• 17:50 Dual-Pol X-Band Pol-InSAR Time Series of a Greenland Outlet Glacier

  Fischer, Georg (1,2); Hajnsek, Irena (1,2) 1: German Aerospace Center (DLR), Microwaves and Radar Institute, Germany; 2: ETH Zurich, Institute of Environmental Engineering, Switzerland

  Show abstract

  The Helheim glacier in south-east Greenland faces a retreat and thinning like many other Greenland outlet glaciers, but showed two particular huge and fast retreats between 2000 and 2005, coinciding with an acceleration of glacier flow [1]. The estimation of glacier retreat and thinning are established procedures based on a variety of remote sensing techniques, e.g. SAR interferometry (InSAR) [1]. However, due to densification processes, e.g. refreezing of meltwater, changes in the mass balance cannot be calculated solely from changes in the volume of a glacier derived from interferometric elevation models [2]. A near surface densification related to refreezing of meltwater leads to a reduction in surface elevation, but does not contribute to a mass loss and thus cannot be attributed to a potential sea level rise. Likewise, the measurement of snow accumulation and its spatial distribution is error prone due to the penetration bias of InSAR measurements. While local data of accumulation or densification give precise information [2], satellite based polarimetric and interferometric SAR (Pol-InSAR) data has the potential to assess these changes in the near surface snow and firn layers with a better spatial and temporal coverage. This study investigates X-band (TanDEM-X) Pol-InSAR data in order to retrieve information about the temporal and spatial variations of surface and near- surface parameters of the Helheim glacier. Although the retrieval of ice scattering extinction with Pol-InSAR at different frequencies (L- and P-band) is possible [3], the modelling of Pol-InSAR signatures of snow and ice is not yet fully understood [4]. Indeed examples exist where the typical Pol-InSAR models (RVuG, OVuG) cannot completely explain the SAR measurements. The expected scattering scenarios are on the one hand surface scattering for wet glacier surfaces in case of melting conditions, which prevents any penetration of the signal and therefore the retrieval of volume parameters. On the other hand, at freezing conditions, snow crystals are not contributing to scattering even at X-band, but introduce differential co-polar phase delays related to the anisotropy of the snow volume, which was successfully modelled for snow depth retrieval in [5]. The scattering in the latter scenario is assumed to originate from ice inclusions or layers with a density transition leading to a change in permittivity. The Helheim super test site is comprised of the longest and most complete time series of dual-pol (HH and VV) TanDEM-X acquisitions over a glacier. Data takes are available almost every 11 days from April 2011 until July 2013 with perpendicular baselines in the order of 200 m. Polarimetric phase differences and Pol-InSAR signatures are used to investigate the temporal evolution of backscattering characteristics, with the goal to find observables which could be related to accumulation or densification processes. In a first assessment, the TanDEM-X measurements show surface scattering in summer during the melt period (May – September), which is confirmed by weather stations in south-east Greenland. Additionally, the interferometric combinations suffer partly from severe decorrelation due to low signal to noise ratio. Over the winter, a build-up of polarimetric HH-VV phase difference is observable, which could be related to snow accumulation. Interestingly, the parts of the glacier which show this evolution of polarimetric co-pol differential phase, exhibit a rather stable behavior of Pol-InSAR signatures, e.g. the height difference of HH+VV and HH-VV channel combinations. If this is interpreted as a negligible elevation spread of different scattering mechanisms, with a simultaneous presence of co-polar differential phase, one could assume a surface like scattering below a certain height of snow, acting as an anisotropic medium. The high accumulation regime in south-east Greenland [6], combined with the typical glacier facies at the ice sheet margin, support this first assessment, which indicates the potential for estimating accumulation patterns. Unfortunately, the sparse availability of in-situ data hampers the interpretation of SAR signatures. Future work will focus on investigating these first hypotheses and the identification of suitable observables of dual-pol TanDEM-X measurements in relation with accumulation or changes in snow structure, which could support a better characterization of glacier mass balance. References: [1] I.M. Howat, I. Joughin, S. Tulaczyk, and S. Gogineni, “Rapid retreat and acceleration of Helheim Glacier, east Greenland”, Geophysical Research Letters, vol. 32, 2005 [2] V. Parry, P. Nienow, D. Mair, J. Scott, B. Hubbard, K. Steffen and D. Wingham, “Investigations of meltwater refreezing and density variations in the snowpack and firn within the percolation zone of the Greenland ice sheet”, Annals of Glaciology, vol 46, no. A252, pp. 61-68, 2007 [3] J.J. Sharma, I. Hajnsek and K.P. Papathanassiou, “Estimation of glacier ice extinction using long-wavelength airborne Pol-InSAR”, IEEE Transactions on Geoscience and Remote Sensing, vol. 51, no. 6, pp. 3715-3732 June 2013. [4] J. Dall, K.P. Papathanassiou and H. Skriver, “Polarimetric SAR interferometry applied to land ice: modeling”, Proceedings of the European Conference on Synthetic Aperture Radar, pp. 247-250, 2004 [5] S. Leinss, G. Parrella and I. Hajnsek, "Snow Height Determination by Polarimetric Phase Differences in X-Band SAR Data," IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. PP, pp. 1-17, 2014, doi: 10.1109/JSTARS.2014.2323199 [6] J. Cappelen, “Greenland - DMI Historical Climate Data Collection 1873-2012”. Danish Meteorological Institute, Technical Report 13-04, Copenhagen, 2013

