Bridging the Gap: A Comprehensive Analysis of Ecosystem Service Models and Stakeholder Perceptions

Abstract

This article synthesizes current research on the critical comparison between data-driven ecosystem service (ES) models and stakeholder perceptions. It explores the foundational theories of this divergence, examines the methodologies for quantifying both models and perceptions, addresses key challenges like validation and uncertainty, and presents empirical evidence from comparative studies. Findings consistently reveal a significant mismatch—with stakeholders often rating ES potential 32.8% higher than models on average—driven by differing knowledge systems and priorities. The article concludes that integrative frameworks, which combine model ensembles with participatory engagement, are essential for credible, salient, and legitimate environmental decision-making. This synthesis provides valuable insights for researchers and practitioners aiming to optimize ecosystem management and policy.

The Science and the Social: Unpacking the Foundation of Ecosystem Service Assessments

Ecosystem Services (ES) are crucial for human well-being and the global economy. The mapping and assessment of these services are imperative for sustainable ecosystem management and informed policy decisions, such as those related to the United Nations Sustainable Development Goals [1]. However, two distinct methodologies dominate ES research: data-driven spatial modeling and assessments based on stakeholder perceptions. A growing body of evidence reveals a significant divide between the outputs of scientific models and the values and perceptions held by local stakeholders [1] [2]. Understanding this divide is critical for researchers and professionals aiming to design effective environmental management and restoration strategies, as failing to consider plural values can create conflicts and result in policy outcomes lacking stakeholder support [2] [3]. These Application Notes provide a structured comparison of these approaches, detailed experimental protocols, and essential tools for conducting comparative research.

Quantitative Data Comparison: Models vs. Perception

A 2024 national-scale study in Portugal provided a direct quantitative comparison between model-based ES potential and stakeholders' perceptions, revealing systematic disparities [1]. The research developed a composite index (ASEBIO) from eight modeled ES indicators and compared it against a matrix-based methodology reflecting stakeholder perceptions for the year 2018.

Table 1: Quantitative Disparities Between Modeled and Perceived Ecosystem Service Potential [1]

Ecosystem Service Indicator	Average Contrast (Stakeholder Perception vs. Model)	Alignment Category
Drought Regulation	Highest contrast	Low alignment
Erosion Prevention	Highest contrast	Low alignment
Climate Regulation	High contrast	Low alignment
Habitat Quality	High contrast	Low alignment
Pollination	Moderate contrast	Moderate alignment
Food Production	Low contrast	High alignment
Recreation	Low contrast	High alignment
Water Purification	Low contrast	High alignment
Overall Average	Stakeholder estimates 32.8% higher	—

The results demonstrate that stakeholders overestimated ES potential for all selected services, with an average overestimation of 32.8% compared to the model outputs [1]. The largest contrasts were observed in regulating services like drought regulation and erosion prevention.

Experimental Protocols for Comparative Research

Protocol 1: Spatial Modeling of Ecosystem Services

This protocol outlines the methodology for calculating multi-temporal ES indicators and a composite index, as applied in the Portuguese case study [1].

• Objective: To quantify and map the spatiotemporal changes of multiple ecosystem services using a spatial modeling approach.
• Workflow:
  • Land Cover Data Collection: Acquire multi-temporal land cover data (e.g., CORINE Land Cover) for the reference years (e.g., 1990, 2000, 2006, 2012, 2018) [1].
  • ES Indicator Selection and Modeling: Select relevant ES indicators. Calculate each indicator using spatial modeling tools (e.g., the InVEST software suite) and ancillary data. The Portuguese study modeled eight indicators: climate regulation, water purification, habitat quality, drought regulation, recreation, food production, erosion prevention, and pollination [1].
  • Index Composition (ASEBIO): Integrate the individual ES indicators into a novel composite index.
    • Multi-Criteria Evaluation: Use a method like the Analytical Hierarchy Process (AHP) to define weights for each ES indicator. These weights should be defined by stakeholders to reflect the relative importance of each service's supply potential [1].
    • Spatial Calculation: Compute the final ASEBIO index value for each land unit based on the weighted ES indicators.
• Output: A series of maps and statistical data depicting the spatiotemporal changes of individual ES and the composite index from 1990 to 2018.

Protocol 2: Eliciting Stakeholder Perceptions via Deliberative Valuation

This protocol is based on the Deliberative Multicriteria Evaluation (DMCE) method used in studies in Mexico and Massachusetts, which formalizes community involvement and helps bridge the gap between individual and shared social values [2] [3].

• Objective: To identify which ES are perceived by different stakeholder groups and elicit their shared social values through a structured deliberative process.
• Workflow:
  • Stakeholder Identification and Recruitment: Identify and recruit stakeholders from various sectors of society (e.g., local residents, farmers, government officials, NGO representatives) ensuring representation of different professions and backgrounds [2] [3].
  • Deliberative Workshops: Conduct virtual or in-person workshops.
    • Individual Surveys: Begin with individual surveys to capture pre-deliberation values based on personal experience and knowledge [3].
    • Group Deliberation: Facilitate in-depth discussions among participants about ES, allowing for social learning and challenging of individual values. Record and transcribe these deliberations for qualitative analysis [3].
    • Swing Weighting Method: Employ the "swing" weighting method within the DMCE framework. This method allows stakeholders to evaluate and assign importance to different ES, which may have different measurement units, in relation to each other [3].
  • Data Analysis:
    • Quantitative Analysis: Analyze individual survey results and the group's collective preferences (shared social values) derived from the swing weighting [3].
    • Qualitative Analysis: Perform applied thematic and co-occurrence analysis on the deliberation transcripts to identify key themes and reasoning behind the values [3].
• Output: Quantitative data on ES prioritization (both individual and group-level) and rich qualitative data explaining the reasoning behind the valuations.

The Scientist's Toolkit: Key Research Reagents & Solutions

Table 2: Essential Materials and Tools for ES Comparison Research

Item Name	Function/Benefit
InVEST (Integrated Valuation of Ecosystem Services and Tradeoffs)	A suite of spatial models used to map and value ecosystem services, widely used for planning and research applications [1].
CORINE Land Cover (Coordination of Information on the Environment)	Provides consistent multi-temporal land cover maps for European countries, essential for analyzing land use changes and their impact on ES [1].
Analytical Hierarchy Process (AHP)	A multi-criteria evaluation method used to define the relative weights of different ES indicators based on stakeholder input for composite index creation [1].
Deliberative Multicriteria Evaluation (DMCE)	A framework that combines structured decision-making with deliberation to elicit shared social values for ES, incorporating local knowledge [3].
"Swing" Weighting Method	A technique used within DMCE that is intuitive for participants and allows for the evaluation of ES with different measurement units [3].
APCA Contrast Calculator (Advanced Perceptual Contrast Algorithm)	A next-generation tool for ensuring color contrast accessibility in data visualizations, critical for creating inclusive charts and graphs for publications and presentations [4].

Discussion & Integration Strategies

The quantitative and methodological disparities highlighted in Sections 2 and 3 underscore a critical challenge in ES research. The significant mismatch, such as the 32.8% average overestimation by stakeholders, reveals that model-based assessments and human perception capture different aspects of reality [1]. Models are grounded in biophysical data but can overlook socio-cultural dimensions, while perceptions are shaped by personal experience, cultural values, and collective discourse, potentially leading to overestimations or differing priorities [1] [2].

A key unsolved issue in ES modeling is the definition of "service sheds"—the appropriate spatial and temporal context for quantifying a service—which, if not properly accounted for, can lead to misleading estimates [5]. Furthermore, the validation of ES models remains a largely overlooked step, raising questions about the credibility of their outcomes and the need for frameworks that include validation against raw field or sensing data [6].

To bridge this divide, integrative strategies are essential. Combining the DMCE protocol with the spatial modeling protocol allows for the creation of scientifically robust and socially legitimate ES assessments. Presenting results both quantitatively and qualitatively, as outlined in the protocols, helps minimize the disconnect between research and policymakers, providing more useful and tangible information for decision-making [3]. This integrated approach ensures that both data-driven models and human perspectives are sufficiently considered, leading to more balanced, inclusive, and effective ecosystem management and policy.

The Critical Role of Integrating Knowledge Systems in ES Management

Ecosystem Services (ES) management has traditionally relied heavily on quantitative, model-driven assessments. However, a growing body of evidence reveals that disconnects between scientific models and the knowledge and priorities of local stakeholders can compromise the effectiveness and sustainability of conservation outcomes [1] [7]. This document outlines application notes and protocols for the systematic integration of diverse knowledge systems—specifically scientific data and Indigenous and Local Knowledge (ILK)—into ES research and management. Grounded in a thesis that compares ES models with stakeholder perceptions, these guidelines are designed to help researchers and practitioners generate more holistic, legitimate, and contextually relevant evidence for decision-making.

Background and Rationale

The Evidence Gap: Models versus Perceptions

Recent research underscores a significant misalignment between model-based ES assessments and stakeholder perceptions. A 2024 national-scale study in Portugal quantified this gap, finding that stakeholders' valuations of ES potential were, on average, 32.8% higher than model-based calculations [1]. The disparity varied by service; while drought regulation and erosion prevention showed the highest contrasts, services like water purification, food production, and recreation were more closely aligned [1]. This mismatch highlights that biophysical models alone are insufficient for capturing the full spectrum of values and benefits that ecosystems provide to people.

The Value of Indigenous and Local Knowledge (ILK)

ILK represents a cumulative body of knowledge, practice, and belief about the relationship of living beings with each other and their environment [8]. Integrating ILK is crucial for several reasons:

• Contextual Understanding: Local communities possess deep, historically rooted knowledge of ecosystem dynamics and functions [8].
• Revealing Divergent Priorities: Studies consistently show that communities grounded in Traditional Ecological Knowledge (TEK) often prioritize tangible provisioning and cultural services (e.g., food, raw materials), whereas experts and models tend to emphasize regulating services (e.g., carbon sequestration, hazard regulation) [7].
• Ethical Imperative and Equity: Including local perspectives is a step toward overcoming power dichotomies and epistemological asymmetry between Western science and other knowledge systems [8].

Table 1: Comparative Analysis of Model-Based and Stakeholder-Perceived ES Potential

Ecosystem Service	Model-Based Potential	Stakeholder-Perceived Potential	Approximate Disparity
Drought Regulation	Lower	Higher	Highest Contrast
Erosion Prevention	Lower	Higher	Highest Contrast
Water Purification	High	High	Closely Aligned
Food Production	Stable	High	Closely Aligned
Recreation	Improved	High	Closely Aligned
Overall Average	-	-	+32.8% [1]

Application Notes: A Framework for Integration

A successful integration process is cyclical and adaptive, ensuring mutual learning and validation. The following framework synthesizes best practices from methodological research [8] [9].

Foundational Principles for Co-Production of Knowledge

• Post-Normal Science Perspective: Acknowledge that ES management operates in contexts characterized by complexity, high stakes, and uncertainty, necessitating an "extended peer community" that includes local stakeholders [8].
• Ethnoecology: Draw from this discipline to revalue cultures and forms of natural resource appropriation [8].
• Iterative Validation: The process must be cyclical, involving constant data validation and agreement-making with communities, rather than a linear, extractive exercise [8].

A Cyclical Integration Workflow

The integration of knowledge systems is not a linear process but a continuous cycle of engagement, analysis, and validation. The following workflow, adapted from methodologies in socio-ecological research, outlines the key phases [8].

Diagram 1: Cyclical Workflow for Knowledge Integration. This process emphasizes reciprocity and adaptive learning between researchers and communities [8].

Experimental Protocols

This section provides detailed, actionable protocols for implementing the key stages of the integration workflow.

Protocol 1: Trust Building and Scoping (Stage 0)

Objective: To establish mutual trust, define common goals, and gain a preliminary understanding of the socio-ecological context.

Steps:

• Initial Contact: Identify and meet with key informants and community leaders to introduce the research objectives.
• Community Meetings: Hold open meetings to discuss the study's purpose, potential benefits, and commitments required from all parties.
• Preliminary Agreements: Collaboratively define the geographical area of study, main themes of interest, and the role of the community in the process.
• Ethical Considerations: Obtain collective and individual informed consent for participation, ensuring transparency about data use and ownership.

Key Outputs: Memorandum of understanding, defined study area, list of key informants.

Protocol 2: Multi-Tool Data Collection (Stage 1)

Objective: To gather rich, qualitative and quantitative data on ES from both individual and group perspectives.

Methods should be applied interdependently, with each tool building on the previous one [8].

Table 2: Suite of Methods for Socio-Cultural ES Assessment

Method	Level of Application	Key Function	Procedure Notes
Semi-Structured Interviews [8]	Individual	Explore individual perceptions, experiences, and relationships with the socio-ecosystem.	Conduct as conversations in homes; use evenly suspended attention and free association. Cover topics: way of life, productive activities, socio-environmental problems.
Participatory Mapping [8]	Group	Visualize collective spatial knowledge and strengthen bonds between participants.	Produce maps with local actors to represent their territory. A moment of collective exchange to understand spatial relationships and values.
"Walking in the Woods" / Field Transects [8]	Individual/Group	Ground-truth information and elicit knowledge in context.	Walk with community members through different ecosystems, discussing uses, names, and changes in vegetation and landscapes.
Structured Priority Surveys [7]	Individual/Group	Quantify perceptions and priorities for different ES across land uses.	Use a two-step design: 1) Assess perception/use level (e.g., 4-point scale). 2) For services with ≥50% recognition, conduct a 100-point allocation task to evaluate relative importance.

Protocol 3: Data Systematization and Analysis (Stage 2)

Objective: To synthesize and analyze the collected data, integrating qualitative narratives with quantitative model outputs.

Steps:

• Qualitative Data Processing: Transcribe interviews and field notes. Use coding techniques (e.g., thematic analysis) to identify key ES, values, and concerns.
• Quantitative Data Analysis: Analyze survey data (e.g., point allocations) to produce descriptive statistics and compare priorities between stakeholder groups (e.g., community vs. experts) and land uses.
• Spatial Data Integration: Georeference participatory maps and integrate them with scientific spatial data (e.g., land cover maps, InVEST model outputs) in a GIS environment.
• Comparative Analysis: Juxtapose the ILK-derived ES priorities and values with the outputs from biophysical and spatial ES models to identify alignments, trade-offs, and gaps.

Protocol 4: Validation and Working Agreements (Stage 3)

Objective: To validate the interpreted results with the community and collaboratively define pathways for action.

Steps:

• Validation Workshops: Return to the community to present the systematized results (e.g., through maps, simple graphs, and narratives) and check for accuracy and interpretation.
• Deliberative Dialogue: Facilitate discussions on the implications of the findings, particularly the identified gaps between model outputs and local perceptions.
• Co-Develop Recommendations: Work with stakeholders to translate the integrated knowledge into management recommendations, policy briefs, or community action plans.
• Establish Long-Term Feedback Mechanisms: Create channels for ongoing communication and monitoring.

This section details essential tools and frameworks for conducting integrated ES assessments.

Table 3: Research Reagent Solutions for Integrated ES Assessments

Tool / Resource	Type	Primary Function	Relevance to Integration
InVEST [1]	Spatial Modelling Suite	Models biophysical and economic production of ES (e.g., carbon storage, erosion control).	Provides the scientific, data-driven baseline for ES supply that can be compared with perceived ES values.
Semi-Structured Interview Guide [8]	Qualitative Instrument	Elicits narratives on way of life, resource use, and environmental change.	Captures ILK and contextualizes quantitative data. The cornerstone of socio-cultural assessment.
Participatory Mapping Kit [8]	Spatial Tool	Engages stakeholders in producing maps of their territory, resources, and values.	Makes local spatial knowledge explicit, allowing it to be visualized and integrated into GIS.
Analytical Hierarchy Process (AHP) [1]	Multi-Criteria Analysis	Structures stakeholder preferences by weighting the relative importance of different ES.	Quantifies and incorporates stakeholder priorities into a composite index (e.g., ASEBIO index [1]).
Final Ecosystem Goods and Services (FEGS) Scoping Tool [10]	Classification Framework	Provides a structured process for identifying stakeholders and the ES benefits relevant to them.	Ensures a comprehensive and inclusive scoping of ES, preventing the omission of locally important services.

Integrated Data Visualization and Decision Support

The ultimate goal of integration is to produce synthesized knowledge that is accessible and useful for decision-makers. Comparative analysis, as conducted in Portugal, can be powerful for highlighting gaps and synergies [1]. The following diagram illustrates a generalized analytical framework for comparing model outputs with stakeholder perceptions.

Diagram 2: Framework for Comparative Analysis of ES Models and Perceptions. This pathway guides the synthesis of quantitative and qualitative knowledge for robust decision-making [1].

In the assessment and management of complex social-ecological systems, such as fisheries or woodland biodiversity, two distinct forms of knowledge are increasingly recognized as essential: data-driven objectivity and contextualized local knowledge. Data-driven objectivity relies on quantitative scientific information generated through formalized processes like monitoring programs, retrospective assessments, and predictive models [11]. Conversely, contextualized local knowledge encompasses the ecological or socioeconomic understanding held by place-based communities and stakeholders, derived from on-the-ground observations, intergenerational experience, and personal perceptions [11]. The integration of these knowledge systems is seen as best practice for decision-making in fields like biodiversity management and ecosystem service assessment, though it presents significant challenges when predictions from these different viewpoints do not align [12]. This document outlines application notes and protocols for researchers aiming to compare and integrate these knowledge forms within ecosystem services and environmental management research.

