Search USGS⌕ Search

SEARCH · Search USGS

Results for “Statistical Methods & Applications”

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 181 records · Page 10Linked to original sources
USGS2012

Thermodynamic method for generating random stress distributions on an earthquake fault

This report presents a new method for generating random stress distributions on an earthquake fault, suitable for use as initial conditions in a dynamic rupture simulation. The method employs concepts from thermodynamics and statistical mechanics. A pattern of fault slip is considered to be analogous to a micro-state of a thermodynamic system. The energy of the micro-state is taken to be the elastic energy stored in the surrounding medium. Then, the Boltzmann distribution gives the probability of a given pattern of fault slip and stress. We show how to decompose the system into independent degrees of freedom, which makes it computationally feasible to select a random state. However, due to the equipartition theorem, straightforward application of the Boltzmann distribution leads to a divergence which predicts infinite stress. To avoid equipartition, we show that the finite strength of the fault acts to restrict the possible states of the system. By analyzing a set of earthquake scaling relations, we derive a new formula for the expected power spectral density of the stress distribution, which allows us to construct a computer algorithm free of infinities. We then present a new technique for controlling the extent of the rupture by generating a random stress distribution thousands of times larger than the fault surface, and selecting a portion which, by chance, has a positive stress perturbation of the desired size. Finally, we present a new two-stage nucleation method that combines a small zone of forced rupture with a larger zone of reduced fracture energy.

Open-File Report↗
USGS1979

Evaluation of methods for estimating ground-water withdrawals in western Kansas

During 1978, methods of estimating ground-water withdrawals in western Kansas were examined and evaluated, using both existing data from the files of the U.S. Geological Survey and newly collected data. Values for annual ground-water withdrawals reported to the Division of Water Resources, Kansas State Board of Agriculture, by water users probably contain substantial errors because most individual users do not possess the means to measure or accurately estimate the discharge rates of their wells. Such reported values are estimated to average 10 to 15 percent higher than the actual discharge rate. Values obtained from discharge-totaling meters are reasonably accurate, but few wells are equipped with these meters, and the cost of equipping all wells with meters may be prohibitive. Measured rates of power consumption can give good estimates of total withdrawal from wells where the power consumption is measured. However, power-consumption values cannot be extrapolated reliably from one well to another. Three techniques are regarded as promising for estimating areal ground-water withdrawals from readily collected data: Examination of a small sample of withdrawal values from metered wells indicates that the statistical approach used by Luckey (1972) may be applied to western Kansas to estimate total withdrawal with acceptable accuracy in a large area from a statistical sample of measured values. Measured values of irrigation application at a small number of selected wells might be used to compute total withdrawal for irrigation by crop type and precipitation zone. Evaluation of reported values of irrigation-water application for selected crops in zones of generally similar precipitation showed that, although the reported values are judged to be erroneously high, the reported values were consistent with each other and with average precipitation. Power-consumption coefficients may be used to calculate an average coefficient for an area. The method described by Luckey (1972) was used to calculate the number of power-consumption coefficients that would have to be determined to estimate the mean power consumption of wells pumped by electric or natural-gas engines. The results indicate that as few as 100 values would produce results accurate to within 10 percent of the true average at the 95-percent confidence level. Thus, power coefficients for a particular type of irrigation system in any given area can be used even though values of power coefficients have little transfer value from well to well. Efforts by the Instrument Development Laboratofy of the U.S. Geological Survey to design improved instruments for measuring ground-water withdrawals have produced two promising results: An electronic running-time meter (Running-time Sentry) that was tested during 1978 in western Kansas and Florida appears to give accurate values of total pumping time. Large-scale testing of this prototype running-time meter is planned for 1979. A low-cost discharge-totaling meter (REELTOT), based on the concept of sensing the velocity head in a pipe with differential-pressure transducers permanently installed in the pipe, is nearly complete. Early prototypes of this instrument may be tested in western Kansas during 1979. For the remainder of this study, the three techniques for estimating areal ground-water withdrawals described above will be tested on a larger scale in Groundwater Management District No. 1, an area of about 1,800 square miles in west-central Kansas. About 150 randomly selected wells will be equipped with Running-time Sentry units and monitored for total withdrawal, instantaneous discharge, and power consumption. A few wells will be equipped with prototype discharge-totaling meters. The performance of the Sentry units and discharge-totaling meters will be evaluated. Areal estimates of withdrawal from measured withdrawal values, from average power-consumption coefficients, and from irrigation application by crop will be calculated and compared with each other and with reported values of withdrawal.

