of 30-50 km/hour within protected areas actually increase total travel time, negating the distance advantage. The route serves approximately 971,000 people but requires the highest construction costs ($4.50 million per kilometer) due to extensive wildlife-friendly infrastructure requirements including animal crossing structures, specialized fencing, wildlife detection systems, and environmental monitoring protocols mandated for protected area construction.\u003c/p\u003e\n\u003cp\u003eThe Northern route spans 342 km connecting Musoma to Arusha via the northern ecosystem boundary, serving the largest population (2.1 million people within a 50km buffer) while intersecting critical wildlife corridors during peak migration periods. Detailed spatial analysis indicates moderate terrain challenges (average slope: 12.3%, maximum: 45%) but substantial ecological risks, with 67% of the route within 50km of protected area boundaries and 34% intersecting designated wildlife corridors used by over 900,000 migrating ungulates annually. Construction costs are high ($3.80 million per kilometer) reflecting the need for wildlife crossing infrastructure, animal detection systems, and specialized construction protocols to minimize wildlife mortality in ecologically sensitive areas.\u003c/p\u003e\n\u003cp\u003eThe Eyasi route follows a 398 km southern trajectory through diverse landscapes including highland forests, agricultural zones, and pastoral areas, serving 847,000 people while maintaining 15 km minimum distance from core protected areas. Terrain analysis indicates moderate construction challenges (average slope: 8.7%, soil stability index: 7.2/10) with strategic opportunities for wildlife-friendly design through careful corridor placement and crossing structures.\u003c/p\u003e\n\u003cp\u003eThe Mbulu route traces a 445 km southeastern boundary path, prioritizing biodiversity conservation while serving primarily rural communities (682,000 people). This route faces challenging terrain conditions (average slope: 15.2%, complex geology requiring specialized engineering solutions) but achieves the lowest construction costs ($2.45 million per kilometer) because it strategically avoids protected areas entirely, eliminating the need for expensive wildlife-friendly infrastructure.\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMulti-Criteria Decision Analysis Identifies Best Compromise Solutions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMCDA analysis integrating stakeholder preferences with spatial datasets revealed clear performance hierarchies among route options while highlighting the complex trade-offs inherent in infrastructure planning within biodiversity hotspots. The analysis employed weighted overlay modeling using stakeholder-derived criteria weights, generating standardized scores from 0-1200 for each route across multiple sustainability dimensions.\u003c/p\u003e\n\u003cp\u003eThe comprehensive results of both MCDA and BBN analyses are presented in Figure 2, which demonstrates the trade-offs between biodiversity conservation and socioeconomic development across all four routes. Panel (a) shows MCDA performance scores, panel (b) illustrates the BBN network structure, and panel (c) presents BBN sustainability scores with uncertainty bounds.\u003c/p\u003e\n\u003cp\u003eThe Eyasi route emerges as the least harmful choice under integrated MCDA evaluation, achieving the highest combined performance with biodiversity conservation score of 1060/1200 and socioeconomic development score of 850/1200. This superior performance reflects strategic positioning that successfully balances multiple competing objectives: substantial population service (847,000 people) without critical habitat intersection, moderate terrain challenges enabling cost-effective construction while avoiding major engineering obstacles, and abundant design opportunities for wildlife-friendly infrastructure through careful corridor placement and strategically positioned crossing structures. The route\u0026apos;s integrated score of 960/1200 represents a 104% performance advantage over the Northern route (470/1200) and substantially outperforms other connectivity-focused alternatives.\u003c/p\u003e\n\u003cp\u003eThe Mbulu route demonstrates exceptional biodiversity protection performance (1200/1200), representing perfect scores across all conservation criteria including protected area avoidance, wildlife corridor preservation, habitat fragmentation minimization, and endangered species protection. This outstanding environmental performance positions the route as the gold standard for ecosystem-sensitive infrastructure development. Its socioeconomic performance (750/1200) reflects moderate development benefits due to serving a smaller total population (682,000 people), but the route achieves superior cost-effectiveness with the lowest construction costs by strategically avoiding protected areas.\u003c/p\u003e\n\u003cp\u003eComprehensive sensitivity analysis across alternative weighting schemes confirms the robustness of these results while revealing decision contexts where different routes might be preferred based on varying stakeholder priorities. When biodiversity weights increase to 70% (development: 30%), Mbulu and Eyasi routes strengthen their comparative advantages with scores increasing by 12% and 8% respectively, while Northern route performance declines substantially with an 18% score decrease.