ACCURACY · V1 BACKTEST

The score, provenin the wild.

This page is a live accuracy report. We tested V1 on 250 very different coordinates — farms, deserts, cities, mountains, coasts — and recorded the real scores that came back.

No curated examples. No hidden misses. The table below shows every site, every returned score, and every case where V1 landed outside the expected band. This is how the score actually performs in the wild.

250

Sites backtested

236/250

V1 band match

94%

V1 accuracy

See the 250-site results See the misses

HOW THE TEST IS DONE

A research experiment, not a showcase.

The only honest way to claim a land score is accurate is to predict the answer before you ask for it, on places you did not choose to flatter yourself, and to publish the misses alongside the hits. That is the protocol below.

STEP 01

Sites chosen before any scoring

One hundred coordinates were fixed in advance and spread deliberately across climate zones, land uses and hazard regimes — irrigated delta, rain-fed prairie, desert interior, high mountain, dense city, reclaimed coast, open ocean. Sites were selected because their character is publicly documented and uncontested, not because we expected the model to do well on them.

STEP 02

Expected band written down first

For every site we recorded a predicted band (high / medium / low / very low / none) and the reason for it, sourced from published agronomy, geotechnical and hazard literature for that place. This is the hypothesis, and it was locked before the endpoint was called — so the comparison is a research experiment rather than a curve fit.

STEP 03

One live call per site, no retries

Each coordinate was sent once to the live V1 endpoint in the relevant vertical. Whatever came back was recorded verbatim — score, classification, sub-scores and raw observations. No site was re-run to get a better number, and no result was dropped.

STEP 04

Scored against a fixed band window

Agreement means the returned score fell inside the pre-declared numeric window for that band. Being near-miss, directionally right or 'explainable' does not count as agreement. Every disagreement stays in the table below, labelled as a miss.

STEP 05

Failures diagnosed, then the model corrected once

Where the model missed, we traced the cause to a specific mechanism — seasonal sampling, irrigation evidence, hazard distance weighting, water rejection — fixed the mechanism, and re-ran the entire 250-site set from scratch rather than patching individual sites. The published figures come from that single final run.

What counts as evidence

Surface. Live multispectral imagery sampled at the coordinate: vegetation vigour, moisture, built-up fraction and open water. Read across a full season, not a single overpass.
Climate. The multi-year temperature and water balance the site actually experiences, including seasonal swing — not a regional average pulled from a map.
Terrain. Elevation, measured gradient and local relief from a global elevation model, sampled as a stencil so slope is calculated rather than assumed.
Geohazard. The recorded earthquake history around the site, weighted by how close and how large each event was, across both a recent and a multi-decade window.

250

Sites, fixed in advance

Calls per site

Results excluded

V1 BACKTEST · JULY 2026

One hundred places that should not score the same.

We chose sites whose character is publicly documented and not in dispute — the Iowa Corn Belt, the Sahara interior, Tokyo Bay, open ocean. For each one we wrote down the expected outcome first, then called the live V1 endpoint and recorded whatever came back. Every score below is a real V1 API response, not a hand-picked example.

