The Coupled GRACE / GIA Evaluation of the Antarctic Ice Mass Loss is Unreliable

GRACE does not measure any ice sheet thickness but only gravity. Similarly to the computation of the global mean sea level (GMSL), the computation of the ice sheet thickness, follows a large number of assumptions. As a result, the actual inaccuracy of the Antarctic ice sheet thickness computation is much larger than any trend proposed. In other words, you can manufacture almost any result you want by using the noisy raw GRACE signal and selected corrections. It is however the further Glacial Isostatic Adjustment (GIA) correction that ultimately produces the reducing Antarctic ice sheet thickness, similarly to the rising Global Mean Sea Level (GMSL). This result is contradicted by other more reliable experimental results as the expanding sea ice extension and the cooling surface air temperature.


INTRODUCTION
While the general understanding of the Antarctic climate offers cooling temperatures and expanding sea ices, [1] propose modelling results contradicting this understanding without even mentioning the other data and analyses, and also downplaying the GRACE result huge inaccuracies making their ice thickness computation much less reliable than the contrasting results for temperatures and sea ice extension.
Antarctica has temperatures so low that over the most part of the region ice melting is simply impossible, as the temperature is always below 0°C and on average -7°C.It is only over the Antarctica Peninsula that protrudes towards South America that temperatures are little bit larger, about -5°C on average, and in a couple of months per year there may also be ice melting.What else could be melting are the huge ice shelves that extend from the coast, even if it is unlikely and unproven that the temperatures of the deep oceans below are increasing.GRACE is a system based on the Earth gravity.If some ice sheet melt, then the gravity directly above it changes and this may change the orbit of a satellite that passes over head.A satellite orbiting above a region of reducing mass experiences a reduced gravity acceleration and therefore has a tendency to increase the altitude and reduce the speed.This orbital effect is what is used by GRACE.
The raw GRACE signal is, as expected, very noisy, and the Antarctic ice thickness is not the raw GRACE signal, but the result of a double computation, one based on GRACE, the other completely decoupled.As soon as a vertical top ice boundary position is computed from GRACE, then there is the issue of computing the vertical bottom ice boundary position.The land beneath is lifting because of the post glacial rebound.This motion is not measured, but computed by a glacial isostatic adjustment (GIA) model that is everything but accurate.This result of two inaccurate computations cannot return a precise estimation of the Antarctic ice sheet thickness.

