As heterogeneous supercomputing architectures leveraging GPUs become increasingly central to high-performance computing (HPC), it is crucial for computational fluid dynamics (CFD) simulations, a de-facto HPC workload, to efficiently utilize such hardware. One of the key challenges of HPC codes is performance portability, i.e. the ability to maintain near-optimal performance across different accelerators. In the context of the REFMAP project, which targets scalable, GPU-enabled multi-fidelity CFD for urban airflow prediction, this paper analyzes the performance portability of SOD2D, a state-of-the-art Spectral Elements simulation framework across AMD and NVIDIA GPU architectures. We first discuss the physical and numerical models underlying SOD2D, highlighting its computational hotspots. Then, we examine its performance and scalability in a multi-level manner, i.e. defining and characterizing an extensive full-stack design space spanning across application, software and hardware infrastructure related parameters. Single-GPU performance characterization across server-grade NVIDIA and AMD GPU architectures and vendor-specific compiler stacks, show the potential as well as the diverse effect of memory access optimizations, i.e. 0.69× - 3.91× deviations in acceleration speedup. Performance variability of SOD2D at scale is further examined on the LUMI multi-GPU cluster, where profiling reveals similar throughput variations, highlighting the limits of performance projections and the need for multi-level, informed tuning.

In this article, an algorithm for snow depth (SD) and snow water equivalent (SWE) retrieval is proposed based on a polarimetric synthetic aperture radar (SAR) decomposition model and field measured snow data. The field campaigns were conducted at the Dhundi observatory (in the Indian Himalaya) in January 2016 and 2018. The field-measured data are used here to build a linear regression between wetness ( w) and the imaginary part of the snow permittivity ( ε''), and the validation of retrieved SD and SWE. The snow density ( ρs) and w are calculated with a generalized volume parameter derived using a theoretical model and SAR data (coherency matrix). These snow parameters and the field-based regression relating w and ε''are eventually used for the SD and SWE retrieval. Three TerraSAR-X scenes of the quad-polarization X-band data acquired in January 2016 are used to study the effect of the snow conditions on the accuracy of the proposed algorithm. The mean absolute error (MAE), root-mean-square error (RMSE), and index of agreement (IOA) for SD are 4.84 cm, 5.12 cm, and 0.71, respectively. On the other hand, for SWE, it is 1.42 cm, 1.53 cm, and 0.71, respectively.

The glaciers of High Mountain Asia form one of the largest assemblies of ice-mass outside the Polar regions. The meltwater originating from these glaciers feed into innumerable river streams acting as an important perennial source of water. Within this region, we study one basin, Tarim which consists of glaciers that show surging activity. Of the 200 glaciers in the selected region of study, six glaciers show strong surge at the terminus. This has been explained with the high velocity observed for these glaciers. However, there are glaciers which are in the transient phase of their surge-cycle which tend to behave differently. Such events when compared to previously reported activities suggest that glaciers have surging cycles that are not the same throughout.

This study analyses the potential of integrating spaceborne radar and LiDAR remote sensing products for above ground biomass estimation. Vegetation height product derived from ICESat-2 was interpolated into a 2-D surface using ordinary kriging. Interpolated forest height was validated with in situ data which resulted in determination coefficient of 0.6221 and RMSE of 3.17 m. Interpolated forest surface height and terrain corrected backscatter from L-band ALOS-2/PALSAR-2 data was used to develop linear regression models for above ground biomass estimation. A combination of HV backscatter and ICESat-2 interpolated height-based regression model performed better with a relative RMSE of 30.67%.

Glaciers are the most important component in the cryosphere and are a reliable indicator of climate change. Thus, observation of glacier dynamics helps understand glacier health and monitor climate change. Glacier velocity, thickness and mass balance are the few important parameters to consider regarding glacier dynamics. The main aim of this study is to estimate these three dynamic parameters of a glacier and deriving mass balance form the velocity. DEM differencing method is the most familiar recent remote sensing technique to estimate mass balance. But for the first time, we are using velocity to derive the mass balance and radar interferometry technique used to estimate the velocity. The potential of Differential Interferometric Synthetic Aperture Radar (DInSAR), a SAR technique, to measure surface movement with millimeter level accuracy is utilized to observe the glacier movement. Laminar flow law used here to derive the thickness from velocity. In this study, we used 15 years' time difference data (1999 and 2015 datasets) to check the health of the largest glacier in Himachal Pradesh, Bara Shigri glacier through dynamic parameters. Bara Shigri glacier velocity and thickness both were decreased in the 2015 year compared with 1999. We used these two thickness maps (1999 and 2015) to estimate ice thickness change and we mainly observed more thinning in the ablation region but in accumulation region, ice thickness was increased from 1999 to 2015 year. Thickness change (2015-1999) used to calculate the mass balance of Bara Shigri glacier, observed this as negative and the yearly thinning rate is 1.18 m. This further used to estimate the specific mass balance using density and area of the glacier.