  
   

• 16:50 Retrieval of sea ice salinity with fully polarimetric SAR

  Grahn, Jakob (1); Brekke, Camilla (1); Eltoft, Torbjørn (1); Dierking, Wolfgang (2) 1: University of Tromsø, Norway; 2: Alfred-Wegener-Institut, Germany

  Show abstract

  The Arctic sea ice cover is one of the components in the climate system that is changing most rapidly at the moment, even faster than predicted by models. This makes observations of sea ice valuable for understanding climate change. In addition, activities such as shipping, oil exploration, tourism and fishery is likely to increase in the Arctic as the sea ice retreats. To minimise the risk of maritime and environmental disasters, accurate information and forecasts about sea ice conditions is needed. The arguably most important parameter of sea ice is the thickness. Previous studies have shown the potential of using polarimetric synthetic aperture radar (SAR) for thickness estimation. For instance, Kwok et al (1995) showed that by inverting a theoretical scattering model (Nghiem et al, 1993) with an explicit thickness dependence, thicknesses in the 0-10 cm range can be retrieved. The retrieval does however seem to saturate at a certain thickness, which limits the approach to thin sea ice. However, Nakamura et al 2005 has indicated that there is an empirical relation between the salinity of the ice and the thickness. Specifically, they show that the co-polarised HH-VV ratio relates ice thicknesses in the 20-120 cm range and argue that this is due to the ratio being sensitive to the amount of brine. Thus, salinity could be a proxy for thickness for thicker sea ice. We therefore consider retrieval of sea ice surface salinity. We examine three fully polarimetric SAR images of varying incidence angle in C-band. The images are acquired by RADARSAT-2 and contains flat first year ice close to the Hopen island, in the Barents sea. We consider the inversion of a simple scattering model for retrieval of the surface permittivity, which in turn is empirically related to the volume fraction of brine (Vant et al, 1978). The scattering model contains a surface scattering contribution from the ice-air interface and a volume scattering contribution from the brine and air structures within the ice. Given the fully polarimetric data, we are able to model properties such as shape and orientation of the brine and air structures. In addition to the polariemtric SAR images, coincident ice samples from the site has been collected. These give us insight on how to model the shape and orientation of the bulk structure and are also used for validating the retrieved salinity estimates. References: Kwok et al, 1995: R Kwok, SV Nghiem, SH Yueh, and DD Huynh. Retrieval of thin ice thickness from multifrequency polarimetric sar data. Remote sensing of environment, 51(3):361–374, 1995. Nakamura et al, 2005: Kazuki Nakamura, Hiroyuki Wakabayashi, Kazuhiro Naoki, Fumihiko Nishio, Teruo Moriyama, and Seiho Uratsuka. Observation of sea-ice thickness in the sea of okhotsk by using dual-frequency and fully polarimetric airborne sar (pi-sar) data. Geoscience and Remote Sensing, IEEE Transactions on, 43(11):2460–2469, 2005. Nghiem et al, 1993: SV Nghiem, R Kwok, Jin Au Kong, and RT Shin. A model with ellipsoidal scatterers for polarimetric remote sensing of anisotropic layered media. Radio science, 28(5):687–703, 1993. Vant et al, 1978: MR Vant, RO Ramseier, and V Makios. The complex-dielectric constant of sea ice at frequencies in the range 0.1–40 ghz. Journal of Applied Physics, 49(3):1264–1280, 1978.

