Landscape Modelling

Slide deck: SETAC xLandscape Training — (1) Landscape Modelling Intro

This page follows the structure of the introductory lecture by Thorsten Schad (Landwerk e.V.).

Subject of our Work: What is a Landscape?

Landscapes are cultural areas — shaped by the combined work of nature and human activities:

UNESCO (1992): "Distinct geographical areas uniquely representing the combined work of nature and of man."
Council of Europe (2000): "An area, as perceived by people, whose character is the result of the action and interaction of natural and/or human factors."

Landscapes provide Ecosystem Services — benefits to society including food production, natural pest regulation, biodiversity conservation, water regulation, and cultural values. In cultivated landscapes, these services emerge from the interaction of environmental processes and human activities, shaped by land use, management practices, and spatial structure.

Landscape modelling provides a means to make ecosystem services explicit, measurable, and comparable — supporting a holistic view on risk, trade-offs, and decision-making in environmental risk assessment and management.

Specific Protection Goals (SPGs) in environmental risk assessment translate ecosystem service values into concrete, spatially explicit, and model-based regulatory protection targets. Landscape modelling is an essential connecting element between Ecosystem Services and SPGs.

Landscape Modelling — A Broad Craft

Landscape modelling in ERA covers a wide range of topics, spanning from the foundational data craft through to digital risk management in precision agriculture and integrated pest management. It can broadly be divided into two categories:

Fundamental Craft — the basic data, geoinformatics, and environmental-fate skills that underpin all landscape modelling work
Modelling-related Topics — the problem-oriented, integrative modelling that connects environmental processes, human activities, and risk endpoints

Aspects in ERA

Landscape modelling fulfils multiple roles in environmental risk assessment:

Aspect	Description
Integration	Links pesticide use, exposure, environmental fate and effect models; integrates risk characterisation with Specific Protection Goals
Holistic risk view	Addresses baseline definition, comparative risk assessment, and multiple stressors
Risk Management	Supports landscape design, mitigation options, integrated pest management
Communication	Stakeholder involvement, transparency, risk communication
EU initiatives	Integration with further EU initiatives and research programmes
Uncertainty reduction	Reduces uncertainty through spatially and temporally explicit representation

Disciplines

Landscape modelling in ERA is inherently interdisciplinary and integrative, linking environmental processes, human activities, and risk modelling:

GIS · Geoinformatics · Remote Sensing · Informatics · Computing · Digital Agriculture · Regulations · Modelling · Data Science · Communication · Exchange · Network · Community

Core cross-disciplinary capabilities include:

Ecological modelling — basis for risk/effect endpoints and integrated risk/benefit analysis
Environmental fate modelling — the basic craft; data is the foundation and daily business
Spatially explicit landscape modelling — the supreme craft and core business; purpose-driven representation of reality
Recovery and connectivity analysis — demonstrates recovery potential for non-target populations (e.g. NTA, insecticide connectivity case studies)
Regional risk management and digital twin — supports digital agriculture, precision application, corporate sustainability

Regulatory-Scientific Challenges in Cultivated Landscapes

Current regulatory science is tackling a range of open questions that require landscape-level approaches:

How can pesticide risk assessment be made more realistic and risk management more effective — across aquatic organisms, arthropods, pollinators, vegetation, birds and mammals?
What are the holistic effects of real-world pesticide use schemes, and how can chemical pest control be minimised in the context of Integrated Pest Management?
Which factors drive the occurrence and composition of plant communities — herbicides, fertilisers, or landscape structure?
How can indirect effects (e.g. of weed control) be assessed and compared across different pest control options?
How can cultivated landscapes be designed to balance ecosystem services (food production, nature, biodiversity)?
How can monitoring and modelling be integrated for decision support and continuous knowledge gain?

The Role of Landscape in Regulatory Documents

While no single dedicated "Landscape Guidance" exists in Europe, landscape aspects appear as a cross-sectional topic across all species groups in regulatory guidance. The only formal European document directly addressing the topic is:

FOCUS Landscape & Mitigation (2008) — Guidance on risk assessments for plant protection products at landscape-scale

Landscape-scale pesticide risk assessment aspects regularly occur across EFSA regulatory guidance documents and scientific opinions, regulatory-scientific development initiatives (SETAC workshops on NTTP, Modelink, MAgPIE, MAD), and higher-tier regulatory-scientific work.

Landscape Aspects in Modern EFSA Documents

More than 95 % of active EFSA guidance documents and scientific opinions on pesticide risk assessment contain landscape modelling aspects. This reflects a mandatory methodological shift over the last decade:

Aspect	Regulatory Basis	Prevalence in Modern Outputs
SPG Framework	Mandatory 5-dimension definition (EFSA 2016)	100 % of documents following the 2016 mandate
Exposure Fate	Shift to spatially distributed PERSAM/numerical models	Standard for soil; increasingly for aquatic/groundwater
Effect Modelling	Use of Agent-Based Models (ApisRAM) and Mechanistic Models	Standard for higher-tier vertebrates and pollinators
Recovery Assessment	Spatial connectivity requirement for ERO	Included in all assessments allowing for population recovery

Key drivers:

2016 EFSA SPG Guidance: Every ERA must define SPGs across five mandatory dimensions — including spatial scale — making landscape thinking a requirement for all new guidance.
PERSAM / Spatial Exposure: The 2017 Soil Exposure Guidance moved to a spatially distributed framework calculating 95th-percentile concentrations across the distribution of a crop within a regulatory zone.
"Action at a Distance": Effects occurring outside the immediate area of application due to hydrological transport or organism movement require landscape-scale connectivity modelling.
Ecological Recovery Option (ERO): A core element of modern EFSA opinions, defined as a spatial process — modelling connectivity between treated fields and untreated refuge areas.
PERA and IPol-ERA: Roadmap for a full transition by 2030 to landscape-scale, population-level assessment tools accounting for multiple stressors.

