USGS Science⌕ Search

SEARCH · USGS Science

Results for “Algorithms”

Search indexed USGS publications on groundwater, aquifers, geologic maps, mineral resources and earthquakes. Explore source records by subject and place.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 523 records · Page 29Linked to original sources
USGS2010

A two-step procedure for calculating earthquake hypocenters at Augustine Volcano: Chapter 7 in The 2006 Eruption of Augustine Volcano, Alaska

This chapter describes a two-step technique for determining earthquake hypocenters at Augustine Volcano. The algorithm, which was originally developed in the mid-1970s, was designed both to overcome limitations in the standard earthquake-location programs available at the time and to take advantage of the detailed seismic-velocity information obtained at Augustine Volcano. Hypocenters are calculated on the basis of a two-dimensional (2D) ray-tracing procedure that accounts for in plane lateral discontinuities within the seismic velocity structure. This algorithm calculates the minimum P- and S-wave travel time between theoretical grid points embedded in the velocity structure to each station in the seismic network. Station corrections that account for the differences between the model and actual velocity structure are derived from a time-term analysis of the 1975 active-source seismic experiment. Each relocated hypocenter is assigned to the grid point with the lowest rms residual between observed and calculated arrival times. Statistical techniques are used to assess the effect of random errors in P-wave-arrival determination on hypocentral location. These tests suggest that the 2D ray-tracing procedure presented here is able to resolve earthquake hypocenter depths to within 0.25 km between the volcano's summit and sea level and within 0.5 km from sea level to depths of 2 km below sea level.

Alaska↗
USGS2009

A Tidally Averaged Sediment-Transport Model for San Francisco Bay, California

A tidally averaged sediment-transport model of San Francisco Bay was incorporated into a tidally averaged salinity box model previously developed and calibrated using salinity, a conservative tracer (Uncles and Peterson, 1995; Knowles, 1996). The Bay is represented in the model by 50 segments composed of two layers: one representing the channel (>5-meter depth) and the other the shallows (0- to 5-meter depth). Calculations are made using a daily time step and simulations can be made on the decadal time scale. The sediment-transport model includes an erosion-deposition algorithm, a bed-sediment algorithm, and sediment boundary conditions. Erosion and deposition of bed sediments are calculated explicitly, and suspended sediment is transported by implicitly solving the advection-dispersion equation. The bed-sediment model simulates the increase in bed strength with depth, owing to consolidation of fine sediments that make up San Francisco Bay mud. The model is calibrated to either net sedimentation calculated from bathymetric-change data or measured suspended-sediment concentration. Specified boundary conditions are the tributary fluxes of suspended sediment and suspended-sediment concentration in the Pacific Ocean. Results of model calibration and validation show that the model simulates the trends in suspended-sediment concentration associated with tidal fluctuations, residual velocity, and wind stress well, although the spring neap tidal suspended-sediment concentration variability was consistently underestimated. Model validation also showed poor simulation of seasonal sediment pulses from the Sacramento-San Joaquin River Delta at Point San Pablo because the pulses enter the Bay over only a few days and the fate of the pulses is determined by intra-tidal deposition and resuspension that are not included in this tidally averaged model. The model was calibrated to net-basin sedimentation to calculate budgets of sediment and sediment-associated contaminants. While simulated net sedimentation in the four basins that comprise San Francisco Bay was correct, the simulations incorrectly eroded shallows while channels deposited because model surface-layer boxes span both shallows and channels, and neglect lateral variability of suspended-sediment concentration. Validation with recent (1983-2005) net sedimentation in South San Francisco Bay was poor, perhaps owing to poorly quantified sediment supply, and to invasive species that altered erosion and deposition processes. This demonstrates that deterministically predicting future sedimentation is difficult in this or any estuary for which boundary conditions are not stationary. The model would best be used as a tool for developing past and present sediment budgets, and for creating scenarios of future sedimentation that are compared to one another rather than considered a deterministic prediction.