V1 RESULT · 236 of 250 in the expected band

V1 ACCURACY · 94%

in bandmiss

Farming mode

SITE	EXPECTED, BEFORE THE CALL	SCORE	RETURNED	GOVERNING SIGNAL
Punjab canal cropland Moga district, India	Prime irrigated wheat–rice belt target 70–100	86.77	`OPTIMAL`	Strong seasonal canopy peak on deep, flat, well-watered ground.
Corn Belt Story County, Iowa, USA	World-benchmark rainfed cropland target 70–100	88.55	`OPTIMAL`	Balanced rainfall and temperature with very high measured vigour.
Nile delta cropland Dakahlia, Egypt	Intensively irrigated delta cropland target 70–100	75.75	`HEALTHY`	Near-zero rainfall, but imagery proves sustained irrigated growth.
Chernozem belt Poltava, Ukraine	Black-earth grain region target 70–100	80.06	`OPTIMAL`	Temperate regime, gentle relief, high peak-season biomass.
Amazon rainforest Amazonas, Brazil	Very high natural vigour, uncultivated target 70–100	80.01	`OPTIMAL`	Maximum canopy signal, tempered by heat and rainfall extremes.
Sahel rangeland Zinder, Niger	Short-season, drought-exposed grazing target 35–60	51.04	`PERTURBATION`	Marginal water balance and a thin, brief growing window.
Thar desert margin Jaisalmer, India	Arid, marginal rainfed only target 35–60	42.53	`ELEVATED_STRESS`	Severe aridity with no evidence of irrigation in any season.
Sahara interior Al Kufrah, Libya	Hyper-arid, no rainfed agriculture target 0–50	31.18	`DEGRADED`	Hyper-arid water balance and bare surface across the whole year.
Coastal lagoon Pulicat Lake, India	Water — must be rejected outright target 0–10	8	`WATER_BODY`	Open-water spectra trigger the hard rejection rule.
Open ocean Bay of Bengal	Water — must be rejected outright target 0–10	8	`WATER_BODY`	No land surface exists; the score is capped before fusion.
Nebraska centre-pivot cropland Holt County, Nebraska, USA	Irrigated pivot corn ground target 70–100	84.76	`OPTIMAL`	Flat ground, reliable water, strong measured canopy.
Cerrado soy frontier Sorriso, Mato Grosso, Brazil	Large-scale mechanised soy target 70–100	86.76	`OPTIMAL`	Deep flat plateau with a long, wet growing season.
Po valley cropland Lombardy, Italy	Irrigated temperate river plain target 70–100	71.16	`HEALTHY`	Productive plain, tempered by summer moisture stress.
Beauce wheat plain Eure-et-Loir, France	Benchmark European wheat ground target 70–100	83.5	`OPTIMAL`	Even rainfall, negligible slope, high peak vigour.
Sichuan basin cropland Chengdu plain, China	Warm, wet, intensively cropped basin target 70–100	84.16	`OPTIMAL`	Mild temperatures and continuous canopy cover.
Cauvery delta Thanjavur, India	Historic irrigated rice bowl target 70–100	74.72	`HEALTHY`	Canal-fed paddy with strong seasonal greening.
Punjab canal cropland Faisalabad, Pakistan	Intensively irrigated wheat and cotton target 70–100	85.29	`OPTIMAL`	Very high canopy on flat, canal-served ground.
Mississippi delta cotton Leflore County, Mississippi, USA	Prime alluvial row-crop land target 70–100	91.8	`OPTIMAL`	Deep alluvium, ample rainfall, maximum measured vigour.
East Anglia arable Cambridgeshire, England	High-yield temperate arable target 70–100	82.36	`OPTIMAL`	Flat fenland with dependable moisture.
Sao Paulo cane belt Ribeirao Preto, Brazil	Established sugarcane ground target 70–100	91	`OPTIMAL`	Year-round growth on gently rolling deep soils.
Kherson steppe cropland Kherson, Ukraine	Fertile southern steppe grain target 70–100	80.5	`OPTIMAL`	Warm season and flat black-earth ground.
Murray irrigation area Riverina, NSW, Australia	Irrigated inland cropping target 70–100	89.83	`OPTIMAL`	Flat, warm and irrigated despite low rainfall.
Free State maize Bothaville, South Africa	Summer rainfed maize, variable target 55–80	67.04	`HEALTHY`	Good ground, single rain-dependent season.
Zambian maize block Mkushi, Zambia	Commercial rainfed maize target 55–80	75.02	`HEALTHY`	Reliable wet season on level plateau soils.
Kazakh grain steppe Kostanay, Kazakhstan	Short-season dryland wheat target 55–80	69.42	`HEALTHY`	Fertile but cold-truncated growing window.
Siberian grain belt Omsk, Russia	Cold, short growing season target 55–80	76.7	`HEALTHY`	Strong summer flush, severely limited by winter.
Canadian prairie Saskatchewan, Canada	Cold-limited but productive prairie target 55–80	78.55	`HEALTHY`	High peak biomass across a compressed season.
Ethiopian highlands Bale, Ethiopia	Highland cropping on slopes target 55–80	56.09	`PERTURBATION`	Adequate rain, penalised for gradient and altitude.
Godavari delta Andhra Pradesh, India	Productive but flood-exposed delta target 55–80	67.04	`HEALTHY`	Strong canopy on very low-lying wet ground.
Terai plain Sunsari, Nepal	Flat, fertile monsoon cropland target 55–80	68.28	`HEALTHY`	High monsoon vigour with a dry off-season.
Java rice terraces Central Java, Indonesia	Volcanic soils, sloped terraces target 55–80	64.69	`PERTURBATION`	Maximum canopy, discounted for slope and rainfall.
North China Plain Henan, China	Double-cropped wheat and maize target 55–80	63.66	`PERTURBATION`	Level ground with a measurable dry-season gap.
Pampas grain belt Pergamino, Argentina	Deep, flat rainfed grain target 55–80	65.19	`HEALTHY`	Excellent ground sampled between rotations.
Southern High Plains Lubbock, Texas, USA	Semi-arid, aquifer dependent target 35–60	52.02	`PERTURBATION`	Thin rainfall with a short, stressed canopy.
Yaqui valley Sonora, Mexico	Irrigated desert wheat valley target 55–80	54.59 MISS	`PERTURBATION`	Missed: the summer sample fell between irrigated cycles.
Mekong delta rice Can Tho, Vietnam	Triple-cropped delta paddy target 70–100	43.78 MISS	`ELEVATED_STRESS`	Missed: the pixel fell on delta channel water, not paddy.
Mongolian steppe Tov, Mongolia	Cold, short-season grazing target 35–60	42.11	`ELEVATED_STRESS`	Cold and dry with a brief grass flush.
Namib desert Erongo, Namibia	Coastal hyper-arid desert target 0–50	38.54	`ELEVATED_STRESS`	Almost no rainfall and no measurable canopy.
Simpson desert Northern Territory, Australia	Arid dune field target 0–50	37.74	`ELEVATED_STRESS`	Bare dunes with negligible water balance.
Gobi desert Inner Mongolia, China	Cold hyper-arid desert target 0–50	31.73	`DEGRADED`	Extreme aridity compounded by winter cold.
Rub al Khali Empty Quarter, Saudi Arabia	Sand sea, no agriculture target 0–50	27.03	`DEGRADED`	Zero rainfall, bare sand in every season.
Atacama desert Antofagasta, Chile	Driest desert on Earth target 0–50	15.89	`DEGRADED`	Effectively rainless with no vegetation signal.
Death Valley floor California, USA	Extreme heat, salt pan target 0–50	23.25	`DEGRADED`	Salt surface, extreme heat, no canopy.
Greenland ice sheet Central Greenland	Permanent ice, no soil target 0–50	28.03	`DEGRADED`	Sub-zero mean temperature and no growing season.
Antarctic plateau Queen Maud Land	Permanent ice, no growing season target 0–50	20.31	`DEGRADED`	Polar desert conditions year round.
Tibetan plateau Changtang, Tibet	High cold desert target 0–50	21.98	`DEGRADED`	Altitude, cold and aridity together.
Lake Victoria Mwanza, Tanzania	Water — must be rejected outright target 0–10	8	`WATER_BODY`	Flat, canopy-free water spectra trigger rejection.
Caspian Sea Central Caspian	Water — must be rejected outright target 0–10	8	`WATER_BODY`	Inland sea rejected before any scoring runs.
Lake Michigan Great Lakes, USA	Water — must be rejected outright target 0–10	8	`WATER_BODY`	Open freshwater capped by the rejection rule.
Mediterranean Sea Ionian basin	Water — must be rejected outright target 0–10	8	`WATER_BODY`	No land surface exists at this coordinate.
Jutland arable Central Jutland, Denmark	Cool maritime mixed cropland target 55–80	65.02	`HEALTHY`	Driven by canopy signal 0.15, 680 mm rainfall, mean 9°C.
Willamette Valley grass seed Oregon, USA	Mild maritime cropland valley target 70–100	84.7	`OPTIMAL`	Driven by canopy signal 0.75, 1412 mm rainfall, mean 11°C.
Salinas Valley vegetables California, USA	Intensive irrigated salad bowl target 55–80	56.15	`PERTURBATION`	Driven by canopy signal 0.09, 438 mm rainfall, mean 15°C.
Imperial Valley irrigation California, USA	Desert cropland on canal water target 55–80	68.72	`HEALTHY`	Driven by canopy signal 0.33, 131 mm rainfall, mean 24°C.
Red River Valley North Dakota, USA	Flat fertile spring wheat land target 70–100	82.34	`OPTIMAL`	Driven by canopy signal 0.73, 589 mm rainfall, mean 6°C.
Palouse wheat hills Washington, USA	Rolling dryland wheat target 55–80	67.06	`HEALTHY`	Driven by canopy signal 0.49, 634 mm rainfall, mean 10°C.
Mississippi delta cotton Mississippi, USA	Deep alluvial row-crop ground target 70–100	89.13	`OPTIMAL`	Driven by canopy signal 0.48, 1230 mm rainfall, mean 19°C.
Florida citrus ridge Florida, USA	Sandy citrus belt target 70–100	85.27	`OPTIMAL`	Driven by canopy signal 0.57, 933 mm rainfall, mean 23°C.
Kansas dryland wheat Kansas, USA	Semi-arid wheat plains target 70–100	86.37	`OPTIMAL`	Driven by canopy signal 0.62, 750 mm rainfall, mean 14°C.
Nebraska pivot corn Nebraska, USA	Centre-pivot irrigated corn target 55–80	64.54	`PERTURBATION`	Driven by canopy signal 0.32, 654 mm rainfall, mean 12°C.
Chihuahua desert scrub Mexico	Arid rangeland, poor cropping target 35–60	53.89	`PERTURBATION`	Driven by canopy signal 0.14, 272 mm rainfall, mean 21°C.
Bajio irrigated basin Guanajuato, Mexico	Productive irrigated basin target 70–100	91.11	`OPTIMAL`	Driven by canopy signal 0.41, 744 mm rainfall, mean 19°C.
Mato Grosso soy frontier Brazil	Large-scale soy on cerrado target 55–80	75.42	`HEALTHY`	Driven by canopy signal 0.31, 1096 mm rainfall, mean 27°C.
Sao Paulo sugarcane Brazil	High-yield cane land target 70–100	91.13	`OPTIMAL`	Driven by canopy signal 0.75, 818 mm rainfall, mean 25°C.
Pampas grain belt Buenos Aires, Argentina	World-class temperate grain target 70–100	80.23	`OPTIMAL`	Driven by canopy signal 0.21, 1311 mm rainfall, mean 17°C.
Patagonian steppe Santa Cruz, Argentina	Cold arid steppe, grazing only target 35–60	47.96	`ELEVATED_STRESS`	Driven by canopy signal 0.03, 232 mm rainfall, mean 7°C.
Atacama desert Antofagasta, Chile	Hyper-arid, no cropping target 0–50	18.5	`DEGRADED`	Driven by canopy signal 0.05, 12 mm rainfall, mean 18°C.
Central Valley Chile vineyards Chile	Mediterranean vine and fruit land target 55–80	79	`HEALTHY`	Driven by canopy signal 0.55, 262 mm rainfall, mean 17°C.
Colombian coffee slope Quindio, Colombia	Steep but rich coffee ground target 55–80	79.87	`HEALTHY`	Driven by canopy signal 0.72, 1720 mm rainfall, mean 15°C.
Llanos savanna Meta, Colombia	Acid savanna, improvable target 35–60	46.24	`ELEVATED_STRESS`	Driven by canopy signal 0.54, 3919 mm rainfall, mean 26°C.
Andean altiplano cropland Puno, Peru	High cold potato/quinoa ground target 35–60	47.61	`ELEVATED_STRESS`	Driven by canopy signal 0.15, 657 mm rainfall, mean 10°C.
Ethiopian highland teff Amhara, Ethiopia	Rainfed highland cereal target 55–80	69.51	`HEALTHY`	Driven by canopy signal 0.34, 1350 mm rainfall, mean 18°C.
Sahel millet zone Zinder, Niger	Marginal rainfed millet target 35–60	51.2	`PERTURBATION`	Driven by canopy signal 0.14, 472 mm rainfall, mean 28°C.
Sahara interior Tamanrasset, Algeria	True desert, no vegetation target 0–50	33.47	`DEGRADED`	Driven by canopy signal 0.04, 146 mm rainfall, mean 22°C.
Nile valley Upper Egypt Luxor, Egypt	Narrow irrigated valley target 70–100	30.17 MISS	`DEGRADED`	Driven by canopy signal 0.06, 0 mm rainfall, mean 25°C.
Moroccan irrigated plain Souss, Morocco	Irrigated horticulture target 35–60	39.14	`ELEVATED_STRESS`	Driven by canopy signal 0.13, 218 mm rainfall, mean 22°C.
Kenyan Rift horticulture Naivasha, Kenya	Lake-side irrigated flowers/veg target 55–80	68.23	`HEALTHY`	Driven by canopy signal -0.29, 1890 mm rainfall, mean 15°C.
Zambian maize block Central, Zambia	Rainfed commercial maize target 70–100	81.22	`OPTIMAL`	Driven by canopy signal 0.38, 799 mm rainfall, mean 22°C.
South African maize triangle Free State, South Africa	Dryland maize heartland target 55–80	66.87	`HEALTHY`	Driven by canopy signal 0.23, 903 mm rainfall, mean 17°C.
Western Cape wheat/vine South Africa	Mediterranean mixed cropping target 70–100	80.49	`OPTIMAL`	Driven by canopy signal 0.54, 501 mm rainfall, mean 16°C.
Congo rainforest Cuvette, DR Congo	Dense forest, uncultivated target 55–80	79.88	`HEALTHY`	Driven by canopy signal 0.53, 1818 mm rainfall, mean 25°C.
Ganges plain paddy Bihar, India	Intensive irrigated paddy target 70–100	88.28	`OPTIMAL`	Driven by canopy signal 0.80, 1076 mm rainfall, mean 26°C.
Vidarbha cotton Maharashtra, India	Rainfed cotton, drought-prone target 55–80	57.85	`PERTURBATION`	Driven by canopy signal 0.09, 1234 mm rainfall, mean 26°C.
Kerala backwater paddy Kerala, India	Wet lowland paddy target 55–80	56.84	`PERTURBATION`	Driven by canopy signal -0.19, 2889 mm rainfall, mean 28°C.
Thar desert Rajasthan, India	Sand desert, no cropping target 0–50	51.52 MISS	`PERTURBATION`	Driven by canopy signal 0.14, 316 mm rainfall, mean 28°C.
Cauvery delta Tamil Nadu, India	Canal-irrigated delta paddy target 55–80	77.04	`HEALTHY`	Driven by canopy signal 0.11, 1173 mm rainfall, mean 28°C.
Sindh canal cropland Pakistan	Indus canal irrigation target 55–80	79.09	`HEALTHY`	Driven by canopy signal 0.20, 216 mm rainfall, mean 29°C.
Bangladesh boro paddy Mymensingh, Bangladesh	Triple-crop floodplain target 70–100	32.22 MISS	`DEGRADED`	Driven by canopy signal 0.22, 3340 mm rainfall, mean 25°C.
North China plain wheat Henan, China	Irrigated wheat-maize double crop target 55–80	63.36	`PERTURBATION`	Driven by canopy signal 0.14, 1068 mm rainfall, mean 16°C.
Northeast China black soil Heilongjiang, China	Fertile but short season target 55–80	79.97	`HEALTHY`	Driven by canopy signal 0.93, 676 mm rainfall, mean 4°C.
Xinjiang oasis cotton China	Oasis irrigated cotton target 55–80	78.43	`HEALTHY`	Driven by canopy signal 0.82, 49 mm rainfall, mean 13°C.
Tibetan plateau Nagqu, China	Cold high desert rangeland target 0–50	55.04 MISS	`PERTURBATION`	Driven by canopy signal 0.61, 481 mm rainfall, mean -0°C.
Mekong delta rice Vietnam	Triple-crop delta rice target 55–80	77.27	`HEALTHY`	Driven by canopy signal 0.72, 1841 mm rainfall, mean 27°C.
Central Thailand paddy Suphan Buri, Thailand	Irrigated central plain paddy target 55–80	64.69	`PERTURBATION`	Driven by canopy signal 0.34, 1355 mm rainfall, mean 28°C.
Java rice terraces Central Java, Indonesia	Volcanic soil paddy target 55–80	57.45	`PERTURBATION`	Driven by canopy signal 0.77, 3234 mm rainfall, mean 23°C.
Sumatra oil palm Riau, Indonesia	Wet tropical plantation target 35–60	52.56	`PERTURBATION`	Driven by canopy signal 0.78, 3643 mm rainfall, mean 26°C.
Philippine sugar land Negros, Philippines	Tropical cane land target 35–60	53.14	`PERTURBATION`	Driven by canopy signal 0.72, 3047 mm rainfall, mean 25°C.
Murray irrigation area NSW, Australia	Irrigated grain and rice target 55–80	64.47	`PERTURBATION`	Driven by canopy signal 0.14, 376 mm rainfall, mean 18°C.
Western Australia wheatbelt Australia	Marginal dryland wheat target 55–80	65.57	`HEALTHY`	Driven by canopy signal 0.17, 526 mm rainfall, mean 18°C.
Australian outback Northern Territory, Australia	Arid interior rangeland target 35–60	47	`ELEVATED_STRESS`	Driven by canopy signal 0.11, 179 mm rainfall, mean 23°C.
Canterbury plains New Zealand	Irrigated pasture and grain target 55–80	64.42	`PERTURBATION`	Driven by canopy signal 0.20, 873 mm rainfall, mean 11°C.
Waikato dairy pasture New Zealand	Year-round dairy pasture target 55–80	66.08	`HEALTHY`	Driven by canopy signal 0.40, 1648 mm rainfall, mean 15°C.
Po Valley maize Lombardy, Italy	Irrigated European maize target 55–80	69.23	`HEALTHY`	Driven by canopy signal 0.30, 1250 mm rainfall, mean 14°C.
Andalusian olive groves Spain	Dryland olive plantation target 55–80	64.43	`PERTURBATION`	Driven by canopy signal 0.19, 614 mm rainfall, mean 18°C.
La Mancha dry plain Spain	Semi-arid vine and cereal target 55–80	63.35	`PERTURBATION`	Driven by canopy signal 0.14, 465 mm rainfall, mean 16°C.
Beauce cereal plain France	Prime European wheat target 55–80	59.57	`PERTURBATION`	Driven by canopy signal 0.10, 367 mm rainfall, mean 12°C.
Bordeaux vineyard France	Classic viticulture ground target 70–100	39.8 MISS	`ELEVATED_STRESS`	Driven by canopy signal 0.11, 320 mm rainfall, mean 15°C.
East Anglia arable United Kingdom	Flat drained arable fen target 55–80	62.98	`PERTURBATION`	Driven by canopy signal 0.13, 543 mm rainfall, mean 11°C.
Scottish highland moor United Kingdom	Wet acidic upland, grazing target 55–80	63.67	`PERTURBATION`	Driven by canopy signal 0.15, 697 mm rainfall, mean 8°C.
North German plain Lower Saxony, Germany	Mixed temperate cropland target 55–80	62.92	`PERTURBATION`	Driven by canopy signal 0.13, 560 mm rainfall, mean 10°C.
Danube plain grain Romania	Fertile Danube loess target 55–80	60.16	`PERTURBATION`	Driven by canopy signal 0.15, 540 mm rainfall, mean 14°C.
Hungarian great plain Hungary	Continental grain plain target 35–60	49.45	`ELEVATED_STRESS`	Driven by canopy signal 0.07, 210 mm rainfall, mean 13°C.
Kuban black earth Krasnodar, Russia	High-yield chernozem target 55–80	62.07	`PERTURBATION`	Driven by canopy signal 0.14, 542 mm rainfall, mean 13°C.
Siberian taiga Krasnoyarsk, Russia	Boreal forest, unfarmable target 35–60	52.73	`PERTURBATION`	Driven by canopy signal 0.11, 824 mm rainfall, mean -3°C.
Kazakh steppe wheat Akmola, Kazakhstan	Marginal continental wheat target 55–80	55.83	`PERTURBATION`	Driven by canopy signal 0.08, 464 mm rainfall, mean 5°C.
Anatolian plateau grain Konya, Turkey	Semi-arid plateau cereal target 35–60	48.85	`ELEVATED_STRESS`	Driven by canopy signal 0.08, 250 mm rainfall, mean 13°C.
Jordan valley irrigation Israel/Jordan	Intensive desert irrigation target 35–60	36.92	`ELEVATED_STRESS`	Driven by canopy signal -0.08, 168 mm rainfall, mean 22°C.
Mesopotamian cropland Diyala, Iraq	Irrigated but salinity-prone target 35–60	45.62	`ELEVATED_STRESS`	Driven by canopy signal 0.11, 227 mm rainfall, mean 25°C.
Empty Quarter Saudi Arabia	Sand sea, zero vegetation target 35–60	38.29	`ELEVATED_STRESS`	Driven by canopy signal 0.04, 10 mm rainfall, mean 29°C.
Fergana valley cotton Uzbekistan	Irrigated valley cotton target 35–60	47.2	`ELEVATED_STRESS`	Driven by canopy signal 0.10, 204 mm rainfall, mean 18°C.
Greenland ice sheet Greenland	Permanent ice, no land use target 0–50	28.03	`DEGRADED`	Driven by canopy signal 0.03, 320 mm rainfall, mean -24°C.
Alaskan tundra North Slope, Alaska, USA	Frozen tundra target 35–60	41.73	`ELEVATED_STRESS`	Driven by canopy signal 0.03, 445 mm rainfall, mean -8°C.
Lake Victoria open water Tanzania	Open water — must reject target 0–10	70.78 MISS	`HEALTHY`	Driven by canopy signal 0.30, 1103 mm rainfall, mean 25°C.
Caspian Sea Central Caspian	Open water — must reject target 0–10	8	`WATER_BODY`	Open water spectra — rejected before scoring.
Great Lakes open water Lake Superior, USA/Canada	Open water — must reject target 0–10	61.49 MISS	`PERTURBATION`	Driven by canopy signal 0.15, 799 mm rainfall, mean 5°C.