Forecasting and detection of avalanche activities in alpine regions are critical for planning safe traverse routes and landing sites for snow operations and environmental change. The recreational activities (skiing and snowmobile) and public commutation during the winter season lead to fatality spatially when avalanche forecasting is not accurate. Visually, the RGB composite image formed using pre- and post-avalanche data can give primary proxy on the avalanche site. However, in the complex terrain (like Himalaya), RGB composite cannot be used as primary a proxy of avalanche activities. In this paper, we are proposing snow characterization and the avalanche detection algorithm using X-band full polarization synthetic aperture radar (SAR) data. The snowpack parameters are considered in the characterization of snow and avalanche detection. The algorithm also utilizes the SAR power decomposition parameters (Model-based, and mathematical decomposition). Furthermore, statistical indices (Canberra distance and Euclidean distance) are used as a change detection parameter.

Over the last decade, many snowpack parameter retrieval algorithms utilizing C- and X-band Synthetic Aperture Radar (SAR) data have been proposed. These algorithms are capable of retrieving snow density and snow wetness accurately, but not Snow Water Equivalent (SWE). In this paper, a novel approach of SWE estimation is proposed based on TerraSAR-X full-polarimetric data. A unique relationship between the extinction coefficient of the snowpack and snow wetness is built and verified with field data. This relationship enables the retrieval of Snow Depth (SD) and SWE using existing snow density algorithm and volume scattering power. The validation of SD and SWE is carried out with near-real-time ground truth measurements. The accuracy assessment for SD showed Mean Absolute (MAE) and Root Mean Squared (RMSE) Errors of 15 cm and 16 cm, respectively. On the other hand, SWE retrievals resulted in MAE and RMSE of 22.7 mm and 30 mm, respectively.

In this study, a methodology to retrieve Snow Water Equivalent (SWE) information for shallow snow coverage is proposed, something where traditional methods fail, as they require more snow depth to function properly. The proposed algorithm uses the thermal resistance of snowpack to retrieve the SWE by linearly relating the thermal resistance to the backscattering ratios of winter and autumn (snow-free) images acquired with C-band SAR (Sentinel-lA). The SWE (absolute value) is then estimated from the thermal resistance. The results are verified using field data collected at Dhundi observatory maintained by the Snow and Avalanche Study Establishment (SASE), located in the Himalaya (Himachal Pradesh, India; 32°21'N and 77°07'). One winter and two autumn images were used to check the feasibility of the algorithm. The RMSE of SWE estimated using the retrieval algorithm is 3 cm for Pair-l (30 Jan 2016 and 02 Oct 2015), and 2.8 cm for Pair-2 (30 Jan 2016 and 20 Oct 2016).

Social groups often exhibit a high degree of dynamism. Some groups thrive, while many others die over time. Modeling group stability dynamics and understanding whether/when a group will remain stable or shrink over time can be important in a number of social domains. In this paper, we study two different types of social networks as exemplar platforms for modeling and predicting group stability dynamics. We build models to predict if a group is going to remain stable or is likely to shrink over a period of time. We observe that both the level of member diversity and social activities are critical in maintaining the stability of groups. We also find that certain 'prolific' members play a more important role in maintaining the group stability. Our study shows that group stability can be predicted with high accuracy, and feature diversity is critical to prediction performance.

The social sciences strive to understand the political, social, and cultural world around us, but have been impaired by limited access to the quantitative data sources enjoyed by the hard sciences. Careful analysis of Web document streams holds enormous potential to solve longstanding problems in a variety of social science disciplines through massive data analysis. This paper introduces the TextMap Access system, which provides ready access to a wealth of interesting statistics on millions of people, places, and things across a number of interesting web corpora. Powered by a flexible and scalable distributed statistics computation framework using Hadoop, continually updated corpora include newspapers, blogs, patent records, legal documents, and scientific abstracts; well over a terabyte of raw text and growing daily. The Lydia Textmap Access system, available through http://www.textmap.com/access, provides instant access for students and scholars through a convenient web user-interface. We describe the architecture of the TextMap Access system, and its impact on current research in political science, sociology, and business/marketing.