SEARCH · USGS Science

Results for “Remote Sensing of the Environment”

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 235 records · Page 13Linked to original sources

USGS2026

Hyperspectral retrieval of phytoplankton absorption and community composition from NASA’s PACE-OCI in estuarine–coastal waters using a hybrid framework combining mixture-of-experts and Variational Autoencoder

Retrieving the phytoplankton absorption coefficient (a phy ; m−1), one of the most spectrally rich inherent optical properties, remains challenging in optically complex coastal waters worldwide. Leveraging NASA's new hyperspectral mission, PACE, we introduce Hyper-MoE-VAE, a deep-learning architecture that integrates a Mixture-of-Experts with a Variational Autoencoder to retrieve high-dimensional a phy and subsequent estimation of phytoplankton community composition (PCC) from PACE-OCI hyperspectral remote sensing reflectance (R rs ). Pre-trained on global hyperspectral bio-optical datasets and fine-tuned using regional field R rs –a phy pairings from inland– estuarine–coastal waters, Hyper-MoE-VAE demonstrated strong transferability and effective adaptation across regions. Validation with in-situ Rrs showed accurate aphy retrievals in Lake Erie (NRMSE = 0.12, ε = 17.10), Lake Pontchartrain (NRMSE = 0.11, ε = 37.12), and the Barataria–Terrebonne Estuary (NRMSE = 0.14, ε = 38.89). Using same-day PACE-OCI Level 2 Rrs, the model achieved comparable performance in Lake Erie (NRMSE = 0.19, ε = 55.19), Lake Pontchartrain (NRMSE = 0.14, ε = 51.39), and the Barataria–Terrebonne Estuary (NRMSE = 0.17, ε = 47.92). Hyper-MoE-VAE derived PACE-OCI hyperspectral aphy was further decomposed against mass-specific absorption spectra to estimate group-specific contributions to total chlorophyll a. The resulting PCC showed strong agreement with HPLC–CHEMTAX in Lake Erie ( R 2 = 0.692) and Gulf estuarine–coastal systems ( R 2 = 0.732). Monte Carlo noise experiments further revealed group-dependent sensitivities, with diatoms and dinoflagellates showing moderate susceptibility to noise, while cyanobacteria and cryptophytes exhibited narrow uncertainty distributions. These results demonstrate Hyper-MoE-VAE's capability for regional, operational water-quality monitoring with PACE-OCI and its adaptability to current and future hyperspectral missions.

Great Lakes, Lake Pontchartrain↗

USGS2026

A framework for integrating spatiotemporal deep learning methods with landsat for annual land cover and impervious surface mapping

Land cover information is essential for understanding Earth’s surface dynamics and how vegetation, water, soil, climate, and terrain interact. The National Land Cover Database (NLCD) has been the authoritative source for consistent U.S. land cover mapping. To extend NLCD’s temporal resolution and reduce production latency, we developed the Land Cover Artificial Mapping System (LCAMS)—a prototype spatiotemporal deep learning framework piloted as the foundation for the new Annual NLCD. LCAMS builds on concepts from legacy NLCD and the U.S. Geological Survey Land Change Monitoring, Assessment, and Projection (LCMAP) initiatives. It employs a loosely coupled two-stage architecture consisting of independent but functionally interdependent spatial and temporal models. Spatial models extract per-year information from Landsat data, while the temporal models refine the spatial outputs to enforce inter-annual consistency—critical for reliable land change monitoring. LCAMS produces annual 30 m resolution land cover and impervious surface outputs, with region-specific fine-tuning to generalize across diverse landscapes and temporal dynamics. Validation was conducted using an independent dataset of 1925 randomly sampled plots from five U.S. Landsat Analysis Ready Data (ARD) tiles spanning 1985-2021, selected for spatial and temporal variability. This dataset was used consistently to evaluate LCAMS, Legacy NLCD, and LCMAP. Using the NLCD legend, LCAMS achieved 72.1 ± 1.60% overall agreement, compared to 71.1 ± 1.7% agreement for Legacy NLCD. Using the LCMAP legend, LCAMS achieved 83.4 ± 1.22% agreement, compared to 84.6 ± 1.11% agreement for LCMAP. Overall, LCAMS delivers comparable accuracy while offering higher thematic resolution, longer temporal coverage, and automated production of annual 30 m CONUS land cover.

