Setup tab — choose NbS, geographic scope, resolution and climate; learn how to use the tool.
Configure your scan
choose what to map, where, at what detail, and the climate lens1 · Nature-based solutions
3 selectedEach selected NbS gets its own opportunity space.
2 · Geographic scope
3 · Analysis resolution
4 · Climate
Review your recipe
the variables and rules derived from your choices aboveVariables & recipe
Agroforestry · StandardA recipe is the set of environmental, socio-economic and hazard variables — and the rules that turn them into a suitability score — used to map where this NbS can work. The Standard recipe is expert-curated; you can switch to your own saved settings in Variable Config.
This recipe pulls 23 variables; after correlation clustering (collapsing variables that measure nearly the same thing, |r| > 0.7) about 16 representatives enter the analysis. Data readiness:
In Variable Config you can inspect each variable and its data source, see the correlation matrix & clusters, adjust cluster/variable weights, and tune how each NbS responds to each variable.
Run the analysis
compute suitability across the opportunity spaceHow to use this tool — guides, videos & methodology
How to use this tool
New here? Start with the quick tour, then dig into the methodology when you need the detail.
Opportunity Space tab — where the selected NbS can plausibly be implemented in Sierra Leone, with suitability classes, scorecard, and what-if constraint toggles.
Soft scenario levers
reversibleOperational constraints the Bank could relax. Toggle one off to see how much opportunity space a road or grid investment would unlock — these are reversible what-ifs, not hard limits.
Hard exclusions
masksAlways applied — these cells are removed, not scored, and can't be relaxed by a lever.
- Built-up & settlement mask
- Open water & wetland mask
- Protected areas mask
- Slope > 30° mask
Context layers
Descriptive overlays — they characterise the footprint, they don't change suitability.
Opportunity fingerprint
Where this NbS could plausibly be implemented.
Coverage
Distributed by suitability class
Country totals split across classes; opp space = VH + H.
Risk to rural livelihoods (climate)
— hazard exposure of people within the opp. space · need, not asset riskOf the selected scope, this is the share of land at each risk level for each hazard. Drought and heat shift the most under SSP2-4.5. This is risk to people / livelihoods — for risk to the NbS asset itself, see the Project Risk tab.
What agroforestry can address
Problems present in the opportunity space — and how strongly this NbS responds. Bars show problem-variable distribution; the Likert chip rates response strength.
NbS Comparison tab — multi-NbS comparison matrix built on Opportunity Space metrics.
Comparing 4 NbS across each one’s opportunity space · Sierra Leone · baseline + SSP2-4.5 / 2050.
Spatial co-occurrence
— where the selected NbS are simultaneously suitableSpatial co-occurrence
— which combinations of NbS are simultaneously suitable (exclusive area, km²)Each column is an exact overlap set. Agroforestry + Water harvesting is the largest shared opportunity (6.8k km²); all four overlap in the south-east hilly belt (Kenema, Kailahun).
Pairwise overlap
— size = shared area · colour = priority scoreAgroforestry × Water harvesting is both the largest overlap and the highest-priority — the strongest combined-investment case.
Priority Hotspots tab — scope filter, conceptual priority weights, hotspot map with bivariate view, ranked districts, and intersection panels (suit ∩ hotspot).
The map shows the intersection — cells that are both suitable for the NbS and high on your priorities.
Investment priorities
conceptual · liveStart from a priority profile, then fine-tune the weights if needed.
Pick a preset or a saved profile as a starting point, then fine-tune the weights below.
Climate hazard4 vars · 35%
NbS response (environment)3 vars · 30%
People & production3 vars · 35%
Equity & inclusionflag · not weighted
Context that describes the area, not problems the NbS resolves. Use them to filter / mask the hotspot view — they never enter the weighted score.
Top hotspots
ranked · Livelihood profileWhy Kenema is #1
focused ADM2- 01Rural poverty headcount 54% — top-quintile nationally (your H priority).
- 02Falls within agroforestry opportunity space at VH suitability across most parcels.
- 03Drought exposure moderate–high under baseline (your M priority).
- 04Production value at risk: $48 M / yr in cocoa & coffee.
- 05Rural pop. density 110 / km² — beneficiary reach is meaningful (your M priority).
No variable double-counted: climate risk used only here, not also in the suitability MCDA.
Opportunity fingerprint
— intersectionWithin the targeted hotspot subset of the opportunity space.
Coverage
Distributed by hotspot class
Opp-space totals split across hotspot classes; intersection = VH+H hotspot.
Risk to rural livelihoods (climate)
— within hotspot intersection · need, not asset riskHazard profile narrows in the 4,200 km² hotspot subset — drought and heat much worse here than in the full opp. space.
