the glacier ice behaves very much like water in a stream or a river, where the shape of the bottom is not shown on the surface? overlying layers of ice simply will flow to the sides, as the upheavals rise from the bottom.
Whoa there. Bottom upheavals raise the ice above them, sometimes many cubic km of ice forced thousands of meters overhead. The nearly flat isochronal surfaces in that 95,000 years of upper ice are bent smoothly upwards well into the Holocene and sometimes into firn.
Indeed the Panton paper uses slope departure from horizontal as a proxy for automatic identification of basal upheavals (which have poor contrast in most radar designs). The angle of bending tapers off smoothly in younger ice. That is what's being measured in Fig.1B.
Bottom upheaval pressure also cause the ice above to flow slightly to the sides, manifested as isochronal lines becoming slightly closer over the bump than before. This got caught to a certain extent in the MacGregor 2015 dataset which did not consider deformations (#346 #358). However the radar archive is so vast they could not consider specific cases at high resolution in their Greenland-wide study.
Ice being incompressible, any new volume injected or forced by folding into the bottom has to be accommodated between these two mechanisms, thinning and uplift. This fractional allocation, maybe 15% thin plus 85% uplift, has never actually been measured in a specific case.
For starters, no one has ever tabulated the volumes of ice displaced. That would require much more focused radar tomography than present-day surveys (as called for in the Panton paper). Many uplifted regions have only a single cross section not aligned with flow nor positioned representatively with respect to the global deformation structure. Think of a low budget CAT scan -- the doctor wants to see if your liver is enlarged but is only given a random abdominal plane.
I would do this by working solely with the two brothers (91 kyr), three sisters (46.5 kyr) and Holocene start *14.5 kyr) -- they're ubiquitous and an adequate proxy for the hundreds of minor isochrons. It's not difficult to pick the layers with the bezier tool of gimp, measure the intervening area, and plot the separation between them.
It's probably better to precede inductively from case studies of experimentally accessible, limited-extent, isolated low-flow deformations like Eqip having many radar cross sections than to initially take on the remote, massive, inadequately gridded Zachariae and Petermann. No one has a clue yet what is going on -- that is why Panton terms them UDRs and why MacGregor sidesteps the whole subject.
A single flight line often provides internal calibration via long uneventful stretches preceding the deformation in which a constant separation defines the unperturbed normal. A few isochrons thus suffice to estimate how thinning thins out with vertical height. Panton gives three reasons to expect linearity -- can't go wrong with that as everything is linear to first order, all that we could aspire to here.
Basal deformations are not always easy to detect, delineate or differentiate from draping over basal topography. Panton's automated method discards sloping isochrons when there is significant bedrock slope but still comes up with 6-7% Greenland affected. However steep bedrock prominences may actually be conducive to deformation initiation in view of ice sheet flow.
Contrast in radargrams is a complex issue. The remedies have some similarities to those for Landsat and Sentinel in that optimal corrections must differ regionally to respect patchiness. Isochron striations bring in linear anisotropy as do crevasse fields that must be exploited by the filters. There are also very pronounced flow anisotropy effects in Petermann fold belt radar as noticed by Mundel.
Image intensity in satellite and ice radar data actually have physical meanings. Those becomes altered, perverted, or even lost in enhancement processes. Some people choke on the whole idea. However it's ultimately about having to optimize information extraction from very expensive data according to the purpose at hand. The original doesn't cut it.
Panton 2014 introduced elliptical band pass filtering empirically adapted to sloped and even curved isochrons. That had a wonderous effect on inter-band noise (and so tracing isochrons) but was ill-adapted to diffuse deformations, even making them worse. Panton 2015 sought to apply these to the voluminous Cresis archive with minimal human intervention.
Image segmentation followed by say ImageJ adaptive contrast applied to deeper ice does much better in pulling out faint upheaval regions. That can be problematic depending on the existence of reflectors and their diffusivity. Tracking isochrons through flares is yet another special situation in radargram enhancement important to isochron continuity In Photoshop or Gimp terms, this is masking the image appropriately before each effect is applied and then reassembling from the parts.
The creative part comes with making an effective mask without tedious manual tracing of boundaries on thousands of individuals radargrams that have built up over 22 years of idleness. Masks can often be defined by semi-automated methods but it's hard to envision these working for more than one year's radar design.
The alternative is optimizing the radar design so that it does a better job on the deformation layer from the get-go. That may be feasible for sled radar at selected sites (eg Eqip or the 2015 Renland ice cap).
The inconvenient truth is a very substantial part of north-central Greenland has significant basal issues, much more than we are acknowledging today. It is complete folly to keep applying the same old classical ice sheet models to the GrIS without first getting to the bottom of these deformations.
Below I relocated some text and links that were previously above. Panton's work was previously covered on these forums at the links below.
http://forum.arctic-sea-ice.net/index.php/topic,400.msg36170.html#msg36170 Zachariae
http://forum.arctic-sea-ice.net/index.php/topic,154.msg34112.html#msg34112 Jakobshavn
http://forum.arctic-sea-ice.net/index.php/topic,909.msg36465.html#msg36465 radon transform
Automated mapping of near bed radio-echo layer disruptions in the Greenland Ice Sheet
C Panton NB Karlsson
http://www.sciencedirect.com/science/article/pii/S0012821X15006597
One of the key processes for modulating ice flow is the interaction between the ice and the bed, but direct observations of the subglacial environment are sparse and difficult to obtain. In this study we use information from an extensive radio-echo sounding dataset to identify areas of the Greenland Ice Sheet where internal layers have been influenced by near-bed processes. Based on an automatic algorithm for calculating the slope of the internal radio-echo layers, we identify areas with disrupted layer stratigraphy. We find that large parts of the northern portion of the ice sheet are influenced by locally confined mechanisms that produce up-warping or folds in the layer stratigraphy inconsistent with the surface and bed topography. This is particularly evident at the onset of ice streams, although less dynamic areas close to the ice divide also contain imprints of layer disturbances. Our results show that the disturbances are found in many different flow and thermal regimes, and underscore the need to understand the mechanisms responsible for creating them.
Tracing Internal Radar Layers in the Greenland Ice Sheet
Christian Panton PhD thesis U Copenhagen 108 pages
http://www.nbi.ku.dk/english/research/phd_theses/phd_theses_2015/christian_panton/Christian_Panton.pdf
Internal layers in radio-echograms from the sounding of ice sheets have long been a valuable resource in glaciology, but their usefulness have been limited by availability of traced (digitized) layers. To speed up this process, we have developed an algorithm for semi-automatic tracing the internal layers and a fully automated algorithm for mapping the layer slope.
The layer slope is inferred by the intensity response to a slanted Gaussian filter, from which layers can be traced using an active contour model. With these techniques we show that it possible to trace internal layers over distances of hundreds kilometers with minimal operator intervention, and the methods have been successfully validated between two Greenland deep ice cores with internal match points.
In order to remove any operator assistance, we show how the layer slope can be used to detect disturbances in the deep radio-stratigraphy of the Greenland Ice Sheet. We find that the disturbances are scattered over the northern part of the ice sheet, with the highest density upstream from the Petermann glacier. The disturbances do not seem to be correlated with surface velocities and can be found on, and close, to the ice divide. These results highlight the need for high resolution mapping of the interior ice sheet to understand the dynamical nature of the basal environment.