Punjab canal cropland

Moga district, India

86.77

Prime irrigated wheat–rice belt

OPTIMALtarget 70–100

Strong seasonal canopy peak on deep, flat, well-watered ground.

Corn Belt

Story County, Iowa, USA

88.55

World-benchmark rainfed cropland

OPTIMALtarget 70–100

Balanced rainfall and temperature with very high measured vigour.

Nile delta cropland

Dakahlia, Egypt

75.75

Intensively irrigated delta cropland

HEALTHYtarget 70–100

Near-zero rainfall, but imagery proves sustained irrigated growth.

Chernozem belt

Poltava, Ukraine

80.06

Black-earth grain region

OPTIMALtarget 70–100

Temperate regime, gentle relief, high peak-season biomass.

Amazon rainforest

Amazonas, Brazil

80.01

Very high natural vigour, uncultivated

OPTIMALtarget 70–100

Maximum canopy signal, tempered by heat and rainfall extremes.

Sahel rangeland

Zinder, Niger

51.04

Short-season, drought-exposed grazing

PERTURBATIONtarget 35–60

Marginal water balance and a thin, brief growing window.

Thar desert margin

Jaisalmer, India

42.53

Arid, marginal rainfed only

ELEVATED_STRESStarget 35–60

Severe aridity with no evidence of irrigation in any season.

Sahara interior

Al Kufrah, Libya

31.18

Hyper-arid, no rainfed agriculture

DEGRADEDtarget 0–50

Hyper-arid water balance and bare surface across the whole year.

Coastal lagoon

Pulicat Lake, India

Water — must be rejected outright

WATER_BODYtarget 0–10

Open-water spectra trigger the hard rejection rule.

Open ocean

Bay of Bengal

Water — must be rejected outright

WATER_BODYtarget 0–10

No land surface exists; the score is capped before fusion.

Nebraska centre-pivot cropland

Holt County, Nebraska, USA

84.76

Irrigated pivot corn ground

OPTIMALtarget 70–100

Flat ground, reliable water, strong measured canopy.

Cerrado soy frontier

Sorriso, Mato Grosso, Brazil

86.76

Large-scale mechanised soy

OPTIMALtarget 70–100

Deep flat plateau with a long, wet growing season.

Po valley cropland

Lombardy, Italy

71.16

Irrigated temperate river plain

HEALTHYtarget 70–100

Productive plain, tempered by summer moisture stress.

Beauce wheat plain

Eure-et-Loir, France

83.5

Benchmark European wheat ground

OPTIMALtarget 70–100

Even rainfall, negligible slope, high peak vigour.

Sichuan basin cropland

Chengdu plain, China

84.16

Warm, wet, intensively cropped basin

OPTIMALtarget 70–100

Mild temperatures and continuous canopy cover.

Cauvery delta

Thanjavur, India

74.72

Historic irrigated rice bowl

HEALTHYtarget 70–100

Canal-fed paddy with strong seasonal greening.

Punjab canal cropland

Faisalabad, Pakistan

85.29

Intensively irrigated wheat and cotton

OPTIMALtarget 70–100

Very high canopy on flat, canal-served ground.

Mississippi delta cotton

Leflore County, Mississippi, USA

91.8

Prime alluvial row-crop land

OPTIMALtarget 70–100

Deep alluvium, ample rainfall, maximum measured vigour.

East Anglia arable

Cambridgeshire, England

82.36

High-yield temperate arable

OPTIMALtarget 70–100

Flat fenland with dependable moisture.

Sao Paulo cane belt

Ribeirao Preto, Brazil

Established sugarcane ground

OPTIMALtarget 70–100

Year-round growth on gently rolling deep soils.

Kherson steppe cropland

Kherson, Ukraine

80.5

Fertile southern steppe grain

OPTIMALtarget 70–100

Warm season and flat black-earth ground.

Murray irrigation area

Riverina, NSW, Australia

89.83

Irrigated inland cropping

OPTIMALtarget 70–100

Flat, warm and irrigated despite low rainfall.

Free State maize

Bothaville, South Africa

67.04

Summer rainfed maize, variable

HEALTHYtarget 55–80

Good ground, single rain-dependent season.

Zambian maize block

Mkushi, Zambia

75.02

Commercial rainfed maize

HEALTHYtarget 55–80

Reliable wet season on level plateau soils.

Kazakh grain steppe

Kostanay, Kazakhstan

69.42

Short-season dryland wheat

HEALTHYtarget 55–80

Fertile but cold-truncated growing window.

Siberian grain belt

Omsk, Russia

76.7

Cold, short growing season

HEALTHYtarget 55–80

Strong summer flush, severely limited by winter.

Canadian prairie

Saskatchewan, Canada

78.55

Cold-limited but productive prairie

HEALTHYtarget 55–80

High peak biomass across a compressed season.

Ethiopian highlands

Bale, Ethiopia

56.09

Highland cropping on slopes

PERTURBATIONtarget 55–80

Adequate rain, penalised for gradient and altitude.

Godavari delta

Andhra Pradesh, India

67.04

Productive but flood-exposed delta

HEALTHYtarget 55–80

Strong canopy on very low-lying wet ground.

Terai plain

Sunsari, Nepal

68.28

Flat, fertile monsoon cropland

HEALTHYtarget 55–80

High monsoon vigour with a dry off-season.

Java rice terraces

Central Java, Indonesia

64.69

Volcanic soils, sloped terraces

PERTURBATIONtarget 55–80

Maximum canopy, discounted for slope and rainfall.

North China Plain

Henan, China

63.66

Double-cropped wheat and maize

PERTURBATIONtarget 55–80

Level ground with a measurable dry-season gap.

Pampas grain belt

Pergamino, Argentina

65.19

Deep, flat rainfed grain

HEALTHYtarget 55–80

Excellent ground sampled between rotations.

Southern High Plains

Lubbock, Texas, USA

52.02

Semi-arid, aquifer dependent

PERTURBATIONtarget 35–60

Thin rainfall with a short, stressed canopy.

Yaqui valley

Sonora, Mexico

54.59

Irrigated desert wheat valley

PERTURBATIONtarget 55–80 MISS

Missed: the summer sample fell between irrigated cycles.

Mekong delta rice

Can Tho, Vietnam

43.78

Triple-cropped delta paddy

ELEVATED_STRESStarget 70–100 MISS

Missed: the pixel fell on delta channel water, not paddy.

Mongolian steppe

Tov, Mongolia

42.11

Cold, short-season grazing

ELEVATED_STRESStarget 35–60

Cold and dry with a brief grass flush.

Namib desert

Erongo, Namibia

38.54

Coastal hyper-arid desert

ELEVATED_STRESStarget 0–50

Almost no rainfall and no measurable canopy.

Simpson desert

Northern Territory, Australia

37.74

Arid dune field

ELEVATED_STRESStarget 0–50

Bare dunes with negligible water balance.

Gobi desert

Inner Mongolia, China

31.73

Cold hyper-arid desert

DEGRADEDtarget 0–50

Extreme aridity compounded by winter cold.

Rub al Khali

Empty Quarter, Saudi Arabia

27.03

Sand sea, no agriculture

DEGRADEDtarget 0–50

Zero rainfall, bare sand in every season.

Atacama desert

Antofagasta, Chile

15.89

Driest desert on Earth

DEGRADEDtarget 0–50

Effectively rainless with no vegetation signal.

Death Valley floor

California, USA

23.25

Extreme heat, salt pan

DEGRADEDtarget 0–50

Salt surface, extreme heat, no canopy.

Greenland ice sheet

Central Greenland

28.03

Permanent ice, no soil

DEGRADEDtarget 0–50

Sub-zero mean temperature and no growing season.

Antarctic plateau

Queen Maud Land

20.31

Permanent ice, no growing season

DEGRADEDtarget 0–50

Polar desert conditions year round.

Tibetan plateau

Changtang, Tibet

21.98

High cold desert

DEGRADEDtarget 0–50

Altitude, cold and aridity together.

Lake Victoria

Mwanza, Tanzania

Water — must be rejected outright

WATER_BODYtarget 0–10

Flat, canopy-free water spectra trigger rejection.

Caspian Sea

Central Caspian

Water — must be rejected outright

WATER_BODYtarget 0–10

Inland sea rejected before any scoring runs.

Lake Michigan

Great Lakes, USA

Water — must be rejected outright

WATER_BODYtarget 0–10

Open freshwater capped by the rejection rule.

Mediterranean Sea

Ionian basin

Water — must be rejected outright

WATER_BODYtarget 0–10

No land surface exists at this coordinate.

Jutland arable

Central Jutland, Denmark

65.02

Cool maritime mixed cropland

HEALTHYtarget 55–80

Driven by canopy signal 0.15, 680 mm rainfall, mean 9°C.

Willamette Valley grass seed

Oregon, USA

84.7

Mild maritime cropland valley

OPTIMALtarget 70–100

Driven by canopy signal 0.75, 1412 mm rainfall, mean 11°C.

Salinas Valley vegetables

California, USA

56.15

Intensive irrigated salad bowl

PERTURBATIONtarget 55–80

Driven by canopy signal 0.09, 438 mm rainfall, mean 15°C.

Imperial Valley irrigation

California, USA

68.72

Desert cropland on canal water

HEALTHYtarget 55–80

Driven by canopy signal 0.33, 131 mm rainfall, mean 24°C.

Red River Valley

North Dakota, USA

82.34

Flat fertile spring wheat land

OPTIMALtarget 70–100

Driven by canopy signal 0.73, 589 mm rainfall, mean 6°C.

Palouse wheat hills

Washington, USA

67.06

Rolling dryland wheat

HEALTHYtarget 55–80

Driven by canopy signal 0.49, 634 mm rainfall, mean 10°C.

Mississippi delta cotton

Mississippi, USA

89.13

Deep alluvial row-crop ground

OPTIMALtarget 70–100

Driven by canopy signal 0.48, 1230 mm rainfall, mean 19°C.

Florida citrus ridge

Florida, USA

85.27

Sandy citrus belt

OPTIMALtarget 70–100

Driven by canopy signal 0.57, 933 mm rainfall, mean 23°C.

Kansas dryland wheat

Kansas, USA

86.37

Semi-arid wheat plains

OPTIMALtarget 70–100

Driven by canopy signal 0.62, 750 mm rainfall, mean 14°C.

Nebraska pivot corn

Nebraska, USA

64.54

Centre-pivot irrigated corn

PERTURBATIONtarget 55–80

Driven by canopy signal 0.32, 654 mm rainfall, mean 12°C.

Chihuahua desert scrub

Mexico

53.89

Arid rangeland, poor cropping

PERTURBATIONtarget 35–60

Driven by canopy signal 0.14, 272 mm rainfall, mean 21°C.

Bajio irrigated basin

Guanajuato, Mexico

91.11

Productive irrigated basin

OPTIMALtarget 70–100

Driven by canopy signal 0.41, 744 mm rainfall, mean 19°C.

Mato Grosso soy frontier

Brazil

75.42

Large-scale soy on cerrado

HEALTHYtarget 55–80

Driven by canopy signal 0.31, 1096 mm rainfall, mean 27°C.

Sao Paulo sugarcane

Brazil

91.13

High-yield cane land

OPTIMALtarget 70–100

Driven by canopy signal 0.75, 818 mm rainfall, mean 25°C.

Pampas grain belt

Buenos Aires, Argentina

80.23

World-class temperate grain

OPTIMALtarget 70–100

Driven by canopy signal 0.21, 1311 mm rainfall, mean 17°C.

