Search USGS⌕ Search

Geology topics

Enrika Hlavacek

Publications and source records attributed to Enrika Hlavacek.

18 recordsLinked to original sources
USGS2024

Deep learning workflow to support in-flight processing of digital aerial imagery for wildlife population surveys

Deep learning shows promise for automating detection and classification of wildlife from digital aerial imagery to support cost-efficient remote sensing solutions for wildlife population monitoring. To support in-flight orthorectification and machine learning processing to detect and classify wildlife from imagery in near real-time, we evaluated deep learning methods that address hardware limitations and the need for processing efficiencies to support the envisioned in-flight workflow. We developed an annotated dataset for a suite of marine birds from high-resolution digital aerial imagery collected over open water environments to train the models. The proposed 3-stage workflow for automated, in-flight data processing includes: 1) image filtering based on the probability of any bird occurrence, 2) bird instance detection, and 3) bird instance classification. For image filtering, we compared the performance of a binary classifier with Mask Region-based Convolutional Neural Network (Mask R-CNN) as a means of sub-setting large volumes of imagery based on the probability of at least one bird occurrence in an image. On both the validation and test datasets, the binary classifier achieved higher performance than Mask R-CNN for predicting bird occurrence at the image-level. We recommend the binary classifier over Mask R-CNN for workflow first-stage filtering. For bird instance detection, we leveraged Mask R-CNN as our detection framework and proposed an iterative refinement method to bootstrap our predicted detections from loose ground-truth annotations. We also discuss future work to address the taxonomic classification phase of the envisioned workflow.

Massachusetts, Wisconsin↗
USGS2024

Identifying conditions where reed canarygrass (Phalaris arundinacea) functions as a driver of forest loss in the Upper Mississippi River floodplain under different hydrological scenarios

Most of the world’s river-floodplain ecosystems are simultaneously undergoing modifications to their hydrological regimes and experiencing species invasions, making it unclear whether invasive species are the main drivers of ecosystem change or simply responding to changes in the hydrological regime. We simulated patterns of forest recruitment and succession in a 2500-ha portion of the Upper Mississippi River floodplain with and without removal of invasive Phalaris arundinacea and under two different future 100-year hydrological scenarios: a future maintaining the average flooding conditions of the past 40 years (random) and a future that projects an observed upward 40-year trend in flooding conditions forward (trending). By comparing scenarios that included Phalaris removal and ones that did not, we were able to identify the conditions where Phalaris was the main driver of forest loss vs. the conditions where hydrology was the main driver of forest loss. Areas where Phalaris was the main driver of forest loss had mean annual flood inundation durations that were similar to areas that did not lose forest cover (60–90 growing season days), while areas where flooding was the main driver of forest loss had longer mean inundation durations (102–124 growing season days). In comparison to the random hydrology scenario, the trending scenario produced a decrease in the area over which Phalaris was identified as the main driver of forest loss and an increase in the area over which flood inundation was identified as the main driver of forest loss. Thus, if the observed trends in flooding continue, our model projects an increase in the area over which eradicating Phalaris is unlikely to result in the maintenance of forest cover. We used the Resist-Accept-Direct (RAD) framework to discuss potential management options to resist changes and maintain forest cover where Phalaris is likely to be the main driver of forest loss and to accept or direct changes in areas where forest loss is likely driven by hydrological change.

Iowa, Minnesota, Wisconsin↗
USGS2023

Reconstructing missing data by comparing interpolation techniques: Applications for long-term water quality data

Missing data are typical yet must be addressed for proper inferences or expanding datasets to guide our limnological understanding and management of aquatic systems. Interpolation methods (i.e., estimating missing values using known values within the dataset) can alleviate data gaps and common problems. We compared seven popular interpolation methods for predicting substantial missingness in a long-term water quality dataset from the Upper Mississippi River, U.S.A. The dataset included 80,000 sampling sites collected over 30 yr that had substantial missingness for total nitrogen (TN), total phosphorus (TP), and water velocity. For all three interpolated water quality variables, random forests had very high prediction accuracy and outperformed the methods of ordinary kriging, polynomial regressions, regression trees, and inverse distance weighting. TP had a mean absolute error (MAE) of 0.03 mg (L-TP) −1 , TN had a MAE of 0.39 mg (L-TN) −1 , and water velocity had a MAE of 0.10 m s −1 . The random forests' error rates were mapped and showed low spatiotemporal variability across the riverscape, indicating high model performance across many habitat types and large spatial scales. In the current era of “big data,” interpolation becomes an imperative step prior to ecological analyses yet remains unfamiliar and underutilized. Our research briefly describes the importance of addressing missingness and provides a roadmap to conduct model intercomparisons of other big datasets. We also share adaptable data analysis scripts, which allows others to readily conduct interpolation comparisons for many limnology applications and contexts.