Remote Sensing of Environment↗

USGS2026

Advances in volcano monitoring driven by the first decade of Sentinel-1 observations

Sentinel-1 has transformed how satellite radar data (SAR and InSAR) are used in volcanology. The systematic, long-term archive and open-access policy means that volcano observatories and research organisations have invested in integrating Sentinel-1 datasets into their monitoring systems. We identify 233 high priority volcanoes and estimate that Sentinel-1 data has been used in peer-reviewed publications for 90 of them. We examine a global archive of 3.3 million automatically processed Sentinel-1 interferograms of volcanoes and use machine learning methods to identify eruptions and periods of unrest. We then review the ways in which InSAR data are being used in different contexts. At frequently erupting basaltic systems in Iceland, Hawaiʻi, the Galápagos , and Piton de la Fournaise, InSAR has become an effective monitoring tool and is integrated with other datasets and models to forecast magma pathways. For large explosive eruptions, deformation measurements often remain challenging, but SAR backscatter is increasingly used to map damaging flows and measure the changing shape of ocean islands. Sentinel-1's long archive provides critical baseline measurements that are vital for measuring slow deformation, capturing new periods of unrest and providing fresh insights into subsurface dynamics. Understanding the drivers of deformation remains challenging and typically relies on integration with external datasets. Future European Space Agency missions have the potential to improve both resolution and coverage providing an even richer dataset to further enhance global volcano monitoring

Erta Ale volcano↗

USGS2026

Mapping winter cover crops using remote sensing of crop phenology

Winter cover crops (WCC) are an effective agricultural conservation practice for improving soil health and water quality. Several U.S. states have established incentive programs to promote planting WCC between cash crop rotations. In recent years, the area planted with WCC has steadily increased due to these efforts, although the planting area varies among states. Since these programs are largely led by individual counties or state programs, there is no centralized data repository that quantifies the total WCC planted annually or their locations. Therefore, there is a practical need to identify fields that have been planted with WCC, as well as to quantify the total planted area both during and after the growing season. Furthermore, incentive programs often require verification that WCC were planted to release payments to participating growers in a timely manner. Satellite remote sensing time series can provide information on WCC growth status. However, it remains unclear what kind of WCC the remote sensing data detects. For example, separating WCC from weeds and winter commodity crops remains a challenge. Additionally, a binary classification of WCC presence and absence cannot accurately describe the category and performance of the WCC. This paper presents a new remote sensing phenology-based approach to map WCC likelihood. We assessed model predictions of WCC likelihood using WCC planting records and ground observations over fields at the Beltsville Agricultural Research Center (BARC) from 2018 to 2024. Regional WCC maps were assessed using the Maryland Department of Agriculture's records of WCC fields enrolled in the 2019 and 2020 incentive programs. Our results show that the balanced overall accuracy for WCC and non-WCC ranges from 75% to 85% but could exceed 90%, with limited training samples. However, the high accuracy of WCC corresponds with low accuracy for non-WCC due to confusion with weeds and perennial grasses. Early-season detection yields promising results, with overall accuracies comparable to late-season detection. The after-season detection can differentiate between winter cereal grain cash crops and incentive WCC based on the ending date of the growing season. With a small set of WCC samples and management information, the phenology-based mapping approach offers a rapid and scalable solution for mapping WCC over a large region during or after each WCC growing season.

Maryland↗

USGS2026

Basal area loss from fire using field-calibrated remote sensing refines western US fire severity measurements

The spatial patterns of fire effects and tree mortality have profound consequences for forest resilience. Cost-effective, medium-resolution, and spatiotemporally extensive fire severity measurements are essential for informing post-fire restoration and improving our understanding of wildfires—from forest stands to continents and from days to decades. Remote sensing advancements have improved burn severity mapping, but methods vary in interpretability, scalability, generalizability, and alignment with field measurements. One meaningful metric of fire effects on forests is proportion basal area loss, but existing methods are limited by a lack of region-specific field reference data and a scalable mapping framework. To address these issues, we compiled 3280 field reference plots from 123 fires in forests across the Western US to calculate the proportion of fire-induced basal area loss. We then used spatially cross-validated machine learning models with concurrent hyperparameter tuning to select a skillful, parsimonious model from a large candidate set of remotely-sensed, climatic, and topographic predictors. Spectral-only measures of severity over- or underestimated basal area loss in dry versus wet years and across aspects, demonstrating the value of incorporating climatic and topographic context. We also tested model performance on a separate holdout dataset in the Southwest US as a demonstration of reproducibility and transparency. We provide a Google Earth Engine tool for estimating proportional basal area loss for any fire perimeter in the Western US, enabling rapid map creation for land management and ecological modeling. All code, model parameters, and training data are released to support reproducibility, community adoption, regional refinement, and adaptation to new regions.