Kansas↗
USGS2017

Short-term response of Holcus lanatus L. (Common Velvetgrass) to chemical and manual control at Yosemite National Park, USA

One of the highest priority invasive species at both Yosemite and Sequoia and Kings Canyon national parks is Holcus lanatus L. (common velvetgrass), a perennial bunchgrass that invades mid-elevation montane meadows. Despite velvetgrass being a high priority species, there is little information available on control techniques. The goal of this project was to evaluate the short-term response of a single application of common chemical and manual velvetgrass control techniques. The study was conducted at three montane sites in Yosemite National Park. Glyphosate spot-spray treatments were applied at 0.5, 1.0, 1.5, and 2.0% concentrations, and compared with hand pulling to evaluate effects on cover of common velvetgrass, cover of other plant species, and community species richness. Posttreatment year 1 cover of common velvetgrass was 12.1% ± 1.6 in control plots, 6.3% ± 1.5 averaged over the four chemical treatments (all chemical treatments performed similarly), and 13.6% ± 1.7 for handpulled plots. This represents an approximately 50% reduction in common velvetgrass cover in chemically- treated plots recoded posttreatment year 1 and no statistically significant reduction in hand pulled plots compared with controls. However, there was no treatment effect in posttreatment year 2, and all herbicide application rates performed similarly. In addition, there were no significant treatment effects on nontarget species or species richness. These results suggest that for this level of infestation and habitat type, (1) one year of hand pulling is not an effective control method and (2) glyphosate provides some level of control in the short-term without impact to nontarget plant species, but the effect is temporary as a single year of glyphosate treatment is ineffective over a two-year period.

California↗
USGS2002

Updated techniques for estimating monthly streamflow-duration characteristics at ungaged and partial-record sites in central Nevada

Techniques for estimating monthly streamflow-duration characteristics at ungaged and partial-record sites in central Nevada have been updated. These techniques were developed using streamflow records at six continuous-record sites, basin physical and climatic characteristics, and concurrent streamflow measurements at four partial-record sites. Two methods, the basin-characteristic method and the concurrent-measurement method, were developed to provide estimating techniques for selected streamflow characteristics at ungaged and partial-record sites in central Nevada. In the first method, logarithmic-regression analyses were used to relate monthly mean streamflows (from all months and by month) from continuous-record gaging sites of various percent exceedence levels or monthly mean streamflows (by month) to selected basin physical and climatic variables at ungaged sites. Analyses indicate that the total drainage area and percent of drainage area at altitudes greater than 10,000 feet are the most significant variables. For the equations developed from all months of monthly mean streamflow, the coefficient of determination averaged 0.84 and the standard error of estimate of the relations for the ungaged sites averaged 72 percent. For the equations derived from monthly means by month, the coefficient of determination averaged 0.72 and the standard error of estimate of the relations averaged 78 percent. If standard errors are compared, the relations developed in this study appear generally to be less accurate than those developed in a previous study. However, the new relations are based on additional data and the slight increase in error may be due to the wider range of streamflow for a longer period of record, 1995-2000. In the second method, streamflow measurements at partial-record sites were correlated with concurrent streamflows at nearby gaged sites by the use of linear-regression techniques. Statistical measures of results using the second method typically indicated greater accuracy than for the first method. However, to make estimates for individual months, the concurrent-measurement method requires several years additional streamflow data at more partial-record sites. Thus, exceedence values for individual months are not yet available due to the low number of concurrent-streamflow-measurement data available. Reliability, limitations, and applications of both estimating methods are described herein.

Open-File Report↗
USGS2007

Statistical Methods and Sampling Design for Estimating Step Trends in Surface-Water Quality

This paper addresses two components of the problem of estimating the magnitude of step trends in surface water quality. The first is finding a robust estimator appropriate to the data characteristics expected in water-quality time series. The J. L. Hodges-E. L. Lehmann class of estimators is found to be robust in comparison to other nonparametric and moment-based estimators. A seasonal Hodges-Lehmann estimator is developed and shown to have desirable properties. Second, the effectiveness of various sampling strategies is examined using Monte Carlo simulation coupled with application of this estimator. The simulation is based on a large set of total phosphorus data from the Potomac River. To assure that the simulated records have realistic properties, the data are modeled in a multiplicative fashion incorporating flow, hysteresis, seasonal, and noise components. The results demonstrate the importance of balancing the length of the two sampling periods and balancing the number of data values between the two periods.

