SEARCH · Search USGS

Results for “State of the Survey”

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

SourceResearch connectionFrom yearThrough year

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

At least 1,117 records · Page 62Linked to original sources

USGS2010

Climate change lessons from a warm world

In the early 1970’s to early 1980’s Soviet climatologists were making comparisons to past intervals of warmth in the geologic record and suggesting that these intervals could be possible analogs for 21st century “greenhouse” conditions. Some saw regional warming as a benefit to the Soviet Union and made comments along the lines of “Set fire to the coal mines!” These sentiments were alarming to some, and the United States Geological Survey (USGS) leadership thought they could provide a more quantitative analysis of the data the Soviets were using for the most recent of these warm intervals, the Early Pliocene.

Transactions of the Leicester Literary & Philosoph↗

USGS2008

Investigating gas hydrate as a factor in accretionary margin frontal ridge slope failures and cold seep biogeochemistry

During August 2008, a research expedition (2008-007-PGC) was carried out offshore Vancouver Island on the northern Cascadia Margin (Figure 1) to study the role of gas hydrate in slope stability and cold seep biogeochemistry. The cruise was organized by the Geological Survey of Canada (GSC) as part of the Earth Science Sector, Natural Gas Hydrate Program, Natural Resources Canada (NRCan). This international collaboration included McGill University, University of Victoria, the U.S. Geological Survey, Florida State University, and the U.S. Department of Energy.

Fire in the Ice: NETL Methane Hydrate Newsletter↗

USGS1992

Remote sensing studies of the geomorphology of Surtsey, 1987-1991

The volcanic island of Surtsey, formed by explosive submarine and effusive subaerial eruptions between November 1963 and June 1967, consists of a complex combination of primary and redeposited tephra and alkaline olivine basalt lava flows in a 2.5 km 2 area (Thorarinsson, 1967; Thorarinsson et al., 1964; Fridriksson, 1975). During the past 24 years, wave and wind erosion of this subaerial mid-ocean ridge (MOR) vent complex have modified Surtsey's coastal morphology, including the deposition of a 0.5 km-long northern peninsula ( ness ) composed of tephra and rounded lava fragments derived from the southern half of the island. Detailed geomorphologic and sedimentologic mapping of the various surface units now present on Surtsey has been accomplished throughout the history of the evolving island, most recently by Calles et al. (1980) and Ingolfsson (1980). On the basis of these studies, an effort to quantify the topographic characteristics of the primary geomorphic units on the island was initiated by the National Aeronautics and Space Administration (NASA) and the United States Geological Survey (USGS) in 1987. The objective has been to directly measure the microtopographic properties of the widest range of surface types possible, with special emphasis on the pristine or dynamic types. While large-scale topographic maps of Surtsey were prepared in 1968 and 1975 (Norrman, 1980; Norrman and Erlingsson, 1991; Calles et al, 1980), and geodetic leveling surveys have been carried out (Moore, 1980), there have been no recent attempts to geodetically determine the local topography of the island. Because of the rapid rates of geomorphic processes, such as erosion and deposition, on a small, geologically isolated volcanic island such as Surtsey, it is desirable to determine the meter-scale topographic character of its surface units and landforms, and later a remeasurement of the same surfaces to further quantify volumetric change, subsidence, and process rates. In addition, precise measurements of sub-meter-scale topography of pristine geologic surfaces provides necessary data for the investigation of whether various geologic processes demonstrate fractal or self-affine behavior at a range of length-scales within the interval 0.1 in to 1 km. Thus Surtsey offers a unique opportunity to apply new remote sensing techniques to the measurement of the evolving surface "roughness" characteristics of pristine geologic surfaces within an historically well-monitored environment.

Surtsey Research Progress Report↗

USGS2011

I3N risk assessment and pathway analysis: Tools for the prevention of biological invasions

Information on invasive alien species from published and unpublished accounts and databases is usually scattered in locations and formats that are not easily accessible. Customized informatics tools for collecting and organizing invasive species information can help resource managers better control biological invasions. The Invasives Information Network (I3N) of the Inter-American Biodiversity Information Network has created a distributed network of databases of invasive species profiles, subject matter experts, projects, and datasets hosted and published online by natural resource institutions throughout the Americas. Invasive species information is documented and published online in a standard format that can be searched by the public. Profiles and occurrence information on invaders can be documented using the I3N Database on Invasive Alien Species, published online using an easy-to-use template, and exchanged in standard formats. The I3N Risk Assessment and Pathway Analysis tools are designed to be used in conjunction with I3N Databases to assist decision-makers in setting priorities for containment. All these tools are freely available online at http://i3n.iabin.net. Coordinated by the United States Geological Survey National Biological Information Infrastructure, this network serves as an example of successful capacity building and regional collaboration on an issue of global significance.