Patagonian steppe

Santa Cruz, Argentina

47.96

Cold arid steppe, grazing only

ELEVATED_STRESStarget 35–60

Driven by canopy signal 0.03, 232 mm rainfall, mean 7°C.

Atacama desert

Antofagasta, Chile

18.5

Hyper-arid, no cropping

DEGRADEDtarget 0–50

Driven by canopy signal 0.05, 12 mm rainfall, mean 18°C.

Central Valley Chile vineyards

Chile

Mediterranean vine and fruit land

HEALTHYtarget 55–80

Driven by canopy signal 0.55, 262 mm rainfall, mean 17°C.

Colombian coffee slope

Quindio, Colombia

79.87

Steep but rich coffee ground

HEALTHYtarget 55–80

Driven by canopy signal 0.72, 1720 mm rainfall, mean 15°C.

Llanos savanna

Meta, Colombia

46.24

Acid savanna, improvable

ELEVATED_STRESStarget 35–60

Driven by canopy signal 0.54, 3919 mm rainfall, mean 26°C.

Andean altiplano cropland

Puno, Peru

47.61

High cold potato/quinoa ground

ELEVATED_STRESStarget 35–60

Driven by canopy signal 0.15, 657 mm rainfall, mean 10°C.

Ethiopian highland teff

Amhara, Ethiopia

69.51

Rainfed highland cereal

HEALTHYtarget 55–80

Driven by canopy signal 0.34, 1350 mm rainfall, mean 18°C.

Sahel millet zone

Zinder, Niger

51.2

Marginal rainfed millet

PERTURBATIONtarget 35–60

Driven by canopy signal 0.14, 472 mm rainfall, mean 28°C.

Sahara interior

Tamanrasset, Algeria

33.47

True desert, no vegetation

DEGRADEDtarget 0–50

Driven by canopy signal 0.04, 146 mm rainfall, mean 22°C.

Nile valley Upper Egypt

Luxor, Egypt

30.17

Narrow irrigated valley

DEGRADEDtarget 70–100 MISS

Driven by canopy signal 0.06, 0 mm rainfall, mean 25°C.

Moroccan irrigated plain

Souss, Morocco

39.14

Irrigated horticulture

ELEVATED_STRESStarget 35–60

Driven by canopy signal 0.13, 218 mm rainfall, mean 22°C.

Kenyan Rift horticulture

Naivasha, Kenya

68.23

Lake-side irrigated flowers/veg

HEALTHYtarget 55–80

Driven by canopy signal -0.29, 1890 mm rainfall, mean 15°C.

Zambian maize block

Central, Zambia

81.22

Rainfed commercial maize

OPTIMALtarget 70–100

Driven by canopy signal 0.38, 799 mm rainfall, mean 22°C.

South African maize triangle

Free State, South Africa

66.87

Dryland maize heartland

HEALTHYtarget 55–80

Driven by canopy signal 0.23, 903 mm rainfall, mean 17°C.

Western Cape wheat/vine

South Africa

80.49

Mediterranean mixed cropping

OPTIMALtarget 70–100

Driven by canopy signal 0.54, 501 mm rainfall, mean 16°C.

Congo rainforest

Cuvette, DR Congo

79.88

Dense forest, uncultivated

HEALTHYtarget 55–80

Driven by canopy signal 0.53, 1818 mm rainfall, mean 25°C.

Ganges plain paddy

Bihar, India

88.28

Intensive irrigated paddy

OPTIMALtarget 70–100

Driven by canopy signal 0.80, 1076 mm rainfall, mean 26°C.

Vidarbha cotton

Maharashtra, India

57.85

Rainfed cotton, drought-prone

PERTURBATIONtarget 55–80

Driven by canopy signal 0.09, 1234 mm rainfall, mean 26°C.

Kerala backwater paddy

Kerala, India

56.84

Wet lowland paddy

PERTURBATIONtarget 55–80

Driven by canopy signal -0.19, 2889 mm rainfall, mean 28°C.

Thar desert

Rajasthan, India

51.52

Sand desert, no cropping

PERTURBATIONtarget 0–50 MISS

Driven by canopy signal 0.14, 316 mm rainfall, mean 28°C.

Cauvery delta

Tamil Nadu, India

77.04

Canal-irrigated delta paddy

HEALTHYtarget 55–80

Driven by canopy signal 0.11, 1173 mm rainfall, mean 28°C.

Sindh canal cropland

Pakistan

79.09

Indus canal irrigation

HEALTHYtarget 55–80

Driven by canopy signal 0.20, 216 mm rainfall, mean 29°C.

Bangladesh boro paddy

Mymensingh, Bangladesh

32.22

Triple-crop floodplain

DEGRADEDtarget 70–100 MISS

Driven by canopy signal 0.22, 3340 mm rainfall, mean 25°C.

North China plain wheat

Henan, China

63.36

Irrigated wheat-maize double crop

PERTURBATIONtarget 55–80

Driven by canopy signal 0.14, 1068 mm rainfall, mean 16°C.

Northeast China black soil

Heilongjiang, China

79.97

Fertile but short season

HEALTHYtarget 55–80

Driven by canopy signal 0.93, 676 mm rainfall, mean 4°C.

Xinjiang oasis cotton

China

78.43

Oasis irrigated cotton

HEALTHYtarget 55–80

Driven by canopy signal 0.82, 49 mm rainfall, mean 13°C.

Tibetan plateau

Nagqu, China

55.04

Cold high desert rangeland

PERTURBATIONtarget 0–50 MISS

Driven by canopy signal 0.61, 481 mm rainfall, mean -0°C.

Mekong delta rice

Vietnam

77.27

Triple-crop delta rice

HEALTHYtarget 55–80

Driven by canopy signal 0.72, 1841 mm rainfall, mean 27°C.

Central Thailand paddy

Suphan Buri, Thailand

64.69

Irrigated central plain paddy

PERTURBATIONtarget 55–80

Driven by canopy signal 0.34, 1355 mm rainfall, mean 28°C.

Java rice terraces

Central Java, Indonesia

57.45

Volcanic soil paddy

PERTURBATIONtarget 55–80

Driven by canopy signal 0.77, 3234 mm rainfall, mean 23°C.

Sumatra oil palm

Riau, Indonesia

52.56

Wet tropical plantation

PERTURBATIONtarget 35–60

Driven by canopy signal 0.78, 3643 mm rainfall, mean 26°C.

Philippine sugar land

Negros, Philippines

53.14

Tropical cane land

PERTURBATIONtarget 35–60

Driven by canopy signal 0.72, 3047 mm rainfall, mean 25°C.

Murray irrigation area

NSW, Australia

64.47

Irrigated grain and rice

PERTURBATIONtarget 55–80

Driven by canopy signal 0.14, 376 mm rainfall, mean 18°C.

Western Australia wheatbelt

Australia

65.57

Marginal dryland wheat

HEALTHYtarget 55–80

Driven by canopy signal 0.17, 526 mm rainfall, mean 18°C.

Australian outback

Northern Territory, Australia

Arid interior rangeland

ELEVATED_STRESStarget 35–60

Driven by canopy signal 0.11, 179 mm rainfall, mean 23°C.

Canterbury plains

New Zealand

64.42

Irrigated pasture and grain

PERTURBATIONtarget 55–80

Driven by canopy signal 0.20, 873 mm rainfall, mean 11°C.

Waikato dairy pasture

New Zealand

66.08

Year-round dairy pasture

HEALTHYtarget 55–80

Driven by canopy signal 0.40, 1648 mm rainfall, mean 15°C.

Po Valley maize

Lombardy, Italy

69.23

Irrigated European maize

HEALTHYtarget 55–80

Driven by canopy signal 0.30, 1250 mm rainfall, mean 14°C.

Andalusian olive groves

Spain

64.43

Dryland olive plantation

PERTURBATIONtarget 55–80

Driven by canopy signal 0.19, 614 mm rainfall, mean 18°C.

La Mancha dry plain

Spain

63.35

Semi-arid vine and cereal

PERTURBATIONtarget 55–80

Driven by canopy signal 0.14, 465 mm rainfall, mean 16°C.

Beauce cereal plain

France

59.57

Prime European wheat

PERTURBATIONtarget 55–80

Driven by canopy signal 0.10, 367 mm rainfall, mean 12°C.

Bordeaux vineyard

France

39.8

Classic viticulture ground

ELEVATED_STRESStarget 70–100 MISS

Driven by canopy signal 0.11, 320 mm rainfall, mean 15°C.

East Anglia arable

United Kingdom

62.98

Flat drained arable fen

PERTURBATIONtarget 55–80

Driven by canopy signal 0.13, 543 mm rainfall, mean 11°C.

Scottish highland moor

United Kingdom

63.67

Wet acidic upland, grazing

PERTURBATIONtarget 55–80

Driven by canopy signal 0.15, 697 mm rainfall, mean 8°C.

North German plain

Lower Saxony, Germany

62.92

Mixed temperate cropland

PERTURBATIONtarget 55–80

Driven by canopy signal 0.13, 560 mm rainfall, mean 10°C.

Danube plain grain

Romania

60.16

Fertile Danube loess

PERTURBATIONtarget 55–80

Driven by canopy signal 0.15, 540 mm rainfall, mean 14°C.

Hungarian great plain

Hungary

49.45

Continental grain plain

ELEVATED_STRESStarget 35–60

Driven by canopy signal 0.07, 210 mm rainfall, mean 13°C.

Kuban black earth

Krasnodar, Russia

62.07

High-yield chernozem

PERTURBATIONtarget 55–80

Driven by canopy signal 0.14, 542 mm rainfall, mean 13°C.

Siberian taiga

Krasnoyarsk, Russia

52.73

Boreal forest, unfarmable

PERTURBATIONtarget 35–60

Driven by canopy signal 0.11, 824 mm rainfall, mean -3°C.

Kazakh steppe wheat

Akmola, Kazakhstan

55.83

Marginal continental wheat

PERTURBATIONtarget 55–80

Driven by canopy signal 0.08, 464 mm rainfall, mean 5°C.

Anatolian plateau grain

Konya, Turkey

48.85

Semi-arid plateau cereal

ELEVATED_STRESStarget 35–60

Driven by canopy signal 0.08, 250 mm rainfall, mean 13°C.

Jordan valley irrigation

Israel/Jordan

36.92

Intensive desert irrigation

ELEVATED_STRESStarget 35–60

Driven by canopy signal -0.08, 168 mm rainfall, mean 22°C.

Mesopotamian cropland

Diyala, Iraq

45.62

Irrigated but salinity-prone

ELEVATED_STRESStarget 35–60

Driven by canopy signal 0.11, 227 mm rainfall, mean 25°C.

Empty Quarter

Saudi Arabia

38.29

Sand sea, zero vegetation

ELEVATED_STRESStarget 35–60

Driven by canopy signal 0.04, 10 mm rainfall, mean 29°C.

Fergana valley cotton

Uzbekistan

47.2

Irrigated valley cotton

ELEVATED_STRESStarget 35–60

Driven by canopy signal 0.10, 204 mm rainfall, mean 18°C.

Greenland ice sheet

Greenland

28.03

Permanent ice, no land use

DEGRADEDtarget 0–50

Driven by canopy signal 0.03, 320 mm rainfall, mean -24°C.

Alaskan tundra

North Slope, Alaska, USA

41.73

Frozen tundra

ELEVATED_STRESStarget 35–60

Driven by canopy signal 0.03, 445 mm rainfall, mean -8°C.

Lake Victoria open water

Tanzania

70.78

Open water — must reject

HEALTHYtarget 0–10 MISS

Driven by canopy signal 0.30, 1103 mm rainfall, mean 25°C.

Caspian Sea

Central Caspian

Open water — must reject

WATER_BODYtarget 0–10

Open water spectra — rejected before scoring.

Great Lakes open water

Lake Superior, USA/Canada

61.49

Open water — must reject

PERTURBATIONtarget 0–10 MISS

Driven by canopy signal 0.15, 799 mm rainfall, mean 5°C.

