Environmental modelling under pressure

Turning the D-USYS model landscape into a shared resource—and a perspective paper

Modelers @ USYS

One initiative, two useful outputs

The model inventory becomes the evidence base for a perspective paper—and the paper gives the inventory a reason to exist and remain useful.

We need to agree on the scientific question before designing the database.

A catalogue alone will not sustain engagement

The weak version

“Which models do D-USYS groups use?”

  • Quickly becomes outdated
  • Produces descriptive counts
  • Has limited external relevance

The stronger version

“What does our diversity reveal about the future of environmental modelling?”

  • Starts from testable propositions
  • Collects only analytically useful fields
  • Makes cross-disciplinary experience visible
  • Builds an argument beyond D-USYS

D-USYS is the microcosm—not the subject

Environmental science contains radically different modelling traditions. D-USYS offers a bounded but unusually diverse test bed for finding shared pressures, bottlenecks and emerging directions.

forest · crop · soil · hydrology · atmosphere · biodiversity · economics · remote sensing · coupled human–environment systems

The paper’s relevance comes from the general argument; local examples make that argument concrete.

A deliberately lightweight model landscape (1/2)

Dimension Examples
Model / framework name, link, group
Domain forest, crop, climate, hydrology, biodiversity, soil, economics…
Purpose understanding, prediction, scenarios, decision support, teaching
Spatial scale organism → plot → landscape → regional → global
Temporal scale hours → centuries
Approach process-based, statistical, ML/AI, hybrid, optimization, agent-based…
Representation deterministic/stochastic, individual/aggregate, spatial/non-spatial

One record should take approximately 10–15 minutes—not become a model documentation project.

A deliberately lightweight model landscape (2/2)

Dimension Examples
Inputs experiments, monitoring, remote sensing, inventories, reanalysis…
Outputs none / external scenario / management / behaviour / coupled
Human dimension understanding, prediction, scenarios, decision support, teaching
Uncertainty whether/how represented
Development status used / adapted / actively developed
Interoperability standalone / coupled / modular
Main current limitation one short free-text answer
Future need “What would make this model substantially more useful in 10 years?”

One record should take approximately 10–15 minutes—not become a model documentation project.

The data collection should expose shared problems across disciplines

Scale

Complexity

Data

Uncertainty

Coupling

Decisions

Software + AI

The intellectual contribution is not that disciplines use different models. It is that different models may be confronting the same structural pressures.

An idea for paper propositions / paper scope / structure

  1. Environmental models are being asked to do more
    The pressures: global change, extremes, decisions, data and AI.
  2. A landscape of modelling approaches
    Briefly introduce the D-USYS evidence base.
  3. Different models, surprisingly similar problems
    Identify recurring cross-disciplinary bottlenecks.
  4. Where modelling is heading
    Evaluate the transitions already underway.
  5. What should change
    Make an opinionated case about priorities and infrastructure.
  6. A research agenda for the next decade
    End with five concrete commitments—not “more research is needed.”

A practical route from mixer idea to publishable work

Phase Collective task Output
1. Frame Agree on 6–8 propositions and paper scope Concept note + provisional author group
2. Pilot Develop ad test the questionnaire with 3 contrasting groups Revised schema + feasibility check
3. Collect One contact per group; structured response + short reflection Searchable inventory
4. Interrogate Descriptive patterns + thematic synthesis + challenge workshop Figures and argument map
5. Write Small writing team; contributors verify interpretation Perspective manuscript
6. Sustain Assign ownership, update cycle and public/private layers Useful living resource

Discussions points for today

  1. Is the combined inventory + perspective paper concept worth pursuing?
  2. Which one central claim is sufficiently important—and contestable?
  3. Which 6–8 propositions should shape the questionnaire?
  4. Who will join a small design group for the pilot?

Start with the scientific argument. Collect only what the argument—and a genuinely useful resource—requires.