Water Resources Bulletin↗
USGS2009

The Digital Shoreline Analysis System (DSAS) Version 4.0 - An ArcGIS extension for calculating shoreline change

The Digital Shoreline Analysis System (DSAS) version 4.0 is a software extension to ESRI ArcGIS v.9.2 and above that enables a user to calculate shoreline rate-of-change statistics from multiple historic shoreline positions. A user-friendly interface of simple buttons and menus guides the user through the major steps of shoreline change analysis. Components of the extension and user guide include (1) instruction on the proper way to define a reference baseline for measurements, (2) automated and manual generation of measurement transects and metadata based on user-specified parameters, and (3) output of calculated rates of shoreline change and other statistical information. DSAS computes shoreline rates of change using four different methods: (1) endpoint rate, (2) simple linear regression, (3) weighted linear regression, and (4) least median of squares. The standard error, correlation coefficient, and confidence interval are also computed for the simple and weighted linear-regression methods. The results of all rate calculations are output to a table that can be linked to the transect file by a common attribute field. DSAS is intended to facilitate the shoreline change-calculation process and to provide rate-of-change information and the statistical data necessary to establish the reliability of the calculated results. The software is also suitable for any generic application that calculates positional change over time, such as assessing rates of change of glacier limits in sequential aerial photos, river edge boundaries, land-cover changes, and so on.

Open-File Report↗
USGS2016

Object-based forest classification to facilitate landscape-scale conservation in the Mississippi Alluvial Valley

The Mississippi Alluvial Valley is a floodplain along the southern extent of the Mississippi River extending from southern Missouri to the Gulf of Mexico. This area once encompassed nearly 10 million ha of floodplain forests, most of which has been converted to agriculture over the past two centuries. Conservation programs in this region revolve around protection of existing forest and reforestation of converted lands. Therefore, an accurate and up to date classification of forest cover is essential for conservation planning, including efforts that prioritize areas for conservation activities. We used object-based image analysis with Random Forest classification to quickly and accurately classify forest cover. We used Landsat band, band ratio, and band index statistics to identify and define similar objects as our training sets instead of selecting individual training points. This provided a single rule-set that was used to classify each of the 11 Landsat 5 Thematic Mapper scenes that encompassed the Mississippi Alluvial Valley. We classified 3,307,910±85,344 ha (32% of this region) as forest. Our overall classification accuracy was 96.9% with Kappa statistic of 0.96. Because this method of forest classification is rapid and accurate, assessment of forest cover can be regularly updated and progress toward forest habitat goals identified in conservation plans can be periodically evaluated.

Arkansas, Louisiana, Mississippi, Missouri↗
USGS2011

Building on crossvalidation for increasing the quality of geostatistical modeling

The random function is a mathematical model commonly used in the assessment of uncertainty associated with a spatially correlated attribute that has been partially sampled. There are multiple algorithms for modeling such random functions, all sharing the requirement of specifying various parameters that have critical influence on the results. The importance of finding ways to compare the methods and setting parameters to obtain results that better model uncertainty has increased as these algorithms have grown in number and complexity. Crossvalidation has been used in spatial statistics, mostly in kriging, for the analysis of mean square errors. An appeal of this approach is its ability to work with the same empirical sample available for running the algorithms. This paper goes beyond checking estimates by formulating a function sensitive to conditional bias. Under ideal conditions, such function turns into a straight line, which can be used as a reference for preparing measures of performance. Applied to kriging, deviations from the ideal line provide sensitivity to the semivariogram lacking in crossvalidation of kriging errors and are more sensitive to conditional bias than analyses of errors. In terms of stochastic simulation, in addition to finding better parameters, the deviations allow comparison of the realizations resulting from the applications of different methods. Examples show improvements of about 30% in the deviations and approximately 10% in the square root of mean square errors between reasonable starting modelling and the solutions according to the new criteria.

Stochastic Environmental Research and Risk Assessm↗
USGS2016

Assessing the role of climate and resource management on groundwater dependent ecosystem changes in arid environments with the Landsat archive

