In recent years, the use of satellite-borne synthetic aperture radar for snow avalanche detection has increased. The ability of SAR to continuously monitor large regions has led to operational use of Sentinel-1 borne avalanche detection, or rapid mapping of extreme events [1]–[3]. It is thus critical to quantify the performance and accuracy of avalanche detection in SAR images. Thus far, validation efforts have either been selective, or have focused on single avalanche cycles.

Knowledge of the spatial and temporal occurrence of avalanche activity is critical for avalanche forecasting and hazard mapping. We present an automatic avalanche monitoring system for Norway. Our system enables us to monitor the release of large avalanches daily. The system monitors the forecasting regions of the Norwegian Avalanche Warning Service continuously throughout the winter. We apply our detection algorithm to an RGB composite of two successive Sentienl-1 images from the same orbit. Detections from several orbits are combined to improve the estimate of the avalanche release time. Our probability of detection depends on the avalanche type and size and is 57% combined.

The Sentinel satellite missions of the European Copernicus programme provide comprehensive data for long-term routine observations of the global environment. A key parameter for climate monitoring, hydrology and water management is the seasonal snow cover. We developed, implemented and tested a novel approach for monitoring snow extent and snowmelt area, exploiting the synergy of imaging radar of the Sentinel-1 mission and multispectral optical imagery of the SLSTR sensor on Sentinel-3. We describe the processing steps and algorithms for generating the synergistic snow cover product and show examples for snow cover maps over Europe based on single mission data as well as the final synergistic product.

The CoReH2O mission design is mature; there are two viable technical configurations, and the breadboards of key hardware components, such as the dual-polarized antenna array feeds and high power amplifiers have been built and tested. The baseline retrieval algorithm has been intensively tested using simulated and experimental data. The tests confirm that the threshold performance for SWE products can be met under a wide range of snow conditions. Data from further field campaigns by the airborne SnowSAR acquired Austria, Canada, and Alaska from November 2012 through April 2013 are being analysed to investigate retrieval performance over alpine, glacier and tundra terrains.

In this paper, we propose a general classification method for determining ice cover on lakes using ENVISAT ASAR data. The method assumes a Gaussian mixture model for feature vectors extracted from a local neighborhood around each pixel. We focus on Lake Päijänne in this work, and the classification results are validated against in situ observations of freeze up dates and a snow cover product from the optical MODIS sensor. The method is fully automatic, but it requires a manual post-processing step to remove poor results. The classification of the selected scenes compared fairly well to both sources of validation data. However, the manual selection step is time-consuming and subjective.

In this paper, we propose a method to monitor the extent of different facies on glaciers on Svalbard. The extent of the firn facies can be used as an indicator for trends in mass balance. We produce an averaged image from Envisat ASAR wide swath mode scenes for an entire winter season. The averaged image is segmented into three classes using a Gaussian Mixture Model. The method is able to detect the area covered by firn for years when the glacier has negative or zero mass balance, while quantitative results are difficult to achieve for years with positive mass balance.

The COld REgions Hydrology High-resolution Observatory (CoReH2O) satellite mission was selected for detailed scientific and technical studies within the Earth Explorer Programme of ESA. The sensor is a dual frequency SAR, operating at 17.2 GHz and 9.6 GHz, VV and VH polarizations The mission will deliver spatially distributed snow and ice observations to improve the representation of the croysphere in hydrological and climate models. Primary parameters are the extent and water equivalent (SWE) of the snow pack and snow accumulation on glaciers. Scientific preparations of the mission include the development and testing of algorithms for retrieval of snow parameters, studies on synergy of CoReH2O-type snow products with passive microwave measurements, the assimilation of satellite snow data in process models, and field experiments. Performance of retrievals for snow extent and SWE was tested with simulated and experimental data, including Ku- and X-band SAR images of the airborne SnowSAR system.

The COld REgions Hydrology High-resolution Observatory (CoReH2O) satellite mission has been selected for detailed scientific and technical studies within the ESA Living Planet Programme. The mission addresses the need for distributed snow and ice observations to improve the representation of the cryosphere in climate models and the prediction of the water cycle. The sensor is a dual frequency SAR, operating at 17.2 GHz and 9.6 GHz, VV and VH polarizations. This configuration enables the decomposition of the scattering signal for retrieving snow mass (SWE) and other physical properties of snow and ice. A major task for mission preparation is the development and testing of algorithms for SWE retrieval. The baseline algorithm applies a constrained minimization approach, matching forward computed and measured backscatter by iterating for SWE and effective grain size of the snow volume. Experimental campaigns on Kuand X-band backscatter of snow deliver important information for retrieval development and validation.

