Comparing line feature morphology with scale specific sinuosity distributions: A modified earth mover’s distance
No abstract available.
Conference Paper↗
Geology topics
Ethan J. Shavers
Publications and source records attributed to Ethan J. Shavers.
GeoAI in the US Geological Survey for topographic mapping
Geospatial artificial intelligence (GeoAI) can be defined broadly as the application of artificial intelligence methods and techniques to geospatial data, processes, models, and applications. The application of these methods to topographic data and phenomena is a focus of research in the US Geological Survey (USGS). Specifically, the USGS has researched and developed applications in terrain feature extraction, hydrographic network extraction, and semantic modeling. This article is a documentation of the recent work and current state of research and development. The article helps define the accomplishments and directions of research and applications in fields of GeoAI for topographic mapping within the USGS and more broadly.
Transactions in GIS↗
Extensibility of U-net neural network model for hydrographic feature extraction and implications for hydrologic modeling
Accurate maps of regional surface water features are integral for advancing ecologic, atmospheric and land development studies. The only comprehensive surface water feature map of Alaska is the National Hydrography Dataset (NHD). NHD features are often digitized representations of historic topographic map blue lines and may be outdated. Here we test deep learning methods to automatically extract surface water features from airborne interferometric synthetic aperture radar (IfSAR) data to update and validate Alaska hydrographic databases. U-net artificial neural networks (ANN) and high-performance computing (HPC) are used for supervised hydrographic feature extraction within a study area comprised of 50 contiguous watersheds in Alaska. Surface water features derived from elevation through automated flow-routing and manual editing are used as training data. Model extensibility is tested with a series of 16 U-net models trained with increasing percentages of the study area, from about 3 to 35 percent. Hydrography is predicted by each of the models for all watersheds not used in training. Input raster layers are derived from digital terrain models, digital surface models, and intensity images from the IfSAR data. Results indicate about 15 percent of the study area is required to optimally train the ANN to extract hydrography when F1-scores for tested watersheds average between 66 and 68. Little benefit is gained by training beyond 15 percent of the study area. Fully connected hydrographic networks are generated for the U-net predictions using a novel approach that constrains a D-8 flow-routing approach to follow U-net predictions. This work demonstrates the ability of deep learning to derive surface water feature maps from complex terrain over a broad area.
Alaska↗
The 4th paradigm in multiscale data representation: Modernizing the National Geospatial Data Infrastructure
The need of citizens in any nation to access geospatial data in readily usable form is critical to societal well-being, and in the United States (US), demands for information by scientists, students, professionals and citizens continue to grow. Areas such as public health, urbanization, resource management, economic development and environmental management require a variety of data collected from many sources to identify problems, monitor trends and propose solutions. Such information needs and demands have driven the coordination of federal and regional government agencies with respective private sector participation to develop national geospatial data infrastructures in many countries.
Book chapter↗
Watersheds and drainage networks
This topic is an overview of basic concepts about how the distribution of water on the Earth, with specific regard to watersheds, stream and river networks, and waterbodies are represented by geographic data. The flowing and non-flowing bodies of water on the earth’s surface vary in extent largely due to seasonal and annual changes in climate and precipitation. Consequently, modeling the detailed representation of surface water using geographic information is important. The area of land that collects surface runoff and other flowing water and drains to a common outlet location defines a watershed. Terrain and surface features can be naturally divided into watersheds of various sizes. Drainage networks are important data structures for modeling the distribution and movement of surface water over the terrain. Numerous tools and methods exist to extract drainage networks and watersheds from digital elevation models (DEMs). The cartographic representations of surface water are referred to as hydrographic features and consist of a snapshot at a specific time. Hydrographic features can be assigned general feature types, such as lake, pond, river, and ocean. Hydrographic features can be stored, maintained, and distributed for use through vector geospatial databases, such as the National Hydrography Dataset (NHD) for the United States.