Conference Paper↗

USGS2001

Planetary Interactive GIS-on-the-Web Analyzable Database (PIGWAD)

The United States Geological Survey (USGS) in Flagstaff, Arizona is producing a Web-based, user-friendly interface that integrates powerful Geographic Information Systems (GIS) statistical and spatial relational tools for analyses of planetary datasets. The interface, known as “Planetary Interactive GIS-on-the-Web Analyzable Database” (PIGWAD), provides database support for the research and academic planetary science communities, particularly for geologic mapping and other surface-related investigations.

Conference Paper↗

USGS2019

Overview of future USGS Gulf of Mexico buoyant storage assessment project

The United States Geological Survey (USGS) is a member of a U.S. Department of Energy-funded partnership headed by the University of Texas Bureau of Economic Geology that is working to assess the feasibility of offshore geologic carbon dioxide (CO2) storage in the Gulf of Mexico. The role of the USGS is to assess the buoyant geologic CO2 storage resource of the western half of the offshore Gulf of Mexico (GoM). Buoyant CO2 storage is the CO2 held in place by a top and lateral seal (either a sealing formation or a sealing fault), that creates a column of CO2 in communication across pore space in a geologic reservoir. This assessment will be similar to the USGS assessment of onshore buoyant geologic CO2 storage [1] and will employ a modified version of existing USGS methodology [2] to assess the buoyant CO2 storage capacity in the GoM.

Gulf of Mexico↗

USGS2018

Describing the distribution and productivity of biota along a nearshore to offshore gradient

The Lake Michigan Lakewide Action and Management Plan (LAMP) proposed adding nutrients (phosphorus) to its “pollutant of concern” list in 2002, given that excessive nutrients were causing impairments in nearshore waters. Since that time, scientists have highlighted the “shunting” of nutrients to the nearshore (Hecky et al. 2004), owing to the ability of invasive dreissenid mussels to capture some portion of allochthonous phosphorus that enters the lake through tributaries. These changes are believed to increase productivity in the nearshore, reflected in increased benthic and pelagic primary production and nuisance Cladophora (Auer et al. 2010). Whether increases in primary productivity lead to concomitant increases for secondary (by zooplankton) and tertiary (by fish) production remains largely untested. Hence, understanding the distribution and abundance of nutrients and biota (e.g., zooplankton, fish) across a nearshore to offshore gradient was identified as a Cooperative Science and Monitoring Initiative (CSMI) priority in 2015. Increased understanding of the Lake Michigan nearshore will also facilitate the development of a Nearshore Strategy by the LAMP, which is called for in the 2012 Great Lakes Water Quality Agreement. Working collaboratively with Environmental Protection Agency (EPA) and National Oceanic and Atmospheric Administration (NOAA), United States Geological Survey (USGS) described the distribution of nutrients and biota across nearshore to offshore transects in 2015 (see Appendix 1). At each transect, we sampled the food web at three sites with differing bottom depths: 18 m, 46 m, and 91-110 m. We purposefully chose transects near tributaries of varying total phosphorus (TP) input (see Figure 1, Dolan and Chapra 2012): three transects that were not associated with any large tributary where total phosphorus would be loaded (Waukegan IL, Frankfort MI, Sturgeon Bay WI), three transects adjacent to tributaries presumed to be relatively low loaders of TP (Pere Marquette MI, Root WI, Muskegon MI), and three transects adjacent to tributaries presumed to be relatively high loaders of TP (St. Joseph MI, Kalamazoo MI, Manitowoc WI). USGS estimated chlorophyll concentrations, zooplankton, Mysis , larval fish, and juvenile and adult fish seasonally (April/May, July, October/November) at eight of these transects (all but Muskegon).

