Multi-year, spatially extensive, watershed-scale synoptic stream chemistry and water quality conditions for six permafrost-underlain Arctic watersheds

Repeated sampling of spatially distributed river chemistry can be used to assess the location, scale, and persistence of carbon and nutrient contributions to watershed exports. Here, we provide a comprehensive set of water chemistry measurements and ecohydrological metrics describing the biogeochemical conditions of permafrost-affected Arctic watersheds. These data were collected in watershed-wide synoptic campaigns in six stream networks across northern Alaska. Three watersheds are associated with the Arctic Long-Term Ecological Research site at Toolik Field Station (TFS), which were sampled seasonally each June and August from 2016 to 2018. Three watersheds were associated with the National Park Service (NPS) of Alaska and the U.S. Geological Survey (USGS) and were sampled annually from 2015 to 2019. Extensive water chemistry characterization included carbon species, dissolved nutrients, and major ions. The objective of the sampling designs and data acquisition was to characterize terrestrial–aquatic linkages and processing of material in stream networks. The data allow estimation of novel ecohydrological metrics that describe the dominant location, scale, and overall persistence of ecosystem processes in continuous permafrost. These metrics are (1) subcatchment leverage, (2) variance collapse, and (3) spatial persistence. Raw data are available at the National Park Service Integrated Resource Management Applications portal (O'Donnell et al., 2021, https://doi.org/10.5066/P9SBK2DZ) and within the Environmental Data Initiative (Abbott, 2021, https://doi.org/10.6073/pasta/258a44fb9055163dd4dd4371b9dce945).

Bush optimization method for high ranked stream networks

Решена задача построения больших потоковых сетей высокого ранга оптимальности на основе кустовой оптимизации, связывающей с каждой вершиной сети её фрагмент, имеющий определенную размерность и позволяющий на всех фрагментах достичь более высокого ранга оптимальности за заданное время решения задачи на компьютере. The problem of constructing high ranked large-scale stream networks is solved by using bush optimization technique. This technique implies connection of each network vertex with its fragment of a certain dimension. That allows reaching a higher ranked optimality for solving the problem in a given amount of time by a computer.

An Efficient Human Instance-Guided Framework for Video Action Recognition

In recent years, human action recognition has been studied by many computer vision researchers. Recent studies have attempted to use two-stream networks using appearance and motion features, but most of these approaches focused on clip-level video action recognition. In contrast to traditional methods which generally used entire images, we propose a new human instance-level video action recognition framework. In this framework, we represent the instance-level features using human boxes and keypoints, and our action region features are used as the inputs of the temporal action head network, which makes our framework more discriminative. We also propose novel temporal action head networks consisting of various modules, which reflect various temporal dynamics well. In the experiment, the proposed models achieve comparable performance with the state-of-the-art approaches on two challenging datasets. Furthermore, we evaluate the proposed features and networks to verify the effectiveness of them. Finally, we analyze the confusion matrix and visualize the recognized actions at human instance level when there are several people.

Local-aware spatio-temporal attention network with multi-stage feature fusion for human action recognition

In the study of human action recognition, two-stream networks have made excellent progress recently. However, there remain challenges in distinguishing similar human actions in videos. This paper proposes a novel local-aware spatio-temporal attention network with multi-stage feature fusion based on compact bilinear pooling for human action recognition. To elaborate, taking two-stream networks as our essential backbones, the spatial network first employs multiple spatial transformer networks in a parallel manner to locate the discriminative regions related to human actions. Then, we perform feature fusion between the local and global features to enhance the human action representation. Furthermore, the output of the spatial network and the temporal information are fused at a particular layer to learn the pixel-wise correspondences. After that, we bring together three outputs to generate the global descriptors of human actions. To verify the efficacy of the proposed approach, comparison experiments are conducted with the traditional hand-engineered IDT algorithms, the classical machine learning methods (i.e., SVM) and the state-of-the-art deep learning methods (i.e., spatio-temporal multiplier networks). According to the results, our approach is reported to obtain the best performance among existing works, with the accuracy of 95.3% and 72.9% on UCF101 and HMDB51, respectively. The experimental results thus demonstrate the superiority and significance of the proposed architecture in solving the task of human action recognition.

From site-focused intervention towards landscape-scale surface water management using Synthetic Stream Networks and Rapid Scenario Screening

This research addresses the need to transform success in technical understanding and practical implementation of surface water management (SWM) interventions at a site-scale towards integrated landscape-scale management. We achieve this through targeting the informative preliminary stages of strategic design, where broad, early and effective exploration of opportunities can enhance and direct a regional SWM perspective. We present a new method, ‘Synthetic Stream Networks’ (SSN), capable of meeting these requirements by taking advantage of easily accessible data, likely to be available during regional screening. We find that results from the SSN are validated by existing, ‘downstream’ focused data (90% of the river network is within 30 m of an associated SSN flow path), with the added advantage of extending understanding of surface water exceedance flow paths and watersheds into the upper catchment, thus establishing a foundational and physically based sub-catchment management unit exploring surface water connectivity at a catchment and landscape scale. We also demonstrate collaborative advantages of twinning the new SSN method with ‘Rapid Scenario Screening’ (RSS) to develop a novel approach for identifying, exploring and evaluating SWM interventions. Overall, we find that this approach addresses challenges of integrating understanding from sub-catchment, catchment and landscape perspectives within surface water management.