INTEGRAL - Objectives

General Objectives • Specific Objectives • State of the Art



 

Project title: Interferometric Evaluation of Glacier Rheology and Alterations
Acronym: INTEGRAL

General Objectives

The general objective of the INTEGRAL initiative is to promote an advanced observation technology for the unsupervised detection, precise measurement and variational analysis of ice motion / deformation on large European glaciers based on the complementary use of radar interferometry and interferometric altimetry, and to support natural exploration, social-economic activities and subsequent surveys in the nival environment with equivalent rheological models and appropriate information on the glacier regime in the form of new value-added INSAR products. Our polar idea is to enhance the detailedness, accuracy, integrity and versatility of glacier interferometric models yet without involving complex process artifices and to demonstrate new utilities of differential radar interferometry to operational users working with SAR data from post-operational, operational and upcoming systems such as E-SAR, ERS-1/2, SRTM, ENVISAT, RADARSAT-1, 2, 3, ALOS and CRYOSAT.


Specific Objectives

The INTEGRAL project is, thus, focused on methodological aspects and empirical issues of glacier interferometry, and major attention is paid to the following specific objectives:

  1. Thorough analysis of existing knowledge about glacier rheology and alterations in different test environments, identification of quality standards, information needs and gaps, critical review of available remote sensing techniques and advocating the design of an advanced approach to the evaluation of glacier rheology and alterations from interferometric data, both air- and spaceborne.
  2. Development of new and optimisation of available algorithms and program tools for reliable modelling of glacier dynamics from differential radar interferometry (DINSAR) including
    • phase gradient approach to modelling glacier morphology and rheology,
    • transferential approach to measuring frontal velocities of tidewater glaciers,
    • SAR offset-tracking approach to measuring large and incoherent glacier displacements,
    • combined approach using phase gradient, transferential, offset-tracking and traditional DINSAR methods.
  3. Design of efficient techniques for geometric processing and fusing radar altimetry data with SAR interferograms aimed at precise geocoding and upgrading the information content of glacier models.
  4. Detection, measurement, interpretation and variational analysis of glacier motion / changes and numerical modelling of the glacier regime at regional and local scale regarding
    • spatial and temporal variations of the magnitude and direction of glacier motion and strain rate,
    • ice-surface velocity structure, ice-flow instability and the detection of glaciers with basal sliding,
    • changes in the accumulation rate at glacier tops and fluctuations of the specific mass balance,
    • measurement of frontal heights of tidewater glaciers and numerical modelling of calving flux,
    • analysis of deformations in the fast-sea-ice cover in the immediate proximity of the glacier face and validation of the transferential approach,
    • assessment of main tendencies in the state of land ice resources in response to climate change.
  5. Verification of results, methodological tests and accuracy control during field surveys and observations in glacial test sites using precise geodetic equipment and different surveying techniques.
  6. Production, demonstration and dissemination of the series of value-added INSAR products including
    • rheological models showing ice-surface velocities (velograms) and glacier strain rate (fluxograms),
    • interferometric "snap-shots" and "movies" of glacier activity, e.g. surges, waves, calving, etc.,
    • slope models representing the morphology of the steady glacier surface (topograms),
    • regional inventory of ice shores and their changes in the Barents Sea region,
    • satellite image maps of glacier rheology, specific mass balance and alterations at different scales.
  7. Integration, generalisation and implementation of the project results, their appraisal and incorporation in the frameworks of the GMES Integrated Projects on land cover, marine environment and global change with the final goal to substantially contribute to the proper maintenance of global environment observation systems and to determine perspectives for future research.

Specific science questions related to geophysics, short-term climatic trends, hydro-thermal regime of glaciers and global change issues will be considered as well, especially with reference to their impact on hydro-power production, safe shipping, environmental management, natural hazard monitoring, regional planning and tourism in the European Arctic & Alpine Sector. The results obtained by different techniques shall be intercompared, verified during field surveys in several arctic & alpine test areas and implemented at large European administrative and industrial organisations. Output products / services will be appraised and incorporated in the GMES frameworks with the final goal to substantially contribute to the proper maintenance of global environment observation systems and to determine perspectives for future research.