Illinois, Iowa, Minnesota, Missouri, Wisconsin↗
USGS2023

U.S. Geological Survey invasive carp strategic framework, 2023–27

U.S. Geological Survey (USGS) research has supported management of Hypophthalmichthys nobilis (bighead carp), Mylopharyngodon piceus (black carp), Ctenopharyngodon idella (grass carp), and H. molitrix (silver carp), hereafter referred to collectively as invasive carps, for over a decade. This strategic framework identifies thematic research areas to guide funding decisions for USGS invasive carp research during 2023–27. The USGS will use this framework and input from research and management agencies to identify leading-edge research to advance control of these highly damaging invasive species.

Circular↗
USGS2023

Challenges and solutions for automated avian recognition in aerial imagery

Remote aerial sensing provides a non-invasive, large geographical-scale technology for avian monitoring, but the manual processing of images limits its development and applications. Artificial Intelligence (AI) methods can be used to mitigate this manual image processing requirement. The implementation of AI methods, however, has several challenges: (1) imbalanced (i.e., long-tailed) data distribution, (2) annotation uncertainty in categorization, and (3) dataset discrepancies across different study sites. Here we use aerial imagery data of waterbirds around Cape Cod and Lake Michigan in the United States to examine how these challenges limit avian recognition performance. We review existing solutions and demonstrate as use cases how methods like Label Distribution Aware Marginal Loss with Deferred Re-Weighting, hierarchical classification, and FixMatch address the three challenges. We also present a new approach to tackle the annotation uncertainty challenge using a Soft-fine Pseudo-Label methodology. Finally, we aim with this paper to increase awareness in the ecological remote sensing community of these challenges and bridge the gap between ecological applications and state-of-the-art computer science, thereby opening new doors to future research.

Remote Sensing in Ecology and Conservation↗
USGS2021

Visualizing proximity of non-native species to protected areas of the United States—A proximity visualization tool for BISON

The Proximity Visualization Tool is a simple lightweight tool that can be placed on web pages that allows users to identify non-native species near Department of Interior lands. The tool works by accessing the more than 400 million species occurrence records in the Biodiversity Information Serving Our Nation (BISON) database using the BISON Application Programming Interface (API).

Fact Sheet↗
USGS2021

Riparian forest cover modulates phosphorus storage and nitrogen cycling in agricultural stream sediments

Watershed land cover affects in-stream water quality and sediment nutrient dynamics. The presence of natural land cover in the riparian zone can reduce the negative effects of agricultural land use on water quality; however, literature evaluating the effects of natural riparian land cover on stream sediment nutrient dynamics is scarce. The objective of this study was to assess if stream sediment phosphorus retention and nitrogen removal varies with riparian forest cover in agricultural watersheds. Stream sediment nutrient dynamics from 28 sites with mixed land cover were sampled three times during the growing season. Phosphorus dynamics and nitrification rates did not change considerably throughout the study period. Sediment total phosphorus concentrations and nitrification rates decreased as riparian forest cover increased likely due to a decline in fine, organic material. Denitrification rates were strongly correlated to surface water nitrate concentrations. Denitrification rate and denitrification enzyme activity decreased with an increase in forest cover during the first sampling period only. The first sampling period coincided with the greatest connectivity between the watershed and in-stream processing, indicating that riparian forest cover indirectly decreased denitrification rates by reducing the concentrations of dissolved nutrients entering the stream. This reduction in load may allow the sediment to maintain greater nitrogen removal efficiency, because bacteria are not saturated with nitrogen. Riparian forest cover also appeared to lessen the effect of agriculture in the watershed by decreasing the amount of fine material in the stream, resulting in reduced phosphorus storage in the stream sediment.