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBayesian Belief Network Analysis Reveals System Dynamics and Uncertainty\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBBN modeling addresses fundamental limitations of deterministic MCDA by incorporating uncertainty, causal relationships, and system dynamics often missing from static evaluation approaches. Our participatory BBN development process engaged the same stakeholder groups in network structure definition and parameter estimation, ensuring methodological consistency with MCDA while adding crucial probabilistic dimensions that capture the inherent unpredictability of infrastructure outcomes in complex socio-ecological systems.\u003c/p\u003e\n\u003cp\u003eThe final BBN architecture includes 35 nodes representing infrastructure development drivers, ecological impact pathways, and socioeconomic outcome mechanisms connected through 100 carefully validated causal relationships, as shown in the network structure component of Figure 2. Network structure emerged through systematic stakeholder workshops where participants mapped causal pathways based on their professional experience, supplemented by literature evidence from 47 analogous developments in similar ecosystems.\u003c/p\u003e\n\u003cp\u003eBBN analysis identified the Mbulu route as optimal under uncertainty conditions, achieving an overall sustainability score of 0.52\u0026plusmn;0.05 (on a 0-1 probability scale). This score represents the probability of achieving positive outcomes across integrated biodiversity and livelihood metrics, with higher values indicating greater likelihood of sustainable development success. The Mbulu route\u0026apos;s superiority reflects its exceptional biodiversity protection probability (0.58, indicating 58% chance of maintaining ecosystem functionality) combined with substantial positive livelihood impacts (0.47, representing moderate but reliable socioeconomic benefits).\u003c/p\u003e\n\u003cp\u003eThe Eyasi route ranked close second place with sustainability score 0.47\u0026plusmn;0.04, demonstrating balanced performance across outcome domains while maintaining low uncertainty. The Northern route ranked third (0.41\u0026plusmn;0.03) despite high positive livelihood impacts (0.50), severely limited by poor biodiversity outcomes (0.31). The Serengeti route performed worst (0.40\u0026plusmn;0.06) with both poor biodiversity (0.37) and moderate livelihood outcomes (0.42), plus high uncertainty reflecting unpredictable protected area consequences.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLong-term Temporal Projections Reveal Sustainability Trajectories\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThirty-year sustainability projections incorporating cumulative environmental impacts through annual decay functions revealed stark differences in route resilience and long-term viability. Our temporal modeling employed route-specific decay rates calibrated to regional deforestation trends: conservation-oriented routes (Mbulu, Eyasi) experience slower ecosystem degradation (4.6-4.9% annually) while connectivity-focused routes (Northern, Serengeti) face accelerated decline (6.2-6.8% annually) due to induced development pressures.\u003c/p\u003e\n\u003cp\u003eThe critical transition analysis revealing permanent sustainability disadvantages for connectivity-focused infrastructure strategies is presented in Figure 3. Panel (a) shows thirty-year trajectories demonstrating initial convergence (0-10 years), critical transition (15-25 years), and permanent divergence (25-30 years), with conservation routes outperforming connectivity routes. Panel (b) shows Northern route components with economic benefits declining while environmental costs compound exponentially after year 15. Panel (c) illustrates performance comparison showing the Northern route crossing critical thresholds with permanent sustainability disadvantages.\u003c/p\u003e\n\u003cp\u003eConservation-oriented routes maintain substantially higher sustainability throughout projection periods, with critical divergence emerging during years 10-15 when cumulative impacts begin overwhelming initial benefits for connectivity-focused alternatives. The Mbulu route sustains performance longest, maintaining 0.32 by year 20 and 0.12 by year 30 (62% and 23% of initial performance). The Northern route drops to 0.22 by year 20 and 0.08 by year 30, reflecting compound environmental degradation.\u003c/p\u003e\n\u003cp\u003eCritical transition analysis reveals precise thresholds where route performances converge initially but diverge permanently after year 15, creating irreversible sustainability disadvantages for connectivity-focused strategies. During the initial decade, the Northern route maintains 79% of the Mbulu route\u0026apos;s performance as immediate economic benefits partially compensate for environmental costs. However, years 15-25 demonstrate accelerating divergence driven by exponential compounding of ecological impacts habitat fragmentation, wildlife population declines, and ecosystem service losses intensifying 40% faster than linear economic benefit decline.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCross-method validation and integration insights\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCross-validation between MCDA and BBN approaches provides critical validation while revealing complementary insights that support decision-making. The relationship between the two analytical approaches and their efficiency implications are illustrated in Figure 4, which shows strong correlation (r=0.89) between MCDA and BBN approaches, with conservation-oriented routes demonstrating superior efficiency per capita served.