Public reading
100 ha · 9.951, -84.157 · Costa Rica
Good bones. The rest is work.
61st percentile for like land · Frontier tier · 18 conservation priority
19 of 30 sources answered, each named and dated ↓ · higher is better on every pillar; pressure reads calm · silent: osm_overpass, sentinel2_ndvi, sentinel1_rtc…
Three things to know
Rainfall is very high.
3,030 mm of rain per year.
clear sky
A meaningful share of this land's tree cover has been cleared since 2000.
About 5% of the parcel lost tree cover since 2000, most recently in 2011.
clear sky
No satellite soil data is available for this parcel.
The SoilGrids survey returned no data for this location.
in fog
First move · before committing
Commission a ground soil survey before any agricultural or restoration decision.
Take samples across the parcel for a lab test of organic carbon, texture, and pH, and dig a pit to read the profile and rooting depth.
then: Confirm tenure and boundaries against the legal record before relying on this reading.
One measurement per instrument, each drawn to its own scale. Hover a mark for the reading behind it.
Rain · 21 years
Up 32.3% over twenty years. Open-Meteo.
The ground and its cover
ESA WorldCover. It cannot tell natural forest from plantation.
Elevation & slope
950–1,011 m across the parcel. AWS terrain tiles.
Climate to 2050
From today's 21.0 °C. Expect is the honest word: this is a projection.
Species records
Log scale. A count this size also measures survey effort, not only life. GBIF.
Access
OpenStreetMap.
Where the parcel sits between securing it and hosting people on it, and the project that fits it best.
Best-fit model
Regenerate working land into productive canopy.
Standalone conservation value, outside the mapped network.
Drawing the film…
The film
The place, the score, the five pillars and three findings, set over contour lines traced from this parcel’s own elevation. Every figure on screen is this reading’s; the last finding is the one it is least sure of. Nothing in it is generated.
vertical · drawn in your browser from this reading
Go deeper
Everything the instruments read about this land, in its own words: the narrative, the maps, the evidence card by card, the 2050 window, what to do first, an analyst to ask, and all 30 sources, each named and dated.
The reading in full
This 100-hectare parcel near San Rafael, Escazú, San José is land with sound foundations and clear regenerative upside. It scores 64 out of 100 on the Dear Wise Earth Land Score.
The risks that most shape a decision here: A meaningful share of this land's tree cover has been cleared since 2000. Long-term loss points to historical or ongoing clearing pressure and a lower conservation baseline; confirm the legal status and whether clearing is still active. This ground has a recorded seismic history. Repeated strong earthquakes within reach shape how anything here should be built — seismic design codes, slope stability under shaking, and insurance all follow from it.
The wider readings: Models project a warmer and drier climate here by 2050. Expect a warmer and drier future here; favour species and infrastructure that tolerate the projected range. Rainfall is very high. Abundant water supports productive growth but raises the risk of flooding, waterlogging, and nutrient leaching, so drainage and slope matter. This parcel is a position in a watershed, not an island. What arrives here is decided upstream, and what happens here is felt downstream — land use on this parcel is water stewardship for everyone below it. Read the upstream land uses before trusting the water, and count the downstream neighbours before changing it. The land is gently rolling. Gentle terrain is generally workable with only minor grading for buildings or roads. The largest buildable bench is 70.95 ha. This is the piece that decides what fits: a single footprint, its access, and its services all have to land on contiguous ground. Everything beyond it is a second site, not an extension of the first. The area has a rich record of documented wildlife. High biodiversity adds conservation value, may bring species-protection rules, and can support eco-tourism or habitat-credit programs. No satellite soil data is available for this parcel. Soil quality cannot be assessed remotely here, so a ground survey is needed before any agricultural or restoration decision. The human fabric around this land is measured, not assumed. Labour, emergencies, schooling, and electricity all run on these distances — they decide what daily life and operations here actually cost. Jurisdiction is part of the parcel. Permits, zoning, water, and roads run through these offices — knowing who governs the land is the first step of every plan that touches it. Official Costa Rican record places this in Pavas, San José, San José. National cadastral and administrative data ground the parcel in the legal record — the basis for title, permitting, and tax diligence.
What the land is telling us: The longer record suggests a land that has been drawn down rather than held: about 5% of its tree cover has been cleared since 2000, most recently in 2011. It reads as opened rather than recovered, but the history is the reason the foundational work comes first. Of the land's core processes, the water cycle reads as adequate to abundant, so the question is less supply than how well the land captures and holds what falls and community dynamics look rich, with a deep record of documented life nearby. What orbit cannot see is what happens below the surface — soil biology, structure, and the carbon held in the ground — so treat those as the first things to confirm underfoot. Taken together, this is land with sound foundations and clear regenerative upside.