Illinois, Michigan, Wisconsin↗

USGS2020

A comparison of groundwater sampling technologies, including passive diffusion sampling, for radionuclide contamination

Using traditional high-flow purge methods for long-term water quality monitoring of deep groundwater wells can be expensive, affect contaminant migration, and produce excessive volumes of discharge water that can be difficult to manage. The use of low-flow pumping methods and depth discrete bailers (DDBs) can reduce the cost of sampling deep groundwater wells. In general, using different pumping methods to obtain reproducible and representative groundwater can be challenging. Passive diffusion samplers (PDSs) have successfully been used in long-term monitoring for volatile organic compounds, major ions, and trace-elements, but application has been limited for stable and radioactive isotopes. Beginning in 2018, the United States Geological Survey (USGS) conducted three sampling events to test the ability of PDSs and DDBs to obtain reproducible and representative groundwater samples. The sampled well is completed in a regional, permeable carbonate aquifer and is one in a set of wells that have historically been used for tritium tracer testing. All samples were obtained at 180 m (590 ft) below land surface (bls) in a 13.97 cm (5.5 inch) uncased well that has a total depth of 202 m (662 ft) bls. The first sampling event deployed a regenerated cellulose dialysis membrane (RCDM) PDS for 14 days with deionized water as the blank. The second sampling event deployed a RCDM for 27 days also with deionized water as the blank. The third sampling event deployed a Dual Membrane (DM) PDS for 65 days using a blank of tritium-dead carbonate water. The DM PDS was used to assess the effect of longer-term deployment on tritium concentrations and address whether or not the PDSs reached equilibrium with ambient groundwater. The day after each of the three passive samplers were retrieved a DDB was used to obtain discrete non-integrated groundwater samples. For each DDB sampling day, the bailer was lowered into the well 10 consecutive times to determine if the water chemistry changed from the first to the last bailed sample. Quality assurance samples including blanks and duplicates were obtained during all three sampling events. All blank waters had tritium concentrations less than 21±33 pCi/L. Major ion (e.g. calcium, chloride, sodium, and sulfate) results were compared between all samples obtained with RCDM and DDB. Major ion concentrations showed a coefficient of variation of less than 6% between all RCDM and DDB samples; however, the coefficient of variation between the different deployment times and the two different methods for trace-element concentrations was much larger, particularly for manganese (82%), lead (41%), and zinc (33%). Stable isotope values were compared between the RCDM and DDB samples. The DDB sample results all fell within analytical uncertainty and were considered representative of the formation groundwater. The stable isotope values from the RCDM samples indicated that a longer deployment time was necessary to gain equilibrium and to obtain representative groundwater samples. Tritium results from groundwater samples obtained from the RCDM, DM, and DDB indicate that groundwaters obtained with PDSs produced tritium concentrations 2 to 3 times higher (between 1,426 and 3,060±87 pCi/L) than groundwaters obtained with a DDB (479 to 1,219±51 pCi/L). The longer the passive diffusion samplers were deployed, the higher the tritium concentration, suggesting that equilibrium with tritium was not reached within a 27-day deployment. DDB samples showed tritium results declining from the first to the last bailed sample for all three sampling events. This research suggests that tritium results from groundwater samples obtained from PDSs are more reproducible than samples obtained from DDBs. Also, PDSs likely do not accumulate isotopes of water but rather equilibrate with the ambient groundwater. On the other hand, both PDSs and DDBs were able to provide representative groundwater samples for major ions and have the potential to produce.