Scientific Investigations Report↗
USGS2010

Empirical models of wind conditions on Upper Klamath Lake, Oregon

Upper Klamath Lake is a large (230 square kilometers), shallow (mean depth 2.8 meters at full pool) lake in southern Oregon. Lake circulation patterns are driven largely by wind, and the resulting currents affect the water quality and ecology of the lake. To support hydrodynamic modeling of the lake and statistical investigations of the relation between wind and lake water-quality measurements, the U.S. Geological Survey has monitored wind conditions along the lakeshore and at floating raft sites in the middle of the lake since 2005. In order to make the existing wind archive more useful, this report summarizes the development of empirical wind models that serve two purposes: (1) to fill short (on the order of hours or days) wind data gaps at raft sites in the middle of the lake, and (2) to reconstruct, on a daily basis, over periods of months to years, historical wind conditions at U.S. Geological Survey sites prior to 2005. Empirical wind models based on Artificial Neural Network (ANN) and Multivariate-Adaptive Regressive Splines (MARS) algorithms were compared. ANNs were better suited to simulating the 10-minute wind data that are the dependent variables of the gap-filling models, but the simpler MARS algorithm may be adequate to accurately simulate the daily wind data that are the dependent variables of the historical wind models. To further test the accuracy of the gap-filling models, the resulting simulated winds were used to force the hydrodynamic model of the lake, and the resulting simulated currents were compared to measurements from an acoustic Doppler current profiler. The error statistics indicated that the simulation of currents was degraded as compared to when the model was forced with observed winds, but probably is adequate for short gaps in the data of a few days or less. Transport seems to be less affected by the use of the simulated winds in place of observed winds. The simulated tracer concentration was similar between model results when simulated winds were used to force the model, and when observed winds were used to force the model, and differences between the two results did not accumulate over time.

Scientific Investigations Report↗
USGS2012

Development of regional skews for selected flood durations for the Central Valley Region, California, based on data through water year 2008

Flood-frequency information is important in the Central Valley region of California because of the high risk of catastrophic flooding. Most traditional flood-frequency studies focus on peak flows, but for the assessment of the adequacy of reservoirs, levees, other flood control structures, sustained flood flow (flood duration) frequency data are needed. This study focuses on rainfall or rain-on-snow floods, rather than the annual maximum, because rain events produce the largest floods in the region. A key to estimating flood-duration frequency is determining the regional skew for such data. Of the 50 sites used in this study to determine regional skew, 28 sites were considered to have little to no significant regulated flows, and for the 22 sites considered significantly regulated, unregulated daily flow data were synthesized by using reservoir storage changes and diversion records. The unregulated, annual maximum rainfall flood flows for selected durations (1-day, 3-day, 7-day, 15-day, and 30-day) for all 50 sites were furnished by the U.S. Army Corps of Engineers. Station skew was determined by using the expected moments algorithm program for fitting the Pearson Type 3 flood-frequency distribution to the logarithms of annual flood-duration data. Bayesian generalized least squares regression procedures used in earlier studies were modified to address problems caused by large cross correlations among concurrent rainfall floods in California and to address the extensive censoring of low outliers at some sites, by using the new expected moments algorithm for fitting the LP3 distribution to rainfall flood-duration data. To properly account for these problems and to develop suitable regional-skew regression models and regression diagnostics, a combination of ordinary least squares, weighted least squares, and Bayesian generalized least squares regressions were adopted. This new methodology determined that a nonlinear model relating regional skew to mean basin elevation was the best model for each flood duration. The regional-skew values ranged from -0.74 for a flood duration of 1-day and a mean basin elevation less than 2,500 feet to values near 0 for a flood duration of 7-days and a mean basin elevation greater than 4,500 feet. This relation between skew and elevation reflects the interaction of snow and rain, which increases with increased elevation. The regional skews are more accurate, and the mean squared errors are less than in the Interagency Advisory Committee on Water Data's National skew map of Bulletin 17B.

California↗
USGS2018

Methods for peak-flow frequency analysis and reporting for streamgages in or near Montana based on data through water year 2015

This report documents the methods for peak-flow frequency (hereinafter “frequency”) analysis and reporting for streamgages in and near Montana following implementation of the Bulletin 17C guidelines. The methods are used to provide estimates of peak-flow quantiles for 50-, 42.9-, 20-, 10-, 4-, 2-, 1-, 0.5-, and 0.2-percent annual exceedance probabilities for selected streamgages operated by the U.S. Geological Survey Wyoming-Montana Water Science Center (WY–MT WSC). These annual exceedance probabilities correspond to 2-, 2.33-, 5-, 10-, 25-, 50-, 100-, 200-, and 500-year recurrence intervals, respectively. Standard procedures specific to the WY–MT WSC for implementing the Bulletin 17C guidelines include (1) the use of the Expected Moments Algorithm analysis for fitting the log-Pearson Type III distribution, incorporating historical information where applicable; (2) the use of weighted skew coefficients (based on weighting at-site station skew coefficients with generalized skew coefficients from the Bulletin 17B national skew map); and (3) the use of the Multiple Grubbs-Beck Test for identifying potentially influential low flows. For some streamgages, the peak-flow records are not well represented by the standard procedures and require user-specified adjustments informed by hydrologic judgement. The specific characteristics of peak-flow records addressed by the informed-user adjustments include (1) regulated peak-flow records, (2) atypical upper-tail peak-flow records, and (3) atypical lower-tail peak-flow records. In all cases, the informed-user adjustments use the Expected Moments Algorithm fit of the log-Pearson Type III distribution using the at-site station skew coefficient, a manual potentially influential low flow threshold, or both. Appropriate methods can be applied to at-site frequency estimates to provide improved representation of long-term hydroclimatic conditions. The methods for improving at-site frequency estimates by weighting with regional regression equations and by Maintenance of Variance Extension Type III record extension are described. Frequency analyses were conducted for 99 example streamgages to indicate various aspects of the frequency- analysis methods described in this report. The frequency analyses and results for the example streamgages are presented in a separate data release associated with this report consisting of tables and graphical plots that are structured to include information concerning the interpretive decisions involved in the frequency analyses. Further, the separate data release includes the input files to the PeakFQ program, version 7.1, including the peak-flow data file and the analysis specification file that were used in the peak-flow frequency analyses. Peak-flow frequencies are also reported in separate data releases for selected streamgages in the Beaverhead River and Clark Fork Basins and also for selected streamgages in the Ruby, Jefferson, and Madison River Basins.