\u003c/p\u003e\n\u003cp\u003eCorrelation analysis demonstrates strong overall agreement (r = 0.89, p \u0026lt; 0.01) between method rankings, confirming that both approaches identify two alternating road options but with very minor differences despite fundamentally different analytical foundations. However, detailed comparison reveals important nuances in method emphases that prove valuable rather than problematic. MCDA favors the Eyasi route\u0026apos;s balanced approach to immediate trade-off optimization, while BBN identifies Mbulu route\u0026apos;s superior uncertainty resilience and long-term sustainability.\u003c/p\u003e\n\u003cp\u003eThe integration reveals critical efficiency insights: Mbulu achieves 0.762 sustainability points per 1,000 people served versus Northern\u0026apos;s 0.195 points per 1,000 people, representing 291% higher efficiency. These findings challenge conventional assumptions that maximizing population connectivity yields optimal outcomes, demonstrating instead that targeted sustainable infrastructure provides superior long-term value per capita served.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003e\u003cstrong\u003eChallenging conventional infrastructure paradigms\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOur results fundamentally challenge prevailing assumptions in road infrastructure planning that prioritize immediate connectivity benefits over integrated sustainability outcomes. The systematic underperformance of the Northern and Serengeti routes selected through traditional economic optimization criteria and currently under implementation demonstrates how narrow decision frameworks can yield globally suboptimal solutions with severe long-term consequences.\u003c/p\u003e\n\u003cp\u003eImportantly, our analysis reveals that routes crossing protected areas (Serengeti and Northern) require the highest construction costs ($4.50 and $3.80 million per kilometer respectively) due to mandatory wildlife-friendly infrastructure including animal crossing structures, detection systems, specialized fencing, and environmental monitoring protocols. Despite these expensive mitigation measures, these routes still generate poor sustainability outcomes, questioning the economic logic of protected area development even when environmental safeguards are implemented.\u003c/p\u003e\n\u003cp\u003eThe superior performance of conservation-oriented routes reveals a critical insight contradicting conventional development wisdom: sustainable road infrastructure can achieve substantial connectivity benefits while providing better long-term outcomes at lower costs. The Mbulu route exemplifies this paradox, achieving the highest sustainability scores while requiring the lowest construction costs ($2.45 million per kilometer) by strategically avoiding protected areas entirely. This challenges fundamental assumptions about cost-effectiveness in infrastructure planning, suggesting that \u0026quot;conservation-by-design\u0026quot; represents the most economically rational approach.\u003c/p\u003e\n\u003cp\u003eThe successful MCDA-BBN integration addresses a fundamental sustainability science challenge: combining stakeholder preference incorporation with uncertainty quantification in complex decision contexts. Our approach demonstrates that methodological integration leverages complementary strengths while addressing individual limitations, providing more robust decision support than single-method approaches.\u003c/p\u003e\n\u003cp\u003eThe participatory stakeholder engagement process reveals sophisticated understanding of infrastructure trade-offs among local experts that conventional consultation processes often fail to capture. The strong consensus across different stakeholder groups (r = 0.73-0.89) demonstrates that meaningful agreement on complex trade-offs is achievable when supported by structured analytical frameworks. This contradicts common assumptions that stakeholder conflicts are irreconcilable and must be resolved through political rather than technical processes.\u003c/p\u003e\n\u003cp\u003eBBN modeling proves particularly valuable for capturing indirect effects and feedback loops that deterministic approaches systematically miss. The probabilistic framework revealed how infrastructure decisions propagate through complex socio-ecological systems, generating cascading consequences that unfold over years or decades. Cross-validation provides crucial robustness testing while revealing complementary insights about different decision aspects.\u003c/p\u003e\n\u003cp\u003eThe framework demonstrates that conservation and development can be synergistic when supported by rigorous analytical tools and participatory planning processes. Our identification of Eyasi and Mbulu routes as superior alternatives provides immediate policy guidance for Tanzania while offering a replicable template for a more sustainable infrastructure planning in biodiversity hotspots globally.\u003c/p\u003e\n\u003cp\u003eSuccessful implementation requires addressing technical capacity constraints; as integrated MCDA-BBN analysis requires specialized skills currently limited in many planning agencies. However, our experience demonstrates that participatory approaches can generate valuable results with modest technical resources when supported by appropriate facilitation.