Application Notes: Comparative Frameworks and Tensions

Table 1: Comparison of Knowledge Types in Environmental Research

Characteristic	Data-Driven (Scientific) Knowledge	Contextualized Local Knowledge	Institutional Expert Knowledge
Primary Source	Scientific monitoring, sensor data, models, species distribution data [11]	On-the-water observations, intergenerational experience, personal perceptions [11]	Management/research experience, colleague communication, personal knowledge [11]
Typical Form	Quantitative, numerical	Qualitative, narrative, experiential	Often a blend of quantitative and qualitative
Key Strength	Provides robust hindcasts/forecasts; generalizable [11]	Offers long historical baselines; rich socio-ecological context [11]	Domain-specific understanding; credible for policy legitimization [11]
Inherent Challenge	Scale mismatch with management; demands long-term monitoring [11]	May be perceived as anecdotal; difficult to standardize [11]	Judgment adequacy for complex challenges [11]
Example in Practice	Species-specific exposure studies; biomass harvest models [11]	Fishers' perceived impacts of climate change [11]	Expert elicitation to rank species' relative vulnerability [11]

Documented Tensions and Synergies in Research Outcomes

Comparative studies consistently reveal tensions between the outcomes of data-driven models and stakeholder-based evaluations. In a Portuguese study on ecosystem services (ES), a significant mismatch was found between ES potential calculated via spatial models and the potential perceived by stakeholders; stakeholder estimates were, on average, 32.8% higher [1]. The degree of contrast also varied by service type. Discrepancies in climate vulnerability assessments (CVAs) for fisheries have been attributed to several factors, including [11]:

• Varying levels of individual familiarity, expertise, and research efforts across species.
• Divergences in the use of assessment indicators and scoring criteria.
• Data and knowledge gaps related to species' biological traits and fisheries socioeconomics.
• Uncertainties stemming from data quality and knowledge confidence.

Despite these tensions, synergies are evident. In woodland management, stakeholder predictions and biodiversity data models showed general similarities in ranking the performance of different management scenarios, though important differences remained [12]. This underscores that these knowledge systems are not mutually exclusive but can provide complementary insights.

Experimental Protocols for Comparative Research

Protocol 1: Integrated Climate Vulnerability Assessment (CVA)

This protocol is adapted from methodologies used in fisheries social-ecological systems [11].

Objective: To assess and compare climate vulnerability findings derived from scientific data, institutional expert knowledge, and local fishermen's knowledge.

Workflow:

Detailed Methodology:

• Phase 1: Desktop Research and Data Compilation (Data-Driven Approach)

  • Action: Compile and pre-process existing scientific data. This includes species biological trait data (e.g., maximum body length, thermal safety margin), long-term fisheries monitoring data (biomass, harvest), and oceanographic data from models or remote sensing [11].
  • Output: A standardized dataset for input into a pre-defined CVA framework. This may involve calculating quantitative vulnerability scores based on exposure, sensitivity, and adaptive capacity indicators.

• Phase 2: Expert Knowledge Elicitation

  • Action: Conduct structured surveys or workshops with a diverse group of institutional experts (e.g., managers, policy-makers, researchers, NGO representatives).
  • Procedure: Use deliberative discussions and repeated scoring techniques [12]. Experts are asked to rank or score species' relative vulnerability based on the same assessment framework used in Phase 1, drawing on their own management or research experience.
  • Output: A set of expert-derived vulnerability scores and qualitative justifications.

• Phase 3: Local Knowledge Collection

  • Action: Perform in-depth, semi-structured interviews with local fishermen and community members.
  • Procedure: Use open-ended questions to gather perceptions about observed climate change impacts, changes in species abundance and distribution, and the socioeconomic vulnerability of their livelihoods [11]. This is a bottom-up, participatory approach.
  • Output: Qualitative and, where possible, quantifiable data on perceived vulnerabilities.

• Phase 4: Data Analysis and Triangulation

  • Action: Systematically compare the results from the three phases.
  • Procedure:
    • Identify areas of convergence and divergence in vulnerability rankings.
    • Analyze the root causes of discrepancies (e.g., differing indicators, data gaps, perception biases) [11].
    • Integrate findings to create a composite vulnerability assessment that leverages the strengths of each knowledge form.

Protocol 2: Ecosystem Services (ES) Modeling vs. Perception Assessment

This protocol is designed to compare model-based ES potential with stakeholder perceptions at a national or regional scale [1].

Objective: To quantify and compare ecosystem service potential as generated by spatial models and as perceived by stakeholders.

Workflow:

Detailed Methodology:

• Spatial ES Modeling Track

  • Action: Calculate multiple multi-temporal ES indicators (e.g., climate regulation, habitat quality, drought regulation, recreation) using a spatial modelling approach like InVEST software [1].
  • Input Data: Land cover cartography (e.g., CORINE Land Cover) over a defined time period.
  • Index Development: Integrate the individual ES indicators into a novel composite index (e.g., the ASEBIO index - Assessment of Ecosystem Services and Biodiversity) using a multi-criteria evaluation method.

• Stakeholder Perception Track

  • Action: Elicit stakeholders' perceptions of ES supply potential.
  • Stakeholder Recruitment: Engage a diverse group of stakeholders from various sectors of society through workshops.
  • Procedure: Use an Analytical Hierarchy Process (AHP) to allow stakeholders to assign weights reflecting the relative importance of each ecosystem service [1]. Additionally, use a matrix-based methodology where stakeholders provide their valuation of ES potential for different land cover classes.

• Comparison and Analysis

  • Action: Quantitatively and spatially compare the model-based composite index (ASEBIO) with the matrix of stakeholder-perceived ES potential for a reference year.
  • Output: Calculation of the average percentage difference between model and perception results, and identification of which specific ES show the highest and lowest alignment [1].

The Scientist's Toolkit: Essential Research Reagents and Materials

Table 2: Key Reagents and Tools for Knowledge Integration Research

Item/Tool	Function in Research	Application Context
Spatial Modelling Software (e.g., InVEST)	A suite of models used to map and value ecosystem services based on land/sea use data. Quantifies ES indicators for comparison with perceptions [1].	Ecosystem Services Assessment
Analytical Hierarchy Process (AHP)	A structured multi-criteria decision-making technique. Used to derive stakeholder-defined weights for the importance of different ES or vulnerability indicators [1].	Stakeholder Elicitation, Index Creation
CORINE Land Cover Data	A standardized geographic land cover inventory. Serves as a primary spatial data input for modelling ES potential and understanding land use changes [1].	Spatial Analysis, ES Modelling
Structured Survey Instruments	Standardized questionnaires for expert elicitation. Ensures consistent and comparable data collection on vulnerability scores and rankings across all expert participants [11].	Expert Knowledge Elicitation
Semi-Structured Interview Guides	Flexible interview protocols with open-ended questions. Allows for the collection of rich, contextualized local knowledge while maintaining a focus on core research themes [11].	Local Knowledge Collection
Design Suitability Score (DSS) Framework	An evaluation framework combining AI metrics and stakeholder validation to quantitatively assess the alignment of designs or models with community values [13].	Model/Design Validation, Co-Design Processes

Stakeholder heterogeneity represents a critical factor in environmental management and ecosystem service (ES) assessment. The divergence in perceptions, priorities, and knowledge systems between local communities and expert stakeholders significantly influences conservation outcomes, policy relevance, and sustainable management practices [7]. Integrating these varied perspectives presents both a challenge and necessity for developing balanced, inclusive, and effective environmental governance frameworks. This document outlines practical protocols and applications for researching and integrating diverse stakeholder viewpoints within ecosystem services research, providing methodologies to systematically capture, analyze, and reconcile differing perceptions between local communities and experts.

Quantitative Data on Stakeholder Perceptions and Priorities

Table 1: Documented Gaps Between Expert and Community Ecosystem Service Priorities

Ecosystem Service Category	Community Priority Level	Expert Priority Level	Documented Discrepancy	Study Context
Food Production (Provisioning)	High [7]	Moderate	Communities prioritize tangible provisioning services [7]	Rural Laos (Bamboo forest, rice paddy, teak plantation)
Raw Materials (Provisioning)	High [7]	Moderate	Communities prioritize tangible provisioning services [7]	Rural Laos (Bamboo forest, rice paddy, teak plantation)
Carbon Sequestration (Regulating)	Low [7]	High	Experts emphasize regulating services [7]	Rural Laos (Bamboo forest, rice paddy, teak plantation)
Hazard Regulation (Regulating)	Low [7]	High	Experts emphasize regulating services [7]	Rural Laos (Bamboo forest, rice paddy, teak plantation)
Biodiversity/Habitat (Habitat)	Low [7]	High	Experts emphasize habitat services [7]	Rural Laos (Bamboo forest, rice paddy, teak plantation)
Drought Regulation	Not Specified	Not Specified	Stakeholder estimates 32.8% higher than models on average; drought regulation showed one of the highest contrasts [1]	Mainland Portugal (National ES assessment)
Erosion Prevention	Not Specified	Not Specified	Stakeholder estimates 32.8% higher than models on average; erosion prevention showed one of the highest contrasts [1]	Mainland Portugal (National ES assessment)
Water Purification	Not Specified	Not Specified	One of the most closely aligned services between stakeholders and models [1]	Mainland Portugal (National ES assessment)
Recreation	Not Specified	Not Specified	One of the most closely aligned services between stakeholders and models [1]	Mainland Portugal (National ES assessment)

Table 2: Modeled vs. Perceived Ecosystem Service Potential (Portugal Case Study)

Assessment Method	Overall ES Potential Estimate	Key Findings on Specific ES	Temporal Coverage
Spatial Modelling (ASEBIO Index)	Quantitative index based on land cover and stakeholder-derived weights [1]	Water purification was the highest contributor to the index; climate regulation was the lowest contributor in recent years [1]	1990 - 2018
Stakeholder Valuation	32.8% higher on average than model outputs [1]	All selected ES were overestimated by stakeholders compared to models; largest contrasts in drought and erosion regulation [1]	2018

Experimental Protocols for Eliciting Stakeholder Perceptions

Two-Step Perception and Priority Survey

This protocol, adapted from research in rural Laos, is designed for comparative analysis of stakeholder groups across different land-use types [7].

• Objective: To identify which Ecosystem Services are recognized by stakeholders and then determine their relative importance.
• Applicability: Useful in data-scarce environments and for comparing perceptions across multiple, distinct stakeholder groups (e.g., community members vs. experts).
• Materials:
  • Structured, interviewer-assisted paper questionnaire.
  • List of pre-defined ES items (e.g., 15 items across provisioning, regulating, cultural, and habitat categories).
  • Translated and back-translated questionnaires in the local language and English.
• Procedure:
  • Step 1 - Perception Assessment:
    • For each land-use type, present respondents with the list of ES items.
    • Ask: “In your opinion, what is the level of use of [land-use type] around the village for [ES item]?”
    • Record responses on a four-point scale (1 = no use; 4 = high use).
    • Data Filtering: Classify responses of 3 or 4 as "high use." An ES item proceeds to Step 2 only if it is rated as "high use" by ≥50% of respondents [7].
  • Step 2 - Priority Evaluation:
    • Present respondents only with the ES items filtered from Step 1.
    • Instruct respondents to allocate a total of 100 points across these items in 10-point increments, reflecting their relative importance.
    • Interviewers must provide detailed instructions and use repeated read-backs and probing questions to ensure understanding and data quality [7].
• Data Analysis:
  • Calculate mean priority scores for each ES item by stakeholder group (community vs. expert) and land-use type.
  • Use statistical tests (e.g., t-tests) to identify significant differences in priority allocations between groups.

Deliberative Workshop and Scenario Scoring

This protocol, based on woodland management studies, uses group discussion and scenario planning to elicit nuanced stakeholder judgments [12].

• Objective: To gather contextualized stakeholder knowledge through deliberative discussions and reach a collective evaluation of different management scenarios.
• Applicability: Effective for integrating local, place-based knowledge with scientific predictions and for exploring trade-offs in future management.
• Materials:
  • Defined scenarios (e.g., "Biodiversity Conservation," "People Engagement," "Low Budget") [12].
  • Scoring sheets for individual and group use.
  • Facilitator guides for managing discussions.
• Procedure:
  • Preparation: Develop 3-4 distinct management scenarios with clear goals and implications.
  • Workshop Execution:
    • Bring stakeholders together in a workshop setting.
    • Present each scenario in detail.
    • Facilitate deliberative discussions where stakeholders debate the potential impacts of each scenario on predefined proxies (e.g., spring flowers, weed species) [12].
    • Repeated Scoring: Ask stakeholders to individually score the effects of each scenario both before and after group discussions. This allows observation of how deliberation influences perception.
  • Data Consolidation: Aggregate scores to rank scenarios and identify consensus views or persistent disagreements.
• Data Analysis:
  • Rank scenarios based on average scores.
  • Analyze the shift in individual scores pre- and post-deliberation to understand the impact of group discussion.

Analytical Hierarchy Process (AHP) for Weighting ES

This protocol, used in Portugal and Mulberry-Dyke systems, employs a structured multi-criteria decision-making technique to derive the relative importance of ES from stakeholder input [1] [14].

• Objective: To quantitatively determine stakeholder-defined weights for multiple ecosystem services in a way that ensures consistency in judgment.
• Applicability: Ideal for integrating diverse stakeholder perspectives into composite indices (e.g., the ASEBIO index) or for spatial optimization where trade-offs must be explicitly weighted [1].
• Materials:
  • AHP pairwise comparison survey.
  • Software for AHP calculation and consistency ratio checks (e.g., Expert Choice, or R/Python packages).
• Procedure:
  • Structure the Problem: Define the goal (e.g., "Assess overall ES potential") and list the relevant ES criteria.
  • Pairwise Comparisons: Present stakeholders with a matrix where they compare each pair of ES criteria. They indicate which is more important and to what degree, using a standard 1-9 scale (1 = equally important, 9 = extremely more important).
  • Data Collection: Conduct surveys with different stakeholder groups (e.g., community members, policymakers, scientists) separately to capture heterogeneous preferences [14].
• Data Analysis:
  • Compute the principal eigenvector of the pairwise comparison matrix to derive the priority weights for each ES.
  • Calculate a consistency ratio (CR) to ensure the stakeholder's judgments are logically coherent. A CR < 0.1 is generally acceptable.
  • Compare the derived weight sets from different stakeholder groups.

Visualization of Research Workflows

Stakeholder Heterogeneity Research Workflow

Two-Step Survey Methodology for ES Perception

The Scientist's Toolkit: Key Research Reagents and Materials

Table 3: Essential Materials for Stakeholder Heterogeneity Research

Item Name	Function/Application	Specifications/Examples
Structured Questionnaire	Core instrument for quantitative data collection on ES perceptions and priorities.	Must be translated and back-translated. Includes sections for demographic data, perception scales (e.g., 4-point), and priority allocation tasks (100-point method) [7].
Land Cover Maps (CORINE)	Base spatial data for modeling ecosystem service potential and relating outputs to stakeholder perceptions.	Used in spatial modeling (e.g., ASEBIO index) to quantify ES supply and changes over time (1990-2018) [1].
Spatial ES Models (InVEST)	Software suite for mapping and valuing ecosystem services to produce biophysical models for comparison with stakeholder views.	Models used to quantify ES like habitat quality, carbon storage, and erosion prevention for comparison with stakeholder perceptions [14].
Analytical Hierarchy Process (AHP) Survey	Tool to derive quantitative weights for different ES from stakeholders via pairwise comparisons.	Used to integrate stakeholder preferences into composite indices, ensuring their values directly influence model outcomes [1] [14].
Management Scenarios	Descriptive frameworks used in workshops to elicit stakeholder evaluations of future options and trade-offs.	Scenarios such as "Biodiversity Conservation," "People Engagement," and "Low Budget" are presented for stakeholder scoring [12].
Predefined ES List	A standardized catalog of ecosystem services ensures consistency in survey and workshop materials.	Typically includes 15-20 items across provisioning, regulating, cultural, and habitat service categories, validated by expert panels [7].

From Theory to Practice: Methods for Modeling and Eliciting Perceptions

Ecosystem Services (ES) models are computational tools that translate ecological and socioeconomic data into quantitative assessments of nature's benefits to people [15]. The mapping and valuation of these services are critically important for sustainable development, environmental planning, and nature-based decision-making processes [15]. This guide provides a detailed examination of three prominent ES modeling platforms—InVEST, ARIES, and Co$ting Nature—which enable researchers to spatially quantify and value natural capital and its associated services. These platforms help balance environmental and economic goals by allowing decision-makers to assess quantified tradeoffs among alternative management choices [16] [17]. Understanding their technical specifications, application methodologies, and appropriate use contexts is essential for researchers conducting comparative analyses of ecosystem services models and stakeholder perceptions.

Table 1: Core Characteristics of Featured Ecosystem Service Modeling Platforms

Feature	InVEST	ARIES	Co$ting Nature
Primary Approach	Production functions [16] [17]	AI-assisted, semantic modeling [18] [19]	Pre-processed global data with spatial models [20] [21]
Key Differentiator	Modular, multi-service suite [16]	Models service supply, demand, and flow [19]	Focuses on "costing" (opportunity cost) over "valuing" [20] [21]
User Interface	Standalone application (Workbench) [16]	Web-based platform (k.Explorer) [18] [19]	Web-based Policy Support System (PSS) [20]
Ease of Use	Requires basic-to-intermediate GIS skills [16]	Varies by version; k.Explorer aims for non-technical users [19] [21]	Simple application with global data; complex with custom data [20]
Global Default Data	Limited [17]	Available [19]	Extensive (140+ input maps) [20] [21]

InVEST (Integrated Valuation of Ecosystem Services and Tradeoffs)

Developed by the Natural Capital Project, InVEST is a suite of free, open-source, spatially explicit models used to map and value the goods and services from nature that sustain human life [16] [17]. It employs a production function approach that defines how changes in an ecosystem's structure and function affect the flows and values of ecosystem services across a landscape or seascape [16] [17]. The toolset includes distinct models for terrestrial, freshwater, marine, and coastal ecosystems, and it returns results in either biophysical or economic terms [16] [22]. Its modular design allows users to select only services of interest without running a full suite of models [16].

ARIES (ARtificial Intelligence for Environment & Sustainability)

ARIES is an open-source, collaborative platform powered by k.LAB technology, which uses artificial intelligence and semantic modeling to rapidly assess ecosystem services [18] [19]. Its key innovation is the use of the semantic web paradigm to automatically select and link the best available models and data for a user's specific query and geographic context [19]. ARIES gives equal emphasis to ecosystem service supply, demand, and flow, allowing it to quantify actual service provision and use by society, rather than just potential benefits [19]. It is designed to be a FAIR (Findable, Accessible, Interoperable, Reusable) knowledge commons, supporting applications from urban to global scales [18].