Montana↗
USGS2021

Optimization of the Idaho National Laboratory water-quality aquifer monitoring network, southeastern Idaho

Long-term monitoring of water-quality data collected from wells at the Idaho National Laboratory (INL) has provided essential information for delineating the movement of radiochemical and chemical wastes in the eastern Snake River Plain aquifer, southeastern Idaho. Since 1949, the U.S. Geological Survey, in cooperation with the U.S. Department of Energy, has maintained as many as 200 wells in the INL water-quality monitoring network. A network design tool, distributed as an R package, was developed to evaluate and optimize groundwater monitoring in the existing network based on water-quality data collected at 153 sampling sites since January 1, 1989. The objective of the optimization design tool is to reduce well monitoring redundancy while retaining sufficient data to reliably characterize water-quality conditions in the aquifer. A spatial optimization was used to identify a set of wells whose removal leads to the smallest increase in the deviation between interpolated concentration maps using the existing and reduced monitoring networks while preserving significant long-term trends and seasonal components in the data. Additionally, a temporal optimization was used to identify reductions in sampling frequencies by minimizing the redundancy in sampling events. Spatial optimization uses an islands genetic algorithm to identify near-optimal network designs removing 10, 20, 30, 40, and 50 wells from the existing monitoring network. With this method, choosing a greater number of wells to remove results in greater cost savings and decreased accuracy of the average relative difference between interpolated maps of the reduced-dataset and the full-dataset. The genetic search algorithm identified reduced networks that best capture the spatial patterns of the average concentration plume while preserving long-term temporal trends at individual wells. Concentration data for 10 analyte types are integrated in a single optimization so that all datasets may be evaluated simultaneously. A constituent was selected for inclusion in the spatial optimization problem when the observations were sufficient to (1) establish a two-range variability model, (2) classify at least one concentration time series as a continuous record block, and (3) make a prediction using the quantile-kriging interpolation method. The selected constituents include sodium, chloride, sulfate, nitrate, carbon tetrachloride, 1,1-dichloroethylene, 1,1,1-trichloroethane, trichloroethylene, tritium, strontium-90, and plutonium-238. In temporal optimization, an iterative-thinning method was used to find an optimal sampling frequency for each analyte-well pair. Optimal frequencies indicate that for many of the wells, samples may be collected less frequently and still be able to characterize the concentration over time. The optimization results indicated that the sample-collection interval may be increased by an of average of 273 days owing to temporal redundancy.

Idaho↗
USGS2022

Synthesizing ground magnetic disturbance using dipole-aligned loop elementary currents and Biot-Savart relationship

This report presents a method for constructing a simplified numerical description of the electric current distributions in the ionosphere and gap region based on dipole-aligned loop elementary currents (DALECs). A theoretical basis for DALECs is presented, along with a prototypical algorithm for constructing an elementary numerical DALEC. The algorithm is verified and validated by combining DALECs with an efficient Biot-Savart solver in order to estimate magnetic disturbance on the Earth’s surface. We examine (1) simple scenarios with known solutions and (2) hemispherical magnetic disturbance fields obtained from a state-of-the-art global geospace circulation model.