Groundwater dependent ecosystems (GDEs) rely on near-surface groundwater. These systems are receiving more attention with rising air temperature, prolonged drought, and where groundwater pumping captures natural groundwater discharge for anthropogenic use. Phreatophyte shrublands, meadows, and riparian areas are GDEs that provide critical habitat for many sensitive species, especially in arid and semi-arid environments. While GDEs are vital for ecosystem services and function, their long-term (i.e. ~ 30 years) spatial and temporal variability is poorly understood with respect to local and regional scale climate, groundwater, and rangeland management. In this work, we compute time series of NDVI derived from sensors of the Landsat TM, ETM +, and OLI lineage for assessing GDEs in a variety of land and water management contexts. Changes in vegetation vigor based on climate, groundwater availability, and land management in arid landscapes are detectable with Landsat. However, the effective quantification of these ecosystem changes can be undermined if changes in spectral bandwidths between different Landsat sensors introduce biases in derived vegetation indices, and if climate, and land and water management histories are not well understood. The objective of this work is to 1) use the Landsat 8 under-fly dataset to quantify differences in spectral reflectance and NDVI between Landsat 7 ETM + and Landsat 8 OLI for a range of vegetation communities in arid and semiarid regions of the southwestern United States, and 2) demonstrate the value of 30-year historical vegetation index and climate datasets for assessing GDEs. Specific study areas were chosen to represent a range of GDEs and environmental conditions important for three scenarios: baseline monitoring of vegetation and climate, riparian restoration, and groundwater level changes. Google's Earth Engine cloud computing and environmental monitoring platform is used to rapidly access and analyze the Landsat archive along with downscaled North American Land Data Assimilation System gridded meteorological data, which are used for both atmospheric correction and correlation analysis. Results from the cross-sensor comparison indicate a benefit from the application of a consistent atmospheric correction method, and that NDVI derived from Landsat 7 and 8 are very similar within the study area. Results from continuous Landsat time series analysis clearly illustrate that there are strong correlations between changes in vegetation vigor, precipitation, evaporative demand, depth to groundwater, and riparian restoration. Trends in summer NDVI associated with riparian restoration and groundwater level changes were found to be statistically significant, and interannual summer NDVI was found to be moderately correlated to interannual water-year precipitation for baseline study sites. Results clearly highlight the complementary relationship between water-year PPT, NDVI, and evaporative demand, and are consistent with regional vegetation index and complementary relationship studies. This work is supporting land and water managers for evaluation of GDEs with respect to climate, groundwater, and resource management.

Remote Sensing of Environment↗
USGS2012

Advances in the simulation and automated measurement of well-sorted granular material: 1. Simulation

1. In this, the first of a pair of papers which address the simulation and automated measurement of well-sorted natural granular material, a method is presented for simulation of two-phase (solid, void) assemblages of discrete non-cohesive particles. The purpose is to have a flexible, yet computationally and theoretically simple, suite of tools with well constrained and well known statistical properties, in order to simulate realistic granular material as a discrete element model with realistic size and shape distributions, for a variety of purposes. The stochastic modeling framework is based on three-dimensional tessellations with variable degrees of order in particle-packing arrangement. Examples of sediments with a variety of particle size distributions and spatial variability in grain size are presented. The relationship between particle shape and porosity conforms to published data. The immediate application is testing new algorithms for automated measurements of particle properties (mean and standard deviation of particle sizes, and apparent porosity) from images of natural sediment, as detailed in the second of this pair of papers. The model could also prove useful for simulating specific depositional structures found in natural sediments, the result of physical alterations to packing and grain fabric, using discrete particle flow models. While the principal focus here is on naturally occurring sediment and sedimentary rock, the methods presented might also be useful for simulations of similar granular or cellular material encountered in engineering, industrial and life sciences.

Journal of Geophysical Research F: Earth Surface↗
USGS2024

Spatio-temporal ecological models via physics-informed neural networks for studying chronic wasting disease

To mitigate the negative effects of emerging wildlife diseases in biodiversity and public health it is critical to accurately forecast pathogen dissemination while incorporating relevant spatio-temporal covariates. Forecasting spatio-temporal processes can often be improved by incorporating scientific knowledge about the dynamics of the process using physical models. Ecological diffusion equations are often used to model epidemiological processes of wildlife diseases where environmental factors play a role in disease spread. Physics-informed neural networks (PINN) are deep learning algorithms that constrain neural network predictions based on physical laws and therefore are powerful forecasting models useful even in cases of limited and imperfect training data. In this paper, we develop a novel ecological modeling tool using PINNs, which fits a feedforward neural network and simultaneously performs parameter identification in a partial differential equation (PDE) with varying coefficients. We demonstrate the applicability of our model by comparing it with the commonly used Bayesian stochastic partial differential equation method and traditional machine learning approaches, showing that our proposed model exhibits superior prediction and forecasting performance when modeling chronic wasting disease in deer in Wisconsin. Furthermore, our model provides the opportunity to obtain scientific insights into spatiotemporal covariates affecting spread and growth of diseases. This work contributes to future machine learning and statistical methodology development by studying spatio-temporal processes enhanced by prior physical knowledge.