  
   

• 18:10 What can polarimetric SAR data tell us about snow accumulation on polar ice sheets?

  Linow, Stefanie (1); Dierking, Wolfgang (1); Rack, Wolfgang (2) 1: Alfred-Wegener-Institut, Helmholtz-Zentrum für Polar- und Meeresforschung, Germany; 2: Gateway Antarctica, Center for Antarctic Studies and Research, University of Canterbury, New Zealand

  Show abstract

  Climate processes in the polar regions affect global sea level variations, and also impact ocean circulation patterns by changing the freshwater flux into the oceans. Those processes need to be carefully monitored and the assessment of the mass balance of the polar ice sheets is crucial in this context. Here, microwave remote sensing instruments can provide observations in regions which are otherwise difficult to access. The mass balance of the polar ice sheets can be described in terms of mass gain by snow accumulation and mass loss by melting and iceberg calving. The effect of melt processes and the export of ice mass beyond the grounding line of the ice shelves can be observed using passive microwave radiometers, scatterometers, imaging radar, and laser and radar altimeters. Also for gathering information on snow accumulation rates, different sensors are in use. Several snow accumulation rates retrieval methods exist, but are still afflicted by large uncertainties. Satellite data from sensors with large spatial coverage and high temporal sampling rate (such as passive microwave radiometers and scatterometers) are used to map snow accumulation on continent-wide scale, and SAR data have been applied on regional and local scales. In the case of SAR data, the backscattering intensity depends on the snow grain size and the thickness of the annual firn* layers, but also on firn density, temperature and wind regime. As large grain sizes and thin annual layers are characteristic for low accumulation, and small grains deposited in thicker layers are typical for high accumulation, radar is used to indirectly measure accumulation rates. Radar backscattering characteristics of firn depend on size, shape, orientation, volume fraction, and absorption loss of the scattering elements, and on the absorption loss of the background medium. Due to the high scatterer density, coherent effects between the scatterers also influence backscattering characteristics. The radar backscattering coefficient additionally exhibits an azimuthal anisotropy which is related to wave-like undulations of the ice surface on scales of decimetres to tens of metres (sastrugi). Since the surface scattering contribution of dry snow and firn is very low, it is assumed that also the surfaces from former years, which are buried under the most recent snow layers, contribute to the radar scattering. The complexity of the radar scattering processes not only requires a number of environmental parameters, such as temperature and wind, to be considered in snow accumulation retrieval approaches, but also calls for an increase in the number of directly observed variables. The objective of our investigation is to analyse whether multi-frequency and multi-polarization SAR data improve the retrieval of accumulation rates and provide additional insight into the interaction between radar waves and firn. We will study polarimetric SAR data from different sensors (Terrasar-X, Radarsat-2) in comparison to conventional SAR systems (Envisat ASAR and possibly Sentinel-1). These data were acquired over the Kottas Traverse in Dronning Maud Land, Antarctica. For this site, a large number of in-situ measurements of snow accumulation are available for validation purposes. We will focus on the differences of the azimuthal anisotropy observed at C- and X-band, taking into account the scene headings and radar penetration depths. We will also examine to which extent the phase difference and correlation between the HH- and VV-channel might give information about snow and ice conditions that is complementary to the radar intensity patterns. * Firn is an intermediate stage between snow and glacial ice, with densities ranging from 400 - 830 kg/m³.