State of the Art

Synthetic aperture radar data and derived products have proven to be a valuable source of information for glaciological studies. The use of spaceborne SAR imagery has increased in recent years following the advancement of satellite radar interferometry, which became especially popular among experts studying glacier rheology, because of its notable data availability and astonishing sensitivity to the ice flow / deformation in the sub-centimetre range (Goldstein et al. 1993, Bamler and Hartl, 1998). The INSAR method ensures the ice-flow detection even over snow-covered glacier surface that cannot be performed with optical observation techniques. Relative velocity accuracy on the order of 2.3 m per year (6 mm per day) was reported to be achievable over interiors of extensive glacier sheets using ERS-1-SAR repeat-pass interferometry (Joughin et al. 1996). A full separation between the impacts of glacier topography and glacier motion on the interferometric phase is a prerequisite for attaining such a high performance.

Differential interferometry (DINSAR) based on differencing between the original SAR interferogram and the reference interferogram, which, in ideal case, does not contain the phase term related to the ice motion, allows the impact of glacial topography to be compensated and the ice-surface displacement to be evaluated. Numerous examples have already been shown of successful DINSAR applications to monitoring ice-sheet motion (e.g. Rignot et al. 1997), identifying surge effects on ice fields (Murray et al. 2002), measuring outlet glacier velocities (Rabus and Fatland 2000), mapping three-dimensional flow of large glaciers (Joughin et al. 1998) and studying flow instability (Dowdeswell et al. 1999). Although the theory of conventional DINSAR is well established (Gabriel et al. 1989, Joughin et al. 1996, Mohr et al. 1998), the technological perfection in converting differential SAR interferograms to a surface-velocity vector field has yet to be completed. The most serious complications and restrictions to the performance of this generally promising technique are due to

  • the significant phase noise and poor coherence of interferometric pictures taken over glacier exteriors and fast moving areas with rough surface,
  • the uneven character of glacier flow and frequently insufficient quality of the reference model,
  • the effect of temporal decorrelation and difficulties in acquiring new INSAR data with short repetition interval,
  • the fact that only the velocity component in satellite look direction can be derived from differential SAR interferograms obtained from parallel orbits,
  • orbital data inaccuracies, algorithmic complexity and processing errors, mostly at the stage of interferometric phase unwrapping and geocoding (Rott and Siegel 1998, Mohr and Madsen 1999, Sharov et al. 2002).

The DINSAR velocity map of any glacier is reputed as a complex and expensive product that requires up to 10 processional steps and quite large volume of computations. The general quality of such product strongly depends on the environment and is commonly low over glacier exteriors (Forster et al. 1999). This explains why reports on using INSAR data for the measurement of frontal glacier velocities are quite few in number. All in all, the presently achieved accuracy of DINSAR rheological modeling, at least in marginal glacier parts, is believed to be far away from the theoretically established limit.

There is thus a natural desire to enhance the accuracy and information contents of glacier interferometric models, and to try other approaches to DINSAR data processing that do not involve complex process artifices and do not require precise topographic reference models. In the INTEGRAL project, we will use the phase gradient approach to rheological modelling of test glaciers that allows the unsupervised detection of glacier motion and modelling of the glacier strain rate to be performed without interferometric phase unwrapping and without inappropriate assumption on the equality of glacier velocities at two different instants of time (Sharov 2003). The surface flow direction and glacier velocity will be estimated in three dimensions at any one time by measuring two components of the flow using dual-azimuth processing (Mohr et al. 1998) or by assuming surface-parallel flow. This assumption is believed to be the best in the absence of measurements of the vertical component of velocity.

Phase gradient approach to the ice motion measurement from several SAR interferograms obtained either from parallel or opposite orbits requires some reference values on the relation between the velocity gradients at different instants of time. We shall use the conventional 2-pass DINSAR technique for the estimation of this relation. The accuracy of reference DEM is, more or less, insignificant in this case. The final integration of the velocity gradients aimed at the determination of absolute glacier velocity will be performed once at the very end of the processing chain, thus reducing the error propagation.

Under long time intervals between SAR surveys, the SAR offset-tracking technique can provide a reasonable alternative to the conventional DINSAR for measuring large and incoherent glacier displacements (Rott et al., 1998; Michel and Rignot, 1999; Strozzi et al., 2002). The offset-tracking procedure in the azimuth direction may be combined with the conventional DINSAR in the slant-range direction in order to retrieve 2-dimensional displacement map, when only SAR data from parallel orbits are available. An efficient albeit very simple transferential technique based on the analysis of the fast-sea-ice motion away from the shore as a result of glacier flow (Sharov et al. 2000) will be applied to measuring frontal velocities of tidewater glaciers from winter SAR interferograms.