Spatial Statistics↗
USGS2025

Estimating spatially explicit survival and mortality risk from telemetry data with thinned point process models

Mortality risk for animals often varies spatially and can be linked to how animals use landscapes. While numerous studies collect telemetry data on animals, the focus is typically on the period when animals are alive, even though there is important information that could be gleaned about mortality risk. We introduce a thinned spatial point process (SPP) modelling framework that couples relative abundance and space use with a mortality process to formally treat the occurrence of mortality events across the landscape as a spatial process. We show how this model can be embedded in a hierarchical statistical framework and fit to telemetry data to make inferences about how spatial covariates drive both space use and mortality risk. We apply the method to two data sets to study the effects of roads and habitat on spatially explicit mortality risk: (1) VHF telemetry data collected for willow ptarmigan in Alaska, and (2) hourly GPS telemetry data collected for black bears in Colorado. These case studies demonstrate the applicability of this method for different species and data types, making it broadly useful in enabling inferences about the mechanisms influencing animal survival and spatial population processes while formally treating survival as a spatial process, especially as the development and implementation of joint analyses continue to progress.

Ecology Letters↗
USGS2024

Peak streamflow trends in South Dakota and their relation to changes in climate, water years 1921–2020

Peak-flow (flood) frequency analysis is essential to water-resources management applications, including the design of critical infrastructure such as bridges and culverts, and floodplain mapping. Federal guidelines for performing peak-flow flood frequency analyses are presented in a U.S. Geological Survey Techniques and Methods Report known as Bulletin 17C. A basic assumption within Bulletin 17C, which documents the guidelines for determining annual peak streamflow frequency, is that, for basins without major hydrologic alterations (for example, regulation, diversion, and urbanization), statistical properties of the distribution of annual peak streamflows are stationary; that is, the mean, variance, and skew are constant through time. Nonstationarity is a statistical property of a peak-flow series such that the long-term (on the order of decades) distributional properties change one or more times either gradually or abruptly through time. Individual nonstationarities may be attributed to one source such as flow regulation, land-use change, or climate but are often the result of a combination of sources, making detection and attribution of nonstationarities challenging. In response to a growing concern regarding nonstationarity in peak streamflows in the region, the U.S. Geological Survey, in cooperation with the Departments of Transportation of Illinois, Iowa, Michigan, Minnesota, Missouri, South Dakota, and Wisconsin; the Montana Department of Natural Resources and Conservation; and the North Dakota Department of Water Resources, assessed the potential nonstationarity in peak streamflows in the north-central United States. This chapter characterizes the effects of natural hydroclimatic shifts and potential climate change on annual peak streamflows in the State of South Dakota. Annual peak and daily streamflow as well as model-simulated gridded climatic data were examined for temporal monotonic trends, change points, and other statistical properties indicative of changing climatic and environmental conditions. Changes in annual peak and daily flows were evaluated among 13, 35, and 81 qualifying U.S. Geological Survey streamgages for the 75-, 50-, and 30-year trend periods through water year 2020 (the period from October 1, 2019, to September 30, 2020) in South Dakota, respectively. No qualifying streamgages were in the 100-year trend period in the State. Statistical tests for autocorrelation (independent and identically distributed assumption), monotonic trends, and change points in the median and scale are analyzed to evaluate potential stationarity violations (nonstationarity) for performing at-site peak-flow flood-frequency analysis. The trends are reported using a likelihood approach as an alternative to simply reporting significant trends with an arbitrary p -value cutoff point. A distinct east-west spatial pattern of likely upward and downward monotonic trends and change points, respectively, was detected in 75- and 50-year trend periods, but an inconsistent spatial pattern was detected in the 30-year trend period. Additionally, change points in the median annual peak streamflows were detected in the late 1970s and early 1980s in the western part of the State, but in the east, the change point was more commonly detected in 1992–93. A similar east-west spatial pattern of likely upward and downward trends was detected in the annual peak-flow timing, the day of the year of the annal peak streamflow. In the western part of the State, the annual peak streamflows are arriving earlier, but in the east, the annual peak streamflows are arriving later. A peaks-over-threshold (POT) analysis where, on average, there are two events per year (POT2) and four events per year (POT4) was also used to evaluate changes in the frequency (count) of daily streamflows exceeding the threshold. Similar to detected changes in the annual peak streamflow, an east-west likely upward or downward change corresponding to an increase or decrease, respectively, in the frequency of daily streamflow greater than a POT2 and POT4 threshold was detected. A monthly water-balance model was used to evaluate hydroclimatic variation in annual and seasonal precipitation, snowfall, potential evapotranspiration, and soil moisture storage for all qualifying streamgages in the 75-, 50-, and 30-year trend periods. Detected trends in the annual hydroclimatic metrics for the 75- and 50-year trend periods indicate a spatially consistent statewide increase in precipitation, decrease in snowfall, increase in potential evapotranspiration, and increase in soil moisture storage. Furthermore, detected trends in seasonal precipitation in the 75- and 50-year trend periods highlight a pronounced change in precipitation in winter and later into the summer season, especially in the 50-year trend period in the eastern part of the State. Statewide increases in seasonal soil moisture storage were also detected, highlighting year-round increasing flood magnitudes, particularly in the eastern part of the State. Based on the results of these stationarity tests for the qualifying streamgages in South Dakota among the 75-, 50-, and 30-year trend periods, consistent temporal and spatial patterns of nonstationarity were detected among the 75- and 50-year trend periods. Furthermore, when nonstationarity is detected in daily streamflow, increased streamflow and volume (increasing frequency in POT), as well as potentially bridge scour, may have implications on culvert and highway design in the eastern part of South Dakota. Thus, when performing at-site peak-flow flood-frequency analyses in South Dakota, potential nonstationarities and alternative approaches are important considerations.