Conference Paper↗

USGS2020

Assessment of methods for soil monitoring in the Adirondack region of New York

Repeated sampling to detect changes in forest soils was rarely used before 1990, but the value of soil monitoring in understanding environmental change is becoming well established. The growing number of resampling studies has shown that sampling designs and procedures must be adapted to the objectives of the monitoring program and the soils being monitored. In the Adirondack region, current priorities include the response of soils to large increases, and more recently, large decreases in acidic deposition, and changes driven by trending climate such as altered pools of soil organic carbon, as well as other unforeseen factors that will occur in the future. To improve methods and assess the feasibility of long-term soil monitoring in the Adirondack region, the United States Geological Survey (USGS) conducted a pilot project to evaluate a new sampling method for characterizing soils on a watershed basis. Results obtained with this new approach, referred to as the ADK sampling method, was compared to methods used in previous sampling conducted in 2004 as part of the Western Adirondack Stream Survey (WASS), and also to previous high-replication pit sampling in the North and South Tributary watersheds of Buck Creek (North Buck and South Buck). The number of sampling locations and spatial distribution of sampling points within watersheds differed among the methods, although pit excavation was used to obtain samples in all cases. In addition, this investigation evaluated the use of small diameter corers as a means to measure forest floor mass with greater accuracy and precision than commonly used methods such as pit excavation. Sufficient statistical power to detect ecologically relevant changes in upper profile horizons (Oe, Oa and upper 10 cm of the B) were achieved with the ADK sampling method that utilized 18 pit excavations per watershed. The sampling locations were organized within each watershed into three study areas (six sampling locations per study area) that represented the primary types of landscape within the watershed. Sampling at 18 locations per watershed was found to be nearly as effective at detecting changes as sampling at 28 locations per watershed. Numerous significant changes (P < 0.10) were detected with both 18 and 28 sampling locations at sampling intervals of 12 to 16 years. The relationship between soil data obtained with the ADK method and stream chemistry at the base of the watershed suggested that this approach adequately characterized soil variability within the watershed for the purpose of studying soil-stream linkages. Significant changes in upper B horizon calcium (P < 0.10) and Oa horizon aluminum (P < 0.01) were detected when the data from the four WASS watersheds were combined with the two Buck Creek watersheds, which suggested that there would be value in resampling other WASS watersheds previously sampled in 2004 to support a regional assessment. Study results support small diameter cores as a useful method to monitor changes in the organic matter mass of the forest floor. This method showed high reproducibility in repeated sampling tests and lower spatial variability in sample data than traditional approaches when compared on a watershed basis. Soil coring is also faster and requires less equipment than pit excavation methods, which makes it more conducive to sampling over large areas. However, organic matter mass of the forest floor determined by coring was consistently less than the values obtained by the ADK sampling method that used pit sampling and vertical horizon measurements, and also literature values of a previous Adirondack study that utilized pit sampling in which the entire horizon was collected over a measured area. However, a high correlation (R2 = 0.87) occurred between organic matter content (expressed as Mg ha-1) determined by coring and the ADK sampling method. Differing methods with regard to where sample could be collected, and how organic matter was collected for chemical analysis were the likely reasons for differences in quantification of forest floor organic mass. Collection of forest floor cores in conjunction with the ADK method is recommended to provide improved sensitivity in detecting changes in the forest floor in proximity of where full analyses of the soil profile are being done. This duel sampling approach represents an optimized method for measuring and understanding how Adirondack soils will change in the future.

New York↗

USGS2021

Risk-informed levee erosion countermeasure site selection and design in the Sacramento area part 1: Soil sampling, testing, and data processing

USACE partnered with the United States Department of Agriculture, Agricultural Research Service, United States Geological Survey, and Texas A&M University to evaluate the erodibility of the river banks and levees to inform probabilistic numerical simulations using the Bank Stability and Toe Erosion Model (BSTEM). This paper discusses the measurement of the intrinsic erosion and geotechnical properties of the soil for improved BSTEM results and is the first of two parts. Sampling and testing were conducted at select sites to obtain the soil stratigraphy and collect samples for estimation of engineering properties. Soil erodibility was measured using the Erosion Function Apparatus test, the Borehole Erosion Test, the Pocket Erodometer Test, and the mini-Jet Erosion Test. Critical evaluation of previously existing datasets and the collection of these new datasets helped to provide better definition of the range in erosion parameters for improved probabilistic erosion estimates using BSTEM for risk-informed levee erosion countermeasure site selection and design.

California↗

USGS2021

Risk-informed levee erosion countermeasure site selection and design in the Sacramento area part 2: Probabilistic numerical simulation of bank erosion

USACE partnered with the United States Department of Agriculture, Agricultural Research Service, United States Geological Survey, and Texas A&M University to evaluate the erodibility of the river banks and levees to inform probabilistic numerical simulations using the Bank Stability and Toe Erosion Model (BSTEM). This paper, the second of two parts, addresses processing the collected data to inform inputs for probabilistic bank erosion estimates in BSTEM. Measuring the intrinsic soil properties for BSTEM is discussed in part one. Soil critical shear stress and soil erodibility coefficients were calibrated by Unified Soil Classification soil type to observed erosion on the American River. Adjustments were made in the probability density functions for these parameters to reflect field-measured variability and carry forward the reduction in error achieved during calibration. The resulting calibrated values were tested at additional sites, validating the resulting critical shear stress and soil erodibility coefficient values and probability density functions for more robust probabilistic bank erosion estimates using BSTEM.