\u003c/p\u003e\n\u003cp\u003eFuture research should expand comparative analysis across additional biodiversity hotspots to test framework generalizability while identifying context-specific adaptations. Integration with emerging technologies including artificial intelligence and automated satellite monitoring could enhance evaluation frameworks while reducing implementation costs.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eStudy area and infrastructure development context\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe Greater Serengeti Ecosystem encompasses approximately 30,000 km\u0026sup2; of interconnected protected areas in northern Tanzania, representing one of Africa\u0026apos;s most significant conservation landscapes\u0026sup2;\u0026sup3;. The ecosystem includes Serengeti National Park (14,750 km\u0026sup2;), Ngorongoro Conservation Area (8,292 km\u0026sup2;), three game reserves (Maswa, Ikorongo-Grumeti, Kijereshi totaling 4,500 km\u0026sup2;), Wildlife Management Areas (Makao, Ikona covering 2,200 km\u0026sup2;), and Lolindo Game Controlled Area (1,000 km\u0026sup2;) \u0026sup2;⁴.\u003c/p\u003e\n\u003cp\u003eThis diverse protected area network hosts the world\u0026apos;s largest terrestrial mammal migration, with approximately 1.5 million wildebeest, 200,000 zebras, and 300,000 Thomson\u0026apos;s gazelles participating in annual movements covering over 1,800 km\u0026sup2;⁵. The ecosystem supports 70+ large mammal species, 500+ bird species, and provides critical habitat for endangered species including African wild dogs, cheetahs, and black rhinoceros. It sequesters approximately 1 million metric tons of carbon annually and contains two UNESCO World Heritage Sites plus one UNESCO Biosphere Reserve\u0026sup2;⁶, \u0026sup2;⁷.\u003c/p\u003e\n\u003cp\u003eHuman populations within and around the GSE total approximately 4.6 million people with annual growth rates of 2.8%, creating increasing pressure for infrastructure development\u0026sup2;⁸. Economic activities include pastoralism (affecting 65% of households), small-scale agriculture (35% of land use), tourism (contributing $1.8 billion annually to national economy), and natural resource extraction, with limited road access constraining market participation and service delivery\u0026sup2;⁹, \u0026sup3;⁰.\u003c/p\u003e\n\u003cp\u003eFour road development options have been proposed since 1990 to improve regional connectivity, each representing different institutional priorities and development philosophies: Northern route (Ministry of Works and Transport, 1995, emphasizing economic connectivity), Serengeti route (tourism industry stakeholders, 2008, prioritizing tourism access), Eyasi route (collaborative planning between Tanzania National Parks Authority and regional development committees, 2012, seeking balanced development), and Mbulu route (conservation organizations and community-based natural resource management groups, 2015, prioritizing biodiversity protection).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStakeholder Engagement and Data Collection\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe employed systematic three-phase data collection designed to ensure methodological rigor while building legitimacy through inclusive participation\u0026sup3;\u0026sup1;, \u0026sup3;\u0026sup2;. The approach integrated established participatory research protocols with structured analytical frameworks to balance stakeholder ownership with technical precision\u0026sup3;\u0026sup3;.\u003c/p\u003e\n\u003cp\u003ePhase 1: Literature Review and Expert Identification involved systematic review of 47 peer-reviewed publications on Serengeti ecology and road impacts using structured search protocols across multiple databases (Web of Science, PubMed, Google Scholar, African Journals Online). Search terms combined ecosystem identifiers with infrastructure and impact keywords, applying quality criteria for peer-reviewed publications and authoritative gray literature from 1990-2023\u0026sup3;⁴. This phase identified key research gaps, established baseline knowledge, and informed expert selection criteria.\u003c/p\u003e\n\u003cp\u003ePhase 2: Expert Consultation occurred at Tanzania Wildlife Research Institute annual conference (December 2022) with 14 specialists selected based on publication record (minimum 5 Serengeti-focused publications), field experience (minimum 10 years\u0026apos; regional work), and disciplinary diversity (ecology, economics, engineering, policy). Structured interviews employed standardized protocols covering impact pathway identification, parameter estimation feasibility, and stakeholder mapping for workshop selection. Individual consultation sessions (15 minutes each) were followed by 2-hour group synthesis to identify consensus areas and knowledge gaps.\u003c/p\u003e\n\u003cp\u003ePhase 3: Collaborative Workshop (January 15-17, 2024, Dar es salaam) included 16 participants representing six stakeholder categories selected using power-interest matrix analysis\u0026sup3;⁵. Participants included tourism operators (2, representing \u0026gt;$50M annual revenue), planning officers (4, from regional and district levels), field ecologists (2, PhD-level with \u0026gt;5 years GSE research), road engineers (4, professionally certified with rural experience), community representatives (2, elected leaders from affected areas), and conservation NGOs (2, with technical expertise and regional presence). Workshop design employed structured facilitation with daily rotating mixed-sector groups to prevent dominance effects and ensure cross-pollination of perspectives.