  
   

• 17:10 Automated sea ice classification based on dual polarimetric TerraSAR-X data

  Ressel, Rudolf Andreas; Lehner, Susanne DLR, Germany

  Show abstract

  In this work, we examine the performance of an automated sea ice classification algorithm based on dual polarimetric TerraSAR-X data. The first step comprises the extraction of an array of polarimetric features for each pixel neighborhood. These features are obtained from HHVV dualpol stripmap images. In a second step, the feature vectors are fed into an artificial neural network to classify each pixel into an ice type. The first part of our analysis addresses the predictive value of different subsets of features for our classification process. To improve classification, we additionally compute second order statistics on these polarimetric feature vectors to investigate the possibility of improving classification by these extended feature vectors. Different neural network configurations are then explored for optimal classification performance. Performance is compared to a decision tree classification. The investigation is carried out on datatakes from various ice-infested regions (Svalbard, Kara Sea, Baffin island coast). The results on these datasets indicate a high reliability of a trained dual polarimetric classifier for different regions and far superior performance compared to a single-pol texture based classifier of ice types. Performance speed and accuracy show that our algorithm is a promising approach for a near real time operational use.

  
   

• 17:30 Polarimetric and Interferometric Methods to determine Snow Depth, SWE, and the Depth of Fresh Snow

  Leinss, Silvan (1); Wiesmann, Andreas (2); Lemmetyinen, Juha (3); Parrella, Giuseppe (4); Hajnsek, Irena (1,4) 1: ETH Zürich, Switzerland; 2: Gamma Remote Sensing, Switzerland; 3: Finnish Meteorological Institute, Finland; 4: Institute of Remote Sensing Technology, German Aerospace Center, DLR, Germany

  Show abstract

  Microwaves interact with snow on a wide variety of effects. The specific effect of interaction depends mainly on liquid water content which determines the penetration depth. The penetration depth changes by orders of magnitude at the transition from dry snow to wet snow. Further affects appear due to roughness of the boundary layers soil-snow and snow-air and discontinuities in the snow pack. Even the microstructural anisotropy of snow affects the interaction with microwaves. Therefore, complementary methods have to be applied to access the medium snow in all it's facets. Here, I will demonstrate different polarimetric and interferometric methods, which show a sensitivity to certain properties of snow. Surface-sensitive methods apply only for wet snow, due to the very limited penetration depth, especially at high frequencies. I will demonstrate, that the height difference of two single-pass interferograms taken by TanDEM-X is equivalent to the height of accumulated snow within one winter season. In contrast to wet snow, I will demonstrate, that it is possible to derive the total snow water equivalent (SWE) of dry snow from a time series of differential interferograms. Going beyond height or SWE sensitive methods, I will demonstrate a polarimetric method, which is sensitive to the microstructure of snow. A simple model will be presented, which shows that the phase delay of the HH vs. the VV polarization is sensitive to the amount of fresh snow and to the temperature-gradient driven metamorphism of snow.

  
   

• 18:30 Round Table Discussion

  Show abstract

  Round Table Discussion

  
   

Highlights

In May 2013, the Biomass mission concept was selected to become the next in the series of satellites developed to further our understanding of Earth.

Released: 21/06/2013

REPLAY EARTH EXPLORER 7 USER CONSULTATION MEETING (PART 1)

Released: 11/03/2013

POLinSAR 2013 Workshop

Released: 22/02/2013