NbS response profile
— within hotspot intersectionLikert response strength stays the same, but the underlying problems are more severe here (that’s what makes them hotspots).
Variable Config — sub-tab of Setup.
Constraint types
— how each suitability variable limits where this NbS can goSuitability variables are grouped by the kind of constraint they impose. Hard exclusions mask cells out entirely; soft scenario levers are reversible what-ifs the Bank could relax (see Opportunity Space).
Farming system
— derived land-system classes that contextualise suitability & prioritiesDerived from land cover, livestock density and irrigation data. Scopes the context layers and tailors which descriptors matter — it does not change the suitability rule itself.
Variable configuration
— how each NbS reads each variable · Reduce / Source / ExplainVariables in the agroforestry formulation
16 of 23 enter the MCDA after correlation clustering. Chosen = cluster representative; collapsed = represented by its rep.
Correlation & clustering
Correlation clustering runs per AOI after standardisation. Variables with |r| > 0.7 collapse into one representative to avoid double-weighting. Hover a label to highlight its cluster; hover a cell for the exact r.
Cluster & variable weights Admin
Each cluster representative carries a weight into the MCDA. Adjust cluster importance, or expand to weight individual collapsed members. Weights normalise to 100%.
Add or replace a variable's dataset
Bring your own raster to replace a standard layer, or add a new variable to the formulation. Uploads are validated against the requirements below.
Upload requirements
- Format — Cloud-Optimised GeoTIFF preferred; NetCDF or zipped shapefile accepted.
- CRS — EPSG:4326 (WGS84) or a defined projection the tool can reproject.
- Extent — must cover the country AOI (Sierra Leone); larger is fine, it will be clipped.
- Resolution — ≤ 5 km; finer is resampled to the analysis grid. Coarser inputs are flagged in the resolution audit.
- Values — single continuous or categorical band; nodata flagged; units documented.
- Size — ≤ 2 GB per file. Larger layers: host the layer in cloud storage and supply a URL / asset ID instead.
- Licence — you must have the right to use the data; record the source & licence.
After upload the tool runs a validation pass (CRS, extent, nodata, range) and shows a preview before the layer is swapped into the recipe. Placeholder — not wired in v0.
Priority / hotspot variables
— where action matters most · feeds Priority Hotspots (M4)A different variable set from the suitability recipe. These score need & impact — climate hazard, the benefit an NbS would deliver, and the people & assets exposed — and are combined with suitability only at the hotspot stage. Keeping the two sets separate is what prevents Mode-A double-counting.
Priority & descriptor variables
11 priorities across 4 pillars + 4 descriptors. Priorities are standardised so higher = higher priority, then weighted into the hotspot index; descriptors are context only.
Gender inequity is held as a national contextual flag — shown alongside the ranking but it does not differentiate the map. Descriptors appear in Priority Hotspots as filter toggles, never as weighted sliders.
Scoring & direction
Each variable is standardised to 0–1. Direction sets which end is "high priority"; thresholds clip extremes. (Suitability response curves do not apply here — priority is a need score, not a fitness score.)
| Variable | Direction | Scoring |
|---|---|---|
| Drought hazard | drier → higher | percentile |
| Soil-erosion risk | more → higher | min–max |
| Rural poverty | poorer → higher | percentile |
| Production gap | larger → higher | percentile |
Correlation & clustering
Clustering runs independently on the priority set (poverty, food insecurity & rural population correlate; one represents the cluster). Kept separate from the suitability matrix so the two stages never share variables.
Theme & variable weights Admin
Theme weights set how much each pillar drives the hotspot index. Normalise to 100%.
Three weighted pillars normalise to 100%. Equity & inclusion (gender inequity) is carried as a contextual flag — shown, not weighted — so it can never silently tilt the ranking.
Drought (SPEI)
Higher = more need; ranked within this country.
Ranked within: this country (AOI)
Right: priority score — ranked within this country (AOI). Both clipped to the AOI.
Next Steps tab — moving from scoping to feasibility and implementation, tailored per NbS / NbS cluster, including cost-benefit.
Pathway for…
Next steps are tailored per NbS, or for a stacked NbS cluster where several overlap.
Placeholder · content per NbS / cluster will be authored with the sector team.
Indicative cost-benefit · AgroforestryT6 economic archetype
Planning-level figures from the T6 economic archetype applied to the opportunity-space extent — not a CBA. A full cost-benefit analysis with local prices is commissioned at feasibility stage (see methods below).
Cost-effectiveness snapshot · Agroforestry indicative
Planning-level unit costs from the T6 economic archetype over the opportunity-space extent — to size the ask and compare options, not a CBA.