Scientific Investigations Report↗
USGS2012

Documentation of the Surface-Water Routing (SWR1) Process for modeling surface-water flow with the U.S. Geological Survey Modular Ground-Water Model (MODFLOW-2005)

A flexible Surface-Water Routing (SWR1) Process that solves the continuity equation for one-dimensional and two-dimensional surface-water flow routing has been developed for the U.S. Geological Survey three-dimensional groundwater model, MODFLOW-2005. Simple level- and tilted-pool reservoir routing and a diffusive-wave approximation of the Saint-Venant equations have been implemented. Both methods can be implemented in the same model and the solution method can be simplified to represent constant-stage elements that are functionally equivalent to the standard MODFLOW River or Drain Package boundary conditions. A generic approach has been used to represent surface-water features (reaches) and allows implementation of a variety of geometric forms. One-dimensional geometric forms include rectangular, trapezoidal, and irregular cross section reaches to simulate one-dimensional surface-water features, such as canals and streams. Two-dimensional geometric forms include reaches defined using specified stage-volume-area-perimeter (SVAP) tables and reaches covering entire finite-difference grid cells to simulate two-dimensional surface-water features, such as wetlands and lakes. Specified SVAP tables can be used to represent reaches that are smaller than the finite-difference grid cell (for example, isolated lakes), or reaches that cannot be represented accurately using the defined top of the model. Specified lateral flows (which can represent point and distributed flows) and stage-dependent rainfall and evaporation can be applied to each reach. The SWR1 Process can be used with the MODFLOW Unsaturated Zone Flow (UZF1) Package to permit dynamic simulation of runoff from the land surface to specified reaches. Surface-water/groundwater interactions in the SWR1 Process are mathematically defined to be a function of the difference between simulated stages and groundwater levels, and the specific form of the reach conductance equation used in each reach. Conductance can be specified directly or calculated as a function of the simulated wetted perimeter and defined reach bed hydraulic properties, or as a weighted combination of both reach bed hydraulic properties and horizontal hydraulic conductivity. Each reach can be explicitly coupled to a single specific groundwater-model layer or coupled to multiple groundwater-model layers based on the reach geometry and groundwater-model layer elevations in the row and column containing the reach. Surface-water flow between reservoirs is simulated using control structures. Surface-water flow between reaches, simulated by the diffusive-wave approximation, can also be simulated using control structures. A variety of control structures have been included in the SWR1 Process and include (1) excess-volume structures, (2) uncontrolled-discharge structures, (3) pumps, (4) defined stage-discharge relations, (5) culverts, (6) fixed- or movable-crest weirs, and (7) fixed or operable gated spillways. Multiple control structures can be implemented in individual reaches and are treated as composite flow structures. Solution of the continuity equation at the reach-group scale (a single reach or a user-defined collection of individual reaches) is achieved using exact Newton methods with direct solution methods or exact and inexact Newton methods with Krylov sub-space methods. Newton methods have been used in the SWR1 Process because of their ability to solve nonlinear problems. Multiple SWR1 time steps can be simulated for each MODFLOW time step, and a simple adaptive time-step algorithm, based on user-specified rainfall, stage, flow, or convergence constraints, has been implemented to better resolve surface-water response. A simple linear- or sigmoid-depth scaling approach also has been implemented to account for increased bed roughness at small surface-water depths and to increase numerical stability. A line-search algorithm also has been included to improve the quality of the Newton-step upgrade vector, if possible. The SWR1 Process has been benchmarked against one- and two-dimensional numerical solutions from existing one- and two-dimensional numerical codes that solve the dynamic-wave approximation of the Saint-Venant equations. Two-dimensional solutions test the ability of the SWR1 Process to simulate the response of a surface-water system to (1) steady flow conditions for an inclined surface (solution of Manning's equation), and (2) transient inflow and rainfall for an inclined surface. The one-dimensional solution tests the ability of the SWR1 Process to simulate a looped network with multiple upstream inflows and several control structures. The SWR1 Process also has been compared to a level-pool reservoir solution. A synthetic test problem was developed to evaluate a number of different SWR1 solution options and simulate surface-water/groundwater interaction. The solution approach used in the SWR1 Process may not be applicable for all surface-water/groundwater problems. The SWR1 Process is best suited for modeling long-term changes (days to years) in surface-water and groundwater flow. Use of the SWR1 Process is not recommended for modeling the transient exchange of water between streams and aquifers when local and convective acceleration and other secondary effects (for example, wind and Coriolis forces) are substantial. Dam break evaluations and two-dimensional evaluations of spatially extensive domains are examples where acceleration terms and secondary effects would be significant, respectively.