Landscape Modelling as a Reference — Managing Complexity

Landscape modelling is more realistic than a standard scenario, but not the reality — still a model of the world. Approaches, comprising models, components, and scenarios, are purpose-driven and problem-oriented, and are typically only as complex as necessary for the question at hand.

Views to Landscape Modelling in ERA

Landscape modelling sits at the intersection of multiple perspectives, each contributing a different lens:

Perspective	Key Attributes
Scientific disciplines	GIS · Geoinformatics · Remote Sensing · Informatics · Computing · Digital Agriculture · Regulations · Modelling · Data Science · Communication; collaborative community and exchange
Emerging holistic view	Holistic risk assessment, multiple stressors, recovery modelling
Regulatory framing	Cross-sectional topic across all species groups; no single 'Landscape Guidance'; explicit Tier-4 ('freestyle') higher-tier approach; Specific Protection Goals as the anchoring framework
ERA applications	Higher-tier (Tier-4) risk assessment; ecosystem services, population and effect modelling; supporting-evidence strategy ("landscape-mechanistically informed"); landscape solutions to EnSa-EFM questions; digital agriculture integration
Landscape as the region	Shaped by ecosystem processes and human activities; region of holistic view to risk and benefits
Management dimension	Landscape management integrating multi-stakeholder interests; integrative modelling platform linking environment, agriculture, exposure, and effect modelling
Discipline base	Environmental, agricultural, landscape ecology, and economy disciplines; pesticide environmental fate data and modelling; Integrated Pest Management

Daily Working Areas

The range of landscape modelling work in ERA practice spans multiple focus areas:

Working Area	Examples and Topics
Higher-tier risk assessment	Aquatic organisms, NTA, NTP, Pollinators, Mammals, Off-field soil, Plant communities, Groundwater
Indirect effects and holistic risk	Weed control indirect effects, holistic pesticide use scheme effects, system approaches, digital twins
Crop/use setting specifics	Orchards, olives, stone/pome fruit, arable, railway verges, industrial vegetation management, golf, greenhouses
Field studies support	Field study site selection and representativeness, study scope, site similarity analysis
Digital agriculture	Precision application, residues in crop, environmental impact reduction, corporate sustainability
Geodata infrastructure	GIS, remote sensing, geodata acquisition, preparation and analysis, field investigation, drone mapping, geodata services (e.g., weather)
Ecological research topics	Insect decline, biodiversity, bee health, runoff and drainage refinement, groundwater vulnerability and monitoring, soil degradation
Communication	Mapping and communication, web services

Work Areas and ERA Applications

Area	Description
Higher-tier (Tier-4) RA	Spatially explicit, population- and ecosystem-level assessment
Ecosystem services	Population and effect modelling in context of ES and SPGs
Evidence strategy	"Landscape-mechanistically informed" supporting evidence
Digital Agriculture	Regional risk management and digital twin/Ag integration
Landscape solutions	Addressing EnSa-EFM questions and landscape-scale management

What Landscape Modelling Can Do — Our Vision

"You have a tool at hand enabling you to build and run processes and models at landscape-level, using real-world data. You can adapt this tool to your problem. It's open source. You can make use of open developments done by your colleagues. You can run the landscape models on your laptop or large cloud systems."

More specifically:

The conceptual foundation is embedded into regulatory-scientific frameworks, with tiered scenario development — start on screening levels and increase detail as needed
Scenario services support users on request
Consistency, freeze, and versioning ensure full long-term availability and reproducibility within a regulatory context

This is why an open, modular, and collaborative landscape modelling framework — xLandscape — is the route to take: models, components and scenarios, developed collaboratively because they are needed and not yet practically available.

Take-Home Messages

Cultivated landscapes — shaped by environmental conditions, biotic processes, and human activities — are the subject of our work. They offer a range of Ecosystem Services, including natural regulation and food provision.
Landscape modelling is a means to realistically measure and objectively quantify services and ecological conditions, for both research and decision support. In its most generic view, it encompasses a wide range of crafts and disciplines.
In pesticide ERA and the current digital transformation, landscape-level risk assessment is becoming part of Integrated Pest Management. Landscape aspects appear as a cross-sectional topic across all species groups.
In the specific ERA context, landscape modelling serves a range of purposes and applications. Spatiotemporally explicit landscape modelling directly addresses SPG requirements and has become a means for more realistic RA and management.
We develop frameworks, models, and scenarios — via the xLandscape platform — because they are needed and do not yet exist or are not practically applicable.