THE GRACE COMPUTATION OF ICE THICKNESSIS NOT RELIABLE
The GRACE estimations are not measurements but actually computations where starting from a noisy signal -the ice upper surface position -the ice thickness is then computed by subtracting the GIA land underneath vertical position -the ice lower surface position -and this result has huge inaccuracies that are being downplayed in the commented paper.Following the principle that any imperfect measure of a climate parameter become perfect as soon as it proves the IPCC narrative of melting ices, rising seas, reducing rainfalls and warming temperatures is true, the authors do not discuss the contrasting results obtained with other techniques that have much better accuracy for temperature and sea ice extent, and do not discuss seriously the accuracy limits of their technique, where the error is much larger than the trend.
After a careful selection of the references to consider -mostly analyses based on direct or indirect climate modelling results -and to neglect -all the analyses built on true measurements, according to [1] "numerous sources" have already confirmed that Antarctica is losing ice at a quickening rate.Hence, they propose a novel and more accurate "evidence".This ultimate evidence is based on the results of the Gravity Recovery and Climate Experiment (GRACE), plus the modelled glacial isostatic adjustment (GIA).According to their purely computational exercise, "over the period 2003-2014West Antarctica has lost ice mass at a rate of -121±8Gtons/year, with an acceleration of the ice mass loss of -18±5 Gtons/year 2 along the Amundsen Sea coast, to produce an overall ice mass loss of -92±10Gtons/year for Antarctica".
If we do consider the information available from multiple other sources for Antarctica, again of expanding sea ice and cooling temperature, we wonder how the authors and the reviewers could ignore the existence of such a large amount of empirical evidence to support one more purely computational exercise.
Measurements and reconstructions are not all the same quality.In the specific, the temperature measurements are the most reliable.Then, the xy measurements of the sea ice extension have relatively good reliability.Both these results show, again, cooling temperatures and expanding sea ices.Finally, we came to the less reliable result, the z ice thickness that strongly depends on the estimation of the position of the top and bottom ice boundaries.This result should not be proposed as the only evidence.
GRACE is certainly an interesting technique but still in its test stage.Sea levels are another possible application of GRACE not requiring an extra information impossible to measure as it is the case of the ice thickness.The first sequence of GRACE gave slightly lowering global mean sea levels.As revealed by [2], this slightly reduction was then immediately corrected to a +2 mm/year global mean sea level (GMSL) rise with just one more round of corrections between the raw signal and the final product claiming the global isostatic adjustment (GIA) was the reason.
With the space gravimeter observations from GRACE it is possible to record changes in the ocean water mass which approximate the mean global sea level changes.Fig. 1 presents this result.The raw data show a slight lowering of the GMSL.Inferring a very questionable global isostatic adjustment (GIA) correction, [3] established a corrected sharply increasing rate.
The difference is significant.The uncorrected data suggest the relative sea level is slightly reducing.The corrected data tell us the absolute sea level is rising significantly.Without the correction, there is no GMSL rise.
The accuracy on the z position of the top ice boundary very unlikely suffer of errors less than ±2 mm/year, as this is the error of the much more computed from the data proposed in possible application of GRACE not requiring an extra information impossible to measure as it is of the ice thickness.The first sequence of GRACE gave slightly lowering global mean sea levels.As revealed by [2], this slightly reduction was then immediately corrected to a +2 mm/year global mean sea level (GMSL) rise with tions between the raw signal and the final product claiming the global isostatic adjustment (GIA) was the reason.
With the space gravimeter observations from GRACE it is possible to record changes in the ocean water mass which approximate the mean 1 presents this result.The raw data show a slight lowering of the GMSL.Inferring a very questionable global isostatic adjustment (GIA) correction, [3] established a corrected sharply increasing rate.
The uncorrected data suggest the relative sea level is slightly reducing.The corrected data tell us the absolute sea level is rising significantly.Without the The accuracy on the z position of the top ice boundary very unlikely suffer of errors less than ±2 mm/year, as this is the error of the much more evolved Global Positioning System (GPS) monitoring of fixed GPS domes (see the SONEL and JPL results for the same domes, [4] and [5]).The accuracy of the z position of the land underneath that is based on a GIA computation is even less accurate (see for example the SONEL and JPL results for the locations where [6] provides the GIA computational results).The GPS based vertical velocities computed by SONEL or JPL for many GPS domes worldwide are very far from the GIA simplistic description with differences of millimeters per year.The GIA correction is indeed very questionable as already noted by [7].A comparison of GPS and GIA results for the key sites of Antarctica is proposed in the Appendix.
The global isostatic adjustment, or GIA, is a model, in which some data are in support (see for example [8]) and other data are in contradiction (for example [9]).The lat and JPL results are also not that supportive for the GIA computation.GIA corrections have been applied to tide gauges, sea level records, satellite altimetry, Ocean and ice mass changes.It seems that without the GIA corrections there is very li room left for a global sea level rise or a reduction of the ice mass [2].

