Saturation Correction Guidance

NSIDC recommends that users apply the saturation correction to ICESat laser elevations for cryospheric applications or for any study of high-albedo targets, especially for the higher laser-energy campaigns (e.g., Laser 1A/B, 2A, 2B, 3A, 3B), before analysis of relative surface elevations or surface elevation change. The saturation correction values are available for each derived elevation but are not automatically applied in the elevation-related data sets from ICESat (products GLA-05, -06, -12 through -15). Refer also to the ICESat Laser Operational Periods table.

NOTE: References to GLAS binary product names GLA01 to GLA15 refer to original GLAS binary data, and are retained here for informational and provenance purposes. Access to GLAS binary data was removed 01 August, 2017. All GLAS data are available in HDF5 format, products GLAH01 to GLAH15.

A significant portion of the ICESat elevation and waveform data are impacted by GLAS detector saturation owing to stronger-than-expected received laser energy. The likelihood of saturation impact on the return waveform increases for campaigns with (1) higher laser transmit energy, and (2) high reflectivity surfaces (such as snow, sea ice, and ice sheets). Conversely, the impact is lessened for either low transmit energy campaigns or low reflectivity land and ocean surfaces, or when atmospheric transmission of laser energy is reduced (i.e. cloudy or hazy atmosphere). In addition, since the gain-setting algorithm is based on the received energy of the previous shot in the profile, the GLAS receiver can have ‘single-shot’ detector saturation due to strong transitions in received energy. In general, saturated GLAS waveforms produce a surface range measurement that is too long, making the elevation of the surface appear lower than it actually is. The scale of the effect can be many 10s of centimeters.  The effect, and the development and validation of a correction, is discussed in detail in Sun et al. (2017, in press to TGARS).

An initial version of the saturation correction algorithm was established early in the mission (Sun et al., 2005). An improved version was developed for Laser 3 campaigns. Applying the correction, now using updated versions in the final releases of the GLA products, compensates for the impact of saturation on the reported range and the derived elevation when the correction is valid. Some unpublished analyses of full-mission data suggest that there is a residual laser-energy-related effect on the altimetry data throughout the mission. This is still under analysis.

If uncorrected, the saturation effect produces biases when comparing elevations between campaigns with very different laser energies. On a hypothetical unchanging Earth surface, elevations measured by the ICESat campaigns with high GLAS laser energy (e.g. L1a/B, L2a, L3a, L3b) will appear lower than the same laser locations measured by weaker-laser campaigns (L2c, L2d, L2e, L2f, L3h, L3i, L3j, L3k).  As a rule of thumb, the saturation effect over bright surfaces is not significant for laser transmit energies below about 20 millijoules, assuming constant GLAS gain settings and surface albedo/reflectivity.

Residual laser-energy-related impacts on elevations over bright surfaces remaining even after application of the saturation correction (if any) can be addressed by applying inter-campaign bias corrections (ICBs), if users are concerned with very precise elevation change detection (centimeter-scale per year) over bright surfaces. NSIDC has published a guide describing ICBs and how to use them. Note that the saturation correction is not applied to any of the products (GLA-05, -06, or GLA 12-15). Guidance on its application is included in the data dictionary entries for the ‘satElevCorr’ parameter, excerpted below.

Binary
i_elev
The saturation elevation correction (i_satElevCorr) has not been applied and needs to be added to this elevation. This can be over a one meter correction. If it is invalid then the elevation should not be used. The saturation correction flag (i_satCorrFlg) is an important flag to understand the possible quality of the elevation data. The saturation index (i_satNdx) can be used for more understanding of concerns on data quality from saturation effects.

i_satElevCorr
Description: Correction to elevation for saturated waveforms. This correction has not been applied to the data so to apply it SUBTRACT the correction from the range estimate. To apply the correction to the elevations it must be ADDED to the elevation estimates.

HDF5
Group: Data_40HZ/Elevation_Corrections
d_satElevCorr
Correction to elevation for saturated waveforms. This correction has NOT been applied to the data. To apply it SUBTRACT the correction from the range estimate. To apply the correction to the elevations it must be ADDED to the elevation estimates.

Group: Data_40HZ/Quality
This group contains data quality flags and related parameters.
sat_corr_flg (should be the same in either data set version)
Saturation Correction Flag; Indicates if the saturation is Not Saturated (i_satNdx<2) or No Signal; Inconsequential (i_satNdx>=2 & i_pctSat<2.0); is Applicable (i_satNdx>=2 & i_pctSat>=2.0 & Full Width*<100ns); is Not Computable; is Not Applicable (i_satNdx>=2 & i_pctSat>=2.0 & Full Width*>=100ns)

flag values      flag_meanings
0, 1, 2, 3, 4, 5  not_saturated, inconsequential, applicable, not_computed, not_applicable

References:

Sun, X., Abshire, J.B., Yi, D. and Fricker, H.A., 2005. ICESat receiver signal dynamic range assessment and correction of range bias due to saturation. In AGU Fall Meeting Abstract C34A-07.

Sun, X., J. Abshire, A. Borsa, H. Fricker, D. Yi, J. DiMarzio, K. Brunt, D. Harding, and G. Neumann, 2017 in press. ICESat/GLAS Altimetry Measurements: Signal Dynamic

Sun, X., J. B. Abshire, D. Yi, H. A. Fricker. 2005. Range and Saturation Correction, Trans. Geoscience and Remote Sensing. , Trans. Geoscience and Remote Sensing TGRS-2016-00302.R1.