California↗

USGS2021

Genetics as a tool for conservation and management of West Indian manatee populations in Brazil

A study conducted by the National Center for Research and Conservation of Aquatic Mammals (CMA), United States Geological Survey Wetland and Aquatic Research Center, and partner researchers found that the marine Antillean manatee ( Trichechus manatus manatus ) has low genetic diversity in regions where there are territorial interfaces with the Amazonian manatee ( Trichechus inunguis ). The study was published as an article in the journal, Frontiers in Marine Science.

Sirenews↗

USGS2016

Learning from the recent Taiwan Meinong Earthquake

This paper highlights the lessons learned following a reconnaissance trip to Tainan, Taiwan two weeks after the February 2016 earthquake. The reconnaissance was conducted by Gilsanz, Murray Steficek engineers (GMS) and an earthquake engineer from the United States Geological Survey (USGS), in collaboration with the Earthquake Engineering Research Institute (EERI), the Applied Technical Council (ATC) and the National Center for Research in Earthquake Engineering (NCREE) of Taiwan. Observations were made of damages, success of base-isolated high-rise and retrofitted school buildings, site effects and ground failures as well as the response of both city agencies and building owners to the earthquake. The motivation for engineers and the public for such a mission is to learn from the observations to be better prepared for the next earthquake.

Meinong↗

USGS2020

Imagery

As part of CRMS, Digital Orthophoto Quarter Quadrangles (DOQQs) for the coastal region of Louisiana are created for years when coastwide land-water classifications are required. A DOQQ is a raster image in which displacement in the image caused by sensor orientation and terrain relief has been corrected. These images combine the image characteristics of a photo with the geometric qualities of a map. The DOQQs generated for this project consist of four components or spectral bands of information: blue, green, red and very-near infrared (VNIR). These images are referred to as Color Infrared (CIR) digital imagery. CRMS site-level assessments of land-water coverage will be based off of color-infrared photography acquired for coastal Louisiana and clipped to each 1-km2 CRMS-Wetlands site. Unless otherwise noted as a specific preliminary condition, all vegetation such as scrub-shrub, emergent vegetation, and forested areas will fall under the land classification, while open water, non-vegetated, regularly flooded mud flats, and aquatic vegetation beds will be characterized as water. CRMS imagery contracts are managed and issued by the United States Geological Survey (USGS) Wetland and Aquatic Research Center.

Louisiana↗

USGS2022

BIA branch of tribal climate resilience regional assessment report

The Bureau of Indian Affairs (BIA) Branch of Tribal Climate Resilience (TCR) was established in 2011 to enable Tribal and BIA Office of Trust Services Programs to incorporate climate considerations into their programs' planning and decision-making. The TCR’s purpose is to fulfill trust responsibility to Federally recognized Tribal Nations by supporting and building Tribal capacity for climate resilience efforts through both technical and financial assistance. That financial assistance includes the TCR Annual Awards Program in which over 700 awards have been distributed totaling more than $74 million as of September 24th, 2021. The TCR has funded over 250 adaptation plans and vulnerability and risk assessments over the last 11 years. Tribes can leverage TCR programs to further support their existing operations, program planning, project design, as well as incorporate Indigenous and Traditional Knowledges into their management and protection of Trust resources and Tribal economies. Due to the availability of TCR awards and other funding, many Tribes now have climate adaptation plans. Thus, both TCR and Tribes are at a new stage compared to a decade ago and have new needs related to implementation of climate adaptation and resiliency projects. This assessment aims to evaluate the effectiveness of TCR’s Annual Awards Program and technical assistance and incorporate direct Tribal input to meet Tribal needs and priorities. To provide more effective assistance to Tribal communities moving forward; BIA staff, in partnership with staff from the United States Geological Survey (USGS), interviewed 16 Tribal employees to gather their feedback on their approaches to climate adaptation as well as their experiences with TCR services. This report aims to summarize guidance from participating Tribes to inform future changes to BIA TCR.

