Ice and snow

Abstract  The Qinghai-Tibet Railway(QTR) is the highest and longer plateau linear structure in the world. the uneven deformation of subgrade along Qinghai-Tibet railway casused by the freeze/thaw processes of frozen soil will result to settlement and pavement cracks; the infrastructure construction and human activities changes environmental conditions and in turn has influenced permafrost evolutions. This article employs a based on discontinuous coherent small baseline subset Interferometric Synthetic Aperture Radar(SBAS-InSAR) technique to monitor the surface deformation over the wind volcano section of Qinghai-Tibet plateau permafrost. Using cyclic deformation model to express the seasonal characteristics of the permafrost in order to better eliminate the effects of topographic residuals and atmospheric artifacts. In total，21 L-band ALOS PALSAR SLC images (acquired from 17 June 2007 to 28 October 2010) and the Modis land surface temperature data were employed to cover the experimental site, Qinghai-Tibet Plateau ,china. Analyzed the permafrost deformation and the temperature, aspect, slope, water content and the relationship between the human activity factors; The motion trend along slopes was complicated due to the geomorphological processes, thus interdisciplinary interpretations were needed. Anthropogenic influences on this frail permafrost environment were significant, as evident from the remarkable surface settlement along the embankment of     Qinghai-Tibet Railway.

the results show that the peak to peak annual deformation over natural surface is 2.8~3.5 cm, while that along the Qinghai-Tibet Highway (QTH) and the Qinghai-Tibet Railway (QTR) is 1.2~2 cm. We also find that the surface deformation over frozen soil is negatively correlated to air temperature, with the correlation being -0.58 to -0.24. By comparing with the field temperatures, it is found that the retrieced deformation sequences conform to physical characteristics of permafrost very well. The method presented in this paper and proposed models is more suitable to analyze the surface deformation in permafrost areas.

Keywords Permafrost, SBAS-InSAR, Qinghai-Tibet, Deformation monitoring, correlaion

The Arctic is a generally challenging environment for reliably measuring surface displacement with InSAR. Deriving long-term permafrost stability from such displacement measurements if they succeed is equally complicated.

Seasonal ground deformation through the freeze-thaw cycle of the active layer usually dominates any long-term displacement trend that may indicate instable permafrost. To retrieve the long-term trend the active layer dynamics must be carefully modeled using assumptions on snow melt completion and meteorological drivers such as temperature. This is to accommodate the possibility of significant temporal aliasing in the InSAR time series and to minimize bias to long term displacement trends introduced by the seasonal component. Current approaches to model the active layer dynamics largely restrict themselves to summer scenes because winter scenes contain other large unknown phase signals from snow and ice cover changes that if uncompensated will contaminate the modeling. However, having to omit all winter scenes due to unknown phase bias severely limits how well the active layer dynamics and long-term displacement trend can be separated. This is because winter interferograms often are of better quality than summer interferograms, where large variations in surface water and vegetation reduce InSAR coherence significantly.

In this paper we demonstrate a novel approach to include coherent winter interferograms in the displacement inversion by simultaneously modeling both the spatial patterns of active layer dynamics and of dry snow deposition. The crux of our approach is that we model the snow deposition iteratively from a combination of meteorological drivers and PCA patterns obtained for the set of coherent winter interferograms. Analogous to temperature driving the active layer dynamics we use precipitation as the main driver to model the temporal variation of snow deposition. However, unlike for the case of temperature, spatial homogeneity cannot be assumed for precipitation and consequently we identify the spatial distribution of snow deposition as one of the winter phase PCA eigen-patterns. We refine the separation of active layer dynamics and the temporal and spatial components of snow deposition through several iterative cycles, each first inverting summer (initial estimate of active layer dynamics), then winter (initial estimate of snow deposition) and last the complete set of summer and winter interferograms (improved estimate of active layer dynamics). The final results of this iteration are (i) a map of “active layer dynamic sensitivity”, which can be interpreted as a combination of active layer thickness and ice-richness of the permafrost useful for permafrost classification, and (ii) a map of “snow accumulation potential”, which can be interpreted as a combination of site aspect and vegetation cover influencing the snow deposition. Finally, (iii) we can retrieve a map of potential long-term displacement trends indicating permafrost instability from the model residual.

First results obtained with the novel approach obtained for a large dataset of both RADARSAT-2 and TerraSAR-X high resolution InSAR time series over two sites in the Canadian Arctic, Inuvik, Northwest Territories and Salluit, Quebec suggests that our proposed new inversion scheme is both more accurate and robust compared to previous results using only summer interferograms. We show examples of subtle trends of permafrost instability that were un-detectable at the accuracy level of previous methods. The by-products of active layer dynamic sensitivity and snow accumulation maps both look reasonable and plausible; detailed verification of these maps on the ground will be carried out in the near future.

Many potential environmental and socioeconomic impacts of global climatic change are associated with permafrost. The seasonally thawed layer (active layer) and the long-term effects of climatic change on permafrost can severely disrupt ecosystems and human infrastructure and potentially intensify global warming. Permafrost degradation may affect slope stability, and in order to evaluate certain geohazards, improved permafrost knowledge is essential.

Satellite SAR Interferometry can provide information about ground deformation related to seasonal and interannual thaw and freeze processes in permafrost environments. It is especially valuable in the Arctic due to the large spatial coverage by SAR satellites, and their insensitivity to weather and light conditions. Work performed at Norut contributed to investigate the magnitude of changes related to permafrost at landscape scale and provide surface deformation measurements to discriminate different periglacial landforms and define how they react toward climate/meteorological variations.

Since 2009, we have acquired Radarsat-2 and TerraSAR-X datasets over study sites in northern Norway and Svalbard. Since 2014, Sentinel-1 provide new valuable data for measurements of fast moving periglacial features and changes at large scale. Thanks to its C-band frequency, its short repeat-pass and its wide swath, Sentinel-1 will revolutionize the monitoring of landscape changes in Arctic regions.

The Sentinel-1 time series are still short, but first results are promising. On Svalbard, deformation patterns from Sentinel-1 and TerraSAR-X have been compared around Longyearbyen and in Kapp Linné and show overall a good fit. In northern Norway, by analyzing single interferograms, interesting features related to freezing/thaw and deformation on creeping permafrost landforms can already be detected. Norut will intensify its work the next years in order to understand the spatio-temporal variability of the deformation related to freeze/thaw, to analyze the advantages and limitations of different SAR sensors and to relate InSAR signal to geomorphological contexts and environmental variables.

Ice and snow