Co$ting Nature

Co$ting Nature is a web-based spatial policy support system focused on natural capital accounting and analyzing the ecosystem services provided by natural environments [20] [21]. Its philosophy centers on "costing nature"—understanding the resource allocation and opportunity costs of protecting nature to produce ecosystem services—rather than merely valuing it in monetary terms [20]. A significant feature is its extensive library of pre-processed global datasets, which allows users to run initial assessments in approximately 30 minutes without specialized GIS skills, though incorporating custom data requires greater technical capacity [20] [21]. It models 18 ecosystem services and incorporates pressures, threats, and conservation priorities [20] [21].

Table 2: Technical Specifications and Resource Requirements

Specification	InVEST	ARIES	Co$ting Nature
Access Model	Downloadable standalone software [16]	Web-based platform (k.Explorer) [18] [19]	Web-based platform (PSS) [20]
Cost	Free, open-source [16]	Free, open-source [18]	Free for non-commercial use [20]
Primary Inputs	Predominantly GIS/map data; information tables [17]	Spreadsheets, databases, maps; global maps available online [19]	Global spatial data at 1 km² or 1 ha; user-uploaded data [21]
Key Outputs	Maps; quantitative ES data; tables/statistics [17]	Maps; quantitative data on ES and environmental assets [19]	Maps; GIS databases; quantitative ES data; economic assessments [21]
GIS Dependency	Required for data prep and viewing results [16] [17]	Not specified	Not required for basic applications [20]
Spatial Resolution	Flexible (local to global) [16]	User-definable [19]	1 km² to 1 ha (10m licensed) [20]
Developer	Natural Capital Project (Stanford University, WWF, TNC) [17]	Basque Centre for Climate Change (BC3) [19]	King's College London, AmbioTEK, UNEP-WCMC [20]

Experimental Protocols and Application Workflows

Protocol: Coastal Flood Risk Assessment Using Co$ting Nature and Suitability Modeling

This protocol is adapted from a study that combined Co$ting Nature outputs with suitability modeling to identify priority areas for Nature-Based Solutions (NBS) in coastal Texas [23].

1. Research Question and Scope Definition

• Objective: Identify coastal areas with high flood risk and low ecosystem service provision to prioritize implementation of nature-based solutions for flood mitigation [23].
• Study Area: Define the geographic boundary (e.g., the Houston-Galveston Metropolitan Statistical Area coastline) [23].

2. Data Acquisition and Preparation

• Primary Data Sources:
  • Access the Co$ting Nature Policy Support System (PSS) [20] [23].
  • Select and download 1-degree tiles at 1-hectare spatial resolution for the study area [23].
  • Obtain complementary regional datasets from local and national agencies (e.g., coastal boundaries, flood risk data, land use/cover) [23].
• Data Integration:
  • Spatially join multiple Co$ting Nature tiles in a GIS environment (e.g., ArcGIS Pro) if the study area spans multiple tiles [23].
  • Harmonize all spatial data to a common coordinate system and resolution.

3. Co$ting Nature Analysis

• Baseline Assessment: Run the Co$ting Nature model to calculate baseline (current) provision of relevant ecosystem services, particularly flood regulation [23].
• Threat and Priority Mapping: Generate maps of conservation priority, ecosystem service provisions, and relative threats from the model outputs [23].

4. Suitability Modeling

• Criterion Identification: Define criteria for NBS implementation based on Co$ting Nature outputs (e.g., low current ecosystem service provision, high conservation priority, high threat levels) [23].
• Suitability Analysis: Conduct a suitability model in GIS software, using the Co$ting Nature outputs as input layers to synthesize areas with the greatest need for NBS [23].
• Flood Risk Overlay: Overlay the suitability output with current flood risk data to identify vulnerable coastal areas that are both high-risk and low-service [23].

5. Validation and Interpretation

• Ground Truthing: Where possible, verify model outcomes with field observations or higher-resolution local data [20].
• Policy Application: Interpret results to inform conservation priorities, target communities with high flood risk and low ecological services, and develop policies for green infrastructure implementation [23].

Protocol: Comparative Assessment of Multiple ES Models

This protocol provides a framework for comparing outputs, data requirements, and usability of InVEST, ARIES, and Co$ting Nature for a single case study, as referenced in global and Turkey-specific reviews [15].

1. Experimental Design

• Objective: Systematically compare the process, resource requirements, and results of multiple ecosystem service models applied to the same geographic context [15].
• Model Selection: Select InVEST, ARIES, and Co$ting Nature for their differing methodological approaches [15].
• Focal Ecosystem Services: Choose 2-3 comparable services (e.g., carbon sequestration, water provisioning, flood regulation) available across all platforms [15].

2. Implementation Phase

• Parallel Modeling: Apply each tool to model the selected ES in the same study area.
• Data Logging: Meticulously document for each tool:
  • Data preparation and acquisition time [17] [20].
  • Model setup and computation time [17].
  • Technical challenges and learning curve [16] [19] [20].
• Output Generation: Produce maps and quantitative values (biophysical or economic) for each service from each model [17] [19] [21].

3. Analysis and Comparison

• Spatial Correlation: Analyze the degree of spatial concordance in ES hotspot maps generated by the different tools [15].
• Quantitative Comparison: Compare the absolute and relative values of ES provision estimated by each model.
• Usability Assessment: Evaluate tools based on required expertise, time investment, and transparency of assumptions [15].

Figure 1: Workflow for comparative assessment of multiple ES modeling platforms, illustrating the sequential stages from research definition through to final synthesis.

The Scientist's Toolkit: Research Reagent Solutions

Table 3: Essential Research Reagents for Ecosystem Services Modeling

Research Reagent	Function	Platform Application Examples
Spatial Data (Land Use/Land Cover)	Serves as primary input representing ecosystem structure; determines service supply potential.	Required by all three platforms [15] [17] [21]
Digital Elevation Model (DEM)	Provides topographic data crucial for hydrologic services (e.g., water yield, flood regulation) and sediment transport.	Core input for InVEST hydrology models; part of Co$ting Nature's base data [17] [23]
Biophysical Tables	CSV files that parameterize ecosystem service production functions by linking land cover to service output.	Critical for InVEST models to define parameters like carbon storage per land cover type [17]
Global Satellite Data Products	Pre-processed, globally consistent datasets (e.g., climate, soil, vegetation indices) that enable rapid initial assessments.	Foundation of Co$ting Nature's global analyses; available in ARIES by default [19] [20] [21]
Economic Valuation Parameters	Data for translating biophysical service flows into monetary values (e.g., shadow prices, damage costs).	Used in InVEST for economic outputs; available in Co$ting Nature V3 [17] [21]
Semantic Model Ontologies	Formal representations of knowledge that define concepts and relationships between ecological and socioeconomic entities.	Core component of ARIES enabling AI-assisted model and data integration [18] [19]

Figure 2: Decision pathway for selecting an appropriate ecosystem services modeling platform based on research objectives, available resources, and technical constraints.

InVEST, ARIES, and Co$ting Nature represent three sophisticated but philosophically distinct approaches to ecosystem service modeling. InVEST offers a structured, modular framework ideal for scenario analysis where ecological processes are well-understood and GIS capacity is available [16] [17]. ARIES leverages artificial intelligence to automate model and data integration, potentially increasing efficiency and reproducibility for users comfortable with its semantic framework [18] [19]. Co$ting Nature provides unparalleled rapid assessment capabilities using its extensive pre-loaded global datasets, making it particularly valuable for initial screenings and conservation priority setting [20] [21]. For researchers conducting comparative studies of ES models and stakeholder perceptions, understanding these technical differences, resource requirements, and appropriate application contexts is fundamental to designing robust experiments and interpreting results critically. The ongoing development of these platforms, particularly towards better integration of global datasets and improved user accessibility, promises to further enhance their utility in both research and policy domains [15] [17].

Integrating stakeholder perceptions with biophysical ecosystem services (ES) models is critical for developing effective environmental management and drug development strategies that are both scientifically robust and socially relevant. Stakeholder perspectives provide essential context, reveal potential conflicts, and help prioritize ecosystem services that might otherwise be overlooked in purely technical assessments [24] [6]. This integration is particularly vital in the context of ecosystem services models, where comparative research between modeled data and human perceptions can validate findings and uncover discrepancies [6]. The following sections provide application notes and detailed protocols for eliciting, quantifying, and integrating these intangible stakeholder priorities into formal research frameworks.

Application Notes: Core Techniques and Their Rationale

Online Stakeholder Engagement Platforms (OSEPs)

Application Note: OSEPs provide a flexible, anonymous environment for engaging geographically and linguistically diverse stakeholders, which is particularly advantageous for discussing potentially polarizing topics in technology and environmental management.

• Key Advantages: These platforms overcome limitations of traditional methods by enabling asynchronous participation, maintaining participant anonymity to reduce power dynamics and encourage open dialogue, and facilitating interactive group discussions through digital diaries and message boards [25]. This is especially valuable for engaging stakeholders across the drug development pipeline or in transnational environmental assessments where in-person gatherings are impractical.
• Integration with ES Models: The qualitative data gathered through OSEPs can be systematically coded and transformed into quantifiable weights for use in multi-criteria decision analysis (MCDA) frameworks, directly linking stakeholder preferences with technical model outputs [25].

Machine Learning-Driven Analysis

Application Note: Machine learning techniques can identify non-linear relationships and key drivers within complex datasets of stakeholder perceptions, moving beyond the limitations of traditional statistical methods.

• Technical Basis: Algorithms such as gradient boosting models excel at processing complex datasets to uncover key patterns and drivers that influence perceptions and priorities [24]. This approach is particularly effective for analyzing large-scale survey data or identifying patterns in open-ended responses that would be difficult to process manually.
• Scenario Planning: Once key drivers are identified, machine learning can inform the design of future scenarios (e.g., natural development, planning-oriented, ecological priority) to model how changes in these drivers might affect stakeholder perceptions and ecosystem service delivery [24].

Integrated Quantitative-Qualitative Data Frameworks

Application Note: Presenting quantitative and qualitative data together provides a cohesive narrative that conveys both measurable trends and the underlying context, creating more compelling and actionable evidence for decision-makers.

• Complementary Roles: Quantitative data (the "what") provides objective metrics and statistical evidence, while qualitative data (the "why" and "how") offers explanatory context, hypotheses, and nuanced understanding [26]. In ecosystem services research, this might involve pairing spatially explicit model outputs from tools like InVEST with narrative data from stakeholder interviews about why certain services are valued.
• Validation Function: Qualitative data can provide crucial contextual validation for quantitative model outputs, helping researchers understand when and why stakeholder perceptions may diverge from biophysical measurements of the same ecosystem services [6].

Experimental Protocols

Protocol: Implementing an Online Stakeholder Engagement Platform

Objective: To systematically elicit, capture, and analyze stakeholder perceptions using a structured online platform.

Table 1: Key Implementation Steps for Online Stakeholder Engagement

Step	Description	Key Considerations
Platform Selection	Select an OSEP (e.g., CMNTY, EngagementHQ) with functionality for surveys, discussion forums, and anonymous interaction.	Ensure the platform complies with data protection regulations (e.g., GDPR, HIPAA).
Stakeholder Identification & Recruitment	Identify candidates through professional networks, literature reviews, and conference participant lists [25].	Target diverse affiliations: government, NGOs, industry, academia, and community representatives.
IRB Approval	Develop and submit all study procedures, including consent forms and data management plans, to the Institutional Review Board [25].	Approval is mandatory for research involving human subjects.
Platform Development & Testing	Create a series of structured surveys and moderated discussion forums focused on the specific topic [25].	Pilot-test the platform to ensure usability and clarity of questions.
Data Collection	Participants engage with platform activities over a defined period (e.g., 2-4 weeks).	Allow asynchronous participation while maintaining moderator presence.
Data Analysis	Employ mixed-methods: quantitative analysis of survey responses and thematic analysis of discussion forum transcripts [26].	Use coding frameworks to identify recurring themes and patterns.

Protocol for Multi-Scenario Ecosystem Service Assessment with Stakeholder Input

Objective: To quantitatively assess ecosystem services under different future scenarios that reflect stakeholder priorities.

Table 2: Phases of Multi-Scenario Ecosystem Service Assessment

Phase	Core Activities	Tools & Methods
1. Baseline ES Assessment	Quantify individual services (water yield, carbon storage, habitat quality, soil conservation) for past and current conditions [24].	InVEST model; comprehensive ES index to assess overall ecological service capacity [24].
2. Driver Identification	Identify key social-ecological drivers influencing ES using machine learning models [24].	Gradient boosting models; analysis of land use, vegetation cover, and socio-economic data [24].
3. Scenario Co-Design	Develop future scenarios (e.g., natural development, planning-oriented, ecological priority) based on stakeholder input and driver analysis [24].	Stakeholder workshops; OSEPs; participatory mapping.
4. Land Use Projection	Project land use changes to a target year (e.g., 2035) under each scenario [24].	PLUS model for simulating complex land-use dynamics at fine spatial scales [24].
5. Future ES Assessment	Evaluate various ecosystem services based on simulated land use for each scenario [24].	InVEST model; trade-off and synergy analysis.

Visualization of Methodological Frameworks

Workflow for Integrated Stakeholder and Ecosystem Services Research

The following diagram illustrates the sequential relationship between stakeholder perception elicitation, data analysis, and ecosystem services modeling, culminating in decision support.

Quantitative and Qualitative Data Integration Workflow

This diagram details the process of collecting, analyzing, and integrating different data types to form a cohesive evidence base.

The Scientist's Toolkit: Essential Research Reagents & Solutions

Table 3: Key Research Reagents and Solutions for Eliciting and Analyzing Stakeholder Perceptions

Tool/Reagent	Type/Category	Primary Function	Application Context
Online Stakeholder Engagement Platform (OSEP)	Software Platform	Provides a structured digital environment for engaging geographically diverse stakeholders asynchronously and anonymously [25].	Eliciting perceptions on complex or potentially polarizing topics (e.g., novel technologies, land use change).
InVEST Model	Biophysical Modeling Suite	Quantifies and maps the biophysical supply and economic value of ecosystem services under different scenarios [24].	Providing quantitative ES data for comparison with stakeholder perceptions; modeling outcomes of different management options.
PLUS Model	Land Use Simulation Tool	Projects land use changes by simulating the interplay between human activities and natural systems under different scenarios [24].	Visualizing potential future landscapes based on stakeholder-driven scenarios.
Machine Learning Algorithms (e.g., Gradient Boosting)	Data Analysis Tool	Identifies non-linear relationships and key drivers within complex datasets of stakeholder perceptions and ecosystem services [24].	Analyzing large-scale survey data; pinpointing the most significant factors influencing perceptions.
RAG Status Indicators	Visualization & Reporting Tool	Uses color-coded status (Red, Amber, Green) to visually demonstrate progress or priority levels in reports [26].	Communicating research priorities or agreement levels distilled from stakeholder data to decision-makers.
Multi-Criteria Decision Analysis (MCDA)	Analytical Framework	Systematically incorporates and weighs different stakeholder views, perceptions, and preferences in a transparent structure [25].	Integrating diverse stakeholder priorities with technical data to support collaborative decision-making.

The Analytical Hierarchy Process (AHP) is a multi-criteria decision analysis (MCDA) method developed by Thomas Saaty in the 1970s that helps individuals and organizations rank alternatives through pairwise comparisons [27]. For ecosystem services research, AHP provides a structured framework to balance the competing demands of technical suitability, stakeholder involvement, and sustainability considerations [28]. This methodology is particularly valuable in complex environmental decision-making contexts such as forest management, where it enables the integration of scientific data with socio-economic values and political considerations [29] [28]. By breaking down complex decisions into a hierarchical structure, AHP facilitates a systematic approach to prioritizing ecosystem services and spatially stratifying management interventions.

The fundamental strength of AHP in ecosystem services assessment lies in its ability to combine quantitative data with qualitative stakeholder judgments. Recent applications in forest ecosystems demonstrate how AHP can capture both scientific foundations and perspectives of various sectors through a stratification model to determine ecosystem service provisions [28]. This integration is crucial for developing management strategies that are not only ecologically sound but also socially acceptable and economically viable, ultimately supporting more sustainable environmental governance.

Fundamental Principles and Workflow of AHP

Core Mathematical Foundation

AHP operates on the principle of decomposing complex decisions into a hierarchical structure and using pairwise comparisons to derive priority scales [27]. The method requires decision-makers to compare elements pairwise at each level of the hierarchy with respect to their parent element at the next higher level. These comparisons are made using a fundamental 1-9 ratio scale where 1 represents equal importance between two elements and 9 represents extreme importance of one element over another [27]. The pairwise comparisons are organized into a reciprocal matrix, and the principal eigenvector of this matrix is computed to derive the priority weights for each element.

The AHP methodology incorporates a consistency measure to validate the logical coherence of judgments. The Consistency Ratio (CR) is calculated to ensure that transitive relationships hold reasonably well across all comparisons. A CR value of 0.1 or less is generally considered acceptable, indicating that the pairwise comparisons are sufficiently consistent to provide meaningful results. This mathematical foundation ensures that subjective judgments are translated into quantitatively derived priorities with known reliability measures.

Hierarchical Decision Structure

The first step in implementing AHP involves structuring the decision problem into a hierarchical model comprising at least three levels [27]:

• Level 1: The overarching decision-making goal
• Level 2: Criteria (and potentially sub-criteria) for evaluating alternatives
• Level 3: The alternatives being considered

For ecosystem services evaluation, this hierarchy typically positions "Sustainable Ecosystem Management" as the goal, with criteria representing different ecosystem services (biodiversity conservation, water protection, timber production, etc.), and spatial units or management scenarios as alternatives [28].

Table: Fundamental AHP Ratio Scale for Pairwise Comparisons

Intensity of Importance	Definition	Explanation
1	Equal importance	Two activities contribute equally to the objective
3	Moderate importance	Experience and judgment slightly favor one activity over another
5	Strong importance	Experience and judgment strongly favor one activity over another
7	Very strong importance	An activity is strongly favored and its dominance demonstrated in practice
9	Extreme importance	The evidence favoring one activity over another is of the highest possible order of affirmation
2,4,6,8	Intermediate values	Used when compromise is needed

Application Notes: AHP for Ecosystem Services Stratification

Case Study Implementation

A recent study in Turkey's Yalnızçam forest area demonstrated the application of AHP for spatial stratification of ecosystem services [28]. Researchers employed a Delphi technique integrated with AHP to capture both scientific grounding and perspectives of various sectors. The iterative framework included ecosystem service identification and prioritization steps, culminating in their spatial stratification of forest stands with geographic information systems. The results revealed a primary focus on biodiversity conservation (78.5%) and water protection (13.3%), with minimal provision for timber production (7.9%) and soil protection (0.04%), and none for climate regulation, eco-tourism, and non-wood forest products [28].