The space gravimeter readings from the GRACE satellites of the global mean sea level
The raw data show a slight lowering.By introducing a very questionable Global Isostatic Adjustment (GIA) correction, [3] established a significant increasing rate.The difference is significant, as the corrected data suggest a trend completely The raw data have not been shown hereafter.The trends are computed from the data proposed in [3] ; Article no.JSRR.2015.205evolved Global Positioning System (GPS) monitoring of fixed GPS domes (see the SONEL ame domes, [4] and [5]).The accuracy of the z position of the land underneath that is based on a GIA computation is even less accurate (see for example the SONEL and JPL results for the locations where the GIA computational results).The based vertical velocities computed by SONEL or JPL for many GPS domes worldwide are very far from the GIA simplistic description with differences of millimeters per year.The GIA correction is indeed very questionable as already of GPS and GIA results for the key sites of Antarctica is proposed The global isostatic adjustment, or GIA, is a model, in which some data are in support (see for example [8]) and other data are in contradiction (for example [9]).The latest SONEL and JPL results are also not that supportive for the GIA computation.GIA corrections have been applied to tide gauges, sea level records, satellite altimetry, Ocean and ice mass changes.It seems there is very little room left for a global sea level rise or a reduction

The space gravimeter readings from the GRACE satellites of the global mean sea level
The raw data show a slight lowering.By introducing a very questionable Global Isostatic Adjustment (GIA) correction, [3] established a significant increasing rate.The difference is significant, as the corrected data suggest a trend completely The trends are Fig. 2 presents the Southern Hemisphere Sea Ice Area Anomaly from [10].The sea ice area is increasing.Fig. 3 presents the Southern Polar Lower Troposphere Temperature from [11].The temperature is decreasing.The ice thickness should certainly not be reducing while the temperatures are cooling and the sea ice is expanding.The data are downloaded from [10] and [11] respectively and analyzed here.
A much ticker than the expected Antarctic ice is also confirmed by novel and more accurate in situ observation [12].Their three-dimensional floe-scale maps of sea-ice draft for ten floes was compiled from two springtime expeditions by an autonomous underwater vehicle to the nearcoastal regions of the Weddell, Bellingshausen, and Wilkes Land sectors of Antarctica.The results show that the floes are much thicker and more deformed than reported by drilling and ship based measurements.The ice thickness in the near coastal and interior pack may therefore be under estimated by existing in situ assessments and the Antarctic sea ice may be thicker than believed before.
The combined GRACE-GIA result is highly speculative, and certainly more to be regarded as a poor computation than a truly accurate measurement.The inaccuracy and unreliability does not emerge at all in the commented work that simply ignores the existence of the other measurements and avoids to mention the actual error of their procedure is orders of magnitude larger than their claim.
If we do consider a reference area for Antarctica of 14•10 and a reference density for ice of 916.7 Kg/m 3 , or 0.9167 tons/m 3 , an error of even ±1 mm/year on the evaluation of the difference between the GRACE based computation of the upper ice boundary movement and the GIA computation of the lower ice boundary movement, by far much less accurate than that, then produce an error on the estimation of the ice loss of many orders of magnitude larger than the claimed ±10 Gtons/year, that is an outrageously small number.
Once more the problem of global warming analyses are the orders of magnitudes, with trends much larger than the legitimate (and sometimes opposite to any logic and other experimental evidence) and errors much smaller than reasonable and the selling as experimental evidence of results that have a very limited experimental support.

CONCLUSION
GRACE is a procedure still under development that suffers of two major sources of inaccuracies, the procedure to compute the vertical position of the ice boundary from the raw mostly noisy satellite signal, that suffer of major uncertainties and somewhat arbitrary corrections, and the procedure that computes the bottom ice boundary with a GIA model, that should never ever have been used for the purpose being only qualitative.
As for the specific of Antarctica the GIA computations certainly overrate the land uplift, not only the error is much larger than the trend, but very likely the trend is of increasing ice thickness rather than reduction, as compatible with all the other measurements.The overall Antarctic ice mass loss since January 2003 is definitively not -92±10 Gtons/year as the actual uncertainty on this result is many order of magnitude larger than the trend and the minus sign is increasingly suspicious.

Fig. 1 .
Fig. 1.The space gravimeter readings from the GRACE satellites of the global mean sea level and their corrected values.The raw data show a slight lowering.By introducing a very questionable Global Isostatic Adjustment (GIA) correction, [3] established a significant increasing rate.The difference is significant, as the corrected data suggest a trend completely different from the raw data.The raw data have not been shown hereafter.computed from the data proposed in