2026
87
Today, current forecast
The same /v1/score number a booking flow consumes. Real-time, not a scenario.
Where this sits today
Goable is founder-led and pre-revenue. Multi-decadal climate viability screening is an exploratory research direction on a T3 horizon, not yet built or shipped. No reports have been produced for clients, nothing here is grant-filing-ready, and there is no contract to buy. Scenarios are plausible pathways, not forecasts: they help compare exposure under different climate assumptions, and every number below is an illustrative sketch at regional scale.
01
The time horizon
One location, four decades
Illustrative. A kitesurfing location in Tarifa scores 87 today. The same engine and the same curves, projected forward via the CMIP6 ensemble under SSP2-4.5 (the middle-of-the-road pathway), would give a scenario trajectory an asset manager could reason about at regional scale. The figures are a sketch, not a delivered projection.
2035
82
Near-term scenario, SSP2-4.5
Slight wind-regime shift in the illustrative scenario. Regional suitability broadly holds; the season may compress by around 2 weeks.
2050
71
Mid-term scenario, SSP2-4.5
Illustrative season compression of roughly 30%. Workable at regional scale; an operator might weigh rebalancing the equipment mix.
2075
58
Long-term scenario, SSP2-4.5
Scenario suggests adaptation questions: site context, season timing, or activity-portfolio change.
These would describe scenario-based changes in climatic suitability at regional scale. They are not deterministic forecasts for an individual beach, business or future season, and are shown here as an illustrative research sketch.
The score drop is not the point: what changes is the confidence interval. As a scenario projects further, the ensemble spread widens. The methodology surfaces that spread directly, so an analyst could compare a portfolio against tail scenarios, not just the central estimate.
02
Illustrative example
Same location, four scenarios
The choice of emissions pathway changes which scenario you are comparing. In the research direction, the engine would return the full grid: scenario score and confidence per (year, scenario, location). Below is the same Tarifa kitesurf location, projected to 2075 across the four canonical IPCC pathways, as labelled illustrative sketches.
SSP1-2.6
82 to 78
2026 to 2075 scenario
Aggressive mitigation
~1.8C by 2100
SSP2-4.5
82 to 71
2026 to 2075 scenario
Middle of the road
~2.7C by 2100
SSP3-7.0
82 to 58
2026 to 2075 scenario
Rivalry, slow tech
~3.6C by 2100
SSP5-8.5
82 to 48
2026 to 2075 scenario
Fossil-fueled development
~4.4C by 2100
Warming figures are IPCC AR6 WG1 best estimates for 2081 to 2100 versus 1850 to 1900, shown for scenario context. Score ranges are illustrative.
Reading the sketch: Under aggressive mitigation (SSP1-2.6) the location screens favorable through 2075 in the scenario. Under the middle-of-the-road pathway (SSP2-4.5) the scenario score drifts into workable territory and adaptation questions become rational to ask. Under SSP5-8.5 the scenario crosses into poor by 2075, where an operator might weigh a shift in timing, location context, or activity. These are illustrative scenario comparisons, not predictions of a specific outcome.
03
Honest status
What is built, and what is not
Readiness is honest by level. None of this is a delivered service today: scenario screening is explored but not built, regional downscaling is not built, and no locally corrected report exists because the correction is not fitted or validated.
Scenario screeningCMIP6 global ensembleExplored, not built
European regional analysisCORDEX / regional downscalingNot built
Locally corrected reportFitted + validated to local recordsNot fitted or validated
Investment / insurance useDecision supportOut of scope today
04
How it could work
From CMIP6 to a labelled scenario sketch
No new physics, no new machine learning. The projection pipeline this direction imagines is composition over what the engine already ships: same providers, same physics formulas, same scoring curves, same confidence model. What would change across real-time and projection is the confidence interval.
1
CMIP6 ensemble
30+ general-circulation models, 4 emission scenarios, multi-decadal series pulled from ESGF. This global ensemble would be the basis for scenario screening at regional scale.
2
Regional downscaling
Higher-resolution regional downscaling (CORDEX EUR-11, 12km) is not built. Until it is, any exploratory request, including ones inside Europe, would be served by the global CMIP6 ensemble and labelled as regional context, never as site-level truth.