This approach enabled a more efficient spatial zoning strategy that balanced technical and socio-cultural factors, streamlining decision-making processes crucial for the sustainable forest management paradigm. The integration of AHP with spatial analysis tools like GIS represents a powerful methodological advancement for translating stakeholder-derived priorities into concrete spatial management recommendations.

Protocol: Structured AHP Implementation for Ecosystem Services

Phase 1: Problem Structuring and Hierarchy Development

• Stakeholder Identification and Engagement: Identify all relevant stakeholder groups (scientists, policymakers, local communities, industry representatives) who will participate in the pairwise comparisons.
• Ecosystem Services Selection: Define the comprehensive set of ecosystem services to be evaluated, ensuring they represent the full range of provisioning, regulating, cultural, and supporting services relevant to the specific context.
• Hierarchical Model Construction: Develop a decision hierarchy with the goal at the top level, ecosystem service categories and sub-categories at intermediate levels, and management alternatives or spatial units at the bottom level.

Phase 2: Data Collection through Pairwise Comparisons

• Structured Questionnaire Development: Create questionnaires for pairwise comparisons using the 1-9 ratio scale.
• Stakeholder Workshops: Conduct facilitated workshops to guide stakeholders through the pairwise comparison process, ensuring understanding of the scale and concepts.
• Data Collection: Collect completed pairwise comparison matrices from all stakeholder participants.

Phase 3: Data Analysis and Priority Derivation

• Matrix Computation: For each stakeholder or stakeholder group, compute the priority weights from the pairwise comparison matrices using the eigenvalue method [27].
• Consistency Verification: Calculate consistency ratios for each set of comparisons, flagging inconsistent responses for potential review.
• Priority Aggregation: Aggregate individual priorities across stakeholder groups using appropriate techniques (geometric mean, weighted average based on stakeholder importance).

Phase 4: Integration with Spatial Planning

• Satial Data Collection: Gather spatial data layers representing each ecosystem service (species distribution maps, water quality data, recreation potential, etc.).
• Priority Mapping: Create composite maps by weighting spatial layers according to the derived AHP priorities.
• Management Zoning: Delineate spatial management zones based on the dominant ecosystem service priorities identified through the AHP process.

Experimental Protocols and Methodologies

Detailed Pairwise Comparison Protocol

The pairwise comparison process represents the core data collection methodology in AHP. The following protocol ensures rigorous implementation:

• Matrix Preparation: Create a pairwise comparison matrix for each level of the hierarchy, with criteria listed in both rows and columns [27].
• Comparison Execution: For each cell in the matrix, pose the question: "With respect to [parent element], how much more important is the element in the row than the element in the column?"
• Scale Application: Use the fundamental 1-9 ratio scale to record responses, with reciprocals (1/2 to 1/9) used when the column element is more important than the row element [27].
• Diagonal Entries: All diagonal elements are automatically 1 (each element compared with itself).
• Reciprocal Enforcement: Ensure that if element A is rated as X times more important than element B, then element B is rated as 1/X times important as element A.

Table: Example Pairwise Comparison Matrix for Forest Ecosystem Services

	Biodiversity Conservation	Water Protection	Timber Production	Soil Protection
Biodiversity Conservation	1	5	7	9
Water Protection	1/5	1	3	5
Timber Production	1/7	1/3	1	3
Soil Protection	1/9	1/5	1/3	1

Priority Calculation Methodology

The calculation of priority weights from pairwise comparison matrices follows this protocol:

• Matrix Squaring: Square the pairwise comparison matrix (multiply the matrix by itself) [27].
• Row Sum Calculation: Sum each row of the squared matrix to obtain row totals.
• Priority Vector Derivation: Sum all row totals, then divide each row total by this sum to obtain the initial priority vector.
• Iteration: Repeat steps 1-3 using the resulting matrix until the priority vector stabilizes (no significant changes to three or four decimal places) [27].
• Consistency Index Calculation: Compute the consistency index (CI) using the formula CI = (λmax - n)/(n - 1), where λmax is the principal eigenvalue and n is the matrix size.
• Consistency Ratio Determination: Calculate CR = CI/RI, where RI is the random index value based on matrix size. A CR ≤ 0.10 indicates acceptable consistency.

Data Presentation and Visualization

Structured Data Tables for Results Communication

Effective presentation of AHP results requires clear tabular formats that enable easy comparison of priorities across stakeholder groups and scenarios. The following table structure is recommended for ecosystem services applications:

Table: Ecosystem Services Priority Weights from AHP Analysis (Yalnızçam Case Study) [28]

Ecosystem Service	Priority Weight	Stratification Percentage	Dominant Stakeholder Perspective
Biodiversity Conservation	0.425	78.5%	Ecological Integrity
Water Protection	0.215	13.3%	Public Health & Safety
Timber Production	0.185	7.9%	Economic Sustainability
Soil Protection	0.095	0.04%	Long-term Productivity
Climate Regulation	0.045	0%	Global Environmental Values
Eco-tourism	0.025	0%	Recreational & Cultural Values
Non-wood Forest Products	0.010	0%	Local Livelihoods

AHP Workflow and Hierarchy Visualization

AHP Implementation Workflow

Ecosystem Services Decision Hierarchy

The Scientist's Toolkit: Essential Research Reagents and Materials

Methodological and Analytical Tools

Table: Essential Research Reagents and Solutions for AHP Implementation

Tool/Resource	Function	Application Context
Pairwise Comparison Survey Instrument	Standardized data collection format for stakeholder judgments	Eliciting consistent preference ratings across all hierarchy elements
AHP Calculation Software (Expert Choice, Super Decisions, R packages)	Matrix computation and priority weight derivation	Performing eigenvalue calculations and consistency verification
Consistency Validation Protocol	Quality control for stakeholder responses	Identifying and addressing logically inconsistent judgments
Stakeholder Analysis Framework	Classification and weighting of different stakeholder groups	Ensuring representative inclusion of diverse perspectives
Spatial Analysis Tools (GIS)	Mapping and visualization of AHP results	Translating priority weights into spatial management zones
Sensitivity Analysis Protocol	Testing robustness of results to judgment variations	Assessing stability of priorities under different scenarios

Advanced Considerations and Methodological Refinements

Recent research in multi-criteria decision making has highlighted approaches that go "beyond consistency" to address the challenges of intransitive preferences in real-world applications [29]. The skew-symmetric bilinear representation offers an alternative mathematical framework for modeling situations where stakeholder preferences may not be perfectly consistent, yet still contain valuable information for decision-making [29]. This is particularly relevant in complex ecosystem services evaluations where different stakeholder groups may have fundamentally different value systems that lead to apparently inconsistent preference structures.

For researchers applying AHP in contested environmental decision contexts, it is advisable to supplement traditional consistency measures with qualitative analysis of the underlying reasons for inconsistency. This may involve facilitated discussions with stakeholders to explore the value tensions that manifest as mathematical inconsistencies in the pairwise comparison matrices, potentially leading to richer insights and more nuanced management recommendations.

Data Management and Structural Considerations

Proper data structuring is essential for effective AHP implementation. Research data should be organized in a structured format with clear rows and columns, where each row represents a single record and each column represents an attribute or variable [30] [31]. This tabular structure enables efficient computation of priority weights and consistency ratios. For ecosystem services applications, it is critical to clearly define the granularity of the data - what each record represents (e.g., individual stakeholder responses, aggregated group preferences, or spatial management units) [30].

When integrating AHP with spatial decision support systems, researchers should maintain clear documentation of data sources, transformation methods, and weighting procedures. This ensures transparency and reproducibility in how stakeholder-derived preference weights are translated into spatial management recommendations. The structured data format also facilitates sensitivity analysis by enabling systematic variation of input parameters to test the robustness of resulting priorities and management zones.

Integrating quantitative ecosystem services (ES) models with qualitative stakeholder perceptions presents a significant opportunity for more holistic environmental assessments. This case study details the methodology behind the ASEBIO (Assessment of Ecosystem Services and Biodiversity) index, a composite index developed for mainland Portugal. The ASEBIO index exemplifies a robust framework for comparing data-driven spatial models with survey-based stakeholder valuations, revealing critical insights for ecosystem management and policy development [1].

Application Notes: The ASEBIO Index Framework

The ASEBIO index was designed to accomplish two primary objectives: first, to monitor the spatiotemporal changes of multiple ecosystem services over a 28-year period (1990-2018), and second, to quantify the divergence between model-based ES indicators and the potential of ES as perceived by stakeholders [1]. This integrative approach is vital for aligning scientific models with human perspectives, thereby fostering more inclusive and effective land-use planning.

Key Findings and Disparities

The comparative analysis between the modelled ASEBIO index and stakeholder perceptions for the year 2018 revealed a significant mismatch. On average, stakeholder estimates of ES potential were 32.8% higher than the model-based calculations [1]. This disparity was not uniform across all ecosystem services, highlighting the nuanced nature of perception versus measured reality.

• Largest Contrasts: The services with the most considerable overestimation by stakeholders were drought regulation and erosion prevention [1].
• Closest Alignment: The services where stakeholder perceptions most closely aligned with the models were water purification, food production, and recreation [1].

This divergence underscores the potential risks of relying exclusively on either modelling or perception-based approaches and validates the need for integrative strategies that leverage the strengths of both methodologies [1].

Experimental Protocols

The development of the ASEBIO index followed a structured, multi-stage protocol involving spatial modelling, stakeholder engagement, and multi-criteria evaluation. The workflow below illustrates the key stages of this process.

Protocol 1: Spatial Modelling of Ecosystem Services (ES) Indicators

This protocol details the calculation of multi-temporal ES indicators, which form the quantitative foundation of the index [1].

• Objective: To quantify the supply of eight distinct ES indicators over time using a spatial modelling approach.
• Input Data: CORINE Land Cover (CLC) maps for the reference years 1990, 2000, 2006, 2012, and 2018 [1].
• ES Indicators Calculated: The eight modelled ES were Climate Regulation, Water Purification, Habitat Quality, Drought Regulation, Recreation, Food Provisioning, Erosion Prevention, and Pollination [1].
• Procedure:
  • Data Preparation: Obtain and preprocess CLC data for mainland Portugal for all five time periods.
  • Model Selection: Employ appropriate spatial models (e.g., similar to functionalities in InVEST software) to translate land cover classes into quantitative ES potential maps for each indicator and year [1].
  • Spatial Analysis: Execute the models in a Geographic Information System (GIS) environment to generate raster or vector maps depicting the spatial distribution and intensity of each ES.
  • Output: A set of eight ES indicator maps for each of the five reference years, providing a time-series dataset for analysis.

This protocol outlines the process for capturing stakeholders' perceptions of ES potential, which provides the qualitative data and weights for the index.

• Objective: To gather expert knowledge on the relative importance and potential supply of different ecosystem services.
• Stakeholder Recruitment: Engage a diverse group of stakeholders from various sectors of society relevant to land management and ecosystem services in Portugal [1].
• Procedure:
  • Survey Design: Develop a survey instrument based on a matrix methodology, where stakeholders evaluate the ES supply potential for different land cover types [1].
  • Analytical Hierarchy Process (AHP): Implement an AHP survey to pairwise compare the relative importance of the eight ES indicators. This structured technique helps derive robust priority weights while checking for consistency in respondent judgments [1].
  • Data Collection: Administer the surveys to the stakeholder group.
  • Weight Calculation: Process the AHP survey results to calculate the final weight for each ES indicator, reflecting its perceived relative importance in the composite index.

Protocol 3: Data Preprocessing and Index Construction

This protocol describes the steps to standardize the diverse data and aggregate them into the final ASEBIO index. The logical flow of data from raw inputs to the final index is shown below.

• Objective: To normalize the modelled ES indicators and combine them into a single, unitless index using stakeholder-derived weights.
• Preprocessing - Variable Standardization:
  • Rationale: Different ES indicators are measured on incompatible scales. Preprocessing brings them to a common measurement scale for appropriate combination [32].
  • Method Selection: Choose a scaling method. The Minimum-Maximum method (scaling between 0-1) is simple and preserves the distribution, but is sensitive to outliers. The Percentile method is robust to outliers and transforms data to a uniform distribution based on rank [32].
  • Directionality Check: Ensure the meaning of high and low values is consistent across all variables. For example, in a vulnerability index, a variable where a high value indicates less vulnerability may need to be reversed so that a high value consistently indicates more vulnerability. This is done by multiplying the values by -1 and rescaling [32].
• Combination - Variable Aggregation:
  • Method: Use a weighted Sum or Mean (additive methods). These methods are straightforward to interpret and allow high values in one variable to compensate for low values in another [32].
  • Calculation: The standardized and direction-corrected values for each ES indicator are multiplied by their respective AHP-derived weights and then summed for each spatial unit (e.g., pixel or polygon) to produce the final ASEBIO index value [1].

Temporal Changes in ES Indicators (1990-2018)

Table 1: Summary of modelled Ecosystem Services (ES) trends in mainland Portugal (1990-2018). "Potential" refers to the relative supply capacity of the service. [1]

Ecosystem Service Indicator	Overall Trend (1990-2018)	Key Regional Changes (NUTS-3)
Climate Regulation	Declined	Notable decline in Alentejo Central; improvement in Alto Minho.
Water Purification	Consistently High	Improved in 10 northern regions; declined in interior and south.
Habitat Quality	Mostly Stable	Increased in the north; declined in Lisbon and Alentejo Central.
Drought Regulation	Improved (Largest)	Significant improvement in central/south; declined in 8 regions.
Recreation	Improved	Improved in Algarve and interior; declined in coastal areas.
Food Provisioning	Mostly Stable	Decreased in Algarve; improved in many interior regions.
Erosion Prevention	Improved	Decreased in Cávado region.
Pollination	Mostly Stable	Remained mostly unchanged with declines in some contiguous regions.

Land Cover Contribution to ASEBIO Index

Table 2: Contribution of selected CORINE Land Cover (CLC) classes to the ASEBIO index (2018), demonstrating the influence of different landscape types. [1]

CLC Level 3 Class	CLC Code	Contribution to Index
Port areas	1.2.3	Least
Road and rail networks	1.2.2	Highest among artificial surfaces
Green urban areas	1.4.1	High among artificial surfaces
Rice fields	2.1.3	Low among agricultural areas
Agro-forestry areas	2.4.4	Substantial (greater than most forests)
Moors and heathland	3.2.2	Highest on average
Wetlands	4.1.1 / 5.1.1	Moderate and nearly equal to water bodies
Water bodies	5.1.2	Moderate and nearly equal to wetlands

The Scientist's Toolkit

Table 3: Essential research reagents and tools for developing a composite index like ASEBIO. [32] [1]

Item	Function / Purpose
Geographic Information System (GIS) Software	The primary platform for spatial data management, model execution, and map creation (e.g., ArcGIS Pro with the Calculate Composite Index tool) [32].
Land Cover/Land Use Maps	Foundational spatial data that serves as a key input for modelling the supply of various ecosystem services (e.g., CORINE Land Cover) [1].
Spatial ES Models (e.g., InVEST)	Software tools containing biophysical models to quantify and map the supply of specific ecosystem services based on input data like land cover [1].
Analytical Hierarchy Process (AHP)	A structured, multi-criteria decision-making technique used to derive consistent and robust weights from stakeholder surveys, reflecting the relative importance of each ES indicator [1].
Standardization & Preprocessing Algorithms	Mathematical procedures (e.g., Minimum-Maximum, Z-score, Percentile) to normalize variables to a common scale before combination, ensuring comparability [32].

Navigating Challenges: Validation, Uncertainty, and Reconciling Differences

Confronting the Validation Gap in Ecosystem Service Models

Ecosystem service (ES) models are crucial tools for translating ecological complexity into actionable insights for policymakers and land managers. However, a significant validation gap often exists, where model outputs are not adequately grounded in or validated against real-world empirical data and stakeholder perceptions [33]. This gap can undermine the credibility and utility of models in critical decision-making processes. This protocol provides a structured framework for confronting this validation gap by integrating robust biophysical modeling with rigorous socio-cultural assessment, ensuring that ES evaluations are both scientifically sound and socially relevant.

Application Notes: Integrated Validation Framework

An effective ES model validation strategy requires a dual-pronged approach that bridges quantitative biophysical simulations and qualitative socio-cultural evaluation. This integration ensures that models reflect not only ecological processes but also the benefits perceived by human populations, thereby closing the loop between model prediction and on-the-ground reality.

• Biophysical Model Calibration: Process-based models like the Soil and Water Assessment Tool (SWAT) provide a quantitative foundation for evaluating ES by simulating complex watershed dynamics, including hydrology, sediment transport, and nutrient cycling [33]. Successful application requires careful calibration and validation against observed data (e.g., streamflow, water quality metrics) to ensure model performance is acceptable before ES quantification is attempted.
• Socio-Cultural Validation: A socio-cultural assessment, framed within ethnoecology and post-normal science, identifies ES from the perspective of local communities, highlighting the relevance of Indigenous and Local Knowledge (ILK) [8]. This participatory process validates model outputs against community-held values and ensures that the ES being modeled are those that are actually perceived and valued by people.
• Iterative Co-Production of Knowledge: The validation process should not be linear but cyclical, involving constant data collection, systematization, and validation with local communities [8]. This iterative dialogue between researchers and stakeholders helps refine models and ensures their ongoing relevance.

Experimental Protocols

Protocol 1: Biophysical Quantification of Ecosystem Services

This protocol details the use of a process-based model to quantify provisional and regulatory ecosystem services, based on the methodology of Logsdon and Chaubey (2013) [33].

Workflow

The diagram below illustrates the sequential stages for the biophysical quantification of ecosystem services.

