The Global Risks Forum (GRF)
Deep uncertainty describes situations where experts and stakeholders cannot agree on the models that link actions to outcomes, the probability of key future conditions, or how to value the results. Treating such conditions as if a single probability were known produces fragile decisions. This page sets out the methods institutions use when probabilities are contested: robustness, adaptive pathways and monitored triggers, and the mandates that must accept an explicit plan for changing course.
GRF · Established · service level R0
Deep uncertainty describes situations where experts and stakeholders cannot agree on the models linking actions to outcomes, the probabilities of key future conditions or how to value results. Decision making under deep uncertainty seeks options that perform acceptably across many futures, and plans that adapt when signposts reach pre-agreed triggers.
For deep uncertainty, in GRF's situational-awareness reading, the map opens on population density, flood water and rain with relief in 3D and clouds now from NASA GIBS, framed on the Arctic; every other NASA layer, live feed and forecast waits in the layer list.
The interactive map needs JavaScript; the layer list and sources are below.
Layers 3
Hazards and exposure, for this page
Keyboard, with the map focused
Hold Ctrl or ⌘ and scroll to zoom
Latest available
NASA GIBS / EOSDIS · ASTER GDEM (METI, NASA) · Blue Marble (NASA) · © OpenStreetMap contributors (ODbL) · Natural Earth
Loading the latest NASA imagery.
Drag to move. Ctrl or ⌘ and scroll, pinch or the buttons to zoom; hover for the values under the pointer; reset returns to the region.
For deep uncertainty, in GRF's situational-awareness reading, the map opens on population density, flood water and rain with relief in 3D and clouds now from NASA GIBS, framed on the Arctic; every other NASA layer, live feed and forecast waits in the layer list.Satellite and model products as NASA publishes them, each with its own date; indicators, not assessments of any place; claim ceiling INTEGRATED.
Layers and dates
The working screen behind this subject: its structure, terms and states. Names are placeholders and values illustrative.
Deep uncertainty · Scenario set B · Policy: POL-19
0 fails to 100 robust
Higher is more robust. Placeholder scores.
Methods when probabilities are contested. The dossier below sets out how the risk develops, what is measured, where the data come from, how it is modelled and who decides.
Long horizons, novel conditions and reflexive human behaviour leave no reliable frequency record. Different credible models give different answers, and choosing one hides the disagreement that decision-makers need to see.
Under deep uncertainty the question shifts from which option is best in an expected future to which option performs acceptably across many futures, and where it fails. Vulnerabilities, not predictions, become the focus.
Plans are sequenced so that early actions keep later options open. Each pathway has an adaptation tipping point, the condition beyond which it no longer meets its objective, and a lead time to switch.
Adaptive plans need monitored signposts, pre-agreed triggers and an institution with authority to act on them. Without that commitment an adaptive plan becomes a static plan with a review clause.
Each term as this page uses it. Where a standard-setting body defines the term, the wording follows that body.
| Term | Definition |
|---|---|
| Regret | The difference between an option's outcome in a given future and the best outcome available in that future. |
| Adaptation tipping point | The point at which a strategy no longer meets its objectives as conditions change. |
| Dynamic adaptive policy pathways | A method that maps sequences of actions and the tipping points at which to switch between them. |
Each indicator carries its unit, its cadence, the kind of source behind it and the evidence label a value would carry. No values are shown: a programme supplies them under its own authority.
| Indicator | Unit | Cadence | Source type | Label |
|---|---|---|---|---|
| Spread of outcomes across a scenario ensemble | units of the decision objective | per assessment | Model ensembles | MODELLED |
| Options meeting performance thresholds across futures | share of sampled futures | per analysis | Robustness analysis | MODELLED |
| Regret of the preferred option | units of the decision objective | per analysis | Regret analysis across scenarios | MODELLED |
| Plans with stated adaptation tipping points | count of plans | per planning cycle | Published plans | OFFICIAL |
| Signposts monitored with pre-agreed triggers | count of signposts | quarterly | Monitoring register | REPORTED |
| Spread of elicited expert judgement | range between elicited quantiles | per elicitation | Structured expert elicitation | ESTIMATED |
| Structural diversity of the model ensemble | count of independent model structures | per assessment | Model intercomparison records | DECLARED |
The topic as a computational graph: named sources feed the methods, an evidence review gates the outputs, and the decisions stay with the authorities that hold them.
Fig. 02Nodes carry their type, two metric rows and a state; connectors show what feeds what; dashed frames name the owners; the timeline is the cadence. Design of the pipeline, not a record of a run.
Named public datasets and institutional records, each with its licence class. Sources by agreement are used only under their terms and never republished.
| Source | What it gives | Licence class |
|---|---|---|
| IPCC Sixth Assessment Report, Working Groups I and II | Assessed ranges, calibrated confidence language and storylines for low-likelihood, high-impact outcomes | IPCC terms, free with attribution |
| Thames Estuary 2100 Plan (Environment Agency) | Adaptive pathways for tidal flood risk with monitored indicators and decision points | Open Government Licence |
| Dutch Delta Programme (Deltaprogramma) | Adaptive delta management, decision points and annual programme reporting | Public records, Dutch government terms |
| NGFS climate scenarios | A structured set of transition and physical risk pathways for financial supervisors | Free with attribution |
| Society for Decision Making under Deep Uncertainty | Published methods, case studies and open tools for robust and adaptive decisions | Terms per publication, many open access |
| CMIP model intercomparison archives (WCRP) | Multi-model climate outputs that show where model structure drives disagreement | Free and open, CC BY 4.0 for most datasets |
Foresight Nexus prepares evidence. The decisions below are taken by the authorities named, under their own rules, and never by the platform.
| Who decides | The question | Evidence this page serves |
|---|---|---|
| Finance ministries and treasuries | Whether appraisal of long-lived public investment should test robustness rather than a single central case | Regret and robustness results beside the central appraisal, with the futures in which each option fails |
| Infrastructure and flood risk authorities | When to commit to the next stage of an adaptive plan | Signpost status against triggers, the lead time to switch, and the tipping point of the current pathway |
| Central banks and financial supervisors | How to set supervisory expectations when climate and transition paths are contested | Scenario ranges with model and assumption differences shown, not averaged away |
| Environment and climate ministries | Which adaptation measures are low-regret now and which should wait for more information | Robustness across ensembles and the value of waiting, stated with its assumptions |
Economists have long distinguished risk, where probabilities are known, from uncertainty, where they are not. Under deep uncertainty, optimizing for a single best estimate can produce strategies that fail badly if the estimate is wrong. Decision science has therefore developed approaches that seek robustness, performing acceptably across many futures, and adaptiveness, changing course as new information arrives.
Methods such as robust decision making, dynamic adaptive policy pathways and exploratory modelling run many plausible futures through models to identify the conditions under which strategies fail. Decisions are then designed with signposts and triggers: observable indicators that would prompt a change of course. These approaches have been used in water management, flood protection and long-term infrastructure planning.
Running models across many assumptions to map where strategies succeed or fail, using complex systems and agent-based modelling.
Comparing strategies by their performance across futures, drawing on decision analysis under uncertainty.
Sequencing actions with signposts and triggers so that plans can change as conditions do.
These methods need significant analytic capacity, depend on the range of futures considered, and still require value judgements about acceptable outcomes.
The current service level is R0: information and enquiries. An enquiry receives a reference and names the next step; applications, payments, compute allocations and registrations open only when their team, terms and review capacity exist.