Construction mode

SITE	EXPECTED, BEFORE THE CALL	SCORE	RETURNED	GOVERNING SIGNAL
Stable interior plain Central Kansas, USA	Flat, dry, low-hazard building ground target 70–100	86.16	`STABLE`	Negligible slope, dry substrate, only distant minor seismicity.
Reclaimed polder Flevoland, Netherlands	Engineered flat land, highly buildable target 70–100	80.04	`STABLE`	Excellent workability, discounted for below-sea-level exposure.
Arid metropolis Business Bay, Dubai, UAE	Dense, serviced, actively built urban ground target 70–100	79.12	`MARGIN_CAUTION`	Serviced flat ground; regional seismicity is real but far away.
Alpine valley Zermatt, Switzerland	Buildable valley floor in constrained mountain terrain target 55–80	77.99	`MARGIN_CAUTION`	Workable valley floor, penalised for altitude and cold season.
Equatorial city Central Business District, Singapore	Buildable, but extreme rainfall loading target 55–80	73.59	`MARGIN_CAUTION`	Very low hazard offset by drainage and rainfall exposure.
Seismic valley fill Kathmandu, Nepal	Buildable but major earthquake exposure target 55–80	69.6	`MARGIN_CAUTION`	Favourable ground conditions pulled down by the hazard record.
Reclaimed bay front Tokyo Bay, Japan	Buildable, but high seismic and low-ground load target 55–80	62.54	`MARGIN_CAUTION`	Dense near-field seismicity on low reclaimed ground.
High Himalaya Zanskar, India	Extreme slope, altitude and frozen ground target 0–50	46.94	`ENGINEERING_REVIEW`	Severe measured gradient, 4,000 m altitude, sub-zero mean.
Gulf coast flatland Houston, Texas, USA	Flat, serviced, low seismicity target 70–100	93.09	`STABLE`	Near-zero gradient with no meaningful hazard record.
Desert metropolis Phoenix, Arizona, USA	Dry, flat, stable ground target 70–100	90.48	`STABLE`	Dry substrate and excellent workability.
Delhi NCR expansion Noida, India	Flat alluvial urban ground target 70–100	85.86	`STABLE`	Level alluvium with only distant seismicity.
Riyadh plateau Riyadh, Saudi Arabia	Dry, flat, aseismic target 70–100	84.06	`STABLE`	Arid, firm ground with negligible hazard.
Doha coastal city Doha, Qatar	Flat, dry, engineered target 70–100	79.61	`MARGIN_CAUTION`	Very workable, discounted for low coastal ground.
Berlin lowland Berlin, Germany	Flat glacial plain, no hazard target 70–100	90.34	`STABLE`	Stable craton with mild climate loading.
Warsaw plain Warsaw, Poland	Stable continental interior target 70–100	92.16	`STABLE`	Flat, dry and seismically quiet.
Paris basin Ile-de-France, France	Gentle relief, low hazard target 70–100	91.03	`STABLE`	Consistent ground with no hazard record.
Madrid meseta Madrid, Spain	Dry plateau, modest hazard target 70–100	87.96	`STABLE`	Dry plateau ground, minor distant seismicity.
Melbourne suburbs Melbourne, Australia	Temperate, stable craton target 70–100	85.82	`STABLE`	Mild climate on very old stable crust.
Perth coastal plain Perth, Australia	Flat sand plain, dry target 70–100	92.02	`STABLE`	Dry, flat and effectively aseismic.
Rio de la Plata city Buenos Aires, Argentina	Very flat, aseismic target 70–100	90.81	`STABLE`	Level ground with no recorded hazard.
Highveld city Johannesburg, South Africa	Dry plateau, stable target 70–100	77.51	`MARGIN_CAUTION`	Stable but with mining-related local seismicity.
Lake plain city Chicago, Illinois, USA	Flat lake plain target 70–100	86.32	`STABLE`	Flat, firm ground with cold-season loading.
Ontario lowland Toronto, Canada	Flat, stable, cold winters target 70–100	86.05	`STABLE`	Very workable ground, winter-adjusted.
Continental cold city Moscow, Russia	Flat, stable, long frozen season target 70–100	84.38	`STABLE`	Excellent ground offset by a long freeze.
Foothills city Calgary, Canada	Elevated but workable ground target 70–100	84.43	`STABLE`	Gentle relief with cold-season penalties.
Baltic cold city Helsinki, Finland	Hard rock, severe winters target 70–100	80.98	`STABLE`	Competent bedrock, discounted for the freeze.
Volcanic island capital Reykjavik, Iceland	Active seismicity and cold target 55–80	65.87	`MARGIN_CAUTION`	Frequent shallow seismicity on cold ground.
Subarctic port Anchorage, Alaska, USA	High seismicity, frozen ground target 35–60	52.88	`ENGINEERING_REVIEW`	Dense near-field seismicity plus deep winter.
High-altitude basin Mexico City, Mexico	Soft lakebed, real seismicity target 55–80	75.01	`MARGIN_CAUTION`	Flat basin floor with regional hazard exposure.
Strait city Istanbul, Turkey	Major fault nearby target 55–80	69.84	`MARGIN_CAUTION`	Workable ground pulled down by fault proximity.
Plateau capital Tehran, Iran	Active faulting, dry ground target 55–80	74.28	`MARGIN_CAUTION`	Dry firm ground with a live hazard record.
Andean valley city Santiago, Chile	Subduction zone exposure target 55–80	71.56	`MARGIN_CAUTION`	Good ground conditions, persistent seismicity.
Coastal desert capital Lima, Peru	Dry ground, high seismicity target 55–80	75.79	`MARGIN_CAUTION`	Arid stable substrate on an active margin.
Bay area city San Francisco, California, USA	Major fault system target 55–80	75.74	`MARGIN_CAUTION`	Dry workable ground with well-documented faults.
Basin metropolis Los Angeles, California, USA	Dry, flat, seismically active target 55–80	76.7	`MARGIN_CAUTION`	Ideal terrain, hazard-limited.
Bay city Manila, Philippines	Seismic and rainfall exposure target 35–60	54.62	`ENGINEERING_REVIEW`	Low ground, heavy rainfall, frequent earthquakes.
Delta megacity Jakarta, Indonesia	Low ground, subsidence, seismic target 55–80	58.11	`ENGINEERING_REVIEW`	Near sea level with high rainfall loading.
Basin capital Taipei, Taiwan	High seismicity, heavy rain target 55–80	62.02	`MARGIN_CAUTION`	Flat basin fill inside an active belt.
Delta port city Osaka, Japan	Reclaimed ground, seismic target 55–80	68.62	`MARGIN_CAUTION`	Level reclaimed ground with real hazard.
Harbour capital Wellington, New Zealand	Fault-line city target 55–80	65.33	`MARGIN_CAUTION`	Constrained terrain astride an active fault.
Andean plateau city Bogota, Colombia	High plateau, moderate hazard target 55–80	64.36	`MARGIN_CAUTION`	Flat plateau floor at altitude, moderate seismicity.
Equatorial highland city Quito, Ecuador	Volcanic, steep, seismic target 35–60	51.32	`ENGINEERING_REVIEW`	Steep volcanic ground with frequent events.
Steep altiplano city La Paz, Bolivia	Extreme slope and altitude target 55–80	63.64	`MARGIN_CAUTION`	Built ground, heavily penalised for gradient.
Andean steep slope Cusco region, Peru	Steep mountain slope, poor access target 35–60	51.56	`ENGINEERING_REVIEW`	Severe gradient and altitude on an active margin.
Rocky Mountain crest Sawatch Range, Colorado, USA	Steep, high, cold target 35–60	56.9	`ENGINEERING_REVIEW`	High relief with a long frozen season.
Mont Blanc massif Haute-Savoie, France	Glacier and rock wall target 35–60	57.3	`ENGINEERING_REVIEW`	Extreme relief, sub-zero mean temperature.
High Andes ridge Aconcagua region, Argentina	Extreme relief, no access target 0–50	38.04	`CRITICAL_HAZARD`	Maximum measured gradient at extreme altitude.
Karakoram range Baltistan, Pakistan	Glaciated extreme terrain target 0–50	41.54	`ENGINEERING_REVIEW`	Glaciated slopes far beyond buildable gradient.
North Sea Central North Sea	Water — must be rejected outright target 0–10	8	`WATER_BODY`	Open sea rejected before scoring.
Pacific Ocean Central Pacific	Water — must be rejected outright target 0–10	8	`WATER_BODY`	No land surface at this coordinate.
Houston coastal plain Texas, USA	Flat, wet, subsidence-prone target 70–100	89.4	`STABLE`	Driven by 0.4% slope, 17 m elevation, 2 recorded quakes in 5 years.
Denver plateau Colorado, USA	Dry stable high plain target 70–100	81.53	`STABLE`	Driven by 0.8% slope, 1612 m elevation, 2 recorded quakes in 5 years.
Chicago lake plain Illinois, USA	Flat glacial till, cold winters target 70–100	84.07	`STABLE`	Driven by 7.5% slope, 265 m elevation, 2 recorded quakes in 5 years.
Phoenix desert basin Arizona, USA	Dry flat desert basin target 70–100	88.23	`STABLE`	Driven by 1.1% slope, 335 m elevation, 5 recorded quakes in 5 years.
Atlanta piedmont Georgia, USA	Firm piedmont bedrock target 70–100	88.92	`STABLE`	Driven by 1.4% slope, 316 m elevation, 0 recorded quakes in 5 years.
New Orleans delta Louisiana, USA	Below sea level, soft soils target 35–60	86.8 MISS	`STABLE`	Driven by 2.4% slope, 21 m elevation, 0 recorded quakes in 5 years.
Miami limestone coast Florida, USA	Low limestone coast, flood risk target 70–100	81.17	`STABLE`	Driven by 1.8% slope, 10 m elevation, 0 recorded quakes in 5 years.
Seattle glacial slopes Washington, USA	Sloped glacial soils, seismic target 55–80	66.18	`MARGIN_CAUTION`	Driven by 7.8% slope, 72 m elevation, 39 recorded quakes in 5 years.
Salt Lake basin Utah, USA	Flat basin near active fault target 55–80	73.82	`MARGIN_CAUTION`	Driven by 2.2% slope, 1302 m elevation, 52 recorded quakes in 5 years.
Toronto glacial plain Canada	Stable till plain, cold target 70–100	86.05	`STABLE`	Driven by 4.8% slope, 87 m elevation, 2 recorded quakes in 5 years.
Vancouver delta Canada	Soft delta ground, seismic target 55–80	77.41	`MARGIN_CAUTION`	Driven by 3.8% slope, 37 m elevation, 28 recorded quakes in 5 years.
Canadian shield rock Sudbury, Canada	Hard shield bedrock target 70–100	86.71	`STABLE`	Driven by 0.5% slope, 261 m elevation, 1 recorded quakes in 5 years.
Yellowknife permafrost NWT, Canada	Permafrost, extreme cold target 55–80	78.03	`MARGIN_CAUTION`	Driven by 1.2% slope, 188 m elevation, 0 recorded quakes in 5 years.
Monterrey foothills Mexico	Firm ground, steep margins target 70–100	89.06	`STABLE`	Driven by 0.6% slope, 542 m elevation, 2 recorded quakes in 5 years.
Panama canal zone Panama	Wet tropical, expansive clays target 55–80	79.79	`MARGIN_CAUTION`	Driven by 1.1% slope, 19 m elevation, 12 recorded quakes in 5 years.
Brasilia plateau Brazil	Flat stable laterite plateau target 70–100	85.27	`STABLE`	Driven by 2.1% slope, 1085 m elevation, 0 recorded quakes in 5 years.
Rio steep granite Brazil	Steep granite hillsides target 55–80	73.16	`MARGIN_CAUTION`	Driven by 53.7% slope, 182 m elevation, 0 recorded quakes in 5 years.
Buenos Aires pampa Argentina	Flat firm pampa, low hazard target 70–100	88.56	`STABLE`	Driven by 2.4% slope, 23 m elevation, 0 recorded quakes in 5 years.
Amazon floodplain Manaus, Brazil	Seasonally flooded soft ground target 70–100	84.38	`STABLE`	Driven by 4.1% slope, 32 m elevation, 0 recorded quakes in 5 years.
Bogota savanna edge Colombia	Soft lacustrine sediments target 55–80	64.44	`MARGIN_CAUTION`	Driven by 0.6% slope, 2556 m elevation, 46 recorded quakes in 5 years.
Chilean coastal port Valparaiso, Chile	Steep coast, subduction hazard target 55–80	67.24	`MARGIN_CAUTION`	Driven by 2.2% slope, 13 m elevation, 388 recorded quakes in 5 years.
Cairo desert plateau Egypt	Dry firm desert ground target 55–80	74.96	`MARGIN_CAUTION`	Driven by 4.0% slope, 126 m elevation, 3 recorded quakes in 5 years.
Lagos coastal lowland Nigeria	Swampy coastal sand, flooding target 70–100	80.95	`STABLE`	Driven by 1.0% slope, 5 m elevation, 0 recorded quakes in 5 years.
Nairobi highland Kenya	Firm volcanic highland target 70–100	81.57	`STABLE`	Driven by 1.0% slope, 1668 m elevation, 3 recorded quakes in 5 years.
Addis highland slope Ethiopia	High volcanic ground, rift edge target 55–80	65.96	`MARGIN_CAUTION`	Driven by 5.2% slope, 2379 m elevation, 342 recorded quakes in 5 years.
Johannesburg reef South Africa	Hard reef rock, mining voids target 55–80	73.82	`MARGIN_CAUTION`	Driven by 2.8% slope, 1743 m elevation, 40 recorded quakes in 5 years.
Cape Town coastal South Africa	Mixed sand and rock coast target 70–100	88.62	`STABLE`	Driven by 4.2% slope, 24 m elevation, 0 recorded quakes in 5 years.
Dubai coastal sand UAE	Flat sand, high salinity target 70–100	82.49	`STABLE`	Driven by 1.8% slope, 0 m elevation, 106 recorded quakes in 5 years.
Riyadh plateau Saudi Arabia	Dry firm limestone plateau target 70–100	81.65	`STABLE`	Driven by 0.6% slope, 623 m elevation, 0 recorded quakes in 5 years.
Kathmandu valley Nepal	Soft valley fill, high seismicity target 35–60	69.17 MISS	`MARGIN_CAUTION`	Driven by 1.9% slope, 1312 m elevation, 52 recorded quakes in 5 years.
Delhi alluvial plain India	Flat alluvium, moderate seismic target 70–100	82.86	`STABLE`	Driven by 1.2% slope, 217 m elevation, 5 recorded quakes in 5 years.
Chennai coastal India	Low coastal ground, flooding target 70–100	81.14	`STABLE`	Driven by 1.0% slope, 10 m elevation, 0 recorded quakes in 5 years.
Bengaluru plateau India	Hard gneiss plateau target 70–100	85.52	`STABLE`	Driven by 1.4% slope, 914 m elevation, 0 recorded quakes in 5 years.
Guwahati seismic zone India	High seismic, flood-prone target 55–80	79.11	`MARGIN_CAUTION`	Driven by 0.4% slope, 55 m elevation, 52 recorded quakes in 5 years.
Kolkata delta India	Soft delta clay, near sea level target 70–100	85.53	`STABLE`	Driven by 0.5% slope, 12 m elevation, 1 recorded quakes in 5 years.
Himalayan slope Uttarakhand, India	Steep unstable mountain ground target 0–50	52.39 MISS	`ENGINEERING_REVIEW`	Driven by 34.3% slope, 1774 m elevation, 26 recorded quakes in 5 years.
Karachi coastal plain Pakistan	Dry flat coast, moderate hazard target 70–100	80.29	`STABLE`	Driven by 0.5% slope, 8 m elevation, 10 recorded quakes in 5 years.
Dhaka floodplain Bangladesh	Soft floodplain, subsidence target 55–80	66.78	`MARGIN_CAUTION`	Driven by 1.3% slope, 9 m elevation, 15 recorded quakes in 5 years.
Colombo coastal Sri Lanka	Wet coastal lowland target 55–80	74.57	`MARGIN_CAUTION`	Driven by 1.7% slope, 8 m elevation, 1 recorded quakes in 5 years.
Bangkok delta Thailand	Soft clay, sinking delta target 35–60	80.67 MISS	`STABLE`	Driven by 0.6% slope, 9 m elevation, 0 recorded quakes in 5 years.
Singapore reclaimed land Singapore	Engineered reclaimed ground target 55–80	76.77	`MARGIN_CAUTION`	Driven by 1.9% slope, 27 m elevation, 1 recorded quakes in 5 years.
Kuala Lumpur karst Malaysia	Limestone karst, sinkhole risk target 55–80	79.36	`MARGIN_CAUTION`	Driven by 5.8% slope, 62 m elevation, 1 recorded quakes in 5 years.
Hanoi red river delta Vietnam	Soft delta ground target 70–100	83.2	`STABLE`	Driven by 1.8% slope, 24 m elevation, 2 recorded quakes in 5 years.
Beijing plain China	Flat dry plain, moderate seismic target 70–100	89.73	`STABLE`	Driven by 0.2% slope, 49 m elevation, 2 recorded quakes in 5 years.
Shanghai soft ground China	Very soft alluvium, subsidence target 55–80	78.58	`MARGIN_CAUTION`	Driven by 0.6% slope, 3 m elevation, 1 recorded quakes in 5 years.
Chongqing steep slope China	Steep hill city, landslides target 70–100	80.23	`STABLE`	Driven by 5.5% slope, 249 m elevation, 43 recorded quakes in 5 years.
Lhasa high plateau China	Extreme altitude, cold, seismic target 55–80	61.98	`MARGIN_CAUTION`	Driven by 0.6% slope, 3656 m elevation, 21 recorded quakes in 5 years.
Hong Kong granite hills China	Steep granite, typhoon rain target 70–100	84.45	`STABLE`	Driven by 0.6% slope, 25 m elevation, 4 recorded quakes in 5 years.
Seoul granite basin South Korea	Firm granite, low seismicity target 70–100	88.03	`STABLE`	Driven by 1.6% slope, 37 m elevation, 2 recorded quakes in 5 years.
Tokyo bay fill Japan	Reclaimed bay fill, high seismic target 55–80	58.85	`ENGINEERING_REVIEW`	Driven by 1.2% slope, 1 m elevation, 500 recorded quakes in 5 years.
Sapporo cold ground Japan	Cold snowy plain, seismic target 55–80	68.73	`MARGIN_CAUTION`	Driven by 0.4% slope, 26 m elevation, 255 recorded quakes in 5 years.
Sydney sandstone Australia	Hard sandstone, mild climate target 55–80	78.51	`MARGIN_CAUTION`	Driven by 11.3% slope, 69 m elevation, 8 recorded quakes in 5 years.
Perth sand plain Australia	Dry sand plain, very low hazard target 70–100	88.33	`STABLE`	Driven by 1.8% slope, 16 m elevation, 9 recorded quakes in 5 years.
Darwin cyclone coast Australia	Wet tropical coast, cyclones target 70–100	85.26	`STABLE`	Driven by 0.6% slope, 27 m elevation, 0 recorded quakes in 5 years.
Auckland volcanic field New Zealand	Volcanic cones, moderate hazard target 70–100	83.24	`STABLE`	Driven by 8.1% slope, 35 m elevation, 37 recorded quakes in 5 years.
Christchurch alluvium New Zealand	Liquefaction-prone alluvium target 55–80	73.12	`MARGIN_CAUTION`	Driven by 0.8% slope, 11 m elevation, 35 recorded quakes in 5 years.
Reykjanes lava field Iceland	Active lava field target 55–80	69.13	`MARGIN_CAUTION`	Driven by 2.3% slope, 58 m elevation, 110 recorded quakes in 5 years.
Oslo bedrock Norway	Hard bedrock, cold winters target 55–80	78.51	`MARGIN_CAUTION`	Driven by 2.8% slope, 9 m elevation, 1 recorded quakes in 5 years.
Stockholm archipelago rock Sweden	Glaciated bedrock, cold target 70–100	86.66	`STABLE`	Driven by 2.6% slope, 18 m elevation, 0 recorded quakes in 5 years.
Amsterdam soft peat Netherlands	Peat and clay below sea level target 55–80	88.52 MISS	`STABLE`	Driven by 1.1% slope, 11 m elevation, 1 recorded quakes in 5 years.
Berlin sand plain Germany	Flat glacial sand, no hazard target 70–100	88.09	`STABLE`	Driven by 2.2% slope, 30 m elevation, 0 recorded quakes in 5 years.
Alpine valley Innsbruck Austria	Narrow valley, steep walls target 70–100	83.91	`STABLE`	Driven by 7.4% slope, 590 m elevation, 7 recorded quakes in 5 years.
Swiss high alps Valais, Switzerland	Glaciated extreme relief target 55–80	56.97	`ENGINEERING_REVIEW`	Driven by 37.0% slope, 3153 m elevation, 20 recorded quakes in 5 years.
Paris basin limestone France	Firm limestone basin target 70–100	88.06	`STABLE`	Driven by 3.0% slope, 50 m elevation, 0 recorded quakes in 5 years.
Madrid plateau Spain	Dry firm plateau target 70–100	85.71	`STABLE`	Driven by 2.5% slope, 667 m elevation, 0 recorded quakes in 5 years.
Lisbon fault coast Portugal	Coastal city, historic seismicity target 70–100	86.03	`STABLE`	Driven by 1.5% slope, 64 m elevation, 3 recorded quakes in 5 years.
Naples volcanic ground Italy	Volcanic hazard, soft tuff target 55–80	76.16	`MARGIN_CAUTION`	Driven by 1.9% slope, 28 m elevation, 42 recorded quakes in 5 years.
Venice lagoon Italy	Lagoon fill, subsiding target 55–80	79.38	`MARGIN_CAUTION`	Driven by 0.7% slope, 5 m elevation, 33 recorded quakes in 5 years.
Athens basin Greece	Dry basin, moderate seismicity target 55–80	69.26	`MARGIN_CAUTION`	Driven by 3.6% slope, 99 m elevation, 164 recorded quakes in 5 years.
Bucharest plain Romania	Flat loess, deep seismic source target 55–80	77.39	`MARGIN_CAUTION`	Driven by 3.8% slope, 83 m elevation, 30 recorded quakes in 5 years.
Moscow plain Russia	Stable cold plain target 70–100	85.89	`STABLE`	Driven by 2.9% slope, 161 m elevation, 1 recorded quakes in 5 years.
Siberian permafrost city Yakutsk, Russia	Continuous permafrost target 55–80	76.66	`MARGIN_CAUTION`	Driven by 0.6% slope, 97 m elevation, 0 recorded quakes in 5 years.
Caucasus mountain slope Georgia	Steep seismic mountain terrain target 35–60	46.07	`ENGINEERING_REVIEW`	Driven by 32.5% slope, 2412 m elevation, 51 recorded quakes in 5 years.
Gulf of Mexico offshore Offshore Louisiana, USA	Open water — must reject target 0–10	8	`WATER_BODY`	Open water spectra — rejected before scoring.
Mediterranean open sea Ionian Sea	Open water — must reject target 0–10	8	`WATER_BODY`	Open water spectra — rejected before scoring.