Book chapter↗
An attention U-Net model for detection of fine-scale hydrologic streamlines
Surface water is an irreplaceable resource for human survival and environmental sustainability. Accurate, finely detailed cartographic representations of hydrologic streamlines are critically important in various scientific domains, such as assessing the quantity and quality of present and future water resources, modeling climate changes, evaluating agricultural suitability, mapping flood inundation, and monitoring environmental changes. Conventional approaches to detecting such streamlines cannot adequately incorporate information from the complex three-dimensional (3D) environment of streams and land surface features. Such information is vital to accurately delineate streamlines. In recent years, high accuracy lidar data has become increasingly available for deriving both 3D information and terrestrial surface reflectance. This study develops an attention U-net model to take advantage of high-accuracy lidar data for finely detailed streamline detection and evaluates model results against a baseline of multiple traditional machine learning methods. The evaluation shows that the attention U-net model outperforms the best baseline machine learning method by an average F1 score of 11.25% and achieves significantly better smoothness and connectivity between classified streamline channels. These findings suggest that our deep learning approach can harness high-accuracy lidar data for fine-scale hydrologic streamline detection, and in turn produce desirable benefits for many scientific domains.
North Carolina↗
Channel cross-section analysis for automated stream head identification
Headwater streams account for more than half of the streams in the United States by length. The substantial occurrence and susceptibility to change of headwater streams makes regular updating of related maps vital to the accuracy of associated analysis and display. Here we present work testing new methods of completely automated remote headwater stream identification using metrics derived from channel Digital Elevation Model (DEM) cross-sections. A jump in standard deviation of curvature (sK) is found to correlate with the presence of stream heads. Field and remotely validated stream and channel initiation points from 4 diverse study areas in North Carolina as well as a simulated surface are used to test the sK findings. The sK value within individual catchments equal to 0.5*Tukey's upper inner fence is found to be a reliable threshold for identifying the upslope extent of channels in varied landscapes.
Environmental Modelling & Software↗
OpenCLC: An open-source software tool for similarity assessment of linear hydrographic features
The National Hydrography Dataset (NHD) is a foundational geospatial data source in the United States that enables extensive and diverse environmental research and supports decision-making in numerous contexts. However, the NHD requires regular validation and update given possible inconsistent initial collection and hydrographic changes. Furthermore, systems or tools that use NHD data must manage regular updates that occur within the high-resolution version of the NHD (NHD HR). This research contributes to filling this gap by establishing an open-source software tool named OpenCLC, which automatically identifies matching and mismatching line features between two sets of hydrographic flowlines. Aside from identifying differences among two version of NHD lines, results can be applied to improve the quality of NHD HR content. OpenCLC significantly outperforms the best available commercial off-the-shelf software in computational scalability, and it is made widely available as part of the CyberGIS Toolkit to benefit broad environmental and geospatial science communities.
SoftwareX↗
Preserving meander bend geometry through scale
Stream meander geometry is a function of hydrologic, geologic, and anthropogenic forces. Meander morphometrics are used in geomorphic classification, ecological characterization, and tectonic and hydrologic change detection. Thus, detailed measurement and classification of meander geometry is imperative to multiscale representation of hydrographic features, which raises important questions. What meander geometries are important to preserve in multi-scale databases? How are geometries measured? How are they preserved? Is the choice between preservation of geometry or use of classification attributes? Questions related to multiscale measurement and representation of hydrographic features continue to emerge with increased spatial and temporal data collection. A key metric for understanding meander bend geometry is sinuosity. The most common measure of sinuosity is the length of a feature divided by the distance between stream head and mouth. The measure relays deviation from a straight line but nothing about meander wavelength. There is not a clear consensus on methods for measuring meander geometry, much less efficiently, at scales made viable with increased data resolution. Here we propose a method for automated characterization of meander wavelength or bend radius. The method, termed Scale-Specific Sinuosity ( S 3 ), is a derivation from the Richardson plot. The Richardson (1961) plot is a classic means of calculating fractal dimension of natural line features and describes feature length (ℓ) given increasing vertex spacing, or step size (S), plotted on a log-log plot. The S 3 metric is defined as negative one times the slope of a Richardson plot for a given stride length. This paper demonstrates utility of S 3 for estimating changes in sinuosity with scale change.