western United States↗

USGS2026

Enhancements to the USGS Landsat Level 2 surface temperature and emissivity product for Collection 3 reprocessing

The Landsat program provides the longest continuous global record of thermal infrared observations of the Earth's surface, underpinning critical applications in climate monitoring, water resources, ecosystem dynamics, urban heat analysis, and natural hazard assessment. The release of a global inventory of Landsat Collection 2 Level 2 surface temperature products by the U.S. Geological Survey (USGS) marked a major milestone in operational provision of Landsat thermal infrared analysis-ready data. Ongoing validations and community uses of Collection 2 have identified opportunities to further improve accuracy, uncertainty characterization, and emissivity correction across diverse atmospheric and surface conditions. In preparation for the planned Landsat Collection 3 reprocessing of the Landsat data record in the late 2020s, the USGS is implementing a coordinated set of enhancements to the Level 2 surface temperature products. These include revised emissivity estimation that leverages external datasets, improved atmospheric characterization and uncertainty propagation, expanded dynamic range for high temperature targets, consideration of split window atmospheric correction algorithm for Landsat 8 and 9, and decoupling of thermal infrared processing from visible to shortwave infrared constraints to enable surface temperature retrievals under low or no solar illumination conditions. These changes are designed to improve product quality and consistency across the Landsat record. Beyond near-term performance gains, the Collection 3 design establishes a scalable processing architecture to accommodate the expanded spectral and radiometric measurement capabilities of the forthcoming Landsat 10 mission. By preserving continuity across the Landsat 4–9 record while enabling future algorithm evolution, Landsat Collection 3 will provide a foundation for long-term, multi-decadal Earth system thermal infrared observations.

Remote Sensing of Environment↗

USGS2000

Classification methods for monitoring Arctic sea ice using OKEAN passive/active two-channel microwave data

This paper presents methods for classifying Arctic sea ice using both passive and active (2-channel) microwave imagery acquired by the Russian OKEAN 01 polar-orbiting satellite series. Methods and results are compared to sea ice classifications derived from nearly coincident Special Sensor Microwave Imager (SSM/I) and Advanced Very High Resolution Radiometer (AVHRR) image data of the Barents, Kara, and Laptev Seas. The Russian OKEAN 01 satellite data were collected over weekly intervals during October 1995 through December 1997. Methods are presented for calibrating, georeferencing and classifying the raw active radar and passive microwave OKEAN 01 data, and for correcting the OKEAN 01 microwave radiometer calibration wedge based on concurrent 37 GHz horizontal polarization SSM/I brightness temperature data. Sea ice type and ice concentration algorithms utilized OKEAN's two-channel radar and passive microwave data in a linear mixture model based on the measured values of brightness temperature and radar backscatter, together with a priori knowledge about the scattering parameters and natural emissivities of basic sea ice types. OKEAN 01 data and algorithms tended to classify lower concentrations of young or first-year sea ice when concentrations were less than 60%, and to produce higher concentrations of multi-year sea ice when concentrations were greater than 40%, when compared to estimates produced from SSM/I data. Overall, total sea ice concentration maps derived independently from OKEAN 01, SSM/I, and AVHRR satellite imagery were all highly correlated, with uniform biases, and mean differences in total ice concentration of less than four percent (sd<15%).

Remote Sensing of Environment↗

USGS2001

Thematic accuracy of MRLC land cover for the eastern United States

One objective of the MultiResolution Land Characteristics (MRLC) consortium is to map general land-cover categories for the conterminous United States using Landsat Thematic Mapper (TM) data. Land-cover mapping and classification accuracy assessment are complete for the eastern United States. The accuracy assessment was based on photo-interpreted reference data obtained from a stratified probability sample of pixels. Agreement was defined as a match between primary or alternate reference land-cover labels assigned to each sample pixel and the mode (most common class) of the map's land-cover labels within a 3×3-pixel neighborhood surrounding the sampled point. At 30-m resolution, overall accuracy was 59.7% at an Anderson Level II thematic detail, and 80.5% at Anderson Level I.

