Start with the site.
Select a parcel in Brazil and define the restoration objective. Where available, add field observations and practical constraints.
Better restoration
starts by understanding
the land.
Arbor is developing an ecological restoration planning platform that combines ecological science with AI to support more informed restoration decisions.
01 — THE CHALLENGEWHY ARBOR EXISTS
Every landscape
needs its own answer.
Ecological restoration is not simply about planting trees. Soil, vegetation, climate and past land use all influence whether a site can recover naturally or needs active intervention. The wrong approach can waste resources and harm the ecosystem a project intends to restore.
Making that choice means interpreting data scattered across different sources — often with important gaps. Arbor is being built to bring the evidence together and use AI-assisted analysis to help professionals compare practical options before work begins.
Local ecological expertise remains essential. The aim is to improve the planning process, not automate the final decision.
THE PROBLEM / COMPLEX LANDSCAPES, FRAGMENTED EVIDENCETHE RESPONSE / CLEARER, TRACEABLE PLANNING
01 / LANDSCAPESEvery place
tells a different story.
tells a different story.
02 / CONTRASTSDifferent conditions.
Different answers.
Different answers.
Left: Brazilian forest landscape · UnsplashRight: dry Brazilian landscape · Unsplash
02 — THE PLATFORMAN AI-ASSISTED WORKFLOW
From environmental data
to better decisions.
Arbor is designed to help restoration teams move from a site boundary to an evidence-based planning brief. Each step has a purpose — and a limit.
Build an environmental picture.
Combine land-cover history, vegetation, terrain, soil estimates and climate information. Highlight meaningful differences within the site, while recording data quality.
Explore realistic approaches.
Use ecological rules and, where justified, AI or statistical models to examine potential natural regeneration, assisted recovery and active restoration.
Prepare the next decision.
Produce a traceable first report with options, assumptions, risks and checks for a professional field assessment — not an automatic planting prescription.
Planned MVP capabilities. Site-specific recommendations and predictive models will only be introduced as their evidence and validation allow.
03 — WHERE WE BEGINBRAZIL / FIRST MVP
THREE ECOSYSTEMS. ONE FIRST PRODUCT.Starting in Brazil.
Starting in Brazil.
Built to go further.
Brazil is the first development and testing ground for Arbor — not the geographical limit of the platform. The initial MVP focuses on three major biomes with fundamentally different restoration needs.
01 / AMAZÔNIA
Amazon
Tropical forest landscapes vary widely in disturbance history, soils, hydrology and proximity to intact vegetation. Those differences affect where natural recovery is possible and where intervention may be necessary.
Arbor's planned contribution
Bring forest-cover history and site conditions into one assessment to help teams investigate regeneration potential and the need for further field checks.
02 / CERRADO
Cerrado
The Cerrado includes native grasslands, savannas and woodlands. Restoring it is not automatically about adding trees: fire regimes, seasonal drought and the original vegetation type matter.
Arbor's planned contribution
Support comparisons that respect the reference ecosystem, so naturally open habitats are not treated as failed forests.
03 / MATA ATLÂNTICA
Atlantic Forest
Remaining Atlantic Forest habitats are often fragmented among agricultural and urban landscapes. Terrain, nearby forest remnants and land-use pressure can substantially change recovery options.
Arbor's planned contribution
Combine landscape context with site-level constraints to help practitioners compare assisted regeneration and active interventions.
One shared technical foundation, with ecological rules and validation adapted to each biome. Expansion beyond these three environments comes after the MVP is tested — not before.
04 — A CLOSER LOOKA useful report.
A useful report.
Not just a map.
Arbor's proposed output is a planning brief a restoration professional can question, refine and take into the field. It should make the evidence behind a decision visible, not hide it behind an AI-generated score.
The preview shows the intended structure only. It does not represent an analysis of a real property or a validated recommendation.
ARBOR / PLANNING BRIEFCONCEPT PREVIEW
WHAT A TEAM COULD RECEIVE
From site conditions to a reviewable plan.
01
Environmental profile
Land-cover history, climate, terrain, soils and nearby native vegetation, with source and resolution noted.
02
Possible restoration approaches
Options to consider according to the ecosystem, project goals and observed constraints — with assumptions clearly stated.
03
Uncertainties and field checks
Missing local evidence, questions for ecologists and observations required before committing to a plan.
05 — TECHNOLOGY & SCIENCEEVIDENCE BEFORE PREDICTIONS
AI should explain
what it knows.
Arbor will combine environmental data with ecological rules and AI-assisted analysis. Machine learning could help identify patterns where suitable training data exist; it cannot establish ecological outcomes that have never been measured.
The first priority is a reproducible environmental assessment. Predictive recommendations will require appropriate field evidence and independent testing.
Environmental evidence
Land cover and land-use history from MapBiomas; satellite observations from Copernicus; soil estimates from SoilGrids; and climate information from documented datasets.
Biome-sensitive reasoning
One technical core will process environmental layers, while biome-specific ecological knowledge helps define which interventions are plausible and which assumptions require checking.
Transparent uncertainty
Satellite observations cannot replace soil measurements, species surveys or local expertise. Arbor is intended to flag missing information instead of presenting unsupported precision.
Validation before claims
Project histories, independent ecologists and field pilots will be needed to test the system. Model performance, transferability across biomes and practical value are still to be demonstrated.
BRAZIL IS OUR STARTING POINTThree biomes.
Three biomes.
Millions of possible
better decisions.
Our first MVP focuses on the Amazon, Cerrado and Atlantic Forest. Three distinct ecosystems in which to develop, test and validate Arbor’s approach — before expanding to other landscapes.
Be part of the journey ↗AMAZON · CERRADO · ATLANTIC FORESTBRAZILIAN LANDSCAPE PHOTOGRAPHY · UNSPLASH
06 — DEVELOPMENT & COLLABORATIONIN DEVELOPMENT
Built with people
who know the land.
Arbor is an initiative of GreenTrust. The planned users are restoration project developers, environmental consultancies and organisations managing restoration projects. We are seeking ecological expertise, data access and conversations with potential pilot partners to shape and test the MVP.
Current priorities are the environmental data pipeline, the first ecological decision rules and an evidence-based validation plan. No operational product, field results or model accuracy are claimed yet.
GreenTrust on LinkedIn ↗Interested in a future pilot or scientific collaboration? Reach out to the team.