Stable interior plain

Central Kansas, USA

86.16

Flat, dry, low-hazard building ground

STABLEtarget 70–100

Negligible slope, dry substrate, only distant minor seismicity.

Reclaimed polder

Flevoland, Netherlands

80.04

Engineered flat land, highly buildable

STABLEtarget 70–100

Excellent workability, discounted for below-sea-level exposure.

Arid metropolis

Business Bay, Dubai, UAE

79.12

Dense, serviced, actively built urban ground

MARGIN_CAUTIONtarget 70–100

Serviced flat ground; regional seismicity is real but far away.

Alpine valley

Zermatt, Switzerland

77.99

Buildable valley floor in constrained mountain terrain

MARGIN_CAUTIONtarget 55–80

Workable valley floor, penalised for altitude and cold season.

Equatorial city

Central Business District, Singapore

73.59

Buildable, but extreme rainfall loading

MARGIN_CAUTIONtarget 55–80

Very low hazard offset by drainage and rainfall exposure.

Seismic valley fill

Kathmandu, Nepal

69.6

Buildable but major earthquake exposure

MARGIN_CAUTIONtarget 55–80

Favourable ground conditions pulled down by the hazard record.

Reclaimed bay front

Tokyo Bay, Japan

62.54

Buildable, but high seismic and low-ground load

MARGIN_CAUTIONtarget 55–80

Dense near-field seismicity on low reclaimed ground.

High Himalaya

Zanskar, India

46.94

Extreme slope, altitude and frozen ground

ENGINEERING_REVIEWtarget 0–50

Severe measured gradient, 4,000 m altitude, sub-zero mean.

Gulf coast flatland

Houston, Texas, USA

93.09

Flat, serviced, low seismicity

STABLEtarget 70–100

Near-zero gradient with no meaningful hazard record.

Desert metropolis

Phoenix, Arizona, USA

90.48

Dry, flat, stable ground

STABLEtarget 70–100

Dry substrate and excellent workability.

Delhi NCR expansion

Noida, India

85.86

Flat alluvial urban ground

STABLEtarget 70–100

Level alluvium with only distant seismicity.

Riyadh plateau

Riyadh, Saudi Arabia

84.06

Dry, flat, aseismic

STABLEtarget 70–100

Arid, firm ground with negligible hazard.

Doha coastal city

Doha, Qatar

79.61

Flat, dry, engineered

MARGIN_CAUTIONtarget 70–100

Very workable, discounted for low coastal ground.

Berlin lowland

Berlin, Germany

90.34

Flat glacial plain, no hazard

STABLEtarget 70–100

Stable craton with mild climate loading.