Wisconsin↗
USGS2021

Mercury and water level management in lakes of northern Minnesota

Water level (WL) fluctuations substantially alter the fauna, flora, and microbial community of nearshore aquatic ecosystems. Water level management therefore has the potential to strongly influence a wide variety of ecosystem processes. Many northern temperate lake food webs experience substantial methylmercury contamination, which is partially mediated by the action of sulfate-reducing bacteria occurring in sediments that are periodically inundated. For lakes with elevated methylmercury, WL management could be designed to reduce methylmercury contamination. At the lake scale, this concept is supported by studies that identified statistical associations between fish mercury content and water level (WL) fluctuations. Here, we compiled a long-term dataset (1997–2015) of mercury content in young-of-year Yellow Perch ( Perca flavescens ) from six lakes on the border of the United States and Canada and examined whether mercury content was associated with WL fluctuation. Many WL metrics covary and appear to have strong associations with Yellow Perch mercury. However, these associations appear to vary by lake, and lake-specific models are needed to identify relationships between WL fluctuation and Yellow Perch mercury content. We used partial least-squares regression (PLSR) to identify the associations between Yellow Perch mercury content and WL metrics, temperature, and annual deposition data for lakes in northern Minnesota. These PLSR models not only showed some variation among lakes, but also supported strong associations between WL fluctuations and annual variation in Yellow Perch mercury content. The study lakes underwent a change in WL management in 2000, when winter WL minimums were increased by about 1 m in five of the six study lakes, which reduced annual WL fluctuation on those lakes. Using the PLSR models, we estimated how this change in WL management would have affected Yellow Perch mercury content. In four of the five study lakes in which annual WL fluctuation was reduced in 2000, the change in WL management likely reduced Yellow Perch mercury content, relative to the previous WL management regime.

Minnesota↗
USGS2020

Preliminary analysis to estimate the spatial distribution of benefits of P load reduction: Identifying the spatial influence of phosphorus loading from the Maumee River (USA) in western Lake Erie

Since the early 2000s, Lake Erie has been experiencing annual cyanobacterial blooms that often cover large portions of the western basin and even reach into the central basin. These blooms have affected several ecosystem services provided by Lake Erie to surrounding communities (notably drinking water quality). Several modeling efforts have identified the springtime total bioavailable phosphorus (TBP) load as a major driver of maximum cyanobacterial biomass in western Lake Erie, and on this basis, international water management bodies have set a phosphorus (P) reduction goal. This P reduction goal is intended to reduce maximum cyanobacterial biomass, but there has been very limited effort to identify the specific locations within the western basin of Lake Erie that will likely experience the most benefits. Here, we used pixel‐specific linear regression to identify where annual variation in spring TBP loads is most strongly associated with cyanobacterial abundance, as inferred from satellite imagery. Using this approach, we find that annual TBP loads are most strongly associated with cyanobacterial abundance in the central and southern areas of the western basin. At the location of the Toledo water intake, the association between TBP load and cyanobacterial abundance is moderate, and in Maumee Bay (near Toledo, Ohio), the association between TBP and cyanobacterial abundance is no better than a null model. Both of these locations are important for the delivery of specific ecosystem services, but this analysis indicates that P load reductions would not be expected to substantially improve maximum annual cyanobacterial abundance in these locations. These results are preliminary in the sense that only a limited set of models were tested in this analysis, but these results illustrate the importance of identifying whether the spatial distribution of management benefits (in this case P load reduction) matches the spatial distribution of management goals (reducing the effects of cyanobacteria on important ecosystem services).

Michigan, Ohio↗
USGS2020

USGS Illinois River monitoring and evaluation

Asian carp monitoring and contract removal will continue throughout the Upper Illinois Waterway system as needed for adaptive management to mitigate, control, and contain Asian carp. Compiling data from monitoring and removal efforts into a centralized database (Illinois River Catch Database application) facilitates data standardization, quality, accessibility, sharing, and analysis to aid in Asian carp removal efforts, evaluations of management actions, and modeling efforts (e.g., SEACarP model). Data summarization, visualization, and modeling supports a better understanding of bigheaded carp life history, behavior, and habitat use. Integrating Asian carp-related data and analyses into decision support tools and products aids in applying control and containment methods in an informed and transparent manner (e.g., improved efficiencies in implementations of the Unified Method, inform targeted removal efforts or deterrent deployments in key locations based on preferential benthic characteristics and environmental conditions).