Techniques and Methods↗
USGS2023

Parameter estimation at the conterminous United States scale and streamflow routing enhancements for the National Hydrologic Model infrastructure application of the Precipitation-Runoff Modeling System (NHM-PRMS)

This report documents a three-part continental-scale calibration procedure and a new streamflow routing algorithm using the U.S. Geological Survey National Hydrologic Model (NHM) infrastructure along with an application of the Precipitation-Runoff Modeling System (PRMS). The traditional approach to hydrologic model calibration and evaluation, which relies on comparing observed and simulated streamflow, is not sufficient for accurately representing the non-streamflow parts of the water budget. If intermediate process variables computed by the hydrologic model are not examined, the variables could be characterized by parameter values that do not replicate those hydrological processes present in the physical system. In answer to this potential problem, alternative hydrologic process variables from the model (in addition to streamflow) are included in a calibration procedure applied to the conterminous United States (CONUS) domain. The three-part calibration procedure presented in this report considers volume (calibration by hydrologic response unit [byHRU]), timing (calibration by headwater watershed [byHW]), and measured streamflow [byHWobs]). The first part, byHRU, is considered a water-balance volume calibration that uses five alternative (non-streamflow) hydrologic quantities (runoff, actual evapotranspiration, recharge, soil moisture, and snow-covered area) as calibration targets for each hydrologic response unit (HRU). These alternative data products were derived, with error bounds, from multiple sources for each of the 109,951 HRUs in the NHM on time scales varying from annual to daily. The second part of the calibration, byHW, is considered a streamflow timing calibration that uses statistically based streamflow simulations developed using ordinary kriging for 7,265 headwater watersheds that had drainage areas of less than 3,000 square kilometers (1,158 square miles) across the CONUS. Two streamflow routing algorithms were tested in this byHW calibration: (1) continuity without attenuation of the flood pulse and (2) a new formulation of the Muskingum routing method, which was added to the PRMS as part of this study. The third part of the calibration, byHWobs, refines the model parameters using available measured streamflow using 1,417 streamgage locations. A multiple-objective, stepwise, automated calibration procedure was used to identify the optimal set of parameters for each calibration procedure. Using a variety of alternative datasets for calibration of the water budget provides users of the NHM-PRMS with improved initial parameters and helps alleviate the equifinality problem (getting the right answer for the wrong reason). Through a community effort, these alternative data products, with error bounds, can be used to improve and expand our understanding of hydrologic-process representation in models. The broader modeling community can use these data products, with error bounds, to calibrate and evaluate hydrologic models using more than streamflow.

Conterminous United States↗
USGS2012

Approaches in highly parameterized inversion - PEST++, a Parameter ESTimation code optimized for large environmental models

An object-oriented parameter estimation code was developed to incorporate benefits of object-oriented programming techniques for solving large parameter estimation modeling problems. The code is written in C++ and is a formulation and expansion of the algorithms included in PEST, a widely used parameter estimation code written in Fortran. The new code is called PEST++ and is designed to lower the barriers of entry for users and developers while providing efficient algorithms that can accommodate large, highly parameterized problems. This effort has focused on (1) implementing the most popular features of PEST in a fashion that is easy for novice or experienced modelers to use and (2) creating a software design that is easy to extend; that is, this effort provides a documented object-oriented framework designed from the ground up to be modular and extensible. In addition, all PEST++ source code and its associated libraries, as well as the general run manager source code, have been integrated in the Microsoft Visual Studio® 2010 integrated development environment. The PEST++ code is designed to provide a foundation for an open-source development environment capable of producing robust and efficient parameter estimation tools for the environmental modeling community into the future.

Techniques and Methods↗
USGS2000

Evaluation of the use of reach transmissivity to quantify leakage beneath Levee 31N, Miami-Dade County, Florida

A coupled ground- and surface-water model (MODBRANCH) was developed to estimate ground-water flow beneath Levee 31N in Miami-Dade County, Florida, and to simulate hydrologic conditions in the surrounding area. The study included compilation of data from monitoring stations, measurement of vertical seepage rates in wetlands, and analysis of the hydrogeologic properties of the ground-water aquifer within the study area. In addition, the MODBRANCH code was modified to calculate the exchange between surface-water channels and ground water using a relation based on the concept of reach transmissivity. The modified reach-transmissivity version of the MODBRANCH code was successfully tested on three simple problems with known analytical solutions. It was also tested and determined to function adequately on one field problem that had previously been solved using the unmodified version of the software. The modified version of MODBRANCH was judged to have performed satisfactorily, and it required about 60 percent as many iterations to reach a solution. Additionally, its input parameters are more physically-based and less dependent on model-grid spacing. A model of the Levee 31N area was developed and used with the original and modified versions of MODBRANCH, which produced similar output. The mean annual modeled ground-water heads differed by only 0.02 foot, and the mean annual canal discharge differed by less than 1.0 cubic foot per second. Seepage meters were used to quantify vertical seepage rates in the Everglades wetlands area west of Levee 31N. A comparison between results from the seepage meters and from the computer model indicated substantial differences that seemed to be a result of local variations in the hydraulic properties in the topmost part of the Biscayne aquifer. The transmissivity of the Biscayne aquifer was estimated to be 1,400,000 square feet per day in the study area. The computer model was employed to simulate seepage of ground water beneath Levee 31N. Modeled seepage rates were usually between 100 and 400 cubic feet per day per foot of levee, but extreme values ranged from about -200 to 500 cubic feet per day (positive values indicate eastward seepage beneath the levee). The modeled seepage results were used to develop an algorithm to estimate seepage based on head differential at selected monitoring stations. The algorithm was determined to adequately predict ground-water seepage.

