What the colours mean
Each coloured square is 30 by 30 metres (90 in Houston, which covers a much bigger area). Red areas have been moving down faster than the area's typical ground, and blue areas have been moving up or sinking more slowly. Pale areas move about the same as everything around them. Blank areas are water, fields or trees, where the radar signal is too scattered to measure reliably.
When you click, the reading uses the whole block of ground around that point (about 450 m across in Frisco, 900 m in Houston), not a single square. Single squares are noisy; blocks are much steadier.
What this map can't tell you
- Whether all of Frisco is sinking or rising. Every reading is compared with the area's typical motion, because the satellite's own reference point is far away and shifts the whole area up and down by centimetres from pass to pass.
- Anything about a single house. These are neighbourhood trends. Cracks in a foundation can have many causes this data can't see.
- Why the ground moved. Groundwater pumping, clay soils, new construction and settling fill can all move the ground. This map shows where and how much, not why.
- Purely up-and-down motion. The satellite looks at the ground at a slant, so it measures mostly vertical motion with a little east-west motion mixed in.
- Differences smaller than about 1.5 mm a year. Below that, the reading says the ground moves about the same as its surroundings.
How I checked it on Houston
Houston's suburbs are one of the best-studied sinking areas in the US. Researchers at the University of Houston found the north, northwest and west, including Katy, Spring and The Woodlands, sinking up to about 2 cm a year, while downtown and the southeast mostly stopped decades ago when groundwater pumping was limited.
Before looking at my Houston map, I wrote down what it had to show to pass: more sinking in the northwest and north than downtown and the southeast. Switch to Houston above to see it. The strongest sinking sits just west of the Addicks and Barker reservoirs, near Katy, and the pattern matches NASA's own displacement portal for the same satellite track.
The first version failed in an instructive way: with only 2.5 years of data, Houston's humid air added smooth fake patterns several millimetres a year strong. Using the full record back to 2016 fixed that, which is why both maps use ten years.
One change I made after seeing results: I had planned to show the tilt-removed map as the main one if Katy still stood out with the tilt removed. It did, but removing the tilt also erased the broad northwest-Houston sinking that the research documents, so it removes real signal too. The plain map is the default and the tilt-removed one is a toggle.
Where the data comes from
The measurements are NASA's OPERA surface displacement product (DISP-S1), made by NASA JPL from the European Space Agency's Sentinel-1 radar satellites and distributed by the Alaska Satellite Facility.
My code downloads each pass, cuts out the area, keeps only the pixels OPERA marks as reliable, links the separate data batches into one ten-year record, removes the area-wide shift on each date, and fits a yearly trend plus a seasonal swing to every pixel. The seasonal term matters here because North Texas clay swells when wet and shrinks when dry.