Conference Paper↗
Scale-specific metrics for adaptive generalization and geomorphic classification of stream features
The Richardson plot has been used to illustrate fractal dimension of naturally occurring landscape features that are sensitive to changes in scale or resolution, such as coastlines and river channels. The Richardson method estimates the length of a path by traversing (i.e., “walking”) the path with a specific stride length. Fractal dimension is determined as the slope of the Richardson plot, which shows path length over a range of stride lengths graphed on log-log axes. This paper describes a variant of the Richardson plot referred to as the Scale-Specific Sinuosity (S 3 ) plot. S 3 is defined as negative one times the slope of the Richardson plot for a given stride length. A plot of S 3 against stride length offers a frequency distribution whose area under the curve reflects total sinuosity, and whose points mark the amount of sinuosity contributed to the total sinuosity at each stride length. Mathematical relations of S 3 with fractal dimension and sinuosity for linear features are described. The S 3 metric is demonstrated and discussed for several linear stream features distributed over the conterminous United States. The S 3 metric can help guide the preservation of stream feature sinuosity during cartographic generalization and may assist automated geomorphic classification of river systems.
Conference Paper↗
Streams do work: Measuring the work of low-order streams on the landscape using point clouds
The mutable nature of low-order streams makes regular updating of surface water maps necessary for accurate representation. Low-order streams make up roughly half the streams in the conterminous United States by length, and small inaccuracies in stream head location can result in significant error in stream reach, order, and density. Reliable maps of stream features are vital for hydrologic modeling, ecosystem research, and boundary monitoring. High resolution digital elevation models derived from lidar data have shown promise in low order stream modeling yet forested high relief landscapes and low relief agricultural areas remain challenging. Here we present early results from research analyzing lidar point clouds to identify features and patterns that may be used in low-order stream identification and classification in challenging geographic conditions. This work has identified characteristics derived from point clouds that correlate with the presence of streams and stream heads and show promise for mapping small streams. In low topographic relief agricultural areas, cross sections collected at regular intervals along drainage channels extracted as 3D lines show a significant jump in value and variance of profile curvature standard deviation at stream heads. In high relief areas, observations show potential for stream mapping by identifying trends in riparian zone structure. Lidar return point density from riparian vegetation under 30 feet tall dips in the vicinity of intermittent stream heads. Also seen is an increase in point density above 60 feet downstream of stream heads. The trends found here likely reflect a change in vegetation structure relative to the presence of streams.
Illinois, Iowa, North Carolina↗
Automated road breaching to enhance extraction of natural drainage networks from elevation models through deep learning
High-resolution (HR) digital elevation models (DEMs), such as those at resolutions of 1 and 3 meters, have increasingly become more widely available, along with lidar point cloud data. In a natural environment, a detailed surface water drainage network can be extracted from a HR DEM using flow-direction and flow-accumulation modeling. However, elevation details captured in HR DEMs, such as roads and overpasses, can form barriers that incorrectly alter flow accumulation models, and hinder the extraction of accurate surface water drainage networks. This study tests a deep learning approach to identify the intersections of roads and stream valleys, whereby valley channels can be burned through road embankments in a HR DEM for subsequent flow accumulation modeling, and proper natural drainage network extraction.
Conference Paper↗
Similarity assessment of linear hydrographic features using high performance computing
This work discusses a current open source implementation of a line similarity assessment workflow to compare elevation-derived drainage lines with the high-resolution National Hydrography Dataset (NHD) surface-water flow network. The process identifies matching and mismatching lines in each dataset to help focus subsequent validation procedures to areas of the NHD that more critically need updates.
Conference Paper↗