Key Procedures

• Step 1: Model Setup and Input

  • Procedure: Utilize the SWAT model to simulate watershed processes [33]. Required spatial data inputs include a Digital Elevation Model (DEM), soil type data, land use/land cover (LULC) maps, and long-term historical climate data (precipitation, temperature, solar radiation, wind speed, relative humidity).
  • Validation: The model must be calibrated and validated against observed streamflow and water quality data (e.g., nitrate, sediment) at the watershed outlet. Statistical parameters such as the Nash-Sutcliffe efficiency (NSE) and coefficient of determination (R²) should be used to evaluate model performance [33].

• Step 2: Ecosystem Service Quantification

  • Procedure: Develop mathematical indices to represent ES provisioning using SWAT outputs as inputs. For example:
    • Fresh Water Provisioning (FWP): Calculate an index that is a function of both water quantity (MF) and quality, the latter represented by a Water Quality Index (WQI_avg) [33]. The formula is structured to increase with greater water yield and better quality.
    • Erosion Regulation (ER): Quantify as the soil loss prevented by the landscape compared to a baseline scenario with no vegetation cover. This can be derived from the model's simulated sediment yield [33].
  • Validation: Run extreme land-use scenarios (e.g., all forest, all urban, all agriculture) to analyze the impact on ES provision and assess if the model outputs align with theoretical expectations [33].

The following table summarizes the ecosystem services quantified and the key model outputs used in their calculation, based on the work of Logsdon and Chaubey (2013) [33].

Table 1: Ecosystem Services and Corresponding SWAT Model Outputs for Quantification

Ecosystem Service	SWAT Model Outputs Used for Quantification	Key Components of Index
Fresh Water Provisioning (FWP)	Water yield, evapotranspiration	Water quantity, water quality (WQI)
Food Provisioning (FP)	Crop yield (e.g., corn grain)	Harvestable biomass
Fuel Provisioning (FuP)	Plant biomass production	Harvestable biomass for biofuel
Erosion Regulation (ER)	Sediment yield	Soil loss prevented
Flood Regulation (FR)	Water yield, surface runoff	Peak flow reduction

Protocol 2: Socio-Cultural Assessment of Ecosystem Services

This protocol outlines a participatory methodology for identifying and validating ES from the perspective of local communities, based on the work of later researchers in the Dry Chaco eco-region [8].

Workflow

The diagram below illustrates the cyclical, multi-stage process for the socio-cultural assessment of ecosystem services.

Key Procedures

• Step 1: Initial Engagement and Data Collection

  • Procedure: This stage employs ethnographic tools at individual and group levels [8].
    • Semi-Structured Interviews: Conduct conversations guided by key themes (e.g., way of life, productive activities, socio-environmental problems, perception of ecosystem changes). Practice evenly suspended attention and free association to access the interviewee's cultural universe [8].
    • Participatory Mapping: Co-produce maps with local actors to visualize their territory, spatialize ES use, and strengthen bonds between participants [8].
    • Participant Observation: Observe and note activities in households and peridomestic areas to understand the practical appropriation of natural resources.

• Step 2: Data Systematization and Validation

  • Procedure: Researchers systematize the collected qualitative data, identifying and categorizing mentioned ES and their perceived contributions to well-being.
  • Validation: Conduct community workshops to present, discuss, and validate the systematized findings. This is a crucial step for ensuring the accuracy and legitimacy of the interpreted data and for reaching working agreements on the identified ES [8].

The following table summarizes the core tools used in the socio-cultural assessment methodology and their primary functions.

Table 2: Core Methodological Tools for Socio-Cultural ES Assessment

Methodological Tool	Level of Application	Primary Function
Semi-Structured Interviews	Individual / Family	Elicit detailed narratives on lifestyle, activities, and perceived ES.
Participatory Mapping	Group / Community	Spatialize ES use and knowledge; foster collective exchange.
Participant Observation	Individual / Community	Contextualize interview data via direct observation of practices.
Community Workshops	Group / Inter-Community	Validate collected data and co-interpret findings with the community.

The Scientist's Toolkit: Research Reagent Solutions

The following table details essential materials and tools required for implementing the integrated validation framework described in this protocol.

Table 3: Essential Reagents and Tools for ES Model Validation

Item	Function/Application	Protocol
SWAT (Soil and Water Assessment Tool)	A process-based, semi-distributed hydrological model used to simulate water, sediment, nutrient, and pesticide transport in watersheds under varying land management practices.	Biophysical Quantification
GIS Software (e.g., ArcGIS, QGIS)	Platform for processing and managing spatial input data (DEM, soils, land use) and visualizing model output and ES spatial distribution.	Biophysical Quantification
Value Equivalent Factor Method	A unit value transfer method that assigns standardized economic values to different land cover types to estimate Ecosystem Service Value (ESV) dynamics.	Biophysical Quantification / ES Valuation
Semi-Structured Interview Guide	A flexible protocol with key themes to guide conversations with local community members, ensuring coverage of relevant topics while allowing for new insights to emerge.	Socio-Cultural Assessment
Participatory Mapping Materials	Physical maps of the study area and markers for participants to visually identify and locate areas of cultural significance, resource use, and ES provision.	Socio-Cultural Assessment

Confronting the validation gap in ES modeling is not merely a technical challenge but an epistemological one. It requires a commitment to methodological pluralism, where the quantitative precision of process-based models like SWAT is continuously informed and validated by the rich, qualitative insights derived from socio-cultural assessments embedded within local communities [33] [8]. The integrated framework and detailed protocols provided here offer a concrete path forward, empowering researchers to produce ES assessments that are not only robust and reproducible but also socially legitimate and decision-relevant.

Ecosystem services (ES) research is crucial for informing sustainable environmental management and policy. However, a significant capacity gap often undermines these efforts in data-poor regions, characterized by limited data availability, insufficient technical infrastructure, and a shortage of specialized expertise. This gap is particularly pronounced in research concerning regulating ecosystem services (RESs)—the benefits derived from the regulatory effects of biophysical processes, such as climate regulation, water purification, and erosion control [34]. In low- and middle-income countries (LMICs), barriers to health research capacity building include insufficient investment, limited coordination between researchers and policymakers, and environments that are not conducive to nurturing future researchers [35]. This application note provides detailed protocols to overcome these barriers, enabling robust ES assessment and comparison even in resource-limited contexts.

Quantifying the Challenge: Data and Expertise Deficits

The capacity gap is not merely anecdotal; it is documented through empirical research. A recent survey of stakeholders across 54 LMICs identified critical barriers, which are summarized in the table below alongside common manifestations in ES research [35].

Table 1: Common Capacity Gaps and Their Manifestations in Ecosystem Services Research

Capacity Barrier Category	Specific Challenges in Health Data Science [35]	Parallel Manifestations in ES Research
Training & Mentoring	Limited training resources; Lack of mentoring	Limited local expertise in spatial modeling (e.g., InVEST) and statistical analysis; lack of advanced ES valuation skills.
Data Infrastructure	Challenges with data quality, infrastructure, and privacy issues	Poor spatial data resolution (e.g., land cover maps); lack of long-term ecological monitoring data; restricted access to specialized software.
Research Environment	Absence of a conducive research environment; insufficient investment	Fragmented and project-based funding for ES studies; weak linkages between ES research and national policy planning.

Furthermore, a national-scale study in Portugal highlighted a critical consequence of this gap: a significant mismatch between ES potential modeled from spatial data and the potential perceived by local stakeholders. On average, stakeholder estimates were 32.8% higher than model outputs, with the largest contrasts in drought regulation and erosion prevention [1]. This disparity underscores the necessity of integrating local expert knowledge with scientific modeling to create a more complete and actionable evidence base for decision-making.

Application Notes and Protocols

The following protocols are designed to be implemented sequentially, building from system understanding to data collection and final analysis.

Protocol 1: Conceptual System Mapping for Guided Data Collection

This protocol provides a rapid, interdisciplinary tool to formalize understanding of a complex social-ecological system and identify the most critical drivers for targeted data collection, thereby optimizing limited resources [36].

Experimental Workflow:

• Expert Panel Formation: Assemble a multidisciplinary group of 8-12 local experts, including ecologists, social scientists, hydrologists, land-use planners, and representatives from local communities.
• Driver Identification and System Mapping: In a facilitated workshop, guide the experts through the following steps:
  • Define the spatial and temporal boundaries of the system and the specific ES to be assessed (e.g., water purification, carbon sequestration, dengue transmission).
  • Brainstorm and agree upon a comprehensive list of direct and indirect drivers (biotic, abiotic, socio-economic, and policy) influencing the selected ES. Record each driver on a separate card or digital node.
  • Establish and document the causal interactions between these drivers. For example, "agricultural expansion" (driver) leads to "deforestation" (driver), which affects "soil erosion" (driver) and "water quality" (ES).
• Network Analysis for Key Driver Identification: Encode the mapped system into a network where nodes are drivers and links are causal interactions. Use established network metrics to analyze the system:
  • Calculate centrality metrics (e.g., degree, betweenness centrality) for each driver to identify those with the most influence within the network.
  • Drivers with high centrality scores are deemed "key drivers" and should be prioritized in subsequent data-gathering efforts.
• Map Maintenance: Treat the conceptual system map as a living document. Revisit and revise the map regularly to incorporate new data and insights, ensuring it remains an accurate reflection of the current understanding [36].

The logical flow of this protocol is visualized in the diagram below.

Protocol 2: Rapid Ecosystem Services Assessment Using Multi-Criteria Evaluation

For a quantitative ES assessment where data is scarce, this protocol outlines a method to develop a composite index of ES potential, combining simplified spatial modeling with stakeholder valuation.

Experimental Workflow:

• Land Cover-Based ES Estimation:
  • Data Acquisition: Obtain a land cover map (e.g., CORINE Land Cover or a locally classified equivalent) for your study area.
  • Indicator Assignment: Assign a relative potential score (e.g., 0-1) for each ES indicator of interest to every land cover class. This can be derived from literature reviews, expert elicitation, or existing look-up tables [1].
• Stakeholder Weighting via Analytical Hierarchy Process (AHP):
  • Structure the Problem: Present the selected ES indicators to a group of stakeholders in a pairwise comparison matrix.
  • Elicit Judgments: Ask stakeholders to rate, for each pair of ES, which is more important for the management objective and by how much (using a standard 1-9 scale).
  • Calculate Weights: Process the pairwise comparison matrices to compute a normalized priority vector (weight) for each ES indicator, ensuring to check for consistency in stakeholder responses [1].
• Compute the Composite Index:
  • Create a weighted linear combination using the formula: ASEBIO_index = (ES1_Potential * ES1_Weight) + (ES2_Potential * ES2_Weight) + ... + (ESn_Potential * ESn_Weight)
  • This calculation is performed for each pixel in the land cover map, resulting in a spatially explicit index of combined ES potential [1].
• Comparison with Perceived Potential:
  • To calibrate your model and account for local knowledge, use a matrix-based methodology to directly capture stakeholders' perceptions of the ES potential for different land cover classes.
  • Quantify and spatially map the differences between the modeled ASEBIO index and the stakeholder-perceived potential to identify areas of consensus and disparity [1].

The following table details key methodological "reagents" and their functions for implementing the above protocols in data-poor contexts.

Table 2: Research Reagent Solutions for Data-Poor Ecosystem Services Research

Research Reagent	Function & Application	Notes for Data-Poor Contexts
Conceptual System Map	Formalizes understanding of the system; identifies key drivers for targeted data collection via network analysis [36].	Overcomes initial data scarcity by leveraging existing expert knowledge; prevents wasted resources on non-critical data.
CORINE Land Cover (or local equivalent)	Foundational spatial dataset used as a proxy for modeling multiple ecosystem services [1].	Widely available; provides a consistent baseline. Can be supplemented with higher-resolution local maps if available.
Analytical Hierarchy Process (AHP)	A multi-criteria decision-making method that derives the relative weights of different ES based on stakeholder perceptions [1].	Systematically incorporates local values and priorities, bridging the gap between technical models and human perspectives.
ASEBIO Index	A novel composite index that integrates multiple ES indicators into a single, comparable metric using stakeholder-defined weights [1].	Simplifies complex, multi-dimensional ES data into a format more accessible for policymakers and non-specialists.
Structured Expert Elicitation	A formal process for obtaining subjective judgments from experts in a structured, repeatable, and defensible manner.	Critical for parameterizing models (e.g., ES scores for land cover) when empirical data is missing or too costly to obtain.

Integrated Workflow for ES Research in Data-Poor Regions

The individual protocols and tools described above are integrated into a cohesive operational workflow below, illustrating the pathway from initial system scoping to final output for decision support.

The Power of Model Ensembles to Increase Accuracy and Fill Data Gaps

Within the domain of ecosystem services (ES) research, accurate predictive modeling is paramount for informing sustainable management policies and understanding the complex trade-offs and synergies between different ecological functions. These models, however, are often challenged by data gaps, nonlinear ecological relationships, and the need for high spatial explicitness [24]. Ensemble modeling has emerged as a powerful technique to overcome these hurdles. By combining multiple models into a single, unified solution, ensemble methods enhance predictive accuracy, improve robustness against overfitting, and increase the reliability of forecasts—a capability as critical in ecological forecasting as it is in healthcare and finance [37]. This document provides application notes and detailed experimental protocols for employing model ensembles to advance ES research, with a specific focus on integrating stakeholder perceptions.

Theoretical Foundation of Ensemble Models

An ensemble model in machine learning refers to the combination of predictions from multiple base models (often called "weak learners") to produce a single, more accurate, and robust prediction [37]. The core principle is that a group of models working together can compensate for individual biases and errors, leading to superior overall performance.

The most common ensemble techniques can be categorized as follows:

• Bagging (Bootstrap Aggregating): This method creates multiple subsets of the original data through bootstrapping (random sampling with replacement) and trains a separate model on each subset. The final prediction is an aggregation (e.g., average or majority vote) of all models' predictions. Random Forest is a prominent example that uses bagging with decision trees and is noted for reducing overfitting [37].
• Boosting: This technique involves training models sequentially, where each new model focuses on correcting the errors made by the previous ones. This creates a strong learner from a series of weakly performing models. AdaBoost and Gradient Boosting (including implementations like XGBoost and LightGBM) are key algorithms in this family [38] [37].
• Stacking (Stacked Generalization): Stacking combines multiple different types of predictive models (e.g., a support vector machine, a random forest, and a neural network) using a meta-model. The base models are trained on the full dataset, and their predictions are then used as input features to train the meta-model, which learns how best to combine them [38] [37].

Application in Ecosystem Services Research: A Case Study

Recent research on the Yunnan-Guizhou Plateau demonstrates the potent application of ensemble learning and related techniques for assessing and predicting ecosystem services [24].

Research Objective and Workflow

The study aimed to quantitatively evaluate key ecosystem services—water yield, carbon storage, habitat quality, and soil conservation—and project their changes under future land-use scenarios. The research integrated traditional assessment models with a machine learning framework to analyze driving factors and design future scenarios [24].

The workflow, summarized in the diagram below, involved data collection, historical assessment, driver identification via machine learning, future land-use simulation, and final ecosystem service projection.

Key Experimental Findings

The study provided critical insights into ES dynamics in the region [24]:

• Fluctuating Services: Ecosystem services on the Yunnan-Guizhou Plateau exhibited significant fluctuations between 2000 and 2020, driven by complex trade-offs and synergies.
• Primary Drivers: Land use and vegetation cover were identified as the most influential factors affecting overall ecosystem services, a relationship effectively captured by the machine learning model's ability to handle nonlinear patterns.
• Scenario Performance: The "ecological priority" scenario, designed using insights from the driver analysis, demonstrated the best performance across all evaluated ecosystem services by 2035.

Quantitative Data Synthesis

The following tables synthesize quantitative data and methodological details from relevant ensemble modeling studies to facilitate comparison and protocol design.

Table 1: Comparative Performance of Ensemble Models in an Educational Context (adapted from [38])

Model / Technique	Key Performance Metric (AUC)	Key Performance Metric (F1-Score)	Notes and Context
LightGBM (Boosting)	0.953	0.950	Best-performing base model for predicting student academic performance.
Stacking Ensemble	0.835	Not Specified	Did not offer a significant performance improvement over the best base model; showed considerable instability.
SMOTE Application	Consistency: 0.907	Not Specified	Technique applied for class balancing; resulted in strong fairness across gender, ethnicity, and socioeconomic status.

Table 2: Ensemble Model Applications Across Disciplines

Field / Application	Ensemble Technique	Purpose and Outcome
Building Energy Prediction [39]	Heterogeneous & Homogeneous Ensembles	Superior prediction accuracy compared to single models by reducing correlation between base models and minimizing overall prediction error.
Ecosystem Services Assessment [24]	Gradient Boosting (for driver analysis)	Accurately identified key drivers (land use, vegetation cover) of ecosystem services by capturing non-linear relationships in complex ecological data.
Disease Diagnostics [37]	Random Forest (Bagging)	Used for identifying diseases from patient data; reduces overfitting, increases accuracy, and makes diagnostic tools more reliable.

Detailed Experimental Protocols

Protocol 1: Implementing a Stacking Ensemble for Predictive Modeling

This protocol outlines the steps for creating a stacking ensemble, as applied in educational and other research contexts [38] [37].

• Problem Formulation and Data Preparation

  • Define Objective: Clearly specify the predictive target (e.g., probability of ecosystem service degradation, student at-risk status).
  • Data Collection and Integration: Gather and clean multimodal data. In ES research, this includes spatial data on land use, topography, climate, and soil properties [24].
  • Feature Engineering: Create relevant features from raw data. For LMS or geospatial data, this may involve creating summary metrics (e.g., course accesses, average slope).
  • Preprocessing: Handle missing values, encode categorical variables, and normalize or standardize numerical features. Crucially, apply data balancing techniques like SMOTE (Synthetic Minority Over-sampling Technique) at this stage if dealing with imbalanced classes (e.g., rare ecological states or at-risk student populations) [38].

• Base Model Selection and Training

  • Select Diverse Algorithms: Choose a diverse set of base learners. A strong combination includes:
    • A tree-based model (e.g., Random Forest - Bagging).
    • A boosting model (e.g., XGBoost or LightGBM - Boosting).
    • A linear model (e.g., Regularized Logistic Regression).
    • A distance-based model (e.g., Support Vector Machine).
  • Train with Cross-Validation: Use k-fold stratified cross-validation (e.g., 5-fold) on the training set to train each base model. This ensures robust performance estimation and prevents data leakage.