3
Bias correction
Local bias-correction would be applied only where it has been fitted and validated against local records. It is not fitted today, so exploratory projections run bias-neutral and say so (bias_correction.applied = false).
4
Profile re-score
The same physics curves the /v1/score endpoint uses. The activity's calibrated suitability function evaluated on the projected (wind, wave, temp, snow) distribution at each decade.
5
Scenario sketch
An illustrative POST /v1/projections/adaptation-report would return a multi-year, multi-scenario summary with confidence intervals from ensemble spread. This is a research sketch, not a report Goable delivers today.
data_lineageregional_context
{
"request": {
"location": "kitesurf-tarifa-balneario",
"scenario": "SSP2-4.5",
"horizon_years": [
2035,
2050,
2075
]
},
"downscaling": {
"method": "cmip6_global_ensemble",
"regional_cordex_applied": false,
"reason": "regional CORDEX downscaling not built"
},
"bias_correction": {
"applied": false,
"reason": "local bias-correction not fitted or validated for this location"
},
"spatial_interpretation": "regional_context",
"uncertainty": {
"source": "ensemble_spread",
"widens_with_horizon": true
},
"methodological_limitations": [
"exploratory research sketch, not a delivered report",
"not a deterministic forecast for an individual site or season",
"regional-scale suitability change, not site-level resolution",
"not hindcast-validated for this region"
]
}
Every response would carry its own lineage. It states the scenario, whether regional CORDEX downscaling was applied (it is not built), whether local bias-correction was applied and why not, the spatial interpretation, and the methodological limitations. Nothing would be presented as site-level truth that the lineage does not support.
05
Validation
Projections should be tested against the past
Train / test split
Fit on one climate period, then test the projection against a held-out earlier period the model did not see. A projection that cannot reproduce the past it was not trained on should not be trusted forward. This validation has not yet been run.
Per-variable, per-region error
Error would be reported for each variable (wind, wave, temperature, snow) and each region separately, because skill is uneven. A region can be reliable for wind and weak for wave.
Exploratory-only, for now
No region has passed hindcast validation, so everything on this page is labelled exploratory: usable for framing questions, not for decisions.
06
Who this could interest
Three personas, one research direction
If this direction were built, it might interest three kinds of reader. The examples below are illustrative uses, not services on offer and not work Goable has delivered.
DMOs, regional development
Tourism boards
Where a scenario roadmap for an activity portfolio could help.
Illustrative: an alpine canton might one day compare five ski seasons ahead across four emission scenarios to weigh which offerings to rethink and which new disciplines (gravel, winter hiking) to push. This describes a possible use of the research direction, not a service on offer today.
Holiday rentals, resorts, marinas
REITs, property
Where asset-level scenario screening across a mortgage horizon could help.
Illustrative: an asset manager with 200 Caribbean units might want to compare which subset screens as more resilient for kitesurf-rentals under a 2050 scenario versus the Aegean. A regional-scale sketch, clearly not a site-level guarantee, and not a deliverable Goable ships today.
EU Horizon, LIFE, national funds
Adaptation grant filers
Where an open, documented scenario methodology could help.
Illustrative: an alpine consortium filing a climate-adaptation project might value an open-citable methodology (CMIP6 ensemble, regional-context interpretation, uncertainty propagation). Goable has not produced or filed any such report, and nothing here is grant-filing-ready.
07
Methodology
Methodology and limitations, in full
Results would describe regional context, not a single beach or business. Until regional CORDEX downscaling is built and validated, a projection is a regionalised location scenario: it inherits the resolution of the global CMIP6 ensemble, and any finer reading is illustrative only.
08
Where this stands
Not for sale, and not on a tier
Posture
Scenario projections are not on any plan or tier, self-serve or otherwise. This is a research direction, not a product with a price, a delivery cadence or a contract. If it is ever built, how it is offered would be decided then, not promised now.
Open methodology
The bias-correction approach and the per-scenario uncertainty propagation are documented on the open Research Frontier page. The contribution here is a transparent, citable methodology, not a black box and not a compliance artifact.
09
Get in touch
If this direction is useful to you, tell us
There is nothing to buy and no report to order. If a multi-decadal scenario direction would be useful to your work, we would welcome a research conversation about the methodology and where it could go. We do not promise a deliverable, a timeline or a rollout.