South Dakota↗
USGS2010

Hierarchical spatial models for predicting pygmy rabbit distribution and relative abundance

Conservationists routinely use species distribution models to plan conservation, restoration and development actions, while ecologists use them to infer process from pattern. These models tend to work well for common or easily observable species, but are of limited utility for rare and cryptic species. This may be because honest accounting of known observation bias and spatial autocorrelation are rarely included, thereby limiting statistical inference of resulting distribution maps. We specified and implemented a spatially explicit Bayesian hierarchical model for a cryptic mammal species (pygmy rabbit Brachylagus idahoensis). Our approach used two levels of indirect sign that are naturally hierarchical (burrows and faecal pellets) to build a model that allows for inference on regression coefficients as well as spatially explicit model parameters. We also produced maps of rabbit distribution (occupied burrows) and relative abundance (number of burrows expected to be occupied by pygmy rabbits). The model demonstrated statistically rigorous spatial prediction by including spatial autocorrelation and measurement uncertainty. We demonstrated flexibility of our modelling framework by depicting probabilistic distribution predictions using different assumptions of pygmy rabbit habitat requirements. Spatial representations of the variance of posterior predictive distributions were obtained to evaluate heterogeneity in model fit across the spatial domain. Leave-one-out cross-validation was conducted to evaluate the overall model fit. Synthesis and applications. Our method draws on the strengths of previous work, thereby bridging and extending two active areas of ecological research: species distribution models and multi-state occupancy modelling. Our framework can be extended to encompass both larger extents and other species for which direct estimation of abundance is difficult. ?? 2010 The Authors. Journal compilation ?? 2010 British Ecological Society.

Journal of Applied Ecology↗
USGS2013

Topological and canonical kriging for design flood prediction in ungauged catchments: an improvement over a traditional regional regression approach?

In the United States, estimation of flood frequency quantiles at ungauged locations has been largely based on regional regression techniques that relate measurable catchment descriptors to flood quantiles. More recently, spatial interpolation techniques of point data have been shown to be effective for predicting streamflow statistics (i.e., flood flows and low-flow indices) in ungauged catchments. Literature reports successful applications of two techniques, canonical kriging, CK (or physiographical-space-based interpolation, PSBI), and topological kriging, TK (or top-kriging). CK performs the spatial interpolation of the streamflow statistic of interest in the two-dimensional space of catchment descriptors. TK predicts the streamflow statistic along river networks taking both the catchment area and nested nature of catchments into account. It is of interest to understand how these spatial interpolation methods compare with generalized least squares (GLS) regression, one of the most common approaches to estimate flood quantiles at ungauged locations. By means of a leave-one-out cross-validation procedure, the performance of CK and TK was compared to GLS regression equations developed for the prediction of 10, 50, 100 and 500 yr floods for 61 streamgauges in the southeast United States. TK substantially outperforms GLS and CK for the study area, particularly for large catchments. The performance of TK over GLS highlights an important distinction between the treatments of spatial correlation when using regression-based or spatial interpolation methods to estimate flood quantiles at ungauged locations. The analysis also shows that coupling TK with CK slightly improves the performance of TK; however, the improvement is marginal when compared to the improvement in performance over GLS.

