Ground Truth: GLOBE

CityCamp Gainesville Hack Day · NASA GLOBE Environmental Data

Ground Truth: GLOBE

NASA's GLOBE Observer program asks volunteers and researchers to measure the world around them, and uses what they record to check what satellites and climate models say. Here is how it works, and where the gaps are.

Protocol cycling
View
Year
all
Layers
GLOBE ERA5

I

Why a satellite needs a witness

A satellite does not see cloud. It sees light, and software turns that light into a number. ERA5, the climate record behind most weather apps, goes further still: it is a computer model that generates estimates every hour, across the globe. It is the most widely used reanalysis dataset in the world — the baseline for climate research, energy forecasting and countless weather services — and it reaches back to 1940.

Estimates have to be checked against something real, and that check is called ground truth. Someone has to be there to construct it. Weather stations do a little of this, but there are few of them and they sit where institutions already are. Most places have none.

GLOBE's answer is to let anyone be the observer. What makes the result usable as science is the method, not the crowd size: every measurement follows a set procedure, the app records exactly where and when, and photographs go with the reading. A cloud report taken as a satellite passes overhead can be compared with what the satellite said about that same spot.

Which is why these readings are used to check the instruments, rather than the other way round.

Why these four. They span the whole range of what a model can do: ERA5 reproduces cloud cover directly, approximates ground wetness, and has no variable at all for tree height or mosquito habitat. They are also the four with real breadth behind them — several Florida measurements have far more records but turn out to come from automated stations, or from one very dedicated volunteer.
The map above shows those four measurements across Florida. Intensity shades each county by how many observations it holds. Observations colours each reading by what it actually recorded. Hover on one to see the record a volunteer filed.
II

How close is the model?

Cloud cover is the one thing here that ERA5 also reports, so it is the one place we can compare satellite data against our surface-based observations.

One ERA5 square — about 30 km across soaked dry Two people, one square, opposite answers. What the model publishes for the whole square: light rain The average is correct. Neither person is in it. What a person records raining here standing water: yes muddy ground: yes 14:20, this exact spot dry here standing water: no dry ground: yes 14:20, two km away

This is determined by resolution; the model is reporting an average over an area bigger than most towns. The closer a question gets to one particular spot, the less an average can tell you — and a mosquito lives at one particular spot.

III

What no instrument records

For three of these four measurements, only humans can record them.

"A satellite can't see it" covers three different problems, and it helps to keep them apart.

One — too small to show up

Every sensor divides the world into squares and reports one number per square. ERA5 uses about 30 km. Landsat manages 30 m, Sentinel‑2 gets to 10 m, and the sharpest commercial cameras reach about 30 cm. Mosquitoes breed in a bucket, an old tyre, a blocked gutter — smaller than the smallest pixel anyone flies.

Two — hidden from above

A satellite can only access the top surface and can't see under a tree, a roof, behind a wall or in the shade — which is exactly where water can sit. Furthermore, clouds can block ordinary cameras completely. And a satellite passes over roughly twice a day, while a Florida afternoon storm and the puddle it leaves can both come and go in between.

Three — light cannot show it at all

Whether there are larvae in the water, which genus they are, how acidic a spring is, whether someone has since tipped the bucket out — no wavelength can carry that, from any height. A sharper camera does not help when there is no signal to sharpen.

Where a partial instrument does exist, the catch is instructive. GEDI, a laser aboard the Space Station, really does measure canopy height — but in scattered spots about 25 m wide, sampling the forest rather than covering it, and blind to the one tree a volunteer can just walk up to.

Why the smallest measurement matters most

Mosquitoes kill more people than any other animal, and dengue is spreading fast: 2024 was the worst year ever recorded, with over 14.6 million cases and more than 12,000 deaths across more than 100 countries, and roughly half the world's population now lives where it can spread. Florida is on that map — the CDC lists it among the handful of US states with limited local transmission, and Aedes aegypti, the mosquito that carries dengue, breeds in exactly the kind of container these records describe.

IV

The blanks still to fill

Florida has sixty-seven counties. Each map shades the ones where anybody has recorded that measurement at all.

has observations nobody has looked

A blank county definitely doesn't signal no mosquitoes or trees, but means that in this county nobody has logged any observations. This map shows exactly where one more person with a phone would add the most.

One note before you start

Do not go looking by record count. Four distinct Florida measurements were noticed while making this page:

A big number of records can mean one machine, or one very dedicated person. Count how many different people contributed before you count rows.

What would make the most difference

  1. Observers in the blank counties. This is the most effective fix on this page. You would only need a phone and the GLOBE Observer app.
  2. More of the measurements only you can make. Mosquito habitat has the fewest people behind it and the most direct link to human health.
  3. The same page, somewhere it matters more. The more data available, the more insights we can gain.
V

Learn more

Want to work with the GLOBE dataset yourself?

UF EMERGE

EMERGE is a NASA-funded program at the University of Florida that teaches anyone — students, teachers, anyone curious — to work with participatory science and satellite data. Its two open textbooks start at "what is citizen science" and go through mapping, satellite indices and risk modelling, using GLOBE's own datasets the whole way. Both are free, and published in English and Spanish.

UF GeoDI Lab

The lab behind EMERGE is GeoDI, the Geospatial Digital Informatics Lab in UF's Department of Geography, directed by Dr. Di Yang. It combines satellite observation, geospatial AI and community-collected data on environmental and public-health questions.

It is currently running the project this page argues for. From 2025 to 2027, under NASA funding with Dr. Yang as the lead researcher: Forecasting Mosquito-Borne Disease Risk in a Changing Climate: Integrating GLOBE Citizen Science and NASA Earth System Modeling.