Warsaw plain

Warsaw, Poland

92.16

Stable continental interior

STABLEtarget 70–100

Flat, dry and seismically quiet.

Paris basin

Ile-de-France, France

91.03

Gentle relief, low hazard

STABLEtarget 70–100

Consistent ground with no hazard record.

Madrid meseta

Madrid, Spain

87.96

Dry plateau, modest hazard

STABLEtarget 70–100

Dry plateau ground, minor distant seismicity.

Melbourne suburbs

Melbourne, Australia

85.82

Temperate, stable craton

STABLEtarget 70–100

Mild climate on very old stable crust.

Perth coastal plain

Perth, Australia

92.02

Flat sand plain, dry

STABLEtarget 70–100

Dry, flat and effectively aseismic.

Rio de la Plata city

Buenos Aires, Argentina

90.81

Very flat, aseismic

STABLEtarget 70–100

Level ground with no recorded hazard.

Highveld city

Johannesburg, South Africa

77.51

Dry plateau, stable

MARGIN_CAUTIONtarget 70–100

Stable but with mining-related local seismicity.

Lake plain city

Chicago, Illinois, USA

86.32

Flat lake plain

STABLEtarget 70–100

Flat, firm ground with cold-season loading.

Ontario lowland

Toronto, Canada

86.05

Flat, stable, cold winters

STABLEtarget 70–100

Very workable ground, winter-adjusted.

Continental cold city

Moscow, Russia

84.38

Flat, stable, long frozen season

STABLEtarget 70–100

Excellent ground offset by a long freeze.

Foothills city

Calgary, Canada

84.43

Elevated but workable ground

STABLEtarget 70–100

Gentle relief with cold-season penalties.

Baltic cold city

Helsinki, Finland

80.98

Hard rock, severe winters

STABLEtarget 70–100

Competent bedrock, discounted for the freeze.

Volcanic island capital

Reykjavik, Iceland

65.87

Active seismicity and cold

MARGIN_CAUTIONtarget 55–80

Frequent shallow seismicity on cold ground.

Subarctic port

Anchorage, Alaska, USA

52.88

High seismicity, frozen ground

ENGINEERING_REVIEWtarget 35–60

Dense near-field seismicity plus deep winter.

High-altitude basin

Mexico City, Mexico

75.01

Soft lakebed, real seismicity

MARGIN_CAUTIONtarget 55–80

Flat basin floor with regional hazard exposure.

Strait city

Istanbul, Turkey

69.84

Major fault nearby

MARGIN_CAUTIONtarget 55–80

Workable ground pulled down by fault proximity.

Plateau capital

Tehran, Iran

74.28

Active faulting, dry ground

MARGIN_CAUTIONtarget 55–80

Dry firm ground with a live hazard record.

Andean valley city

Santiago, Chile

71.56

Subduction zone exposure

MARGIN_CAUTIONtarget 55–80

Good ground conditions, persistent seismicity.

Coastal desert capital

Lima, Peru

75.79

Dry ground, high seismicity

MARGIN_CAUTIONtarget 55–80

Arid stable substrate on an active margin.

Bay area city

San Francisco, California, USA

75.74

Major fault system

MARGIN_CAUTIONtarget 55–80

Dry workable ground with well-documented faults.

Basin metropolis

Los Angeles, California, USA

76.7

Dry, flat, seismically active

MARGIN_CAUTIONtarget 55–80

Ideal terrain, hazard-limited.

Bay city

Manila, Philippines

54.62

Seismic and rainfall exposure

ENGINEERING_REVIEWtarget 35–60

Low ground, heavy rainfall, frequent earthquakes.

Delta megacity

Jakarta, Indonesia

58.11

Low ground, subsidence, seismic

ENGINEERING_REVIEWtarget 55–80

Near sea level with high rainfall loading.

Basin capital

Taipei, Taiwan

62.02

High seismicity, heavy rain

MARGIN_CAUTIONtarget 55–80

Flat basin fill inside an active belt.

Delta port city

Osaka, Japan

68.62

Reclaimed ground, seismic

MARGIN_CAUTIONtarget 55–80

Level reclaimed ground with real hazard.

Harbour capital

Wellington, New Zealand

65.33

Fault-line city

MARGIN_CAUTIONtarget 55–80

Constrained terrain astride an active fault.

Andean plateau city

Bogota, Colombia

64.36

High plateau, moderate hazard

MARGIN_CAUTIONtarget 55–80

Flat plateau floor at altitude, moderate seismicity.

Equatorial highland city

Quito, Ecuador

51.32

Volcanic, steep, seismic

ENGINEERING_REVIEWtarget 35–60

Steep volcanic ground with frequent events.

Steep altiplano city

La Paz, Bolivia

63.64

Extreme slope and altitude

MARGIN_CAUTIONtarget 55–80

Built ground, heavily penalised for gradient.

Andean steep slope

Cusco region, Peru

51.56

Steep mountain slope, poor access

ENGINEERING_REVIEWtarget 35–60

Severe gradient and altitude on an active margin.

Rocky Mountain crest

Sawatch Range, Colorado, USA

56.9

Steep, high, cold

ENGINEERING_REVIEWtarget 35–60

High relief with a long frozen season.

Mont Blanc massif

Haute-Savoie, France

57.3

Glacier and rock wall

ENGINEERING_REVIEWtarget 35–60

Extreme relief, sub-zero mean temperature.

High Andes ridge

Aconcagua region, Argentina

38.04

Extreme relief, no access

CRITICAL_HAZARDtarget 0–50

Maximum measured gradient at extreme altitude.

Karakoram range

Baltistan, Pakistan

41.54

Glaciated extreme terrain

ENGINEERING_REVIEWtarget 0–50

Glaciated slopes far beyond buildable gradient.

North Sea

Central North Sea

Water — must be rejected outright

WATER_BODYtarget 0–10

Open sea rejected before scoring.

Pacific Ocean

Central Pacific

Water — must be rejected outright

WATER_BODYtarget 0–10

No land surface at this coordinate.

Houston coastal plain

Texas, USA

89.4

Flat, wet, subsidence-prone

STABLEtarget 70–100

Driven by 0.4% slope, 17 m elevation, 2 recorded quakes in 5 years.

Denver plateau

Colorado, USA

81.53

Dry stable high plain

STABLEtarget 70–100

Driven by 0.8% slope, 1612 m elevation, 2 recorded quakes in 5 years.

Chicago lake plain

Illinois, USA

84.07

Flat glacial till, cold winters

STABLEtarget 70–100

Driven by 7.5% slope, 265 m elevation, 2 recorded quakes in 5 years.

Phoenix desert basin

Arizona, USA

88.23

Dry flat desert basin

STABLEtarget 70–100

Driven by 1.1% slope, 335 m elevation, 5 recorded quakes in 5 years.

Atlanta piedmont

Georgia, USA

88.92

Firm piedmont bedrock

STABLEtarget 70–100

Driven by 1.4% slope, 316 m elevation, 0 recorded quakes in 5 years.

New Orleans delta

Louisiana, USA

86.8

Below sea level, soft soils

STABLEtarget 35–60 MISS

Driven by 2.4% slope, 21 m elevation, 0 recorded quakes in 5 years.

Miami limestone coast

Florida, USA

81.17

Low limestone coast, flood risk

STABLEtarget 70–100

Driven by 1.8% slope, 10 m elevation, 0 recorded quakes in 5 years.

Seattle glacial slopes

Washington, USA

66.18

Sloped glacial soils, seismic

MARGIN_CAUTIONtarget 55–80

Driven by 7.8% slope, 72 m elevation, 39 recorded quakes in 5 years.

Salt Lake basin

Utah, USA

73.82

Flat basin near active fault

MARGIN_CAUTIONtarget 55–80

Driven by 2.2% slope, 1302 m elevation, 52 recorded quakes in 5 years.

Toronto glacial plain

Canada

86.05

Stable till plain, cold

STABLEtarget 70–100

Driven by 4.8% slope, 87 m elevation, 2 recorded quakes in 5 years.

Vancouver delta

Canada

77.41

Soft delta ground, seismic

MARGIN_CAUTIONtarget 55–80

Driven by 3.8% slope, 37 m elevation, 28 recorded quakes in 5 years.

Canadian shield rock

Sudbury, Canada

86.71

Hard shield bedrock

STABLEtarget 70–100

Driven by 0.5% slope, 261 m elevation, 1 recorded quakes in 5 years.

Yellowknife permafrost

NWT, Canada

78.03

Permafrost, extreme cold

MARGIN_CAUTIONtarget 55–80

Driven by 1.2% slope, 188 m elevation, 0 recorded quakes in 5 years.

Monterrey foothills

Mexico

89.06

Firm ground, steep margins

STABLEtarget 70–100

Driven by 0.6% slope, 542 m elevation, 2 recorded quakes in 5 years.

Panama canal zone

Panama

79.79

Wet tropical, expansive clays

MARGIN_CAUTIONtarget 55–80

Driven by 1.1% slope, 19 m elevation, 12 recorded quakes in 5 years.

Brasilia plateau

Brazil

85.27

Flat stable laterite plateau

STABLEtarget 70–100

Driven by 2.1% slope, 1085 m elevation, 0 recorded quakes in 5 years.

Rio steep granite

Brazil

73.16

Steep granite hillsides

MARGIN_CAUTIONtarget 55–80

Driven by 53.7% slope, 182 m elevation, 0 recorded quakes in 5 years.

Buenos Aires pampa

Argentina

88.56

Flat firm pampa, low hazard

STABLEtarget 70–100

Driven by 2.4% slope, 23 m elevation, 0 recorded quakes in 5 years.

Amazon floodplain

Manaus, Brazil

84.38

Seasonally flooded soft ground

STABLEtarget 70–100

Driven by 4.1% slope, 32 m elevation, 0 recorded quakes in 5 years.

Bogota savanna edge

Colombia

64.44

Soft lacustrine sediments

MARGIN_CAUTIONtarget 55–80

Driven by 0.6% slope, 2556 m elevation, 46 recorded quakes in 5 years.

Chilean coastal port

Valparaiso, Chile

67.24

Steep coast, subduction hazard

MARGIN_CAUTIONtarget 55–80

Driven by 2.2% slope, 13 m elevation, 388 recorded quakes in 5 years.

Cairo desert plateau

Egypt

74.96

Dry firm desert ground

MARGIN_CAUTIONtarget 55–80

Driven by 4.0% slope, 126 m elevation, 3 recorded quakes in 5 years.

Lagos coastal lowland

Nigeria

80.95

Swampy coastal sand, flooding

STABLEtarget 70–100

Driven by 1.0% slope, 5 m elevation, 0 recorded quakes in 5 years.

Nairobi highland

Kenya

81.57

Firm volcanic highland

STABLEtarget 70–100

Driven by 1.0% slope, 1668 m elevation, 3 recorded quakes in 5 years.

Addis highland slope

Ethiopia

65.96

High volcanic ground, rift edge

MARGIN_CAUTIONtarget 55–80

Driven by 5.2% slope, 2379 m elevation, 342 recorded quakes in 5 years.

Johannesburg reef

South Africa

73.82

Hard reef rock, mining voids

MARGIN_CAUTIONtarget 55–80

Driven by 2.8% slope, 1743 m elevation, 40 recorded quakes in 5 years.

Cape Town coastal

South Africa

88.62

Mixed sand and rock coast

STABLEtarget 70–100

Driven by 4.2% slope, 24 m elevation, 0 recorded quakes in 5 years.

Dubai coastal sand

UAE

82.49

Flat sand, high salinity

STABLEtarget 70–100

Driven by 1.8% slope, 0 m elevation, 106 recorded quakes in 5 years.

Riyadh plateau

Saudi Arabia

81.65

Dry firm limestone plateau

STABLEtarget 70–100

Driven by 0.6% slope, 623 m elevation, 0 recorded quakes in 5 years.

Kathmandu valley

Nepal

69.17

Soft valley fill, high seismicity

MARGIN_CAUTIONtarget 35–60 MISS

Driven by 1.9% slope, 1312 m elevation, 52 recorded quakes in 5 years.

Delhi alluvial plain

India

82.86

Flat alluvium, moderate seismic

STABLEtarget 70–100

Driven by 1.2% slope, 217 m elevation, 5 recorded quakes in 5 years.

Chennai coastal

India

81.14

Low coastal ground, flooding

STABLEtarget 70–100

Driven by 1.0% slope, 10 m elevation, 0 recorded quakes in 5 years.

Bengaluru plateau

India

85.52

Hard gneiss plateau

STABLEtarget 70–100

Driven by 1.4% slope, 914 m elevation, 0 recorded quakes in 5 years.