The ground and its cover
The verdict, the five pillar readings, how much of the land was read, and the headline figures.
The verdict
San Rafael, Escazú, San José
61st percentile for like land · Frontier tier · 18 conservation priority
Four of five readings
Weakest: EcologyNot read: Soil
All five read on one axis: higher is better. Pressure scores calm, so a high number means little human pressure on this land. The unread pillars had their weight redistributed across the rest — reweighted, never guessed.
Standing
Subtropical · Neotropic · against 38 readings · 62nd of all 124
Coverage
19 of 30 layers answered
Nearly complete — gaps reweighted, not guessed.
Silent: the river and road trace, Sentinel-2, the radar +8.
Where the capital goes
Frontierhospitality fit 49
Best-fit model
Regenerative agroforestry estate
Fit from published rules over the indicators.
The land in numbers
Headline figures. The Why tab carries each reading in full, with its source and its date. “Not read” means the source had nothing for this parcel: a gap is named, never filled with a guess.
How this land works
About 3,030 mm of rain falls here in a year. 20% of the parcel is under tree cover, and the ground is nearly flat (3° on average), so water lingers rather than runs. By 2050 the models expect +0.9 °C and -6% rain, so every flow above runs a little differently than it does today.
Holding — the canopy record shows no meaningful loss in a generation.
Regenerate working land into productive canopy. Gentle working ground (0% already cropland) with 3,030 mm of rain a year — well set up to convert into multi-strata agroforestry that carries yield and canopy on the same hectare. To help: stop the clearing at the streams first: every metre of canopy kept along water holds the loop together.
The score is the footnote. What matters is the pattern above it: what arrives, what the ground does with it, what leaves, and which way that is moving.
This place
100 hectares at San Rafael of rolling land around 988 m. The ground reads 20% under canopy, 48% built. Water shapes it — 3,030 mm of rain a year. 326,919 people within 5 km.
Regenerate working land into productive canopy. Gentle working ground (0% already cropland) with 3,030 mm of rain a year — well set up to convert into multi-strata agroforestry that carries yield and canopy on the same hectare.
This is an outsider's reading, assembled from instruments and public records. It is where understanding begins, not where it ends — the land, its neighbours, and the people who know it hold the rest.
Occurrence records (count + top species)
Recorded historic/heritage features within 10 km — generalized to counts, classes and distance bands; never coordinates
Named places and features of the surrounding landscape (titles + distances)
Administrative units, cadastral parcels, agrological land-capability, protected areas, Holdridge life zones (WFS)
Declared biological corridors (bundled screening reference of official corridors; approximate extents — confirm against SINAC/PNCB)
Dengue notifications by cantón 2013–2018 (suspected + probable + confirmed) + national series to 2024; bundled OpenDengue V1.3 extract — historical burden geography, not the current season
Nominatim reverse geocode
Terrarium elevation tiles
Archive 2004–2024 + MPI-ESM1-2-XR to 2050
WorldCover 2021 v200
Global Surface Water v1.4 — occurrence & seasonality
Global Forest Change (tree cover 2000 + annual loss)
Top-decile hotspots of predicted AM + ECM fungal richness (Van Nuland et al. 2025, Nature; Zenodo 10.5281/zenodo.14871588)
Per-cell slope and aspect (Horn), buildable share, largest contiguous bench
Upstream basin + downstream flowpath (screening hydrography)
Gridded population 2020 (parcel + 5 km ring sums)
M ≥ 5 events within 100 km, past 50 years
Last burn year and burn count per 500 m pixel over the record since 2001
Administrative hierarchy + structured facts of the governing unit
Every number above is public, open data — named, dated, and licensed. What TERRA does not publish is the lens: the weights, curves, and verdict thresholds that turn these layers into a Land Score. That judgment is calibrated from land walked on the ground, and it stays proprietary.
Unavailable for this parcel: the river and road trace, Sentinel-2, the radar, the infrastructure reach, osm adaptation, the soil survey, the governance record, the thermal read, the biomass map, the hazard screen, the territory register — the affected pillars were reweighted, not guessed.
Generated Sun, 20 Sep 2026 22:35:22 GMT · 19.1s · synthesized by TERRA.