\u003c/p\u003e\n\u003cp\u003eDaily sessions followed systematic protocols: Day 1 focused on criteria development through individual brainstorming (30 minutes), small group synthesis (90 minutes), and plenary integration (120 minutes). Day 2 addressed weight estimation using Analytical Hierarchy Process training (2 hours) followed by individual implementation with consistency checking (5 hours). Day 3 covered Bayesian network development through collaborative structure mapping (3 hours) and parameter elicitation using probability wheels and structured scenarios (4 hours). All sessions were recorded with participant consent, with multiple independent note-takers ensuring accurate documentation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMulti-Criteria Decision Analysis Implementation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMCDA implementation integrated established Analytical Hierarchy Process protocols with spatial analysis using standardized geographic datasets\u0026sup3;⁶⁻\u0026sup3;⁸. The approach followed systematic protocols for criteria standardization, weight determination, and spatial evaluation to ensure methodological rigor and reproducibility.\u003c/p\u003e\n\u003cp\u003eCriteria Development and Standardization employed participatory identification of 16 potential indicators across four domains (engineering feasibility, biodiversity conservation, socioeconomic development, political implementation), refined to 13 spatially-analyzable indicators based on data availability and measurement feasibility. Each indicator received operational definitions, measurement protocols, and standardization procedures to ensure consistent evaluation across route alternatives. Standardization employed Min-Max normalization to 0-100 scales with higher values indicating better performance, accounting for indicator directionality (cost vs. benefit measures).\u003c/p\u003e\n\u003cp\u003eWeight Determination used Analytical Hierarchy Process with structured pairwise comparisons at domain and indicator levels\u0026sup3;⁹. Individual participants completed comparison matrices using established 9-point scales, with consistency checking using Consistency Ratio thresholds (CR \u0026lt; 0.1). Group weight aggregation employed geometric mean approaches to maintain AHP consistency properties while preserving individual preferences. Final weights underwent sensitivity analysis across alternative aggregation methods and stakeholder subgroups to test robustness.\u003c/p\u003e\n\u003cp\u003eSpatial Analysis used QGIS 3.28 with 13 standardized 1km datasets covering terrain (elevation, slope, aspect from 30m SRTM DEM), geology (Tanzania Geological Survey 1:250,000 maps), soils (Harmonized World Soil Database v2.0), land use (ESA WorldCover 2021), protected areas (Tanzania National Parks Authority boundaries), wildlife corridors and migration routes (Tanzania Wildlife Authority GPS collar data), settlements (Global Human Settlement Layer 2020), and infrastructure (OpenStreetMap with ground-truthing). All datasets underwent comprehensive preprocessing including coordinate standardization (EPSG:32736), resolution harmonization, extent standardization, and quality validation through field verification at 50 stratified random points.\u003c/p\u003e\n\u003cp\u003eRoute-specific analysis evaluated performance within 1km buffer zones using zonal statistics, accounting for construction impact areas and immediate effects. Final suitability scores derived through weighted linear combination: S = \u0026Sigma; (wi \u0026times; si), where S = overall score, wi = criterion weight, si = standardized score. Comprehensive sensitivity analysis tested alternative weighting schemes, normalization approaches, and buffer distances to assess result stability.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBayesian Belief Network Development and Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBBN development employed participatory modeling approaches combining stakeholder knowledge with literature-based calibration to create robust probabilistic frameworks for uncertainty analysis⁴⁰,⁴\u0026sup1;. The methodology integrated established BBN development protocols with participatory research methods to ensure technical rigor while maintaining stakeholder ownership.\u003c/p\u003e\n\u003cp\u003eNetwork Structure Development used systematic group processes ensuring stakeholder ownership while maintaining technical feasibility. Participatory workshops employed structured techniques including individual variable identification (10-15 variables per participant), small group pathway mapping using sticky-note visualization on large-format charts, systematic validation of causal relationships using three criteria (logical consistency, empirical support, expert consensus), and complexity management through iterative simplification while preserving essential dynamics.