Regional Assessment Report↗

USGS2022

Applying consequence-driven scenario selection to lifelines

We present a new consequence-driven framework for earthquake scenario selection. For emergency managers, utility operators, policy makers, and other stakeholders, a scenario-based seismic risk assessment is often necessary for the purpose of emergency management and planning. In developing a scientifically defensible scenario, stakeholders can simulate a realistic event in order to pre-identify vulnerabilities in the system and support action to address these vulnerabilities. Selecting scenarios is particularly challenging for important population centers and critical infrastructure in stable tectonic environments, such as in the central and eastern United States, where uncertain long-term seismicity and unknown faults offer inadequate constraints. Notably, significant events in these so-called stable regions do occur (e.g., Nahanni, Canada, 1985, M6.9; Tennant Creek, Australia, 1998, M6.7). In regions of low seismicity, even moderate events can be consequential due to the higher vulnerability of buildings typical of such regions when compared to regions of higher seismicity. Furthermore, communicating seismic risk to stakeholders and the general public in these regions can be especially challenging due to the complexities of characterizing the hazard level. This framework has been developed to address these challenges for scenario selection in low seismic hazard regions. In this new approach, the analysis begins instead with the explicit definition of a consequence of concern to the specific stakeholder. This can range from a definition of loss (in lives, dollars, or another metric of interest), or a performance metric for critical infrastructure. The framework leverages United States Geological Survey software to run the hazard and consequence analysis. Driven by this stakeholder-defined consequence, an inversion analysis generates a complete event set of candidate scenarios that could breach this consequence. The final selection of a scenario, or family of scenarios, is then scientifically informed, but not limited by our lack of constraints in characterizing the hazard.

Conference Paper↗

USGS2024

A journey to the center of the USGS National Strong-motion Project processing and beyond

The United States Geological Survey (USGS) National Strong Motion Project (NSMP) has the primary U.S. government responsibility to acquire, process, and disseminate significant strong-motion earthquake ground motion records measured at surficial free-field stations, structures (buildings, dams, and bridges, and geotechnical arrays to the earthquake engineering community. As a result of the deployment of modern seismic instrumentation and growth of tools such as web-services, earthquake data from U.S. and international seismic networks are more accessible than ever. Our mission is to provide raw and processed strong-motion waveforms with PGA values greater than 0.1%g for M3.0 earthquakes and larger in California and M4.0 and larger within the conterminous US, Hawaii, Puerto Rico, and Alaska. Datasets of interest to the engineering and geophysics communities, such as event sequences in areas of induced seismicity and significant global events, are also processed and posted at the Center for Engineering Strong Motion Data (CESMD) at strongmotioncenter.org when available through collaboration with the international strong-motion data community. Here we outline (1) the NSMP’s current workflow to acquire, process, and distribute data at CESMD; (2) our new endeavours and collaborations focusing on comparison and integration of waveform processing software, development of techniques for metadata quality checks before and after earthquakes, and construction of a dynamic site characterization repository; and (3) our topics for possible collaboration topics across the global strong-motion community.

Conference Paper↗

USGS2024

Adaptive fine-tuning for transferring a U-net hydrography extraction model using K-means

The United States Geological Survey (USGS) coordinates the collection of hydrographic features derived from remotely sensed interferometric synthetic aperture radar (IfSAR) elevation and intensity data in Alaska. Hydrographic features are cartographic representations of surface water features such as stream, rivers, lakes, ponds, canals, etc. Collection and validation procedures involve complex automated and manual techniques that furnish snapshots of hydrographic vector data that exist during the IfSAR surveys. The dynamic nature of fluvial conditions warrants monitoring and updating hydrographic data, but extraction procedures for updates can be cost prohibitive. This paper overviews progress on automated workflows to extract hydrography from IfSAR data using deep learning methods trained and tested with USGS collected hydrography data. This research tests transfer learning methods on a well-performing U-net model trained on a 4600-square kilometer (sq km) base model area in northcentral Alaska. The base model is transferred and fine-tuned to regions in the target domain covering roughly 127,000 sq km. The target domain is subdivided into areas with similar hydrogeomorphic conditions using principal components and k-means clustering, and the base model is adaptively fine-tuned to each hydrogeomorphic class by selecting training watersheds from each cluster within the target domain. Results are compared with transfer learning that is fine-tuned with a random sample of watersheds in the target domain.

Conference Paper↗