Illinois↗
USGS2020

USGS Telemetry Project

Telemetry of acoustically tagged bigheaded carp (i.e., bighead carp Hypophthalmichthys nobilis and silver carp H. molitrix ) and surrogate fish species has become an invaluable tool in management for these species in the upper Illinois Waterway Systems (i.e., upper Illinois River, lower Des Plaines River, and Chicago Area Waterway System). For example, movement probabilities between adjacent navigation pools need to be estimated to parameterize the Spatially Explicit Asian Carp Population Model (SEAcarP). SEAcarP is a population model used in scenario planning by the Monitoring and Response Workgroup (MRWG) to evaluate alternative management actions. These movement probabilities are estimated from the telemetry data obtained from a longitudinal network of strategically placed receivers that detect bigheaded carp that have been implanted with acoustic transmitters. In addition, fish removal by contracted fishers has become the primary method of controlling bigheaded carp in the upper Illinois and lower Des Plaines Rivers. Variable patterns in bigheaded carp distribution, habitat, and movement, influenced by seasonal and environmental conditions, make targeting bigheaded carp for removal and containment challenging and costly. Understanding these movement patterns for bigheaded carp through modeling and real-time telemetry applications informs removal efforts and facilitates monitoring and contingency actions based on fish movements. To develop a better understanding of fish movement dynamics to meet management objectives, an existing network of real-time and data-logging acoustic receivers in the upper Illinois Waterway Systems is collaboratively managed by a multi-agency team (see Participating Agencies section above). A Telemetry Workgroup has been established by the MRWG to ensure that the multi-agency telemetry efforts are coordinated to efficiently and effectively meet the MRWG goals. This workgroup plans and executes the placement of receivers, tagging of bigheaded carp with acoustic tags, and management of the telemetry data. Three primary objectives to meet MRWG goals identified by the Telemetry Workgroup included (1) development of a common standardized telemetry database with visualization and analysis tools, (2) transitioning from Program MARK (http://www.phidot.org/software/mark/) to a custom Bayesian multi-state model for estimating movement probabilities needed for SEAcarP and (3) deploying, maintaining, and serving data from real-time acoustic receivers to inform contingency planning and fish removal. A telemetry database and visualization tools (FishTracks) will facilitate standardization, archiving, sharing, quality assurance, visualization and analysis of the telemetry data needed for management. Modifications and additions to FishTracks will facilitate more problem-free use of the database and associated applications, as well as useful extraction of information to meet management goals. The transition to a custom Bayesian multi-state model to estimate movement probabilities will support more efficient, effective, and robust population modeling with SEAcarP by overcoming short comings of Program MARK for this purpose. These shortcomings include lack of customizability and extensibility, problems of singularities and poor-convergence, software crashes, parameter exclusion from models, an inability to consistently generate estimates of movement probability, and a lack of uncertainty estimates for movement probabilities. A real-time receiver network that is maintained and tested annually will ensure reliability and accuracy of the real-time alerts to bigheaded carp movements that can be used by management to plan contingency actions.

Illinois↗
USGS2017

National Park Service Vegetation Mapping Inventory Program: Appalachian National Scenic Trail vegetation mapping project