• Meta-Model Training

  • Generate Predictions: Use the cross-validated models to generate out-of-fold predictions on the training set. These predictions become the new input features for the meta-model.
  • Train Meta-Model: Train a typically simpler, interpretable model (e.g., Logistic Regression, Linear Model) on the new dataset of base model predictions. This model learns the optimal way to combine the base models' outputs.

• Model Evaluation and Interpretation

  • Performance Assessment: Evaluate the final stacked model on a held-out test set using relevant metrics (e.g., AUC, F1-Score, Mean Absolute Error).
  • Interpretability: Use explainability techniques like SHAP (SHapley Additive exPlanations) to understand the contribution of each base model and the original features to the final prediction, which is vital for stakeholder communication [38].

The logical flow of data and models in a stacking ensemble is visualized below.

Protocol 2: Integrating Ensemble Models with Stakeholder Perception Analysis

Integrating quantitative model outputs with qualitative social data is essential for comprehensive ecosystem service assessments [40]. The following protocol, drawing from the SolVES model framework, details this process.

• Data Collection and Preparation

  • Spatial and Environmental Data: Compile GIS data on environmental variables (e.g., elevation, slope, distance to water, land cover type).
  • Social Survey Data: Design and conduct surveys to collect georeferenced data on stakeholder perceptions. Use a virtual currency allocation approach where respondents assign a limited budget to different social value types (e.g., aesthetic, biodiversity, cultural) [40].

• Spatial Modeling of Social Values

  • Run SolVES Model: Input the survey and environmental data into the SolVES model. The model will:
    • Calculate a Value Index for each social value type at each survey point.
    • Use machine learning (typically MaxEnt) to derive relationships between the reported social values and the environmental variables.
    • Generate predictive maps of social value intensity across the entire landscape [40].

• Spatial and Statistical Analysis

  • Hotspot Analysis: Perform spatial cluster analysis (e.g., Getis-Ord Gi*) on the SolVES output maps to identify statistically significant hotspots and coldspots of each social value.
  • Correlation Analysis: Analyze spatial correlations between different social value types (e.g., between aesthetic and recreational values) and between social value maps and biophysical ecosystem service maps (e.g., from the InVEST model) [40].

• Multi-Scenario Prediction and Synthesis

  • Future Land-use Simulation: Use land-use change models (e.g., PLUS model) to project future land-use patterns under different scenarios (e.g., natural development, planning-oriented, ecological priority) [24].
  • Future ES and SV Projection: Use the InVEST model to project future biophysical ES under these land-use scenarios. Use the relationships learned by SolVES to project how social value distributions would change under these future landscapes.
  • Synthesis for Decision-Making: Compare the spatial overlap and trade-offs between future biophysical ES capacity and social value hotspots to inform spatial planning and optimize resource allocation [24] [40].

The Scientist's Toolkit: Research Reagent Solutions

Table 3: Essential Tools and Models for Ensemble-based Ecosystem Services Research

Tool / Model Name	Type / Category	Primary Function in Research
InVEST Model	Ecosystem Services Assessment Suite	Quantifies and maps multiple ecosystem services (e.g., carbon storage, water yield, habitat quality) based on land-use and biophysical data [24].
SolVES Model	Social Value Assessment Tool	Spatially maps perceived social values of ecosystem services (aesthetic, cultural, etc.) by integrating survey data with environmental variables [40].
PLUS Model	Land Use Simulation Model	Projects future land-use changes by simulating the interactions between human activities and natural systems under various scenarios [24].
XGBoost / LightGBM	Machine Learning Algorithm (Boosting)	High-performance algorithms for regression and classification tasks; effective for identifying key drivers and making predictions from complex datasets [38] [24].
SHAP	Explainable AI Library	Provides interpretability for complex models by quantifying the contribution of each input feature to a single prediction, crucial for stakeholder trust [38].
SMOTE	Data Pre-processing Technique	Addresses class imbalance in datasets by generating synthetic samples of the minority class, improving model fairness and performance [38].

Ecosystem services are defined as the benefits that people obtain from natural ecosystems [41]. These services are fundamental to human survival and economic activity, yet they often create management conflicts due to competing stakeholder interests. The central challenge lies in reconciling the immediate, market-driven demand for tangible provisioning services with the long-term, often public benefits provided by regulating services [42] [41] [43].

Provisioning services include harvestable goods such as food, timber, fiber, and freshwater [42] [41]. These services generate direct economic value and are easily quantifiable in market terms, making them a primary focus for many economic stakeholders. In contrast, regulating services include processes like climate regulation, water purification, flood control, and pollination [41] [43]. These services provide critical life-support functions but their value is often not captured by markets, leading to their underrepresentation in decision-making processes.

This divergence creates fundamental tensions in environmental management, where the overharvesting of provisioning services can degrade the regulatory functions that sustain ecosystem health and, ultimately, the continued supply of both service types.

Quantitative Comparison of Ecosystem Services

The following tables provide a systematic, quantitative comparison of key provisioning and regulating services, detailing their outputs, measurement units, and economic valuation contexts. This structured overview facilitates direct comparison and highlights the distinct characteristics of each service category.

Table 1: Quantitative Profile of Key Provisioning Services

Service Category	Specific Output Examples	Typical Quantitative Measures	Economic Valuation Context
Food Production	Fish, crops, livestock [41]	Yield (tons/ha), Maximum Sustainable Yield (MSY) [43]	Market price, export revenue (e.g., ~$129B from fisheries exports) [42]
Raw Materials	Timber, wood fuel, fiber [41]	Stock (m³/ha), sustainable harvest volume [43]	Market value for construction, paper, and energy [41]
Water Resources	Freshwater for drinking/irrigation [41]	Volume extracted (m³), aquifer recharge rates [43]	Cost of alternative supply (e.g., desalination), agricultural output value [43]
Biochemicals	Natural medicines, cosmetics [41]	Quantity discovered/extracted, bioassay activity	R&D investment saved, potential pharmaceutical revenue

Table 2: Quantitative Profile of Key Regulating Services

Service Category	Specific Function	Typical Quantitative Measures	Economic Valuation Context
Climate Regulation	Carbon sequestration and storage [41]	t CO₂ stored/ha/year, carbon stocks [43]	Social cost of carbon, carbon market price [43]
Water Purification	Pollutant filtration and nutrient regulation [41]	kg of pollutants (N, P) removed/ha/year [43]	Cost savings from human-built water treatment plants [43]
Erosion Regulation	Soil retention and landslide prevention [41]	Tons of soil retained/ha/year	Cost of dredging, lost agricultural productivity, fertilizer replacement cost
Pollination	Support for crop reproduction [41]	Crop yield increase (%) attributable to wild pollinators, pollinator abundance	Value of agricultural output dependent on pollination [43]
Air Quality Regulation	Capturing/filtering of dust and chemicals [41]	Particulate Matter (PM2.5/PM10) removal per unit area	Public health cost avoidance (e.g., reduced asthma cases)

Experimental Protocols for Assessing and Mapping Services

To effectively reconcile provisioning and regulating services, researchers require robust, standardized methodologies for their assessment. The following protocols provide detailed procedures for quantifying, mapping, and analyzing these services.

Protocol 1: Biophysical Assessment of Regulating Services

Objective: To quantify the biophysical supply of key regulating services (e.g., water purification, carbon storage) within a defined study area using direct measurement, indirect sensing, and modeling approaches [41].

Materials:

• GPS Unit
• Water Quality Testing Kits (for nitrates, phosphates, turbidity)
• Soil Cores
• Vegetation Survey Equipment
• Remote Sensing Imagery (e.g., satellite, aerial photos)
• GIS Software (e.g., ArcGIS, QGIS)
• Statistical Analysis Software (e.g., R, Python)

Procedure:

• Site Delineation: Define the ecosystem boundary (e.g., wetland, forest) using a combination of field verification and remote sensing data [41].
• Direct Measurement:
  • For Water Purification: Collect water samples at inflow and outflow points of the ecosystem. Analyze for key pollutants (e.g., sediments, nitrates, phosphates) using standardized water quality kits. Calculate the removal efficiency as (Input Concentration - Output Concentration) × Water Volume over a specified time period [43].
  • For Carbon Storage: Collect soil cores and vegetation (above-ground and below-ground) samples from randomly located plots within the ecosystem. Analyze samples in a lab for organic carbon content. Calculate total carbon stocks by summing soil and biomass carbon per unit area [43].
• Indirect Measurement & Modeling:
  • Utilize spatial proxy models where direct measurement is impractical. For example, model flood protection capacity based on land cover classification (e.g., from satellite imagery) and known water absorption coefficients for different vegetation types [41].
  • Apply process-based models that simulate ecosystem functions, such as nutrient cycling or evapotranspiration, using input data on climate, soil, and vegetation [41].
• Data Integration: Compile all direct and modeled data into a GIS. Create maps illustrating the spatial distribution and intensity of the regulating service across the study area.

Protocol 2: Stakeholder Perception Analysis Using the 3i Framework

Objective: To systematically identify and analyze the perceptions of different stakeholder groups regarding their interest in, influence on, and perceived impact from specific ecosystem services and wildlife species [44].

Materials:

• Pre-defined list of stakeholder groups (e.g., crop farmers, livestock farmers, tourism operators, protected area managers) [44]
• Pre-defined list of relevant ecosystem services or wildlife species categories [44]
• Survey questionnaires or interview guides
• Data analysis software (e.g., SPSS, R)

Procedure:

• Stakeholder and Species/Service Selection: Identify all major stakeholder groups involved in or affected by the management of the ecosystem. Select the provisioning services (e.g., fish, timber) and regulating services (e.g., pest control, water purification) or wildlife species (e.g., wolves, pollinators) that are central to the management conflict [44].
• Data Collection - Expert Elicitation: First, consult with local experts to get an initial assessment of the 3i (Interest, Influence, Impact) for each stakeholder-service combination [44].
• Data Collection - Stakeholder Surveys: Conduct structured surveys or interviews with a representative sample of individuals from each stakeholder group (e.g., n=251 across 7 groups) [44]. For each service/species, ask respondents to rate on a scale (e.g., 1-5):
  • Interest: How important is this service/species to your livelihood/well-being?
  • Influence: How much control or influence do you have over the management of this service/species?
  • Impact: How strongly are you impacted (positively or negatively) by the current state of this service/species?
• Data Analysis:
  • Analyze data for within-group and between-group heterogeneity. Look for significant differences in perceptions across groups and sites [44].
  • Compare expert perceptions with the actual perceptions of the stakeholder groups. Note where experts underestimate or misjudge stakeholder views [44].
  • Identify key conflict points, such as when the perceived impact of a service/species (e.g., wolves damaging livestock) far exceeds the stakeholder's perceived influence over its management, indicating a sense of powerlessness [44].
• Interpretation: Use the results to map conflict landscapes and identify stakeholder groups that may require targeted engagement strategies or compensation mechanisms to facilitate coexistence and sustainable management.

Integrated Assessment and Decision-Making Workflows

Effective reconciliation of ecosystem services requires a structured process that integrates biophysical data with socio-economic analysis. The following diagrams, generated using Graphviz, illustrate key workflows for this integration.

Integrated ecosystem services management decision pathway

Stakeholder-service perception mapping for conflict identification

The Scientist's Toolkit: Essential Research Reagents and Materials

This section details the key tools, datasets, and analytical resources required for conducting comprehensive ecosystem services research as outlined in the preceding protocols.

Table 3: Essential Research Tools for Ecosystem Services Assessment

Tool/Category	Specific Examples & Specifications	Primary Function in Research
Field Measurement Equipment	Water quality testing kits (Nitrate, Phosphate, pH), Soil corers, Dendrometers, GPS units	Direct biophysical data collection for quantifying service supply (Protocol 1) [41]
Remote Sensing Data	Satellite imagery (Landsat, Sentinel-2), Aerial photography, LIDAR data	Large-scale, indirect measurement of ecosystem extent and condition (e.g., forest cover, wetland area) [41]
GIS Software	ArcGIS, QGIS (Open Source), GRASS GIS	Spatial analysis, data integration, and mapping of ecosystem service supply and demand [41]
Statistical & Modeling Software	R (with packages like vegan, lme4), Python (with pandas, scikit-learn), IN-VEST model	Statistical analysis of stakeholder data (Protocol 2), predictive modeling of service flows under different scenarios [41] [44]
Stakeholder Engagement Tools	Structured survey questionnaires, Semi-structured interview guides, participatory mapping exercises	Eliciting and quantifying stakeholder perceptions of the 3i factors (Interest, Influence, Impact) [44]
Classification Frameworks	CICES (Common International Classification of Ecosystem Services), TEEB frameworks	Standardizing definitions and categories of ecosystem services for consistent reporting and comparison across studies [41]

Evidence and Outcomes: Empirical Comparisons and Real-World Implications

Ecosystem services (ES), the benefits humans derive from ecosystems, form the foundation for sustainable development and human well-being [45] [46]. Accurate assessment of these services is imperative for effective environmental management and policy-making, particularly under escalating anthropogenic pressures [1]. The comparative analysis between model-based quantifications and stakeholder perceptions of ES has emerged as a critical research frontier, revealing significant discrepancies that challenge integrated assessment frameworks. This application note synthesizes statistical evidence from national-scale comparisons, documenting a pervasive and quantifiable mismatch between scientific models and human valuation of ecosystem benefits. We present standardized protocols for quantifying these disparities, enabling researchers to systematically evaluate and bridge the gap between data-driven assessments and stakeholder perspectives in environmental decision-making.

Conceptual Framework of Ecosystem Service Mismatches

Defining Mismatch Dimensions

Ecosystem service mismatches manifest across three primary dimensions, each requiring distinct methodological approaches for quantification [47]:

• Spatial Mismatches: Occur when the geographical distribution of ES supply does not align with the spatial pattern of demand or perception
• Temporal Mismatches: Arise from discrepancies in the timing of ES provision, assessment, or valuation
• Functional-Conceptual Mismatches: Emerge from fundamental differences in how ecosystem functions are perceived, defined, or valued by different groups

Table 1: Classification of Ecosystem Service Mismatch Types

Mismatch Dimension	Definition	Measurement Approach
Spatial	Geographical disconnection between supply and demand/perception	GIS spatial analysis, hotspot mapping
Temporal	Timing discrepancies in provision, assessment, or valuation	Time-series analysis, trend comparison
Functional-Conceptual	Differences in perception, definition, or valuation of ES	Stakeholder surveys, expert elicitation, AHP

Theoretical Foundations

The conceptual framework for understanding ES mismatches integrates social-ecological systems theory with perception studies [46] [47]. This framework acknowledges that both biophysical reality and human perception constitute valid dimensions of ecosystem service assessment, with mismatches arising from complex interactions between ecological systems and social valuation processes. The framework further distinguishes between:

• Ecological mismatches: Discrepancies between ecological components or processes
• Social mismatches: Disconnects between social components, knowledge systems, or institutions
• Social-ecological mismatches: Fundamental disalignments between ecological capacity and social demand or perception

Diagram 1: Conceptual framework for ES mismatch analysis (82 characters)

National-Scale Case Study: Portugal

Study Design and Temporal Framework

The Portugal national-scale assessment provides a comprehensive template for quantifying model-perception mismatches [1]. This research employed a spatiotemporal approach analyzing eight key ecosystem services across a 28-year period (1990-2018) using five reference years (1990, 2000, 2006, 2012, 2018). The study integrated:

• Spatial modelling of ES indicators using CORINE Land Cover data
• Multi-temporal analysis of land use changes and ES trade-offs
• Stakeholder engagement through Analytical Hierarchy Process (AHP) weighting
• Integrated assessment via the novel ASEBIO index (Assessment of Ecosystem Services and Biodiversity)

Quantitative Mismatch Results

The comparative analysis revealed statistically significant disparities between model outputs and stakeholder perceptions across all ecosystem services assessed [1]:

Table 2: Quantified Mismatches Between Models and Stakeholder Perceptions in Portugal

Ecosystem Service	Stakeholder Overestimation (%)	Mismatch Severity	Key Findings
Drought Regulation	Highest contrast	Extreme	Largest perception-model gap
Erosion Prevention	Highest contrast	Extreme	Significant stakeholder overestimation
Water Purification	Lower overestimation	Moderate	Closest alignment between approaches
Food Production	Lower overestimation	Moderate	Relatively close alignment
Recreation	Lower overestimation	Moderate	Similar valuation patterns
All Selected ES	32.8% (average)	Substantial	Consistent stakeholder overestimation trend

The comprehensive assessment demonstrated that stakeholder estimates exceeded model-based calculations by an average of 32.8% across all ecosystem services, with drought regulation and erosion prevention showing the most pronounced disparities [1].

