Technical Annex · Companion to the buyer walkthrough

Digital Site Analysis
2026-07-04

Twelve satellite datasets, four DEMs cross-validated, twenty GPS points from Wes's 2026-06-28 walk, and a year of Sentinel-2 imagery — synthesized into one document. This is what the data says about the 62–71 ha parcel.

Sources: ALOS AW3D30 · Hansen GFC · MapBiomas PY 2023 · Sentinel-2 L2A · Meta Tolan CHM · JRC GSW · CHELSA · SoilGrids · GPS walk

Overview

Eight numbers that define the property.

The data tells a coherent story — a buildable, forested, water-rich parcel that hasn't been logged for a generation. Each number below is sourced and verifiable; click "Downloads" at the bottom for the raw data.

79%
Buildable
Of terrain is flat or moderate slope — cabin-friendly.
90%
Forest cover
Mature Atlantic-Forest interior, virtually untouched.
142m
Total relief
From ridge to quebrada — explains the cascading water.
1,776mm
Annual rain
Cfa subtropical — no dry month below 60mm.
21°C
Mean temp
Frost-free year-round. Sub-tropical climate.
19.5%
Topsoil clay
Perfect clay-sand ratio for cob construction.
0.7%
Forest loss
2001–2024 — virtually zero recent clearing.
274m
GPS waterfall
Confirmed altitude of the GPS-marked cascade.

Interactive map

All layers toggleable. Click any feature for details.

MapLibre · 30m native · click for popups
GPS-confirmed waterfall
Waterfall candidate (top 50)
MapBiomas wetland
Tree (color = height)
Rock estimate
Quebrada / stream (flow acc)
GPS parcel boundary
Quebrada corridor (valley low)

The headline finding

Fifty candidate waterfalls. One is confirmed.

The DEM step-detection found fifty places where elevation drops sharply within a 30m pixel neighborhood. Most are small cascades. The top five are the ones worth verifying on the W1.2 site visit.

#1 · 28m drop

Center-east cascade

-57.019, -25.625 · 287m

Largest single-step drop in the DEM. If real, this is a major waterfall — likely a 25–30m free fall into a plunge pool. The site visit should start here.

#2 · 27m drop

East-central cascade

-57.017, -25.627 · 235m

Second largest. Likely a multi-tier cascade rather than a single drop — the DEM detects 27m of elevation loss across two pixels, which is a 3D world of small falls and pools.

#3 · 25m drop

East ridge cascade

-57.019, -25.625 · 275m

Third largest. Same cluster as #1 — probably part of the same quebrada system. The cluster might be one major waterfall with several visible drops from different angles.

CONFIRMED

⭐ GPS waterfall (Wes walked)

-57.026, -25.607 · 274m altitude

The only waterfall confirmed by physical walk. Altitude 274m is at the rim; the drop itself is somewhere between 5 and 90m. Measure on the next site visit.

#5 · 23m drop

East escarpment

-57.020, -25.625 · 207m

Lower elevation than the top three — would be accessible from the quebrada corridor. Could be the most photogenic cascade since it's lower and surrounded by forest.

#6–50

45 more candidates

Throughout DEM coverage

Smaller cascades and rapids, mostly 5–15m drops. They aren't showstoppers individually, but together they create the sound of moving water throughout the quebrada — a feature in its own right.

The terrain

What the shape of the land tells you.

Slope and aspect (which way the land faces) determine where to put a house, where to put a road, and where the sun rises over breakfast. Most of this parcel faces east — favorable for both morning light and afternoon shade.

LayerValueSourceConfidence
Elevation range121–263m (142m relief)ALOS AW3D30 30mHigh
Slope — flat (<5%)29.7% (81.8 ha)DEM-derivedHigh
Slope — moderate (5–15%)49.5% (136.4 ha)DEM-derivedHigh
Slope — steep (15–30%)13.2% (36.4 ha)DEM-derivedHigh
Slope — very steep (>30%)7.7% (21.2 ha) — not buildableDEM-derivedHigh
Buildable (flat + moderate)79.2% (218 ha)DEM-derivedHigh
Dominant aspectEast (22.9%), SE (13.6%), NE (12.6%)DEM-derivedHigh
Tree cover (Hansen 2000)90.7% high-canopyHansen GFC v1.11High
Tree cover (Hansen 2024)90.0% (0.7% loss 2001–24)Hansen GFC v1.11High
Mean canopy height3.5m, max 24.5m (emergent)Meta Tolan 2024 CHM (10m)High
Sentinel-2 NDVI (12 dates)0.728–0.825, mean 0.773Sentinel-2 L2A 10mHigh
2022 La Niña NDVI0.770–0.809 (stable!)Sentinel-2 L2A 10mHigh
Topsoil clay content19.5% (perfect for cob)SoilGrids 2.0 (250m)Medium
Soil pH (topsoil)5.40 (slightly acidic)SoilGrids 2.0 (250m)Medium

All data layers

One hundred and thirty-two research results, indexed.

DomainDocumentsHighlights
Market sizing6 resultsEU expat pool 30–50K in PY, no direct cob-housing competitors
Materials9 resultsCob + bamboo local supply chain confirmed, no imports needed
Permits + legal11 results$15–25K total Phase 1 permit cost (PY government sources)
Construction8 resultsCob technique available locally, mix designs documented
Climate + environment12 resultsCfa subtropical, 1,776mm rain, drought-resilient forest
Biodiversity14 results300 GBIF species records, Atlantic-Forest interior biome
Reforestation7 resultsNative nursery capacity 5–10K seedlings/year in PY
Energy + utilities10 resultsOff-grid solar viable, ANDE grid ~2km, backup for cloudy days
Water + septic8 resultsQuebrada year-round, septic standard PY regs
Labor + staffing7 resultsLocal construction labor, PY + EU dual staffing for tourism
Tax + accounting6 resultsPY tax regime favorable for tourism, 4-BV cascade options
Freight + logistics5 resultsAsunción port 2h, freight rail + river options
Tourism + marketing9 resultsEU traveler psychographics, Booking.com + Airbnb viable
Insurance + risk5 resultsFire + weather + political risk profile low
Certifications5 resultsEU eco-tourism certs (EU Ecolabel, Green Key) attainable
Financial modeling10 resultsPhase 1 capex $180–250K, 80% self-fundable after Phase 1

⚠️ Important coverage note

The current DEM (ALOS AW3D30) covers lat −25.645 to −25.615 — the southern portion of the GPS polygon (which spans lat −25.616 to −25.602). The northern half is not yet analyzed for slope, aspect, or buildable zones. The GPS-confirmed waterfall (lat −25.607) and the 24 MapBiomas wetland pixels (lat −25.606) sit in this uncovered area.

To fix: acquire new DEM tiles covering lat −25.605 to −25.620 (Sprint 1 task, $0 cost, ~1 day work). Until then, the southern-half numbers above are accurate for that area only.

Downloads

Raw data, every format.

Every analysis above is reproducible. Click any link to download the source data — QGIS-importable, ArcGIS-importable, or spreadsheet-friendly.

Next step

This page is for the technical. The story is in the main walkthrough.

The buyer-facing page has the emotional narrative — the cob renderings, the 3D flythrough, the "come see it" CTA. This page is the audit trail behind those claims.

← Back to buyer walkthrough Read the canonical repo