Florida↗
USGS2001

Hydrologic conditions and budgets for the Black Hills of South Dakota, through water year 1998

The Black Hills are an important recharge area for aquifers in the northern Great Plains. The surface-water hydrology of the area is highly influenced by interactions with the Madison and Minnelusa aquifers, including large springs and streamflow loss zones. Defining responses of ground water and streamflow to a variety of hydrogeologic influences is critical to development of hydrologic budgets for ground- and surface-water systems. Hydrographs for 52 observation wells and 1 cave site are used to show ground-water response to cumulative precipitation departures. Aquifers considered include the Precambrian, Deadwood, Madison, Minnelusa, Minnekahta, and Inyan Kara aquifers, with wells completed in the Inyan Kara aquifer generally showing small response to precipitation patterns. Many wells completed in the other aquifers have large short- and long-term fluctuations in water levels. Madison and Minnelusa wells in the southern Black Hills show a general tendency for smaller water-level fluctuations than in other areas. Streamflow characteristics and relations with precipitation are examined for 33 gaging stations representative of five different hydrogeologic settings that are identified. The ?limestone headwater? setting occurs within outcrops of the Madison Limestone and Minnelusa Formation along the ?Limestone Plateau,? where direct runoff is uncommon and streamflow consists almost entirely of base flow originating as ground-water discharge from headwater springs. Thus, variability in daily, monthly, and annual flow is small. Annual streamflow correlates poorly with precipitation; however, consideration of ?moving averages? (involving up to 11 years of annual precipitation data for some stations) improves relations substantially. The ?crystalline core? area is encircled by the outcrop band of the Madison and Minnelusa Formations and is dominated by igneous and metamorphic rocks. Base flow ranges from about 41 to 73 percent for representative streams; however, monthly flow records demonstrate shortterm response to precipitation, which probably indicates a relatively large component of interflow. Streamflow generally correlates well with annual precipitation, with r2 values ranging from 0.52 to 0.87. Downgradient from the crystalline core area is the ?loss zone? setting, where streamflow losses occur to outcrops of the Madison and Minnelusa Formations. Relations between streamflow and annual precipitation are defined by a power equation for the only two representative gages in this setting. The loss zone and ?artesian spring? areas are combined because many artesian springs are located along stream channels that are influenced by streamflow losses and several artesian springs are within outcrops of the Minnelusa Formation. Streamflow characteristics for artesian springs generally have small variability and poor correlations with annual precipitation because of large influence from relatively stable ground-water discharge. The ?exterior? setting is located downgradient from the outcrop of the Inyan Kara Group, which coincides with the outer extent of the loss zone/artesian spring setting. Large flow variability is characteristic for this setting, and base flow generally is smaller than for other settings. Basin yields are highly variable, with the largest yields occurring in high-altitude areas of the northern Black Hills that receive large annual precipitation. Relations between annual yield efficiency and precipitation were applied by previous investigators in developing a method for estimating annual precipitation recharge, based on annual precipitation. The resulting ?yield-efficiency algorithm? compares spatial distributions for annual precipitation, average annual precipitation, and efficiency of basin yield. This algorithm is applied in estimating precipitation recharge on aquifer outcrops and in estimating streamflow yield from various outcrop areas, for purposes of developing average hydrologic budgets

Water-Resources Investigations Report↗
USGS1992

Geohydrology of, and simulation of ground-water flow in, the Milford-Souhegan glacial-drift aquifer, Milford, New Hampshire

