When Everything Got Coordinates

THE JOIN KEY

A soil sample and a satellite pixel have nothing in common as measurements. One is wet chemistry in a laboratory, the other is reflected energy recorded from orbit. What lets a model use them together is that both can be tagged with the same pair of numbers, and those numbers mean the same thing to both.

Position is the join key of planetary data. Almost every combined product in this book, soil maps trained against imagery, flood forecasts fed by river gauges, species records matched to habitat, exists because two unlike measurements could be lined up by where they were taken.

GEOREFERENCING

Latitude and longitude are less obvious than they look

A coordinate is meaningless without a model of the shape of Earth to measure it against. The planet is not a sphere and not quite an ellipsoid, and different national surveys historically fitted different reference surfaces, called datums, to their own region because a local best fit was more accurate than a global compromise. The consequence is that the same physical spot can carry coordinates that differ by hundreds of meters between systems.

WGS 84, the reference frame GPS uses, is what made a single global set of coordinates practical. That is a quieter achievement than the satellites themselves: an agreement that one particular mathematical Earth would be the Earth everyone's numbers refer to. Every dataset that lines up with every other dataset in this book is standing on it.

Map projection is the second half of the problem, and it is unsolvable in principle. A curved surface cannot be flattened without distorting area, shape, distance or direction, so every projection chooses which distortion to accept. A land-cover statistic computed in one projection and compared against one computed in another can disagree for no reason other than the flattening.

GPS AND ITS SUCCESSORS

Positioning became something anything could carry

GPS reached full operational capability in 1995, and its receivers work by trilateration: measuring how long signals from several satellites took to arrive, and solving for the one position consistent with all of them. Because the calculation depends on timing, every receiver also has access to a precise shared clock, which turns out to matter as much as the position does.

The systems multiplied from there. Europe's Galileo, Russia's GLONASS and China's BeiDou all provide independent constellations at similar medium-altitude orbits, and modern receivers use several at once. The practical effect is that positioning stopped being an instrument and became an assumption: a tractor, a phone, a shipping container, a collared animal and a field researcher all know where they are without anyone treating that as a measurement task.

That shared clock deserves its own note. A timestamp is the second coordinate, and without agreement on it, sequences fall apart. Deciding whether a flood alert preceded or followed a dam release, whether a fire detection preceded a lightning strike, whether an animal moved before or after a road opened, all depend on unrelated systems having recorded time the same way.

WHAT COORDINATES HIDE

Six decimal places do not make a claim true

A coordinate printed to six decimal places suggests precision of about a tenth of a meter. The observation attached to it is frequently much vaguer: a species recorded at the center of a district, a soil sample located at the nearest track junction, a historical record georeferenced from a place name that has since moved. The coordinate format carries no room to say approximately.

This is one of the most common quiet failures in planetary data. Precision in the position field gets read as precision in the observation, and a record that honestly means somewhere in this valley ends up being used as though it meant this exact rock.