Guwahati seismic zone

India

79.11

High seismic, flood-prone

MARGIN_CAUTIONtarget 55–80

Driven by 0.4% slope, 55 m elevation, 52 recorded quakes in 5 years.

Kolkata delta

India

85.53

Soft delta clay, near sea level

STABLEtarget 70–100

Driven by 0.5% slope, 12 m elevation, 1 recorded quakes in 5 years.

Himalayan slope

Uttarakhand, India

52.39

Steep unstable mountain ground

ENGINEERING_REVIEWtarget 0–50 MISS

Driven by 34.3% slope, 1774 m elevation, 26 recorded quakes in 5 years.

Karachi coastal plain

Pakistan

80.29

Dry flat coast, moderate hazard

STABLEtarget 70–100

Driven by 0.5% slope, 8 m elevation, 10 recorded quakes in 5 years.

Dhaka floodplain

Bangladesh

66.78

Soft floodplain, subsidence

MARGIN_CAUTIONtarget 55–80

Driven by 1.3% slope, 9 m elevation, 15 recorded quakes in 5 years.

Colombo coastal

Sri Lanka

74.57

Wet coastal lowland

MARGIN_CAUTIONtarget 55–80

Driven by 1.7% slope, 8 m elevation, 1 recorded quakes in 5 years.

Bangkok delta

Thailand

80.67

Soft clay, sinking delta

STABLEtarget 35–60 MISS

Driven by 0.6% slope, 9 m elevation, 0 recorded quakes in 5 years.

Singapore reclaimed land

Singapore

76.77

Engineered reclaimed ground

MARGIN_CAUTIONtarget 55–80

Driven by 1.9% slope, 27 m elevation, 1 recorded quakes in 5 years.

Kuala Lumpur karst

Malaysia

79.36

Limestone karst, sinkhole risk

MARGIN_CAUTIONtarget 55–80

Driven by 5.8% slope, 62 m elevation, 1 recorded quakes in 5 years.

Hanoi red river delta

Vietnam

83.2

Soft delta ground

STABLEtarget 70–100

Driven by 1.8% slope, 24 m elevation, 2 recorded quakes in 5 years.

Beijing plain

China

89.73

Flat dry plain, moderate seismic

STABLEtarget 70–100

Driven by 0.2% slope, 49 m elevation, 2 recorded quakes in 5 years.

Shanghai soft ground

China

78.58

Very soft alluvium, subsidence

MARGIN_CAUTIONtarget 55–80

Driven by 0.6% slope, 3 m elevation, 1 recorded quakes in 5 years.

Chongqing steep slope

China

80.23

Steep hill city, landslides

STABLEtarget 70–100

Driven by 5.5% slope, 249 m elevation, 43 recorded quakes in 5 years.

Lhasa high plateau

China

61.98

Extreme altitude, cold, seismic

MARGIN_CAUTIONtarget 55–80

Driven by 0.6% slope, 3656 m elevation, 21 recorded quakes in 5 years.

Hong Kong granite hills

China

84.45

Steep granite, typhoon rain

STABLEtarget 70–100

Driven by 0.6% slope, 25 m elevation, 4 recorded quakes in 5 years.

Seoul granite basin

South Korea

88.03

Firm granite, low seismicity

STABLEtarget 70–100

Driven by 1.6% slope, 37 m elevation, 2 recorded quakes in 5 years.

Tokyo bay fill

Japan

58.85

Reclaimed bay fill, high seismic

ENGINEERING_REVIEWtarget 55–80

Driven by 1.2% slope, 1 m elevation, 500 recorded quakes in 5 years.

Sapporo cold ground

Japan

68.73

Cold snowy plain, seismic

MARGIN_CAUTIONtarget 55–80

Driven by 0.4% slope, 26 m elevation, 255 recorded quakes in 5 years.

Sydney sandstone

Australia

78.51

Hard sandstone, mild climate

MARGIN_CAUTIONtarget 55–80

Driven by 11.3% slope, 69 m elevation, 8 recorded quakes in 5 years.

Perth sand plain

Australia

88.33

Dry sand plain, very low hazard

STABLEtarget 70–100

Driven by 1.8% slope, 16 m elevation, 9 recorded quakes in 5 years.

Darwin cyclone coast

Australia

85.26

Wet tropical coast, cyclones

STABLEtarget 70–100

Driven by 0.6% slope, 27 m elevation, 0 recorded quakes in 5 years.

Auckland volcanic field

New Zealand

83.24

Volcanic cones, moderate hazard

STABLEtarget 70–100

Driven by 8.1% slope, 35 m elevation, 37 recorded quakes in 5 years.

Christchurch alluvium

New Zealand

73.12

Liquefaction-prone alluvium

MARGIN_CAUTIONtarget 55–80

Driven by 0.8% slope, 11 m elevation, 35 recorded quakes in 5 years.

Reykjanes lava field

Iceland

69.13

Active lava field

MARGIN_CAUTIONtarget 55–80

Driven by 2.3% slope, 58 m elevation, 110 recorded quakes in 5 years.

Oslo bedrock

Norway

78.51

Hard bedrock, cold winters

MARGIN_CAUTIONtarget 55–80

Driven by 2.8% slope, 9 m elevation, 1 recorded quakes in 5 years.

Stockholm archipelago rock

Sweden

86.66

Glaciated bedrock, cold

STABLEtarget 70–100

Driven by 2.6% slope, 18 m elevation, 0 recorded quakes in 5 years.

Amsterdam soft peat

Netherlands

88.52

Peat and clay below sea level

STABLEtarget 55–80 MISS

Driven by 1.1% slope, 11 m elevation, 1 recorded quakes in 5 years.

Berlin sand plain

Germany

88.09

Flat glacial sand, no hazard

STABLEtarget 70–100

Driven by 2.2% slope, 30 m elevation, 0 recorded quakes in 5 years.

Alpine valley Innsbruck

Austria

83.91

Narrow valley, steep walls

STABLEtarget 70–100

Driven by 7.4% slope, 590 m elevation, 7 recorded quakes in 5 years.

Swiss high alps

Valais, Switzerland

56.97

Glaciated extreme relief

ENGINEERING_REVIEWtarget 55–80

Driven by 37.0% slope, 3153 m elevation, 20 recorded quakes in 5 years.

Paris basin limestone

France

88.06

Firm limestone basin

STABLEtarget 70–100

Driven by 3.0% slope, 50 m elevation, 0 recorded quakes in 5 years.

Madrid plateau

Spain

85.71

Dry firm plateau

STABLEtarget 70–100

Driven by 2.5% slope, 667 m elevation, 0 recorded quakes in 5 years.

Lisbon fault coast

Portugal

86.03

Coastal city, historic seismicity

STABLEtarget 70–100

Driven by 1.5% slope, 64 m elevation, 3 recorded quakes in 5 years.

Naples volcanic ground

Italy

76.16

Volcanic hazard, soft tuff

MARGIN_CAUTIONtarget 55–80

Driven by 1.9% slope, 28 m elevation, 42 recorded quakes in 5 years.

Venice lagoon

Italy

79.38

Lagoon fill, subsiding

MARGIN_CAUTIONtarget 55–80

Driven by 0.7% slope, 5 m elevation, 33 recorded quakes in 5 years.

Athens basin

Greece

69.26

Dry basin, moderate seismicity

MARGIN_CAUTIONtarget 55–80

Driven by 3.6% slope, 99 m elevation, 164 recorded quakes in 5 years.

Bucharest plain

Romania

77.39

Flat loess, deep seismic source

MARGIN_CAUTIONtarget 55–80

Driven by 3.8% slope, 83 m elevation, 30 recorded quakes in 5 years.

Moscow plain

Russia

85.89

Stable cold plain

STABLEtarget 70–100

Driven by 2.9% slope, 161 m elevation, 1 recorded quakes in 5 years.

Siberian permafrost city

Yakutsk, Russia

76.66

Continuous permafrost

MARGIN_CAUTIONtarget 55–80

Driven by 0.6% slope, 97 m elevation, 0 recorded quakes in 5 years.

Caucasus mountain slope

Georgia

46.07

Steep seismic mountain terrain

ENGINEERING_REVIEWtarget 35–60

Driven by 32.5% slope, 2412 m elevation, 51 recorded quakes in 5 years.

Gulf of Mexico offshore

Offshore Louisiana, USA

Open water — must reject

WATER_BODYtarget 0–10

Open water spectra — rejected before scoring.

Mediterranean open sea

Ionian Sea

Open water — must reject

WATER_BODYtarget 0–10

Open water spectra — rejected before scoring.

What the backtest changed

The first pass was not clean, and the failures were instructive. A single-overpass reading called prime Punjab cropland mediocre because the satellite happened to pass over a fallow bed between harvests — the model now reads vigour across the whole season. The Nile delta was penalised as desert until the model learned to accept measured irrigation as proof of water supply. Downtown Singapore was misread as water; the rejection rule now requires far stronger evidence. Dubai was over-penalised for earthquakes hundreds of kilometres offshore, so hazard is now weighted by distance — while Kathmandu, whose governing event predates a five-year window, is now measured against a forty-year record. Each row above is the result after those corrections.

Want to check it? Pick any site above, drop the same location into the live demo, and compare what you get against the row.

THE MISSES

The 14 sites V1 got wrong.

These are the rows where the returned score fell outside the band we declared before the call. They are marked MISS in the tables above as well. Every one is a known, named failure mode rather than random noise — which is what makes them fixable in V2.

Yaqui valley

Sonora, Mexico · farming

54.59

expected medium · target 55–80 · returned PERTURBATION

Missed: the summer sample fell between irrigated cycles.

Mekong delta rice

Can Tho, Vietnam · farming

43.78

expected high · target 70–100 · returned ELEVATED_STRESS

Missed: the pixel fell on delta channel water, not paddy.

Nile valley Upper Egypt

Luxor, Egypt · farming

30.17

expected high · target 70–100 · returned DEGRADED

Driven by canopy signal 0.06, 0 mm rainfall, mean 25°C.

Thar desert

Rajasthan, India · farming

51.52

expected very low · target 0–50 · returned PERTURBATION

Driven by canopy signal 0.14, 316 mm rainfall, mean 28°C.

Bangladesh boro paddy

Mymensingh, Bangladesh · farming

32.22

expected high · target 70–100 · returned DEGRADED

Driven by canopy signal 0.22, 3340 mm rainfall, mean 25°C.

Tibetan plateau

Nagqu, China · farming

55.04

expected very low · target 0–50 · returned PERTURBATION

Driven by canopy signal 0.61, 481 mm rainfall, mean -0°C.

Bordeaux vineyard

France · farming

39.8

expected high · target 70–100 · returned ELEVATED_STRESS

Driven by canopy signal 0.11, 320 mm rainfall, mean 15°C.

Lake Victoria open water

Tanzania · farming

70.78

expected none · target 0–10 · returned HEALTHY

Driven by canopy signal 0.30, 1103 mm rainfall, mean 25°C.

Great Lakes open water

Lake Superior, USA/Canada · farming

61.49

expected none · target 0–10 · returned PERTURBATION

Driven by canopy signal 0.15, 799 mm rainfall, mean 5°C.

New Orleans delta

Louisiana, USA · construction

86.8

expected low · target 35–60 · returned STABLE

Driven by 2.4% slope, 21 m elevation, 0 recorded quakes in 5 years.

Kathmandu valley

Nepal · construction

69.17

expected low · target 35–60 · returned MARGIN_CAUTION

Driven by 1.9% slope, 1312 m elevation, 52 recorded quakes in 5 years.

Himalayan slope

Uttarakhand, India · construction

52.39

expected very low · target 0–50 · returned ENGINEERING_REVIEW

Driven by 34.3% slope, 1774 m elevation, 26 recorded quakes in 5 years.

Bangkok delta

Thailand · construction

80.67

expected low · target 35–60 · returned STABLE

Driven by 0.6% slope, 9 m elevation, 0 recorded quakes in 5 years.

Amsterdam soft peat

Netherlands · construction

88.52

expected medium · target 55–80 · returned STABLE

Driven by 1.1% slope, 11 m elevation, 1 recorded quakes in 5 years.

Why we leave them in

Dropping these rows would put the headline at 100% and make the whole exercise worthless. The pattern in them is consistent: sites where a single-date pixel disagrees with the season (irrigated valleys between cycles, delta channels, brief steppe flushes) or where a genuine physical penalty is applied more strongly than local practice warrants. Both are sampling and calibration problems, and both are the first targets for V2.

NEXT STEP

Test it on land you already know.

The fastest way to trust a score is to run it somewhere you can already judge. Try your own coordinates, then talk to us about access.

Run the demo Contact Us