Abstract—The unprecedented accuracy of elevations retrieved

from the Ice Cloud and Land Elevation Satellite (ICESat) laser

altimeter is investigated and used to characterize the range errors

in the Environmental Satellite (Envisat) and European Remote

Sensing 2 Satellite (ERS-2) radar altimeters over the continental

ice sheets. Cross-mission crossover analysis between time-coincident

ERS-2-, Envisat-, and ICESat-retrieved elevations and comparisons

to an ICESat-derived digital elevation map are used to
quantify the radar elevation error budget as a function of surface

slope and to investigate the effectiveness of a method to account

for the radar altimeter slope-induced error. The precision and

accuracy of the elevations retrieved from the ICESat Geoscience

Laser Altimeter System and the European Space Agency radar

altimeters on ERS-2 and Envisat are calculated over the Greenland

and Antarctic ice sheets using a crossover analysis. As a result

of this work, the laser precision is found to vary as a function of

surface slope from 14 to 59 cm, and the radar precision varies

from 59 cm to 3.7 m for ERS-2 and from 28 cm to 2.06 m for

Envisat. Envisat elevation retrievals when compared with ICESat

results over regions with less than 0.1◦ surface slopes show a mean

difference of 9 ± 5 cm for Greenland and −40 ± 98 cm over
Antarctica. ERS-2 elevation retrievals over these same low surface

slope regions differ from ICESat results by −56 ± 72 cm over

Greenland and 1.12 ± 1.16 m over Antarctica. At higher surface

slopes of 0.7◦ to 0.8◦, the Envisat/ICESat differences increase

to −2.27 ± 23 m over Greenland and to 0.05 ± 26 m over
Antarctica.

Accuracy

Accuracy and precision have very different meanings.

The average value of many readings taken with an accurate instrument will be close to the true value of the quantity being measured.

(If the instrument is not very precise, many repeated measurements will have to be made to reduce the uncertainty in the average value, whereas a precise and accurate instrument requires few repeated measurements to obtain a good measurement.)

Conversely, an inaccurate instrument is one which produces measurements whose average value is not the true value.

A famous example of the difference between precision and accuracy (and the influence of human nature in science) occurred a few years ago with the Hubble space telescope.

The manufacturers of the concave mirror for the telescope used a machine that could very precisely cut the glass so that its radius of curvature was known precisely to about 0.5 µm (0.0005 mm).

An astronomer who was going to be using the telescope soon after it was launched into orbit offered to confirm the measurement by using a much simpler technique that was not as precise as the instruments used by the manufacturer.

The manufacturers told him to stop worrying, and that their instruments were far more precise than the astronomer's simple equipment.

The telescope was launched into orbit, and as you may remember the images were fuzzy, very fuzzy ... in fact much worse than telescopes on earth. When the mirror manufacturers were asked to check their equipment, they found that they had forgot to include a certain spacer that added 2 mm to the measurement of the radius of curvature.

They had manufactured a radius of curvature that was precise to 0.5 µm but was inaccurate by 2 mm. The repercussions of this little "mistake" were enormous - NASA had a substantial amount of egg on its face, and, more to the point, the cost of correcting the telescope was enormous in terms of money (millions of dollars) and lost time.

The correction was achieved by inserting a corrective lens into the telescope, which sharpened the image but unfortunately meant that it could never be quite as good as was originally hoped.

http://www.patana.ac.th/secondary/science/anrophysics/ntopic1/resources/accuracy...