Hydrology and Earth System Sciences↗
USGS2024

Propensity score matching mitigates risk of faulty inferences in observational studies of effectiveness of restoration trials

Determining effectiveness of restoration treatments is an important requirement of adaptive management, but it can be non-trivial where only portions of large and heterogeneous landscapes of concern can be treated and sampled. Bias and non-randomness in the spatial deployment of treatment and thus sampling is nearly unavoidable in the data available for large-scale management trials, and the biophysical landscape characteristics underlying the bias are key but rare considerations in analyses of treatment effects. Treatment effects from large-scale management trials are typically estimated with multivariable regression (MVR) models. However, this method is unsuited to reliable estimations of treatment effects when treated and untreated areas differ in their underlying biophysical variability. An alternative to conventional regression is to use propensity score (PS) matching, which can limit the differences in confounding variables among treatment groups and assure the data collected or selected for analysis are more consistent with a randomized and unconfounded experiment. Thus, PS is expected to identify treatment effects more accurately. We used data from a large-scale monitoring effort of a megafire to evaluate the efficacy of PS matching in making inferences on treatment effects when treatments are applied non-randomly over a large heterogeneous area. We compared the resulting inference to both traditional MVR methods and to “naïve” methods that do not consider treatment allocation bias. Treatment effects varied between the different statistical methods for controlling selection bias and confounding biophysical factors. The PS-matched model revealed a weaker treatment effect of drill seeding and a greater effect of herbicide spraying on the cover of perennial bunchgrasses when compared to MVR or naïve modelled estimates. The inferences from the PS-matched model are considered more reliable because the treated and untreated plots are more similar in their underlying biophysical characteristics. Synthesis and applications . Failure to consider the non-random and selective deployment of restoration treatments by managers leads to faulty inference on their effectiveness. However, tools such as propensity-score matching can be used to remove the bias from analyses of the outcomes of management trials or to devise sampling plans that efficiently protect against the bias.

Journal of Applied Ecology↗
USGS2008

Watershed Regressions for Pesticides (WARP) for Predicting Annual Maximum and Annual Maximum Moving-Average Concentrations of Atrazine in Streams

Regression models were developed for predicting annual maximum and selected annual maximum moving-average concentrations of atrazine in streams using the Watershed Regressions for Pesticides (WARP) methodology developed by the National Water-Quality Assessment Program (NAWQA) of the U.S. Geological Survey (USGS). The current effort builds on the original WARP models, which were based on the annual mean and selected percentiles of the annual frequency distribution of atrazine concentrations. Estimates of annual maximum and annual maximum moving-average concentrations for selected durations are needed to characterize the levels of atrazine and other pesticides for comparison to specific water-quality benchmarks for evaluation of potential concerns regarding human health or aquatic life. Separate regression models were derived for the annual maximum and annual maximum 21-day, 60-day, and 90-day moving-average concentrations. Development of the regression models used the same explanatory variables, transformations, model development data, model validation data, and regression methods as those used in the original development of WARP. The models accounted for 72 to 75 percent of the variability in the concentration statistics among the 112 sampling sites used for model development. Predicted concentration statistics from the four models were within a factor of 10 of the observed concentration statistics for most of the model development and validation sites. Overall, performance of the models for the development and validation sites supports the application of the WARP models for predicting annual maximum and selected annual maximum moving-average atrazine concentration in streams and provides a framework to interpret the predictions in terms of uncertainty. For streams with inadequate direct measurements of atrazine concentrations, the WARP model predictions for the annual maximum and the annual maximum moving-average atrazine concentrations can be used to characterize the probable levels of atrazine for comparison to specific water-quality benchmarks. Sites with a high probability of exceeding a benchmark for human health or aquatic life can be prioritized for monitoring.

Open-File Report↗
USGS2021

Occurrence, fate, and transport of aerially applied herbicides to control invasive buffelgrass within Saguaro National Park Rincon Mountain District, Arizona, 2015–18