\u003c/p\u003e\n\u003cp\u003eThe consensus network comprised 35 nodes organized in three hierarchical levels: infrastructure drivers (4 nodes: route characteristics, construction timeline, traffic volume, investment level), intermediate impacts (8 nodes: habitat degradation, wildlife mortality, corridor disruption, migration interference, edge effects, poaching access, pollution impact, agricultural expansion), and outcome variables (5 nodes: market access, healthcare access, education access, employment generation, tourism revenue). Each node received operational definitions with 2-3 discrete states and clear measurement protocols. The final architecture included 100 causal relationships, each validated using established criteria and documented with supporting rationale.\u003c/p\u003e\n\u003cp\u003eParameter Estimation involved structured expert elicitation for conditional probability tables using established protocols for environmental decision-making⁴\u0026sup2;. Individual elicitation employed systematic four-step approaches: node state definition with operational criteria, scenario presentation covering all parent node combinations, numerical estimation using calibrated probability wheels with confidence intervals, and consistency checking through probability rules and logical constraints. Group consensus building used equal-weight linear pooling with outlier identification (\u0026gt;2 standard deviations), structured discussion of reasoning behind divergent estimates, and final consensus with documentation of persistent disagreements.\u003c/p\u003e\n\u003cp\u003eProbabilistic Inference and Analysis used R statistical software with specialized packages: bnlearn for network structure learning and validation, gRain for probabilistic inference and belief updating, and Rgraphviz for network visualization. Uncertainty propagation employed Monte Carlo analysis with 1,000 iterations incorporating parameter uncertainty (normal distributions around expert estimates), structural uncertainty (alternative network specifications), and scenario analysis (systematic infrastructure development variations).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTemporal Analysis and Cross-validation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTemporal modeling employed differential decay functions calibrated to regional empirical data to project 30-year sustainability trajectories⁴\u0026sup3;, ⁴⁴. Cross-validation used multiple approaches to assess framework robustness and methodological reliability.\u003c/p\u003e\n\u003cp\u003eDecay Rate Calibration used regional deforestation studies and infrastructure impact meta-analyses to establish route-specific decay parameters. Calibration employed 15-year Landsat analysis of forest loss around existing Serengeti roads with spatial buffers (1km, 5km, 10km), wildlife population monitoring data from Serengeti Research Institute (1990-2020), tourism revenue trends from Tanzania National Parks Authority, and ecosystem service assessment studies. Statistical analysis used exponential decay functions St = S0 \u0026times; e^(-\u0026alpha;t) fitted to empirical data with route-specific parameter estimation. This was done on the assumption that the current past 20-year trend (2000 \u0026ndash; 2020) of changes will remain the same. Conservation-oriented routes (as defined by stakeholders) received decay rates of 4.6-4.9% annually, while connectivity-focused routes showed 6.2-6.8% rates based on induced development pressure patterns.\u003c/p\u003e\n\u003cp\u003eUncertainty Analysis employed Monte Carlo simulation with 1,000 iterations using Latin Hypercube Sampling for efficient parameter space coverage. Analysis incorporated parameter uncertainty (\u0026plusmn;10-20% variation around mean estimates), model uncertainty (ensemble modeling of alternative decay functions), scenario uncertainty (multiple development intensity pathways), and stochastic variation (historical environmental fluctuations). Convergence testing verified simulation adequacy while sensitivity analysis identified critical parameters requiring additional research attention.\u003c/p\u003e\n\u003cp\u003eCross-validation employed multiple approaches including correlation analysis (rank correlations, score correlations, uncertainty overlap assessments), bootstrap validation (n=500) for robustness testing across stakeholder subsamples and parameter variations, external validation through independent expert reviews, and comparison with outcomes from analogous infrastructure developments. Statistical significance testing used non-parametric methods (Kruskal-Wallis for multiple groups, Mann-Whitney U for pairwise comparisons) with Bonferroni correction. Effect sizes reported Cohen\u0026apos;s d for continuous variables and Cliff\u0026apos;s delta for ordinal comparisons.\u003c/p\u003e\n\u003cp\u003eIntegration Analysis compared MCDA and BBN rankings using correlation coefficients, assessed efficiency metrics (sustainability per capita served), and evaluated complementary insights from different analytical approaches. Results demonstrated strong overall agreement (r = 0.89) while revealing valuable methodological complementarity for different decision aspects.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eData Availability\u003c/h2\u003e\n\u003cp\u003eAll research data supporting the conclusions are available in the supplementary materials. Comprehensive results tables including route performance metrics and economic analysis are provided in Supplementary Tables S1-S4. Detailed methodology protocols, additional results tables, and extended sensitivity analyses are provided in Supplementary Materials S1-S6.