Feasibility methods & tools
Guidance on the valuation & assessment methods to commission downstream for agroforestry — signposts, not prescriptions or endorsements. Choose by decision stage (see proportionality).
Tool names are quiet references to common practice, not recommendations. The appropriate method set is confirmed with the sector economist at feasibility.
Design step · right-tree-right-place
Tree-based NbS need a species & establishment design before implementation — scoping suitability is necessary but not sufficient.
Recommended next steps
Safeguards & FPIC hand-off
This tool flags the trigger and hands off to the safeguards team at pre-feasibility — it does not perform the screening.
Confirm data before feasibility
Two data actions to close out before committing to sites.
- 12 m national DEM (SLE-NSDI) — sharpen slope on steep SE terrain (default is 30 m SRTM).
- 2024 district poverty surface — replace the 2018 default once the survey round is released.
- Updated local road / market network — refine the market-access & road-access layers.
Tailoring to context
Placeholder · country- and NbS-specific guidance to be authored with the sector & country teams.
Project Risk tab — where disasters could damage or destroy the NbS investment itself (asset risk), the WB Climate & Disaster Risk Screening lens. Distinct from risk to rural livelihoods.
Asset risk for…
Asset risk is NbS-specific — a wetland and an agroforestry plot are vulnerable to different hazards.
Highest-risk units
ADM2 · rankedHazards threatening the asset
From T3 — the hazards this NbS does not mitigate but that can damage the asset. Bars show share of the opportunity space at high+ for each.
Exposure
The rating weights hazard intensity by what is exposed:
- NbS assetwhere this NbS would be established (its suitable extent)
- Roadsaccess for establishment & maintenance
- Powersupporting infrastructure near the asset
These describe whether an NbS can be implemented and maintained — they filter / flag the opportunity space to help rule sites in or out. They are deliberately kept out of the hotspot score so a feasibility blocker can't be silently traded off against need.
Methods · rating construction
Rating = Σ(hazardᵢ × weightᵢ) × exposure, classified into Low → Very High. Hazards are the complement of the NbS mitigation set drawn from T3; weights are per-NbS. Exposure combines asset presence with proximity to roads & power.
| Hazard (T3) | Weight | Exposure layer |
|---|
Mock weights. Confirmed with the DRM specialist at pre-feasibility; not a substitute for the project-level screening.
Danger Zone tab — restricted, global, destructive methodology and cache controls.
Restricted actions
audit-logged- 01Invalidate cached suitability surfaces — forces full re-run for the whole teamCACHE
- 02Add / remove variables from the master schema — affects every recipeSCHEMA
- 03Change global default weights — resets every TTL's starting pointMETHODOLOGY
- 04Publish a standard recipe version — promotes a draft to the org defaultMETHODOLOGY
- 05Delete a saved profile shared with the teamDESTRUCTIVE
All changes are logged in the methodology version log and update T0.last_updated. Per-NbS curve edits do not appear here — they are saved as personal/custom settings in Variable Config.
v0 · controls not yet wired
Visible architectural boundary. Wire to the protected schema-mutation API when the back-end exists.
To Do / Dev Notes — internal living checklist of remaining medium/major work. Not part of the TTL workflow.
- Priorities vs Descriptors split introduced across Variable Config & Priority Hotspots — descriptors filter, never weight.
- Priority variables reorganised into 4 weighted pillars (Climate hazard · NbS response · People & production · Equity & inclusion) + a Descriptors group; gender moved to a national contextual flag.
- Suitability variables grouped by constraint type (Biophysical · System · Operational) with hard-exclusion vs soft-lever marks.
- Per-variable data-source panel + provenance chips (dataset+grain · tier · sensitivity · quote); country-endorsed flag on population / poverty / production; farming-system derived classes.
- Opportunity Space: soft levers vs hard masks, production-gap & farming-system context, area-vs-line footprint note.
- Project Risk split into Part A (asset hazards) + Part B (operational feasibility filters incl. FPIC) — filters / flags, not scored.
- NbS Comparison: like-with-like guard + indicative cost-effectiveness snapshot. Next Steps: cost-effectiveness, country-endorsed-data prompt, FPIC hand-off, data-gap list.
- Dropped the “native GEE” framing → “Runs in Python”; tab set + order ratified in
AGENTS.md.
- Project Risk tab added at position 03; risk-to-rural-livelihoods relabel; dual scope bands on Priority Hotspots & NbS Comparison; project-risk rating confined to the opportunity-space footprint.
risk_role + asset_fragility, T2 risk_lens + asset_exposure (see M2b spec). Blocks the Project-Risk hotspot scope. Owner: Brayden.