Hydrogeologic data collected since 1990 were assessed and a ground-water-flow model was refined in this study of the Milford-Souhegan glacial-drift aquifer in Milford, New Hampshire. The hydrogeologic data collected were used to refine estimates of hydraulic conductivity and saturated thickness of the aquifer, which were previously calculated during 1988-90. In October 1990, water levels were measured at 124 wells and piezometers, and at 45 stream-seepage sites on the main stem of the Souhegan River, and on small tributary streams overlying the aquifer to improve an understanding of ground-water-flow patterns and stream-seepage gains and losses. Refinement of the ground-water-flow model included a reduction in the number of active cells in layer 2 in the central part of the aquifer, a revision of simulated hydraulic conductivity in model layers 2 and representing the aquifer, incorporation of a new block-centered finite-difference ground-water-flow model, and incorporation of a new solution algorithm and solver (a preconditioned conjugate-gradient algorithm). Refinements to the model resulted in decreases in the difference between calculated and measured heads at 22 wells. The distribution of gains and losses of stream seepage calculated in simulation with the refined model is similar to that calculated in the previous model simulation. The contributing area to the Savage well, under average pumping conditions, decreased by 0.021 square miles from the area calculated in the previous model simulation. The small difference in the contrib- uting recharge area indicates that the additional data did not enhance model simulation and that the conceptual framework for the previous model is accurate.

Water-Resources Investigations Report↗
USGS1981

An examination of techniques for reformatting digital cartographic data/part 1: The raster-to- vector process.

Current graphic devices suitable for high-speed computer input and output of cartographic data are tending more and more to be raster-oriented, such as the rotating drum scanner and the color raster display. However, the majority of commonly used manipulative techniques in computer-assisted cartography and automated spatial data handling continue to require that the data be in vector format. This situation has recently precipitated the requirement for very fast techniques for converting digital cartographic data from raster to vector format for processing, and then back into raster format for plotting. The current article is part one of a two-part paper concerned with examining the state-of-the-art in these conversion techniques. In part one, algorithms to perform all phases of the raster-to-vector process are systematically outlined, and then compared in general terms. Examples of existing implementations of the raster-to-vector process are also described and evaluated. Part two will outline and compare algorithms to perform the reverse process, vector-to-raster conversion. Part two will also discuss raster-to-vector and vector-to-raster conversion as an integrated process.

Cartographica: The International Journal for Geogr↗
USGS1997

Daily air temperature interpolated at high spatial resolution over a large mountainous region

Two methods are investigated for interpolating daily minimum and maximum air temperatures ( T min and T max ) at a 1 km spatial resolution over a large mountainous region (830000 km 2 ) in the U.S. Pacific Northwest. The methods were selected because of their ability to (1) account for the effect of elevation on temperature and (2) efficiently handle large volumes of data. The first method, the neutral stability algorithm (NSA), used the hydrostatic and potential temperature equations to convert measured temperatures and elevations to sea-level potential temperatures. The potential temperatures were spatially interpolated using an inverse-squared-distance algorithm and then mapped to the elevation surface of a digital elevation model (DEM). The second method, linear lapse rate adjustment (LLRA), involved the same basic procedure as the NSA, but used a constant linear lapse rate instead of the potential temperature equation. Cross-validation analyses were performed using the NSA and LLRA methods to interpolate T min and T max each day for the 1990 water year, and the methods were evaluated based on mean annual interpolation error (IE). The NSA method showed considerable bias for sites associated with vertical extrapolation. A correction based on climate station/grid cell elevation differences was developed and found to successfully remove the bias. The LLRA method was tested using 3 lapse rates, none of which produced a serious extrapolation bias. The bias-adjusted NSA and the 3 LLRA methods produced almost identical levels of accuracy (mean absolute errors between 1.2 and 1.3°C), and produced very similar temperature surfaces based on image difference statistics. In terms of accuracy, speed, and ease of implementation, LLRA was chosen as the best of the methods tested.

Columbia River basin↗
USGS2024

Automatic identification and quantification of volcanic hotspots in Alaska using HotLINK: The hotspot learning and identification network