The National Park Service (NPS) Vegetation Mapping Inventory (VMI) Program classifies, describes, and maps existing vegetation of national park units for the NPS Natural Resource Inventory and Monitoring (I&M) Program. The NPS VMI Program is managed by the NPS I&M Division and provides baseline vegetation information to the NPS Natural Resource I&M Program. The U.S. Geological Survey Upper Midwest Environmental Sciences Center, NatureServe, NPS Northeast Temperate Network, and NPS Appalachian National Scenic Trail (APPA) have completed vegetation classification and mapping of APPA for the NPS VMI Program. Mappers, ecologists, and botanists collaborated to affirm vegetation types within the U.S. National Vegetation Classification (USNVC) of APPA and to determine how best to map the vegetation types by using aerial imagery. Analyses of data from 1,618 vegetation plots were used to describe USNVC associations of APPA. Data from 289 verification sites were collected to test the field key to vegetation associations and the application of vegetation associations to a sample set of map polygons. Data from 269 validation sites were collected to assess vegetation mapping prior to submitting the vegetation map for accuracy assessment (AA). Data from 3,265 AA sites were collected, of which 3,204 were used to test accuracy of the vegetation map layer. The collective of these datasets affirmed 280 USNVC associations for the APPA vegetation mapping project. To map the vegetation and land cover of APPA, 169 map classes were developed. The 169 map classes consist of 150 that represent natural (including ruderal) vegetation types in the USNVC, 11 that represent cultural (agricultural and developed) vegetation types in the USNVC, 5 that represent natural landscapes with catastrophic disturbance or some other modification to natural vegetation preventing accurate classification in the USNVC, and 3 that represent nonvegetated water (non-USNVC). Features were interpreted from viewing 4-band digital aerial imagery using digital onscreen three-dimensional stereoscopic workflow systems in geographic information systems (GIS). (Digital aerial imagery was collected each fall during 2009–11 to capture leaf-phenology change of hardwood trees across the latitudinal range of APPA.) The interpreted data were digitally and spatially referenced, thus making the spatial-database layers usable in GIS. Polygon units were mapped to either a 0.5-hectare (ha) or 0.25-ha minimum mapping unit, depending on vegetation type or scenario; however, polygon units were mapped to 0.1 ha for alpine vegetation. A geodatabase containing various feature-class layers and tables provide locations and support data to USNVC vegetation types (vegetation map layer), vegetation plots, verification sites, validation sites, AA sites, project boundary extent and zones, and aerial image centers and flight lines. The feature-class layer and related tables of the vegetation map layer provide 30,395 polygons of detailed attribute data covering 110,919.7 ha, with an average polygon size of 3.6 ha; the vegetation map coincides closely with the administrative boundary for APPA. Summary reports generated from the vegetation map layer of the map classes representing USNVC natural (including ruderal) vegetation types apply to 28,242 polygons (92.9% of polygons) and cover 106,413.0 ha (95.9%) of the map extent for APPA. The map layer indicates APPA to be 92.4% forest and woodland (102,480.8 ha), 1.7% shrubland (1866.3 ha), and 1.8% herbaceous cover (2,065.9 ha). Map classes representing park-special vegetation (undefined in the USNVC) apply to 58 polygons (0.2% of polygons) and cover 404.3 ha (0.4%) of the map extent. Map classes representing USNVC cultural types apply to 1,777 polygons (5.8% of polygons) and cover 2,516.3 ha (2.3%) of the map extent. Map classes representing nonvegetated water (non-USNVC) apply to 332 polygons (1.1% of polygons) and cover 1,586.2 ha (1.4%) of the map extent.

Appalachian National Scenic Trail↗
USGS2017

National Park Service Vegetation Mapping Inventory Program: Jean Lafitte National Historical Park and Preserve Vegetation Mapping Project

The National Park Service, Natural Resource Stewardship and Science office in Fort Collins, Colorado, publishes a range of reports that address natural resource topics. These reports are of interest and applicability to a broad audience in the National Park Service and others in natural resource management, including scientists, conservation and environmental constituencies, and the public. The Natural Resource Report Series is used to disseminate comprehensive information and analysis about natural resources and related topics concerning lands managed by the National Park Service. The series supports the advancement of science, informed decision-making, and the achievement of the National Park Service mission. The series also provides a forum for presenting lengthier results that may not be accepted by publications with page limitations. All manuscripts in the series receive the appropriate level of peer review to ensure that the information is scientifically credible, technically accurate, appropriately written for the intended audience, and designed and published in a professional manner. This report received formal peer review by subject-matter experts whose background and expertise put them on par technically and scientifically with the authors of the information. The peer review was led according to the Fundamental Science Practices of the U.S. Geological Survey. Views, statements, findings, conclusions, recommendations, and data in this report do not necessarily reflect views and policies of the National Park Service, U.S. Department of the Interior. Mention of trade names or commercial products does not constitute endorsement or recommendation for use by the U.S. Government. This report is available in digital format from the Gulf Coast Network website and the Natural Resource Publications Management website.

Louisiana↗
USGS2015

Automated lidar-derived canopy height estimates for the Upper Mississippi River System

Land cover/land use (LCU) classifications serve as important decision support products for researchers and land managers. The LCU classifications produced by the U.S. Geological Survey’s Upper Midwest Environmental Sciences Center (UMESC) include canopy height estimates that are assigned through manual aerial photography interpretation techniques. In an effort to improve upon these techniques, this project investigated the use of high-density lidar data for the Upper Mississippi River System to determine canopy height. An ArcGIS tool was developed to automatically derive height modifier information based on the extent of land cover features for forest classes. The measurement of canopy height included a calculation of the average height from lidar point cloud data as well as the inclusion of a local maximum filter to identify individual tree canopies. Results were compared to original manually interpreted height modifiers and to field survey data from U.S. Forest Service Forest Inventory and Analysis plots. This project demonstrated the effectiveness of utilizing lidar data to more efficiently assign height modifier attributes to LCU classifications produced by the UMESC.

Minnesota, Wisconsin↗