In this concern, an upcoming launch of the Phased Array type L-band Synthetic Aperture Radar (PALSAR) on board the ALOS Japanese satellite in 2004 might provide a valuable source of data for the INTEGRAL project. The PALSAR sensor will be capable of acquiring high-resolution radar scenes with alternated polarisation at 1.5-month repetition interval within the stripe from 81° S to 81° N latitude. NASDA's simulations of ALOS orbit showed that the spatial repeatability of orbits (after every 45 days) is assumed to be quite good and the length of perpendicular baselines is expected to be shorter than 500 m in nearly 89% of cases. This means that most PALSAR interferometric pairs will be suitable for the change detection and surface motion estimation. Besides, it is believed that the PALSAR using the L-band frequency and operating at oblique looking angles might provide INSAR pairs with higher coherence comparing to the available C-band SARs.

Studies in the past have indicated that the combined use of spaceborne multi-pass INSAR and radar altimetry data might be very expedient for reliable modelling of glacier morphology and dynamics (Markham & Morris 2002). The launch of the ERS series of altimeters in 1992 introduced a new measurement system that has completely altered our knowledge and understanding of ice mass fluxes. Five years of ERS data were sufficient to reduce the mass imbalance of 64% of the Antarctic ice sheet to ± 60 km³ per year (Wingham et al. 1998a), 9 years of data provided the first evidence of dynamic thinning in western Antarctica (Shepherd et al. 2002). Recent advances have extended measurements to include the floating ice shelves at the ice sheet margin and to estimate rheological parameters (Testut et al. 2000). Although much progress has been made in the determination of ice mass fluxes, the ERS altimeters are reaching the end of their life, and three new altimeter satellites will take over their role: ENVISAT (launched March 2002), ICESat (launched January 2003) and CryoSat (to be launched mid-2004). These satellites will provide continuation of the ERS time series and increased spatial coverage.

The Cryosat mission has been defined in order to determine fluctuations in the mass of the Earths major ice fields, including the small ice caps and glaciers, as present observations are deficient in both space and time (Wingham et al., 1998b). Improved spatial resolution will be achieved using novel synthetic aperture- and interferometric-radar altimeter (SARIN) operation modes that are capable of resolving elevation trends to within 3.3 cm per year over small (10000 km²) areas, with an earth footprint of as little as 250 m and a ground track spacing of ~ 2 km at a latitude of 60°. For advanced algorithm development, data simulation, and calibration and validation experimentation, a Cryosat orbit, radar-echo and control-system simulator has been developed (Wingham et al., 2001) for provision of synthesised level 1b and 2 data products from each instrument operation mode. The INTEGRAL objectives show essential thematic overlaps with the CRYOSAT specific tasks and indicate promising opportunities for SAR data fusion. The most important methodological and practical causes for fusing interferometric image and altimetry data can be summarised as follows:

  • joint interpretation of SAR data and getting new glaciological knowledge on rheological variables at both regional and local scale, e.g. measuring frontal heights of tidewater glaciers / ice walls above the fast-sea-ice surface along all ice shores in the Barents Sea region. It is worth noting that the information about frontal heights of tidewater glaciers and ice shores given in available topographic maps and hydrographic charts is extremely scarse, obsolete and inaccurate (Sharov et al. 2003);
  • providing basic vertical control for enhanced geometric processing of single SAR interferograms (baseline refinement, phase-offset determination, topographic phase removal, etc.) and INSAR block adjustment, e.g. precise geocoding and co-registration of multi-pass SAR data for the reliable assessment of ice-shore changes. In the European Arctic, frontal heights of tidewater glaciers reach 100 meters and, in radar images, the effect of foreshortening makes up several hundred meters (Figure 1);
  • c) "intervalidation" and interpolation of "INSARIN" data / products and analysis of factors influencing the accuracy of interferometric modelling, e.g. atmospherical effects, declined coherence of the interferometric signal backscattered from the rough glacier surface, deformation of the fast-sea-ice cover at glacier fronts or radar penetration into dry snow.

In the INTEGRAL frameworks, special methods for interlayer fusion of INSAR and altimetry data using optimal gradient filters, slope-matching procedures and inverse co-ordinate transformations will be designed and tested.


Figure 1. Fusing SAR image-and altimetry data (a) over European ice coasts (b)



Home • Pre-history • Consortium • Objectives • Test Sites • Methodology • Data
Workplan • Results • Publications • References • Gallery • Events • Links • FTP

© INTEGRAL Coordination, Joanneum Research. Last update: 01.04.2004