The spread of the invasive and fire-adapted buffelgrass ( Cenchrus ciliaris L.) threatens desert ecosystems by competing for resources, increasing fuel loads, and creating wildfire connectivity. The Rincon Mountain District of Saguaro National Park addressed this natural resource threat with the use of glyphosate-based herbicides (GBHs). In 2010, the Rincon Mountain District initiated an aerial restoration plan to control dense buffelgrass patches in remote areas and implemented a trial project to evaluate the effects of aerial restoration techniques that included the helicopter application of GBHs. In 2014, more than 250 acres of buffelgrass in the Rincon Mountain District were treated with the aerial application of GBHs. This widespread aerial application of GBHs continued through 2018, but the potential transport and effects to aquatic ecosystems were unknown. In 2015–18, the U.S. Geological Survey, in cooperation with the National Park Service, studied the occurrence, distribution, fate, and transport of glyphosate in surface water and sediments derived from areas that were treated during past and current aerial herbicide applications. Three watersheds, treated with different regimens of GBHs, were sampled for glyphosate and the primary metabolite of glyphosate, aminomethylphosphonic acid (AMPA), during various hydrologic flow conditions. Water and aquatic sediment were collected from three watersheds, each in a different stage of application during the U.S. Geological Survey study. The unnamed watershed above the Loma Verde Trailhead referred to by the National Park Service as “Loma Verde canyon” had received no aerial treatment since 2014, whereas the Box Canyon watershed was aerially treated every year beginning in 2014. The Madrona Canyon watershed was first sprayed in 2016 and aerial application continued once a year though the entirety of the study. In addition, terrestrial soil samples were sampled from areas sprayed to understand dissipation rates and herbicide transport via sediments washing away during rainfall runoff. The concentrations present in water and sediment samples were compared to ecological benchmarks and characterized within the context of the environmental conditions of the park setting. Of the 48 water samples collected and analyzed for glyphosate and AMPA, 10.4 percent and 14.6 percent were detected above the laboratory minimum detection limit, respectively. Mean water concentrations, calculated using specific statistical methods for non-detects, were equal to the laboratory minimum detection limit of 0.02 microgram per liter for samples collected in all the watersheds. In aquatic sediments, glyphosate and AMPA were detected in 10.7 and 25.0 percent of the samples, whereas 89.5 and 100 percent of the terrestrial soil samples had detections for glyphosate and AMPA, respectively. Mean aquatic sediment concentrations were 1.13 and 4.42 micrograms per kilogram (μg/kg) for glyphosate and AMPA, respectively. Mean terrestrial soil concentrations were orders of magnitude greater than water and aquatic sediment with concentrations of 678 μg/kg for AMPA and 1,240 μg/kg for glyphosate. Hours after glyphosate-based herbicide was applied, the concentrations of glyphosate and AMPA were present in terrestrial soil samples near or above the laboratory maximum detection limit of 5,000 μg/kg. The Box Canyon watershed was the most intensively treated watershed in terms of total land area treated, total amount of GBH applied, and number of years treated. The frequent and large volume of treatment resulted in the highest number of detections of glyphosate and AMPA in water (3 and 7 detections, respectively) and in aquatic sediment (2 and 6 detections, respectively) samples. In comparison, the other two watersheds had two or fewer detections for glyphosate and AMPA in water and aquatic sediment. Glyphosate detected in pools was associated with increased rainfall closer in time to the last herbicide treatment. Glyphosate and AMPA concentration ratios above one, along with stable-isotope and tritium results, indicated that runoff processes were the primary transport mechanism for the two compounds when found in streams and pools rather than subsurface recharge or deeper flow paths. One pool in a small tributary of Box Canyon consistently had detections of glyphosate and AMPA in aquatic sediments, but these frequent concentrations were likely related to the intensive application upstream, near the steep terrain above the head of the channel that supplies the downstream pool. Intense flows during summer rainfall events move treated sediments into this channel where vegetation and the incised bedrock banks of the pool retained those sediments and ultimately led to frequent detections of both compounds. Isotope results in most of the pools and tinajas indicated that the water source had residence time representative of recently recharged waters, on the order of years. No water concentrations exceeded published criteria for human health or aquatic life. Median and maximum glyphosate and AMPA water concentrations were lower than those reported in other national assessments, but maximum concentrations observed in individual runoff samples were higher than median concentrations measured in the national assessments. A similar finding was observed with aquatic sediment concentrations measured in the Rincon Mountain District. Results from the study were compared and assessed in the context of other studies examining GBHs and their effects on amphibians, fish, and macroinvertebrates. This comparison was used to generalize the potential risk to aquatic species similar to those species in the Rincon Mountain District. Concentrations of published effect levels were several orders of magnitude greater than the highest concentration detected in water at the Rincon Mountain District. Most published studies evaluate acute and chronic toxicity for glyphosate and GBHs, and these criteria may not be representative of environmental conditions in the Rincon Mountain District. The classic lethal dose studies conducted in a controlled laboratory setting may not be suitable for comparison to the longer, variable, low-dose exposure conditions in the pools and tinajas in the Rincon Mountain District. However, this study determined that the fate of GBHs transported from treated areas to potential aquatic habitat was highly variable in occurrence, timing, and concentrations. This variability in glyphosate concentrations was too high, and the potential exposure was determined to be far too complex to directly compare with the results from controlled studies. This study provides the first information collected on GBHs used to control invasive buffelgrass in a remote, mountainous, and semiarid setting. The information about the transport and fate of herbicide application near aquatic habitat will help to inform managers about the broader ecosystem implications and provide useful information to other agencies implementing buffelgrass remediation strategies near aquatic habitat.

Arizona↗