\u003c/p\u003e\n\u003ch2\u003eAcknowledgements\u003c/h2\u003e\n\u003cp\u003eWe thank Tanzania Commission for Science and Tanzania Wildlife Research Institute for research permits, and data access. We acknowledge the 30 workshop participants for sharing their expertise and time, and the communities of the Greater Serengeti Ecosystem for their insights and local knowledge. Funding provided by the Federal Ministry of Research, Technology and Space of Germany under the LANUSYNCON Project.\u003c/p\u003e\n\u003ch2\u003eAuthor Contributions\u003c/h2\u003e\n\u003cp\u003eP.J.M., A.E., and L.B.-F. conceived the study and developed the methodological framework. P.J.M., Q.M.N., and V.M. designed and conducted stakeholder engagement activities. P.J.M., T.K., and C.B. performed spatial data analysis and MCDA implementation. P.J.M. and A.E. developed BBN models and conducted uncertainty analysis. P.J.M., and Q.M.N. coordinated field activities and stakeholder relations. P.I. and L.B.-F. supervised methodology development and provided conceptual guidance. P.J.M. led manuscript writing with substantial contributions from all authors. All authors reviewed and approved the final manuscript.\u003c/p\u003e\n\u003ch2\u003eCompeting Interests\u003c/h2\u003e\n\u003cp\u003eThe authors declare no competing financial or non-financial interests related to this research.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eGlobal Infrastructure Hub. Global Infrastructure Outlook, Infrastructure Investment Needs 50 Countries, 7 Sectors to 2040 (2017).\u003c/li\u003e\n \u003cli\u003eCeballos, G. et al. Accelerated modern human-induced species losses: entering the sixth mass extinction. Sci. Adv. 1, e1400253 (2015).\u003c/li\u003e\n \u003cli\u003eMyers, N. et al. Biodiversity hotspots for conservation priorities. Nature 403, 853-858 (2000).\u003c/li\u003e\n \u003cli\u003eLaurance, W.F. et al. A global strategy for road building. Nature 513, 229-232 (2014).\u003c/li\u003e\n \u003cli\u003eFearnside, P.M. \u0026amp; Gra\u0026ccedil;a, P.M.L. BR-319: Brazil\u0026apos;s Manaus-Porto Velho highway and the potential impact of linking the arc of deforestation to central Amazonia. Environ. Manage. 38, 705-716 (2006).\u003c/li\u003e\n \u003cli\u003eLaurance, W.F. et al. Reducing the global environmental impacts of rapid infrastructure expansion. Curr. Biol. 25, R259-R262 (2015).\u003c/li\u003e\n \u003cli\u003eFlyvbjerg, B. et al. How common and how large are cost overruns in transport infrastructure projects? Transp. Rev. 23, 71-88 (2003).\u003c/li\u003e\n \u003cli\u003eGeneletti, D. Reasons and options for integrating ecosystem services in strategic environmental assessment of spatial planning. Int. J. Biodivers. Sci. Ecosyst. Serv. Manag. 7, 143-149 (2011).\u003c/li\u003e\n \u003cli\u003eKaneko, S. et al. Method of designing forest road construction scenario with GIS-based calibration using sustainable forestry model. Results Eng. 24, 103078 (2024).\u003c/li\u003e\n \u003cli\u003e\u0026Ccedil;alışkan, E. Geographic Information System(GIS) and Roadeng usage to determine the environmentally sensitive forest road route in mountainous terrain. Eur. J. Interdiscip. Stud. 2, 18-25 (2016).\u003c/li\u003e\n \u003cli\u003eMalczewski, J. \u0026amp; Rinner, C. Multicriteria Decision Analysis in Geographic Information Science (Springer, 2015).\u003c/li\u003e\n \u003cli\u003eYannis, G. et al. State-of-the-art review on multi-criteria decision-making in the transport sector. J. Traffic Transp. Eng. Engl. Ed. 7, 413-431 (2020).\u003c/li\u003e\n \u003cli\u003eKiker, G.A. et al. Application of multicriteria decision analysis in environmental decision making. Integr. Environ. Assess. Manag. 1, 95-108 (2005).\u003c/li\u003e\n \u003cli\u003eMendoza, G.A. \u0026amp; Martins, H. Multi-criteria decision analysis in natural resource management: a critical review of methods and new modelling paradigms. For. Ecol. Manag. 230, 1-22 (2006).\u003c/li\u003e\n \u003cli\u003eMcCann, R.K. et al. Bayesian belief networks: applications in ecology and natural resource management. Can. J. For. Res. 36, 3053-3062 (2006).\u003c/li\u003e\n \u003cli\u003eUusitalo, L. Advantages and challenges of Bayesian networks in environmental modelling. Ecol. Model. 203, 312-318 (2007).\u003c/li\u003e\n \u003cli\u003eAguilera, P.A. et al. Bayesian networks in environmental modelling. Environ. Model. Softw. 26, 1376-1388 (2011).\u003c/li\u003e\n \u003cli\u003eBrooks, T.M. et al. Global biodiversity conservation priorities. Science 313, 58-61 (2006).\u003c/li\u003e\n \u003cli\u003eOlson, D.M. \u0026amp; Dinerstein, E. The Global 200: priority ecoregions for global conservation. Ann. Mo. Bot. Gard. 89, 199-224 (2002).\u003c/li\u003e\n \u003cli\u003eCarpenter, S.R. et al. Science for managing ecosystem services: beyond the Millennium Ecosystem Assessment. Proc. Natl. Acad. Sci. 106, 1305-1312 (2009).\u003c/li\u003e\n \u003cli\u003eFolke, C. et al. Resilience thinking: integrating resilience, adaptability and transformability. Ecol. Soc. 15, 20 (2010).\u003c/li\u003e\n \u003cli\u003eAfrican Development Bank. East Africa Regional Integration Strategy 2018-2022 (2018).\u003c/li\u003e\n \u003cli\u003eSinclair, A.R.E. et al. Serengeti IV: Sustaining Biodiversity in a Coupled Human-Natural System (University of Chicago Press, 2015).\u003c/li\u003e\n \u003cli\u003eWalelign, S.Z. et al. Roads and livelihood activity choices in the Greater Serengeti Ecosystem, Tanzania. PLOS ONE 14, e0213089 (2019).\u003c/li\u003e\n \u003cli\u003eThirgood, S. et al. Can parks protect migratory ungulates? The case of the Serengeti wildebeest. Anim. Conserv. 