An increase in volcanic thermal emissions can indicate subsurface and surface processes that precede, or coincide with, volcanic eruptions. Space-borne infrared sensors can detect hotspots—defined here as localized volcanic thermal emissions—in near-real-time. However, automatic hotspot detection systems are needed to efficiently analyze the large quantities of data produced. While hotspots have been automatically detected for over 20 years with simple thresholding algorithms, new computer vision technologies, such as convolutional neural networks (CNNs), can enable improved detection capabilities. Here we introduce HotLINK: the Hotspot Learning and Identification Network, a CNN trained to detect hotspots with a dataset of −3,800 satellite-based, Visible Infrared Imaging Radiometer Suite (VIIRS) images from Mount Veniaminof and Mount Cleveland volcanoes, Alaska. We find that our model achieves an accuracy of 96% (F1-score 0.92) when evaluated on −1,700 unseen images from the same volcanoes, and 95% (F1-score 0.67) when evaluated on −3,000 images from six additional Alaska volcanoes (Augustine Volcano, Bogoslof Island, Okmok Caldera, Pavlof Volcano, Redoubt Volcano, Shishaldin Volcano). In comparison with an existing threshold-based hotspot detection algorithm, MIROVA (Coppola et al., Geological Society, London, Special Publications, 2016, 426, 181–205), our model detects 22% more hotspots and produces 12% fewer false positives. Additional testing on −700 labeled Moderate Resolution Imaging Spectroradiometer (MODIS) images from Mount Veniaminof demonstrates that our model is applicable to this sensor’s data as well, achieving an accuracy of 98% (F1-score 0.95). We apply HotLINK to 10 years of VIIRS data and 22 years of MODIS data for the eight aforementioned Alaska volcanoes and calculate the radiative power of detected hotspots. From these time series we find that HotLINK accurately characterizes background and eruptive periods, similar to MIROVA, but also detects more subtle warming signals, potentially related to volcanic unrest. We identify three advantages to our model over its predecessors: 1) the ability to detect more subtle volcanic hotspots and produce fewer false positives, especially in daytime images; 2) probabilistic predictions provide a measure of detection confidence; and 3) its transferability, i.e., the successful application to multiple sensors and multiple volcanoes without the need for threshold tuning, suggesting the potential for global application.

Alaska↗
USGS2021

A novel framework to predict relative habitat selection in aquatic systems: Applying machine learning and resource selection functions to acoustic telemetry data from multiple shark species

Resource selection functions (RSFs) have been widely applied to animal tracking data to examine relative habitat selection and to help guide management and conservation strategies. While readily used in terrestrial ecology, RSFs have yet to be extensively used within marine systems. As acoustic telemetry continues to be a pervasive approach within marine environments, incorporation of RSFs can provide new insights to help prioritize habitat protection and restoration to meet conservation goals. To overcome statistical hurdles and achieve high prediction accuracy, machine learning algorithms could be paired with RSFs to predict relative habitat selection for a species within and even outside the monitoring range of acoustic receiver arrays, making this a valuable tool for marine ecologists and resource managers. Here, we apply RSFs using machine learning to an acoustic telemetry dataset of four shark species to explore and predict species-specific habitat selection within a marine protected area. In addition, we also apply this RSF-machine learning approach to investigate predator-prey relationships by comparing and averaging tiger shark relative selection values with the relative selection values derived for eight potential prey-species. We provide methodological considerations along with a framework and flexible approach to apply RSFs with machine learning algorithms to acoustic telemetry data and suggest marine ecologists and resource managers consider adopting such tools to help guide both conservation and management strategies.

Buck Island Reef National Monument, St Croix↗
USGS2022

Medium spatial resolution mapping of global land cover and land cover change across multiple decades from Landsat

Land cover maps are essential for characterizing the biophysical properties of the Earth’s land areas. Because land cover information synthesizes a rich array of information related to both the ecological condition of land areas and their exploitation by humans, they are widely used for basic and applied research that requires information related to land surface properties (e.g., terrestrial carbon models, water balance models, weather, and climate models) and are core inputs to models and analyses used by natural resource scientists and land managers. As the Earth’s global population has grown over the last several decades rates of land cover change have increased dramatically, with enormous impacts on ecosystem services (e.g., biodiversity, water supply, carbon sequestration, etc.). Hence, accurate information related to land cover is essential for both managing natural resources and for understanding society’s ecological, biophysical, and resource management footprint. To address the need for high-quality land cover information we are using the global record of Landsat observations to compile annual maps of global land cover from 2001 to 2020 at 30 m spatial resolution. To create these maps we use features derived from time series of Landsat imagery in combination with ancillary geospatial data and a large database of training sites to classify land cover at annual time step. The algorithm that we apply uses temporal segmentation to identify periods with stable land cover that are separated by breakpoints in the time series. Here we provide an overview of the methods and data sets we are using to create global maps of land cover. We describe the algorithms used to create these maps and the core land cover data sets that we are creating through this effort, and we summarize our approach to accuracy assessment. We also present a synthesis of early results and discuss the strengths and weaknesses of our early map products and the challenges that we have encountered in creating global land cover data sets from Landsat. Initial accuracy assessment for North America shows good overall accuracy (77.0 ± 2.0% correctly classified) and 79.8% agreement with the European Space Agency (ESA) WorldCover product. The land cover mapping results we report provide the foundation for robust, repeatable, and accurate mapping of global land cover and land cover change across multiple decades at 30 m spatial resolution from Landsat.

Frontiers in Remote Sensing↗