Snow is a critical component of the global water cycle and climate system, and a major source of water supply in many parts of the world. There is a lack of spatially distributed information on the accumulation of snow on land surfaces, glaciers, lake ice, and sea ice. Satellite missions for systematic and global snow observations will be essential to improve the representation of the cryosphere in climate models and to advance the knowledge and prediction of the water cycle variability and changes that depend on snow and ice resources. This paper describes the scientific drivers and technical approach of the proposed Cold Regions Hydrology High-Resolution Observatory (CoReH2O) satellite mission for snow and cold land processes. The sensor is a synthetic aperture radar operating at 17.2 and 9.6 GHz, VV and VH polarizations. The dual-frequency and dual-polarization design enables the decomposition of the scattering signal for retrieving snow mass and other physical properties of snow and ice.

The COld REgions Hydrology High-resolution Observatory (CoRe-H2O) satellite mission has been selected for scientific and technical studies within the ESA Earth Explorer Programme. The mission addresses the need for spatially detailed snow and ice observations in order to improve the representation of the cryosphere in climate models and to improve the knowledge and prediction of water cycle variability and changes. CoRe-H2O will observe the extent, water equivalent and melting state of the snow cover, accumulation and diagenetic facies of glaciers, and properties of sea ice and lake ice. The sensor is a dual frequency SAR, operating at 17 GHz and 9.6 GHz, VV and VH polarizations. This configuration enables the decomposition of the scattering signal for retrieving physical properties of snow and ice. Scientific preparation activities include experimental field campaigns, improvement of radar backscatter models, and the development of inversion algorithms.

Southern Norway has strong climatic gradients and is well suited for studying snow melt and greening patterns. In this study we combine snow cover maps with phenological maps for the early May period for the years 2004 and 2006. The snow cover area maps are based on both microwave (ASAR) and optical (MODIS) sensors. The phenological maps are based on NDVI thresholds from the MODIS sensor. The onset of growing season 2004 was among the earliest recorded over the last 30 years. On May 8, 2004 30% of the study area was covered by snow and 29% had unfolded leaves on trees. The onset of the growing season in 2006 was slightly later than average, and on May 4, 2006 64% of southern Norway was covered by snow and the phenophase of unfolded leaves had not started yet.

The standard method for retrieving snow-covered area with C-band SAR is based on thresholding the ratio between a wet snow SAR image and a dry snow reference scene. A new approach is suggested here, where the snow cover fraction is retrieved by using a gradual transition between snow free and snow covered conditions. The paper discusses the method and applies it on a set of Radarsat data acquired in Norway in May 2003. A near simultaneous aerial optical image and data from field campaigns are used to verify the results.

In this study, we have developed an approach for fusion of optical and SAR data for snow cover fraction (SCF) retrieval that avoids the typical blending effects when combining independently retrieved geophysical data from different sensors. Instead of undertaking the sensor fusion at the geophysical parameter level, the fusion is done at the electromagnetic signal level. A state model, based on hidden Markov model theory, has been developed for the simultaneous signal from the optical and the SAR sensors. The model goes through a given set of states through the snowmelt season where transition probability distribution functions of time have been determined for each state transition. A coupling between corresponding models for optical and SAR observations has been developed in order to make a more reliable model of the sensor co-variation.

The repeat pass interferometric phase change between a summer reference and a dry snow SAR image is proportional to the snow water equivalence. The technique is, however, com- plicated due to rapid de-correlation and low coherence. Phase wrapping is also a problem, and occurs at typically 20 cm dry snow. Phase unwrapping can be remedied by using a delta-K processing technique. By taking into account refined masks for land cover and folding, averages at catchment level may be per- formed to retrieve SWE and finally a field value for the total water content in a catchment. A carefully controlled experiment was performed at our test site at Altevatn, in northern Norway (N68o42', E18o50') the winter 2004. A total of 63 ASAR SLC scenes with high incidence angle (IS1 and IS2) were collected during winter and summer (refer- ence-scenes) following 10 repeated geometries. Two corner reflec- tors were deployed to provide calibration. Three field campaigns were carried out during the winter to validate the SWE results.