7, 113-120 (2004).\u003c/li\u003e\n \u003cli\u003eDobson, A.P. et al. Road will ruin Serengeti. Nature 467, 272-273 (2010).\u003c/li\u003e\n \u003cli\u003eHoldo, R.M. et al. Predicted impact of barriers to migration on the Serengeti wildebeest population. PLoS ONE 6, e16370 (2011).\u003c/li\u003e\n \u003cli\u003eMsoffe, F.U. et al. A new approach to the classification of land use in Africa: the case of the Greater Serengeti-Mara ecosystem. Afr. J. Ecol. 49, 98-111 (2011).\u003c/li\u003e\n \u003cli\u003eHomewood, K. et al. Staying Maasai? Livelihoods, Conservation and Development in East African Rangelands (Springer, 2009).\u003c/li\u003e\n \u003cli\u003eNelson, F. et al. The evolution and reform of Tanzanian wildlife management. Conserv. Soc. 5, 232-261 (2007).\u003c/li\u003e\n \u003cli\u003eReed, M.S. Stakeholder participation for environmental management: a literature review. Biol. Conserv. 141, 2417-2431 (2008).\u003c/li\u003e\n \u003cli\u003eReed, J. et al. A theory of participation: what makes stakeholder and public engagement in environmental management work? Restor. Ecol. 25, S7-S17 (2017).\u003c/li\u003e\n \u003cli\u003eCornwall, A. \u0026amp; Jewkes, R. What is participatory research? Soc. Sci. Med. 41, 1667-1676 (1995).\u003c/li\u003e\n \u003cli\u003ePetticrew, M. \u0026amp; Roberts, H. Systematic Reviews in the Social Sciences: A Practical Guide (Blackwell Publishing, 2006).\u003c/li\u003e\n \u003cli\u003eBryson, J.M. What to do when stakeholders matter: stakeholder identification and analysis techniques. Public Manag. Rev. 6, 21-53 (2004).\u003c/li\u003e\n \u003cli\u003eSaaty, T.L. The Analytic Hierarchy Process (McGraw-Hill, 1980).\u003c/li\u003e\n \u003cli\u003eSaaty, T.L. Decision making with the analytic hierarchy process. Int. J. Serv. Sci. 1, 83-98 (2008).\u003c/li\u003e\n \u003cli\u003eIshizaka, A. \u0026amp; Labib, A. Review of the main developments in the analytic hierarchy process. Expert Syst. Appl. 38, 14336-14345 (2011).\u003c/li\u003e\n \u003cli\u003eBeier, P. \u0026amp; Noss, R.F. Do habitat corridors provide connectivity? Conserv. Biol. 12, 1241-1252 (1998).\u003c/li\u003e\n \u003cli\u003eCastelletti, A. \u0026amp; Soncini-Sessa, R. Bayesian networks and participatory modelling in water resource management. Environ. Model. Softw. 22, 1075-1088 (2007).\u003c/li\u003e\n \u003cli\u003eVoinov, A. \u0026amp; Bousquet, F. Modelling with stakeholders. Environ. Model. Softw. 25, 1268-1281 (2010).\u003c/li\u003e\n \u003cli\u003eO\u0026apos;Hagan, A. et al. Uncertain Judgements: Eliciting Experts\u0026apos; Probabilities (John Wiley \u0026amp; Sons, 2006).\u003c/li\u003e\n \u003cli\u003eForman, R.T.T. et al. Road Ecology: Science and Solutions (Island Press, 2003).\u003c/li\u003e\n \u003cli\u003eJaeger, J.A.G. et al. Predicting when animal populations are at risk from roads: an interactive model of road avoidance behavior. Ecol. Model. 185, 329-348 (2005).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Bayesian networks, conservation, road infrastructure, multi-criteria analysis, sustainable development","lastPublishedDoi":"10.21203/rs.3.rs-7419361/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7419361/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eGlobal infrastructure investment will reach \u003cspan\u003e$\u003c/span\u003e94 trillion by 2040, posing critical threats to biodiversity in rural regions newly exposed to road access. Traditional planning approaches have failed to sufficiently incorporate stakeholder preferences and account for uncertainty, leading to decisions that often compromise ecosystems. To overcome these limitations, we introduce an integrated decision-support framework that fuses Multi-Criteria Decision Analysis (MCDA) with Bayesian Belief Networks (BBN), allowing explicit incorporation of stakeholder-derived preferences and probabilistic modeling of ecological and socioeconomic outcomes. Applying this framework to four proposed roads in Tanzania's Greater Serengeti Ecosystem, we show that conservation-oriented options (Mbulu and Eyasi) offer 15\u0026ndash;25% better sustainability performance than conventional connectivity-maximizing routes, while serving 70% of the population with 291% higher per-capita efficiency. Thirty-year projections confirm the long-term benefits of ecosystem-sensitive designs, and cross-validation between MCDA and BBN confirms methodological robustness (r, 0.89). This study provides a scalable, evidence-based planning framework for sustainable road development.\u003c/p\u003e","manuscriptTitle":"An Integrated Decision Support Framework for Sustainable Road Planning in Biodiversity Hotspots","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-22 08:44:38","doi":"10.21203/rs.3.rs-7419361/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"bb844c8b-b12f-4303-aa3c-b254df1a1c3b","owner":[],"postedDate":"September 22nd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":54965301,"name":"Earth and environmental sciences/Ecology/Ecosystem ecology"},{"id":54965302,"name":"Earth and environmental sciences/Environmental sciences/Environmental impact"}],"tags":[],"updatedAt":"2026-04-21T13:28:03+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-22 08:44:38","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7419361","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7419361","identity":"rs-7419361","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}