GPS accuracy... or lack thereof

It is increasingly common to have access to GPS measurements (Global Positioning S ystem). Initially, the goal is to be able to locate any place on the planet. Practical ways to get the position are automatic stations and portable instruments. The general public has become familiar with this technology, for example, car navigation systems.

Technically, GPS tracking consist on receiving signals broadcast by satellites in low orbit. The satellites transmit position and clock (time information) signals. To ensure a good location, simultaneous reception from several satellites is necessary, and it requires a direct line of sight to the satellites (which is why the GPS is not directly usable in tunnels or underground parking).

Location usually requires only rough accuracy, a decimeter is sufficient in most cases (for example, a car or a plane). This means that the exact value is not known, the value calculated and displayed is an approximation.

With a precision of a decimeter, for example, if the height displayed is 203.12 meters, this means that in reality the exact value is between 203.02 and 203.22 meters.

To monitor the evolution of a volcano-tectonic or tectonic phenomenon, GPS is a useful tool. But greater accuracy is required. The specific methods for this increased accuracy are beyond the scope of this article, but you can quickly note several ways:

• comparison of the information received from nearby stations (with the assumption that some of the errors is due to the path of the waves and therefore, different from stations to stations)
• comparison of the location for the same station as a function of time (the assumption being that a station moves linearly or is "fixed", the moving speed is low compared to the positional error)
• corrections (often retrospectively) taking into account several other parameters

Some advanced software like BERNESE (University of Bern), GIPSY-OASIS (NASA JPL) ... use these methods to obtain more accurate values, and one can commonly get the following error margins:

• horizontal (latitude and longitude): a few millimeters
• vertically (height): about 2 centimeters

Let us now consider some concrete examples based on information available from El Hierro.

The following two graphs are derived from analyzes of prof. Sagiya (Nagoya University). We know that BERNESE is used in this case, so the precision specified above applies.

Let us first look at the horizontal evolution for a GPS station (of Tenerife):

Fig 1.Horizontal of station TEIT (Tenerife)

The first graph corresponds to the east-west trend (east is positive) and the second the north-south trend (north is positive). And now compare with the vertical plane:

Fig 2. Vertical evolution of station TEIT (Tenerife)

It is clear that the dispersion of the measuring points is greater in the vertical plane, but overall there is no clear trend. By adding the error margin (red lines), we get:

Fig 3. Vertical evolution of station TEIT (Tenerife), with error margin.

Now we examine the Frontera station:

Fig 4. Horizontal evolution of Fontera station

We compare with the vertical evolution:

Fig 5. Vertical evolution of Fontera station.

We can see that the trend is similar (several movements) and dispersion is again much greater is the vertical plane. By adding a visualization of the error margin (red curves), we obtain:

Fig 6. Vertical evolution of the Fontera station (with error margin).

It seems that the margin of error is larger for Frontera than for Tenerife.
If we compare with the CVCB station (La Palma):

Fig 7. Vertical evolution of the CVCB station

The dispersion of the measurement points is even more important here (maybe 3 cm). It may be a local effect. Too often, non - informed personsdo not take into account the error margin and, looking at the changes from day to day and "zooming in" the last days, conclude that a station "jumps" two centimeters per day (sometimes in one direction and sometimes in the other). In fact, the rate of change is generally much lower, except for sudden movement of a tectonic fault. It is possible to obtain millimeter accuracy in the vertical direction, but it is extremely complex (for instance, you must take into account the effect of tidal height variation due to the attraction of the sun and moon on the ground. Yup, there is not only the sea that undergoes this!)