Frequent mapping of snow parameters, like snow cover area (SCA) and snow surface wetness (SSW), is important for applications in hydrology, meteorology and climatology. In this study, we have developed a few general multi-sensor/time-series approaches for such monitoring. The objective is to analyze, on a daily basis, a time series of optical and Synthetic Aperture Radar (SAR) data together producing sensor-independent products. A few algorithms for multi-sensor/time-series processing have been developed and are compared in this study. A typical approach is to analyze each image individually and combine them into a day product. How each image contributes to the day products is controlled by a pixel-by-pixel confidence value that is computed for each image analyzed. The confidence algorithm may take into account information about the local observation angle/IFOV size, probability of clouds, prior information about snow state, etc. The time series of day products are then combined into a multi-sensor/multi-temporal product. The combination of products is done on a pixel-by-pixel basis and controlled by each individual product/pixel's confidence and a decay function of time. The "multi product" should then represent the most likely status of the monitored variable. The sensors applied in this algorithm study are MODIS for optical data and ENVISAT ASAR for SAR data. The study area is South Norway, and the study focuses on the snowmelt seasons in 2003 and 2004

The glacial firn line of the Svartisen Glacier has been detected using ERS II SAR and ENVISAT ASAR data from 1995 and up to today. The firn line is detected by first correcting the image backscatter intensity for topographic and geometric contributions using the Muhleman backscattering model. Then we discriminate between firn and ice facies based on the backscatter intensity since frozen firn has a much higher backscatter than ice. Transects across different areas of the glacier were chosen based on requirements of smoothness of topography, precipitation zones, as well as the availability of field data for validation and comparison. From having quite stable conditions during the nineties we have observed a substantial retreat of the firn line over the last few years. The equilibrium line derived from field measurements show a similar trend as the firn line changes, but has a much larger year to year variability. This indicates that the firn line may be a better indicator of climate change than the equilibrium line due to the smaller variance

It has previously been shown that wet snow can be detected using ERS SAR repeat pass imagery where a reference image is captured during cold dry snow conditions and subtracted from the image one want to classify. We have extended and validated this technique for retrieving snow covered area (SCA) using Envisat ASAR wide swath data (500 by 500 km swath coverage, 100 meter resolution), covering the mountainous regions of Southern Norway. The algorithm has also been extended to postulate dry snow above areas with wet snow, thus giving a total snow covered area that is comparable to SCA from optical sensors. A sliding window technique has been applied to facilitate the dry snow classification. The method has been implemented in a near-real time environment and has been run pre-operationally in Norway in 2004. The relatively large coverage allows SAR to become an operational tool for snow monitoring, as opposed to standard modes used in previous works. In order to improve snow classification we have used air temperature data from the Norwegian meteorological station network to create high-resolution surface air temperature maps. These maps are used to filter wet snow from reference images and prevent incorrect classification of dry snow. Snow covered area maps for South Norway has been derived for the spring melt season with a one-week temporal resolution. The results are validated against optical sensor retrievals (MODIS, ASTER) and high accuracy field measurements.

Generally, the contrast between water and land in SAR images decreases with decreasing incidence angle. Thus, surface water detection by intensity thresholding requires high incidence angle data. In this work we demonstrate a texture based surface water detector that produces accurate results, independently of the incidence angle.

We have investigated whether synthetic aperture radar images can be used to observe and position the fish farms and polymer type fish cages along the Norwegian coast. In order to maximize sea-object backscatter difference, the images selected are Radarsat F5 with 10 m/spl times/10 m resolution, 45-48/spl deg/ incidence angle and HH-polarization. We found that it was possible to detect and position all fish cages within the study area, even single ones of polymer-type. It was also possible to determine the overall shape of larger collections of fish cages, and count the number of fish cages within the collections. The wind condition was favorable during the satellite pass (less then 5 m/s). It is concluded that Radarsat F5 images are well suited to detect and position fish farms.

Estimates of snow cover area (SCA) can be obtained from spaceborne SAR sensors such as ERS, Radarsat and Envisat ASAR. A number of algorithms using backscattering images and interferometric coherence have been proposed and documented. It has also been shown that multifrequency and polarization SAR instruments gives enhanced snow mapping capabilities. In this paper we suggest improvements to the Nagler SCA-algorithm and apply them to Radarsat data in the Norwegian mountains. The effect of reduced backscattering due to increased incidence angle is discussed. The effect of mixed pixels (wet snow and other constituents) is addressed, and a sub-pixel method for snow classification is suggested. We also suggest classifying snow on lakes and in forests with different algorithms and separate masks, since the backscattering properties vary from class to class. Finally we try to assess the classification errors by statistical methods.