365 Days of Climate Awareness 179 – Satellite Altimetry 1: Remote Sensing

Much of science is conducted by remote sensing, using a machine to detect energy signals which originated somewhere else, and inferring physical conditions at the point of origin based on the signal. There are two types: passive, where the machine simply monitors an aspect of its environment (like a telescope, microphone or magnetometer), or active, where it broadcasts a signal and then receives a return, whether reflected (like an MRI, police radar or satellite altimeter). There are several advantages to remote sensing.

The basic idea. Passive remote sensors (left) receive a signal (energy) from the environment. Active sensors (right) emit and then receive an energy signal. The difference between what was emitted and received makes it possible to infer things—height, velocity--about the target.

In medical applications, it’s non-invasive (despite the barium sulfate you have to drink before a CT scan), enabling scans of sensitive organs like the brain. In natural science applications, it simplifies the problem of working in a hostile environment (towing instruments near the sea surface, as opposed to sending them down to the sea floor), and can provide far wider spatial coverage (satellite view of large swathes of the planet surface versus in-situ point measurements).

The equipment is costly, of course, and can be tricky to maintain and interpret. Spatial resolution is much lower from space. And distinguishing signal (the part of the data you want) from noise (the part you don’t) can be difficult. Noise is a relative term, because the very same data can contain several different signals, all combined into one record. Removing each separate data signal is delicate work and our methods must be constantly refined.

A relatively simple example is using a magnetometer for seafloor engineering work (one aspect of what I do). Magnetometers are passive sensors: a magnetic mechanism inside the unit is responsive to changing magnetic fields, and those responses are recorded as changing voltages in the data record. The changing voltages are then used to infer the strength of the magnetic field the sensor passed through.

Across much of the planet, Earth’s magnetic field has a strength between 25-80 uT (microteslas). While looking for engineering hazards—old pipes or cables, shipwrecks, or unexploded ordnance (UXO)—a magnetometer must be able to sense magnetic fields as small as 1 nT (nanoteslas: 1 nT = 0.001 uT). To do this, the earth’s magnetic field—a signal in its own right, but useless to the engineers and therefore noise in this context—must be subtracted from the record to make it useful for construction plans. In this case, the planet’s magnetic field is an overwhelming source of noise which obscures the information we want (small local fields).

The same happens with an active sensor, which broadcasts a signal (usually sound or radio wave). Energy is emitted from the sensor, strikes the target and returns to the sensor. In the meantime a number of other environmental factors can scramble the desired information and must be processed out. In the example of using sonar to map seafloor depths, monitoring changing ocean temperatures with depth is critical to account mathematically for how the sound will refract (bend) as it passes through the water. Without this, the sonar signal will be useless. 

In the case of a satellite measuring heights of the ocean, land or ice, one important factor to control for is varying levels of moisture throughout the atmosphere, which likewise bend and scatter radiation. (While at the same time, that scattering at various altitudes is itself a signal other scientists want. One data series can have an incredible amount of value to many different people! One person’s noise is another’s signal.)

Tomorrow: satellite altimetry 2: how it works.

Be brave, and be well.