Methodological Protocols

Ecosystem Services Modelling Workflow

Diagram 2: ES modelling workflow (25 characters)

Data Collection and Preprocessing

• Land Cover Data: Utilize CORINE Land Cover or equivalent datasets with standardized classification (minimum 30m resolution recommended)
• Temporal Framework: Establish at least 3-5 reference years across the study period to capture trends
• Spatial Unit Definition: Implement grid-based analysis (1km² recommended for regional/national assessments) [45]
• Normalization Procedures: Apply appropriate statistical normalization to enable cross-comparison

ES Indicator Calculation Protocol

For each ecosystem service, employ established modelling approaches:

• Habitat Quality: Use InVEST Habitat Quality model or equivalent
• Carbon Sequestration: Estimate via net primary productivity data and land cover coefficients
• Water Purification: Apply nutrient retention models based on land cover and topographic parameters
• Erosion Prevention: Utilize RUSLE-based models with land cover and soil parameters
• Recreation Potential: Implement accessibility and natural attractiveness indices

Stakeholder Perception Assessment

Sampling and Recruitment

• Stakeholder Stratification: Include representatives from key sectors (government, academia, NGOs, local communities, industry)
• Sample Size: Target minimum 50 participants for national-scale assessments
• Geographic Representation: Ensure proportional regional coverage within study area

Analytical Hierarchy Process (AHP) Implementation

The AHP provides a structured framework for eliciting and quantifying stakeholder preferences [1]:

• Hierarchy Construction: Decompose decision problem into ES objectives and criteria
• Pairwise Comparisons: Present stakeholders with systematic comparisons of all ES pairs
• Consistency Validation: Calculate consistency ratios (<0.1 acceptable) to ensure logical judgment
• Weight Aggregation: Synthesize individual judgments into group priorities

Mismatch Quantification Protocol

Statistical Comparison Methods

• Percentage Difference Calculation: Mismatch (%) = [(Stakeholder valuation - Model valuation) / Model valuation] × 100
• Spatial Concordance Analysis: Hotspot mapping and spatial autocorrelation (Local Moran's I) [45]
• Temporal Trend Analysis: Regression analysis of parallel time-series data
• Correlation Assessment: Pearson/Spearman correlation coefficients between model outputs and perception scores

ASEBIO Index Computation

The ASEBIO index provides an integrated measure of ES potential [1]:

• Normalization: Transform all ES indicators to common scale (0-1)
• Weight Application: Apply AHP-derived stakeholder weights to each ES indicator
• Aggregation: Compute weighted sum across all ES indicators
• Validation: Compare ASEBIO results with independent perception assessments

Research Reagent Solutions

Table 3: Essential Research Tools for ES Mismatch Studies

Research Tool	Function	Application Context
CORINE Land Cover Data	Provides standardized land cover classification	Base spatial data for ES modelling
InVEST Software Suite	Integrated ecosystem service modelling	Biophysical ES quantification
AHP Survey Instruments	Structured stakeholder preference elicitation	Weight assignment for ES importance
GIS Platforms (QGIS, ArcGIS)	Spatial analysis and mapping	Spatial mismatch identification
WhiteStripe Normalization	Intensity normalization for spatial data	Preprocessing of heterogeneous data sources [48]
Geographically Weighted Regression	Spatial regression analysis	Quantification of local relationships [48]

Discussion and Implementation Guidelines

Interpretation of Statistical Evidence

The consistent overestimation of ES potential by stakeholders (32.8% average in Portugal) underscores fundamental differences in valuation frameworks between scientific and perceptual approaches [1]. This mismatch pattern reveals several critical insights:

• Cognitive Biases: Stakeholders may overweight visually prominent or culturally valued services
• Scale Disconnects: Localized stakeholder experiences may not align with regional-scale model outputs
• Knowledge Gaps: Technical understanding of ecosystem processes varies between experts and laypersons
• Value Incorporation: Models may inadequately capture cultural, spiritual, or non-material values

Methodological Recommendations

Based on the empirical evidence from national-scale assessments, we recommend:

• Integrated Assessment Frameworks: Combine biophysical modelling with structured stakeholder engagement in iterative cycles
• Explicit Uncertainty Quantification: Document and communicate uncertainties in both model outputs and perception data
• Contextual Interpretation: Consider socio-economic, cultural, and institutional factors when interpreting mismatch patterns
• Policy Integration: Develop decision-support tools that explicitly acknowledge and address identified mismatches

The protocols presented herein provide a standardized methodology for quantifying and analyzing ecosystem service mismatches, enabling more transparent and replicable research in this emerging field. By adopting these comprehensive assessment frameworks, researchers and practitioners can better understand the complex interplay between ecological reality and human perception, ultimately supporting more effective and inclusive ecosystem governance.

The alignment between computational models and human perception represents a critical frontier in artificial intelligence and environmental science. This alignment ensures that models act towards human-intended goals, preferences, and ethical principles, serving as a proxy measure for AI safety and reliability [49]. Within the specific context of ecosystem services (ES) research, a significant gap exists between data-driven model assessments and human perspectives, creating potential challenges for effective environmental decision-making [1]. This protocol systematically addresses perception-model alignment through standardized assessment frameworks, experimental methodologies, and validation techniques that bridge technological and ecological applications.

Quantitative Data Synthesis

Comparative Analysis of Model-Human Alignment in Ecosystem Services

Table 1: Disparities between model-generated and stakeholder-perceived ecosystem service potential in mainland Portugal (2018) [1]

Ecosystem Service Indicator	Stakeholder Overestimation (%)	Alignment Level
Drought Regulation	Highest contrast	Low
Erosion Prevention	Highest contrast	Low
Water Purification	32.8% average across all services	Medium-High
Food Production	32.8% average across all services	Medium-High
Recreation	32.8% average across all services	Medium-High
Climate Regulation	Not specified	Low
Habitat Quality	Not specified	Medium
Pollination	Mostly unchanged	Medium

Vision-Language Model Merging for Perception-Reasoning Alignment

Table 2: Layer-wise contribution analysis after cross-modal model merging [50]

Model Layer Group	Primary Function Before Merging	Contribution After Merging
Early Layers	Visual perception encoding	Perception unchanged; begins to contribute to reasoning
Middle Layers	Transition processing	Significant increase in reasoning contribution
Late Layers	Reasoning facilitation	Enhanced reasoning capabilities
All Layers Combined	Separate functionalities	Unified perception-reasoning integration

Experimental Protocols

Protocol 1: Ecosystem Services Perception-Model Alignment Assessment

Purpose and Scope

This protocol details a methodology for quantifying and comparing ecosystem service potential between spatial models and stakeholder perceptions, enabling the identification of alignment disparities for improved environmental decision-making [1].

Materials and Equipment

• Geographic Information System (GIS) software with spatial analysis capabilities
• Land cover cartography data (CORINE Land Cover recommended)
• Multi-criteria evaluation framework
• Stakeholder survey instruments (digital or physical)
• Statistical analysis software (R, Python, or equivalent)

Procedure

Step 1: Temporal ES Indicator Modeling

• Select multiple reference years (e.g., 1990, 2000, 2006, 2012, 2018) for longitudinal analysis
• Calculate eight distinct ES indicators using spatial modeling approaches:
  • Climate regulation
  • Water purification
  • Habitat quality
  • Drought regulation
  • Recreation
  • Food provisioning
  • Erosion prevention
  • Pollination
• Generate maps illustrating changes to ES potential across administrative regions

Step 2: Composite Index Development

• Develop an Assessment of Ecosystem Services and Biodiversity (ASEBIO) index
• Integrate ES data models with stakeholder-defined weights
• Apply Analytical Hierarchy Process (AHP) for weight assignment
• Calculate index values across the temporal range

Step 3: Stakeholder Perception Assessment

• Recruit stakeholders from various sectors of society
• Administer perception surveys using matrix-based methodology
• Collect valuations of ES potential for the target year (2018)
• Aggregate and normalize perception data

Step 4: Alignment Quantification

• Conduct statistical comparison between ASEBIO index and stakeholder perceptions
• Calculate percentage overestimation/underestimation for each ES indicator
• Identify services with highest alignment disparities
• Generate spatial distribution maps of differences

Data Analysis

• Perform ANOVA to test significant differences in ES distribution levels across years
• Calculate mean values and interquartile ranges for ES indicators
• Compute contrast ratios between model outputs and perception values
• Conduct regional comparative analysis of alignment patterns

Protocol 2: Cross-Modal Model Merging for Perception-Reasoning Alignment

Purpose and Scope

This protocol describes a training-free method for transferring reasoning capabilities from Large Language Models (LLMs) to Vision-Language Models (VLMs) through parameter merging, enhancing visual reasoning performance while maintaining interpretability [50].

Materials and Equipment

• Pre-trained Vision-Language Models (VLMs)
• Pre-trained Large Language Models (LLMs)
• Computational resources for model inference and analysis
• Visual reasoning benchmark datasets
• Interpretability analysis tools (e.g., feature visualization, attribution methods)

Procedure

Step 1: Model Selection and Preparation

• Select base VLM with strong perceptual capabilities but limited reasoning
• Identify LLM with advanced reasoning capacities
• Ensure architectural compatibility for parameter alignment
• Initialize models in evaluation mode

Step 2: Cross-Modal Parameter Merging

• Implement model merging techniques across modalities
• Connect parameters between visual and linguistic components
• Apply layer-wise merging strategies:
  • Early layers: preserve visual perception encoding
  • Middle layers: enable transition capabilities
  • Late layers: enhance reasoning facilitation
• Validate architectural integrity post-merging

Step 3: Ability Transfer Validation

• Administer visual reasoning puzzles to baseline VLM and merged model
• Assess performance on:
  • Chart interpretation
  • Mathematical problem solving from images
  • Complex visual question answering
• Quantify accuracy improvements while monitoring perceptual capability retention

Step 4: Interpretability Analysis

• Analyze internal mechanisms of perception and reasoning
• Map perception capabilities across model layers
• Identify reasoning facilitation patterns
• Document how merging affects functional distribution

Data Analysis

• Compare pre- and post-merging performance on benchmark tasks
• Conduct layer-wise contribution analysis
• Calculate statistical significance of improvement metrics
• Generate visualizations of functional distribution changes

Visualization of Methodological Frameworks

Ecosystem Services Perception-Model Alignment Workflow

Three-Layer Framework for Embodied Intelligence Alignment

Cross-Modal Model Merging for Enhanced Reasoning

The Scientist's Toolkit: Research Reagent Solutions

Table 3: Essential materials and computational tools for perception-model alignment research

Research Reagent/Tool	Function	Application Context
CORINE Land Cover Data	Provides standardized land use/cover classification for spatial analysis	Ecosystem services modeling and mapping [1]
InVEST Software	Spatial modeling tool for estimating ecosystem services and tradeoffs	Integrated valuation of multiple ES indicators [1]
Analytical Hierarchy Process (AHP)	Multi-criteria decision-making method with stakeholder-defined weights	ASEBIO index development and priority weighting [1]
Multimodal Large Models (MLMs)	Cross-modal Transformer architectures for unified representation learning	Embodied intelligence systems and visual-linguistic alignment [51]
World Models (WMs)	Internal environment simulation for state prediction and causal reasoning	Adaptive decision-making in embodied systems [51]
Target Confusability Competition (TCC)	Model combining probabilistic memory with psychophysical similarity	Formal unification of individual item and ensemble memory [52]
VisAlign Dataset	Benchmark for measuring AI-human visual alignment in image classification	Validation of perception alignment in visual AI systems [49]
Perceptual Summation Model	Framework where ensemble representations reflect global sum of activations	Computational modeling of ensemble perception [52]
Feature-Conditioned Modal Alignment (F-CMA)	Mechanism for task-semantics-guided perceptual fusion	Enhanced cross-modal information alignment in embodied systems [51]

Application Notes

Ecosystem services (ES) assessments are vital for sustainable ecosystem management, yet a significant challenge persists in reconciling the divergent perspectives of scientific models and stakeholder communities. These differences can impact the effectiveness of environmental policies and conservation strategies. The following notes outline the core findings and methodological approaches for understanding and bridging this gap.

Key Evidence of Divergent Priorities

Quantitative data from a comprehensive 2024 national-scale study in Portugal reveals a consistent pattern where stakeholder perceptions of ecosystem service potential significantly exceed model-based valuations. The table below summarizes the average divergence for key ecosystem services [1].

Table 1: Comparative Valuation of Ecosystem Service Potential: Models vs. Stakeholders [1]

Ecosystem Service	Level of Contrast (Avg. Stakeholder Overestimation)	Key Contextual Factors Influencing Divergence
Drought Regulation	Highest Contrast	Technical complexity of hydrological models; abstract nature of the service.
Erosion Prevention	Highest Contrast	Difficulty in observing long-term, cumulative benefits; reliance on visible landscape features.
Climate Regulation	High Contrast	Global/abstract nature of the service versus local/tangible stakeholder concerns.
Habitat Quality	High Contrast	Differing definitions of "habitat quality" (biodiversity metrics vs. general "greenness").
Pollination	Moderate Contrast	Dependence on specialized ecological knowledge versus general appreciation of pollinators.
Food Production	Low Contrast	Direct market valuation and tangible outputs make the service more easily understood.
Recreation	Low Contrast	Direct user experience and cultural valuation align closely with modeled accessibility.
Water Purification	Low Contrast	High visibility in policy and media can align scientific and public understanding.
Overall Average	32.8% Higher (Stakeholder Valuation)	Mismatch between spatially-explicit data and aggregated human valuation.

The core divergence of 32.8% in the Portuguese study underscores a fundamental mismatch; stakeholders tended to assign a higher overall importance and potential supply to ES bundles compared to data-driven models [1]. This suggests that while models account for biophysical constraints and land cover efficiencies, community perceptions are shaped by cultural values, direct experiences, and the visibility of services.

Furthermore, divergent priorities are not limited to ES potential. A European study on wildlife interactions found that expert perceptions often underestimated stakeholder interest and occasionally misjudged their sense of influence and impact from species like brown bears, wolves, and eagles. This highlights a broader pattern where expert assessments can fail to capture the nuanced, contextualized realities of local stakeholders [44].

Conceptual Framework for Divergence

The workflow below illustrates the parallel processes of scientific modeling and stakeholder perception assessment, highlighting points where methodological differences can lead to divergent outcomes.

Experimental Protocols

To systematically investigate the divergence between community and expert valuations of ES bundles, researchers can employ the following detailed protocols. These methodologies are designed for parallel implementation to facilitate direct comparison.

Protocol 1: Model-Based ES Assessment with the ASEBIO Index

This protocol outlines a quantitative, spatial modeling approach to establish a baseline of biophysical ES potential [1].

1.1 Data Collection and Preparation

• Primary Data Inputs: Secure CORINE Land Cover (CLC) maps or equivalent local LULC data for the study area for multiple time points to assess temporal trends [1].
• Supplementary Geospatial Data: Collect national or regional data on topography, soil type, climate (precipitation, temperature), and administrative boundaries.
• Data Standardization: Pre-process all spatial data to a consistent coordinate system and raster resolution (e.g., 1 ha grid cells) to ensure compatibility.

1.2 Calculation of Individual ES Indicators

• Service Selection: Define a suite of ES indicators relevant to the study context (e.g., climate regulation, water purification, habitat quality, drought regulation, recreation, food production, erosion prevention, pollination) [1].
• Spatial Modeling: Utilize established modeling frameworks (e.g., the InVEST suite of models) or develop simplified spatial models based on literature-derived look-up tables linking LULC classes to ES capacity [1].
• Temporal Analysis: Run models for each CLC time-slice (e.g., 1990, 2000, 2018) to generate maps and time series data for each ES, revealing trade-offs and synergies over time [1].

1.3 Integration into a Composite Index (ASEBIO Index)

• Multi-Criteria Evaluation (MCE): Implement an Analytical Hierarchy Process (AHP) survey with a panel of technical experts to derive weights reflecting the relative importance of each ES for the composite index [1].
• Index Calculation: Use a weighted linear combination in a GIS environment: ASEBIO_index = (w1 * ES1) + (w2 * ES2) + ... + (wn * ESn), where w is the AHP-derived weight and ES is the standardized value of each ecosystem service [1].
• Validation: Perform sensitivity analysis on the AHP weights and validate model outputs against field-measured data where available.

Protocol 2: Stakeholder-Based Perception Assessment

This protocol details methods for capturing community and stakeholder perceptions of ES potential, enabling a direct comparison with model outputs [1] [44].

2.1 Stakeholder Mapping and Recruitment

• Identification: Categorize key stakeholder groups (e.g., crop farmers, livestock farmers, foresters, hunters, tourism operators, protected area managers, local residents, environmental NGOs) to ensure a representative sample of perspectives [44].
• Recruitment: Use purposive sampling to recruit participants from each group. Aim for a sample size sufficient for robust qualitative and quantitative analysis (e.g., n > 30 per group where possible) [53].

2.2 Data Collection Techniques

• Structured/Semi-Structured Interviews: Conduct in-depth and key informant interviews using pre-tested guides. Explore perceptions of ES potential, their relative importance, and observed changes over time. Audio-record and transcribe interviews verbatim for analysis [53].
• Perception Matrix Surveys: Implement a matrix-based methodology. Present stakeholders with a matrix where rows are land cover types and columns are ES. Ask them to score the potential of each LULC to provide each ES (e.g., on a scale of 0-10) [1].
• Analytical Hierarchy Process (AHP) with Stakeholders: Administer an AHP survey to a different subgroup of stakeholders (e.g., community representatives) to derive a separate set of weights for the same ES used in the ASEBIO index [1].
• 3i Method (Interest, Influence, Impact): For studies involving wildlife-related ES, employ the modified 3i method. Have stakeholders rate their interest in, influence on, and how they are impacted by specific wildlife species or ES categories [44].

2.3 Data Analysis

• Qualitative Analysis: Analyze interview transcripts using thematic analysis with inductive coding in software like NVivo. Identify emergent themes related to ES valuation, knowledge, and perceived threats [53].
• Quantitative Analysis: Calculate mean perception scores from the matrix surveys. Compute descriptive statistics and use non-parametric tests (e.g., Mann-Whitney U) to compare perceived ES potential across different stakeholder groups. Compare stakeholder-derived AHP weights with expert-derived weights.

Protocol 3: Integrative Analysis and Divergence Assessment

This protocol describes how to synthesize data from Protocol 1 and 2 to quantitatively and qualitatively assess divergent priorities.

3.1 Quantitative Comparison

• Standardization: Re-scale both modeled ES values (from Protocol 1) and mean perceived ES values (from Protocol 2, matrix survey) to a common scale (e.g., 0-1).
• Divergence Metric: Calculate the percentage difference for each ES: ((Stakeholder_Value - Model_Value) / Model_Value) * 100. Aggregate to find an average divergence across all ES [1].
• Spatial Concordance Analysis: Use spatial statistics (e.g., correlation analysis, difference maps) to identify geographic areas where model and perception valuations show the highest and lowest agreement.

3.2 Qualitative Synthesis

• Triangulation: Integrate qualitative interview data with quantitative divergence results. Use stakeholder quotes and themes to explain why certain ES (e.g., drought regulation) show high levels of divergence, providing context to the numbers [1] [44].
• Barrier Identification: Analyze transcripts from all stakeholder groups to identify systemic, logistical, and knowledge-based barriers that may contribute to divergent valuations, such as traditional beliefs or lack of access to information [53].

The Scientist's Toolkit: Research Reagent Solutions

The following table lists key "reagents" – essential datasets, software, and methodological tools – required for executing the comparative research on ES valuations.

Table 2: Essential Research Tools for ES Valuation Studies

Tool Name / Type	Function in Research	Application Notes
CORINE Land Cover (CLC)	Provides standardized, temporally consistent land use/land cover (LULC) maps as the foundational spatial data layer for modeling.	European standard; for other regions, use national LULC datasets or classified satellite imagery (e.g., Landsat, Sentinel-2).
InVEST Model Suite	A suite of spatially explicit models for quantifying and mapping multiple ES (e.g., carbon storage, nutrient retention, habitat quality).	Enables a standardized, comparable approach to ES modeling. Requires biophysical data as inputs (e.g., biomass, water quality).
GIS Software	The primary platform for processing spatial data, running models, mapping ES outputs, and performing spatial analyses.	Essential for handling raster and vector data, and for calculating indices like ASEBIO.
Analytical Hierarchy Process (AHP)	A multi-criteria decision-making method used to derive objective weights for different ES based on expert or stakeholder input.	Critical for creating composite indices. Reduces subjective bias in weighting. Can be administered via surveys.
NVivo / Qualitative Software	Facilitates the organization, coding, and thematic analysis of unstructured qualitative data from interviews and focus groups.	Allows for systematic analysis of stakeholder perceptions, identifying emergent themes and contextual factors behind valuations.
3i Method Framework	A structured method to assess stakeholder Interest in, Influence on, and perceived Impact from specific ecosystem elements or services.	Particularly useful for dissecting complex human-nature relationships and conservation conflicts involving specific species [44].
Perception Matrix	A survey tool (matrix table) to rapidly capture stakeholder estimations of ES potential for different land cover types.	Provides directly quantifiable data that can be compared pixel-by-pixel with model outputs. Simple to administer.

Application Note 1: Quantitative Assessment of Perception Gaps in Ecosystem Services

Objective

To quantify and compare stakeholder perceptions against modeled data for ecosystem services, establishing a baseline for identifying conflict zones and informing co-management strategies.

Table 1: Comparative analysis of model-based versus stakeholder-perceived ecosystem service potential in Portugal (2018)

Ecosystem Service Indicator	Model-Based Value (ASEBIO Index)	Stakeholder Perception Value	Percentage Difference
Drought Regulation	0.28	0.62	+121.4%
Erosion Prevention	0.25	0.55	+120.0%
Climate Regulation	0.22	0.41	+86.4%
Habitat Quality	0.38	0.65	+71.1%
Pollination	0.31	0.52	+67.7%
Food Production	0.35	0.53	+51.4%
Recreation	0.42	0.61	+45.2%
Water Purification	0.49	0.58	+18.4%
Overall Average	0.34	0.56	+64.7%

Source: Adapted from Scientific Reports volume 14, Article number: 25995 (2024) [1]

Experimental Protocol

Protocol 1.1: Integrated Ecosystem Services Assessment

Purpose: To systematically measure discrepancies between quantitative ecosystem models and stakeholder perceptions for conflict identification.

Materials:

• CORINE Land Cover data or equivalent land use classification system
• GIS software with spatial analysis capabilities
• Standardized stakeholder survey instruments
• Analytical Hierarchy Process (AHP) templates for weight assignment

Methodology:

• Spatial Modeling Phase:
  • Calculate eight multi-temporal ES indicators using spatial modeling approaches
  • Integrate indicators into a composite index (e.g., ASEBIO index) using multi-criteria evaluation
  • Generate spatial distribution maps for each ecosystem service

• Stakeholder Perception Elicitation:

  • Recruit diverse stakeholder groups (community members, experts, policymakers)
  • Conduct perception assessment using four-point scale (1 = no use; 4 = high use)
  • Apply ≥50% recognition threshold to filter locally relevant services
  • Implement priority evaluation through 100-point allocation task

• Comparative Analysis:

  • Calculate percentage differences between model outputs and perceived values
  • Identify services with highest perception gaps as priority conflict areas
  • Map spatial distribution of discrepancy hotspots

Duration: 8-12 weeks for data collection and analysis Sample Size: Minimum 30 experts and 500 community members for statistical significance [1] [7]

Application Note 2: Conflict Trajectory Analysis and Intervention Mapping

Objective

To characterize conflict escalation stages and identify appropriate co-management interventions at each conflict level using standardized assessment frameworks.

Table 2: Conflict escalation stages and corresponding management interventions based on Glasl's model

Conflict Stage	Escalation Level	Key Characteristics	Recommended Interventions	Co-Management Applicability
Early Stage	1-3	Hardening, Debates, Actions	Negotiation, Mediation, Joint fact-finding	High - Direct stakeholder engagement effective
Middle Stage	4-6	Images/Coalitions, Loss of Face, Strategies	Facilitated dialogue, Confidence-building measures	Moderate - Requires third-party facilitation
Late Stage	7-9	Limited Destruction, Fragmentation, Together into the Abyss	Arbitration, Legal intervention, High-level policy changes	Low - Co-management difficult without prior de-escalation

Source: Adapted from Glasl's conflict escalation model (1999) as applied in Oryx (2018) [54]

Experimental Protocol

Protocol 2.1: Conflict Stage Diagnostic Assessment

Purpose: To classify conflict stages and select appropriate co-management interventions using standardized diagnostic tools.

Materials:

• Conflict assessment interview protocols
• Stakeholder mapping templates
• Conflict escalation indicator checklist
• Trust and transparency assessment scales

Methodology:

• Stakeholder Analysis:
  • Identify all relevant parties using legitimacy, resources, connections framework
  • Map relationships and dependencies between stakeholders
  • Assess interests and values related to the conflict issues

• Conflict Stage Diagnosis:

  • Conduct semi-structured interviews with key informants (minimum 22 interviews)
  • Organize focus group discussions (26 discussions across stakeholder categories)
  • Apply Glasl's impairment criteria to identify conflict stage
  • Document manifestations and subjective perceptions of impairment

• Intervention Matching:

  • Select conflict management approaches appropriate to escalation stage
  • Design deliberative processes matching conflict intensity
  • Establish monitoring indicators for de-escalation progress

Duration: 4-6 weeks for comprehensive conflict assessment Outputs: Conflict stage classification, stakeholder network map, intervention roadmap [54] [55]

Application Note 3: Co-Management Implementation Framework

Objective

To establish participatory co-management structures that integrate local knowledge with scientific models for conflict resolution and sustainable ecosystem governance.

Table 3: Stakeholder priority differences in ecosystem services across land-use types in Laos (2025)

Ecosystem Service Category	Community Priority (Bamboo Forest)	Expert Priority (Bamboo Forest)	Priority Gap	Policy Implication
Provisioning Services	68.2%	42.5%	+25.7%	Communities prioritize tangible benefits
Regulating Services	18.4%	36.8%	-18.4%	Experts emphasize regulatory functions
Cultural Services	9.8%	11.2%	-1.4%	Relative alignment in cultural values
Habitat Services	3.6%	9.5%	-5.9%	Experts prioritize biodiversity conservation

Source: Adapted from Forests 2025, 16(10), 1581 [7]

Experimental Protocol

Protocol 3.1: Participatory Co-Management Establishment

Purpose: To implement adaptive co-management structures that formally integrate local knowledge with scientific expertise.

Materials:

• Focus group discussion guides
• Multi-criteria decision analysis tools
• Participatory mapping resources
• Monitoring and evaluation frameworks

Methodology:

• Stakeholder Activation:
  • Conduct moderated focus groups (8-10 participants per group)
  • Facilitate interaction between conflicting parties using structured dialogues
  • Identify shared concerns and mutual interests

• Knowledge Integration:

  • Combine traditional ecological knowledge with scientific models
  • Develop shared conceptual models of social-ecological systems
  • Co-produce management plans integrating both knowledge systems

• Institutional Design:

  • Establish Protected Areas Management Advisory Units
  • Create Community Resource Management Areas
  • Define clear roles, responsibilities, and benefit-sharing mechanisms
  • Implement conflict resolution mechanisms within governance structure

• Adaptive Management:

  • Establish continuous monitoring of both ecological and social indicators
  • Create feedback mechanisms for plan adjustment
  • Schedule regular review sessions for adaptive learning

Duration: 6-9 months for full implementation Success Indicators: Reduced conflict incidents, improved trust metrics, enhanced livelihood benefits [54] [55] [7]

Visualization: Integrated Conflict-to-Co-Management Pathway

Figure 1: Integrated pathway from conflict assessment to co-management implementation, showing sequential phases and key transition points.

The Scientist's Toolkit: Essential Research Reagents and Materials

Table 4: Key research reagents and methodological tools for integrated ecosystem assessment and conflict management research

Tool/Reagent	Specifications	Application	Key Considerations
CORINE Land Cover Data	Minimum mapping unit: 25 hectares; Thematic accuracy: >85%	Baseline ecosystem service modeling; Land use change analysis	Requires compatibility with local classification systems
Analytical Hierarchy Process (AHP)	Pairwise comparison scale: 1-9; Consistency ratio threshold: <0.1	Weight assignment for multi-criteria evaluation; Stakeholder preference integration	Effective for managing complex decision criteria with multiple stakeholders
Stakeholder Survey Instrument	Four-point perception scale; 100-point allocation task; Back-translation protocol	Eliciting ecosystem service perceptions; Quantifying priority differences	Requires cultural adaptation and pre-testing in local context
Conflict Assessment Framework	Glasl's 9-stage escalation model; Impairment criteria; Stakeholder mapping templates	Diagnosing conflict stage; Identifying appropriate intervention levels	Dependent on trust-building for accurate information sharing
Focus Group Discussion Guide	Semi-structured protocol; Conflict-sensitive facilitation techniques; Recording and transcription protocols	Eliciting qualitative insights; Building shared understanding among stakeholders	Requires skilled moderators familiar with conflict dynamics
Spatial Analysis Software	GIS capabilities; Multi-criteria evaluation modules; Spatial statistic tools	Mapping ecosystem services; Identifying spatial mismatch hotspots	Computational resources must match analysis scale
Adaptive Co-Management Monitoring Framework	Social-ecological indicators; Mixed-methods assessment; Participatory evaluation	Tracking co-management effectiveness; Enabling iterative improvement	Must balance scientific rigor with practical feasibility

Sources: Adapted from multiple studies [1] [54] [55]

Conclusion

The comparison between ecosystem service models and stakeholder perceptions is not merely an academic exercise but a fundamental challenge in environmental management. The evidence consistently shows that while discrepancies are significant and systematic—often rooted in the contrast between generalized scientific data and localized, traditional knowledge—they are not insurmountable. Key takeaways are the demonstrated superiority of model ensembles for improving predictive accuracy, the critical need for robust model validation, and the importance of institutionalizing participatory processes that integrate local perspectives. Future efforts must focus on developing transparent, accessible frameworks that actively bridge the certainty and capacity gaps. For researchers and policymakers, the path forward lies in embracing these integrative strategies, which promise not only more accurate ecosystem assessments but also more inclusive, equitable, and sustainable outcomes for both people and nature.

References

• 1. Humans versus models: a comparative assessment of ... [https://www.nature.com/articles/s41598-024-76600-w]
• 2. The relevance of stakeholders' perceptions of ecosystem ... [https://www.sciencedirect.com/science/article/abs/pii/S221204161730548X]
• 3. Exploring stakeholders' ecosystem services perceptions ... [https://www.frontiersin.org/journals/environmental-science/articles/10.3389/fenvs.2023.1214879/full]
• 4. It’s time for a more sophisticated color check for contrast ... data [https://www.datawrapper.de/blog/color-contrast-check-data-vis-wcag-apca]
• 5. An overview of methods to enhance ecosystem services ... [https://aries.integratedmodelling.org/aries-hub/an-overview-of-methods-to-enhance-ecosystem-services-assessments-within-the-global-biodiversity-framework/]
• 6. Ecosystem services mapping and modelling. Where is the ... [https://www.sciencedirect.com/science/article/pii/S2666683925000252]
• 7. Stakeholder Perception and Priority Gaps in Ecosystem ... [https://www.mdpi.com/1999-4907/16/10/1581]
• 8. A methodological approach for the analysis of ecosystem ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC9989104/]
• 9. Selecting methods for ecosystem service assessment [https://www.sciencedirect.com/science/article/pii/S2212041617300487]
• 10. Ecosystem Services Research | US EPA [https://www.epa.gov/eco-research/ecosystem-services-research]
• 11. Integrating scientific data, local knowledge, and expert ... [https://ecologyandsociety.org/vol30/iss4/art14/]
• 12. Comparing stakeholder perspectives and biodiversity ... [https://www.sciencedirect.com/science/article/pii/S0301479725005171]
• 13. AI-driven cultural urbanism: A data-integrated model for ... [https://www.sciencedirect.com/science/article/pii/S2590291125007399]
• 14. Integrating ecosystem services, stakeholders' perspective ... [https://link.springer.com/article/10.1007/s10980-024-01886-0]
• 15. İstatistik Araştırma Dergisi » Submission » Doğal Sermaye ve... [https://dergipark.org.tr/en/pub/jsstr/issue/93965/1703557]
• 16. InVEST - Stanford Natural Capital Project [https://naturalcapitalproject.stanford.edu/software/invest]
• 17. InVEST (Integrated Valuation of Ecosystem Services and ... [https://ecosystemsknowledge.net/resources/tool-assessor/invest-integrated-valuation-of-ecosystem-services-and-trade-offs/]
• 18. ARIES - ARtificial Intelligence for Environment & Sustainability [https://aries.integratedmodelling.org/]
• 19. ARIES (ARtifical Intelligence for Environment & ... [https://ecosystemsknowledge.net/resources/tool-assessor/aries-artifical-intelligence-for-environment-and-sustainability/]
• 20. Co$tingNature [https://www.policysupport.org/costingnature]
• 21. $ Co - Ecosystems Knowledge Network ting Nature [https://ecosystemsknowledge.net/resources/tool-assessor/costing-nature/]
• 22. InVEST - Marine Planning [https://marineplanning.org/tools/software/invest/]
• 23. Combining Co$ting Nature and Suitability Modeling to ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC8443209/]
• 24. Assessment and multi-scenario prediction of ecosystem ... [https://www.frontiersin.org/journals/ecology-and-evolution/articles/10.3389/fevo.2025.1539547/full]
• 25. Eliciting Stakeholder Perceptions Using a Novel Online ... - NCBI [https://www.ncbi.nlm.nih.gov/books/NBK589558/]
• 26. How to Present Quantitative & Qualitative Data in Reports [https://www.clearpointstrategy.com/blog/qualitative-and-quantitative-data]
• 27. What is the Analytic Hierarchy Process (AHP)? [https://www.1000minds.com/decision-making/analytic-hierarchy-process-ahp]
• 28. Integrating Technical, Socio-Economic, and Sustainability Dimensions... [https://link.springer.com/article/10.1007/s00267-025-02224-z]
• 29. Multi-criteria decision making beyond consistency [https://www.sciencedirect.com/science/article/pii/S0360835224007836]
• 30. Structure Data for Analysis [https://help.tableau.com/current/pro/desktop/en-us/data_structure_for_analysis.htm]
• 31. Structured Data vs Unstructured Data - Difference Between ... [https://aws.amazon.com/compare/the-difference-between-structured-data-and-unstructured-data/]
• 32. How Calculate Composite Index works—ArcGIS Pro [https://pro.arcgis.com/en/pro-app/latest/tool-reference/spatial-statistics/how-calculate-composite-index-works.htm]
• 33. A quantitative approach to evaluating ecosystem services [https://www.sciencedirect.com/science/article/abs/pii/S0304380013000902]
• 34. Progress in regulating ecosystem services research [https://www.nature.com/articles/s40494-025-01674-2]
• 35. Frontiers | Exploring the landscape of essential health data science skills and research challenges: a survey of stakeholders in Africa, Asia, and Latin America and the Caribbean [https://www.frontiersin.org/journals/public-health/articles/10.3389/fpubh.2025.1523873/full]
• 36. Conceptual system maps to guide research and ... [https://www.sciencedirect.com/science/article/pii/S2530064425000197]
• 37. Ensemble Models in ML: Techniques and Benefits - SoluLab [https://www.solulab.com/ensemble-modeling/]
• 38. Evaluating ensemble models for fair and interpretable ... [https://www.nature.com/articles/s41598-025-15388-9]
• 39. A comprehensive review and future research directions of ... [https://www.sciencedirect.com/science/article/abs/pii/S0378778825003196]
• 40. Social valuation of urban ecosystem services using the ... [https://www.nature.com/articles/s41599-025-05946-x]
• 41. Ecosystem services [https://www.coastalwiki.org/wiki/Ecosystem_services]
• 42. Ecosystem Services [https://oceanwealth.org/ecosystem-services/]
• 43. Ecosystem Services: AP® Environmental Science Review [https://www.albert.io/blog/ecosystem-services-ap-environmental-science-review/]
• 44. Comparative stakeholder perceptions of wildlife ... [https://www.sciencedirect.com/science/article/pii/S0301479725021620]
• 45. A case study in mountainous areas of northwest Yunnan, China [https://pmc.ncbi.nlm.nih.gov/articles/PMC11168320/]
• 46. In the eye of the stakeholder: Changes in perceptions ... [https://www.sciencedirect.com/science/article/abs/pii/S2212041614000321]
• 47. Mismatches in the Ecosystem Services Literature—a ... [https://link.springer.com/article/10.1007/s40823-021-00063-2]
• 48. Quantifying T2-FLAIR Mismatch Using Geographically ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC8757555/]
• 49. VisAlign: Dataset for Measuring the Alignment between AI ... [https://openreview.net/forum?id=Wz2BJNQlyI]
• 50. Bring Reason to Vision: Understanding Perception and ... [https://openreview.net/forum?id=ntCAP6tMoX]
• 51. A review of embodied intelligence systems: a three-layer ... [https://www.frontiersin.org/journals/robotics-and-ai/articles/10.3389/frobt.2025.1668910/full]
• 52. A quantitative model of ensemble perception as summed ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC10810262/]
• 53. Assessing stakeholder perceptions and contextual factors in ... [https://bmchealthservres.biomedcentral.com/articles/10.1186/s12913-025-13577-w]
• 54. From conflict to collaboration: the contribution of co ... [https://www.cambridge.org/core/journals/oryx/article/from-conflict-to-collaboration-the-contribution-of-comanagement-in-mitigating-conflicts-in-mole-national-park-ghana/92E1A64118694BB4C37DD7D63253E49A]
• 55. Adaptive co-management of conservation conflicts [https://pmc.ncbi.nlm.nih.gov/articles/PMC6218403/]