Remote Sensing of Environment↗

USGS2001

Effects of Landsat 5 Thematic Mapper and Landsat 7 Enhanced Thematic Mapper plus radiometric and geometric calibrations and corrections on landscape characterization

The Thematic Mapper (TM) instruments onboard Landsats 4 and 5 provide high-quality imagery appropriate for many different applications, including land cover mapping, landscape ecology, and change detection. Precise calibration was considered to be critical to the success of the Landsat 7 mission and, thus, issues of calibration were given high priority during the development of the Enhanced Thematic Mapper Plus (ETM+). Data sets from the Landsat 5 TM are not routinely corrected for a number of radiometric and geometric artifacts, including memory effect, gain/bias, and interfocal plane misalignment. In the current investigation, the effects of correcting vs. not correcting these factors were investigated for several applications. Gain/bias calibrations were found to have a greater impact on most applications than did memory effect calibrations. Correcting interfocal plane offsets was found to have a moderate effect on applications. On June 2, 1999, Landsats 5 and 7 data were acquired nearly simultaneously over a study site in the Niobrara, NE area. Field radiometer data acquired at that site were used to facilitate crosscalibrations of Landsats 5 and 7 data. Current findings and results from previous investigations indicate that the internal calibrator of Landsat 5 TM tracked instrument gain well until 1988. After this, the internal calibrator diverged from the data derived from vicarious calibrations. Results from this study also indicate very good agreement between prelaunch measurements and vicarious calibration data for all Landsat 7 reflective bands except Band 4. Values are within about 3.5% of each other, except for Band 4, which differs by 10%. Coefficient of variation (CV) values derived from selected targets in the imagery were also analyzed. The Niobrara Landsat 7 imagery was found to have lower CV values than Landsat 5 data, implying that lower levels of noise characterize Landsat 7 data than current Landsat 5 data. It was also found that following radiometric normalization, the Normalized Difference Vegetation Index (NDVI) imagery and classification products of Landsats 5 and 7 were very similar. This implies that data from the two sensors can be used to measure and monitor the same landscape phenomena and that Landsats 5 and 7 data can be used interchangeably with proper caution. In addition, it was found that difference imagery produced using Landsat 7 ETM+ data are of excellent quality.

Remote Sensing of Environment↗

USGS2001

Cloud characterization and clear-sky correction from Landsat-7

Landsat, with its wide swath and high resolution, fills an important mesoscale gap between atmospheric variations seen on a few kilometer scale by local surface instrumentation and the global view of coarser resolution satellites such as MODIS. In this important scale range, Landsat reveals radiative effects on the few hundred-meter scale of common photon mean-free-paths, typical of scattering in clouds at conservative (visible) wavelengths, and even shorter mean-free-paths of absorptive (near-infrared) wavelengths. Landsat also reveals shadowing effects caused by both cloud and vegetation that impact both cloudy and clear-sky radiances. As a result, Landsat has been useful in development of new cloud retrieval methods and new aerosol and surface retrievals that account for photon diffusion and shadowing effects. This paper discusses two new cloud retrieval methods: the nonlocal independent pixel approximation (NIPA) and the normalized difference nadir radiance method (NDNR). We illustrate the improvements in cloud property retrieval enabled by the new low gain settings of Landsat-7 and difficulties found at high gains. Then, we review the recently developed “path radiance” method of aerosol retrieval and clear-sky correction using data from the Department of Energy Atmospheric Radiation Measurement (ARM) site in Oklahoma. Nearby clouds change the solar radiation incident on the surface and atmosphere due to indirect illumination from cloud sides. As a result, if clouds are nearby, this extra side-illumination causes clear pixels to appear brighter, which can be mistaken for extra aerosol or higher surface albedo. Thus, cloud properties must be known in order to derive accurate aerosol and surface properties. A three-dimensional (3D) Monte Carlo (MC) radiative transfer simulation illustrates this point and suggests a method to subtract the cloud effect from aerosol and surface retrievals. The main conclusion is that cloud, aerosol, and surface retrievals are linked and must be treated as a combined system. Landsat provides the range of scales necessary to observe the 3D cloud radiative effects that influence joint surface-atmospheric retrievals.

Remote Sensing of Environment↗

USGS2001

The Landsat 7 mission: terrestrial research and applications for the 21st century

The Landsat Earth observation approach introduced in 1972 created a new way of monitoring land cover and land use globally. The Landsat 7 mission, successfully launched on April 15, 1999, continues those observations and demonstrates significant progress in precise numerical radiometry, spectral differentiation, and seasonally repetitive monitoring. Substantial improvements in calibration procedures, both prior to launch and during normal operations, have also been made to ensure long-term stability in the acquired spectral radiometry. Landsat 7 data acquisitions are being driven by a long-term data acquisition plan that was designed to ensure that substantially cloud-free, seasonal coverage would be recorded and archived in the US for all land areas of the globe. NASA competitively selected a Landsat Science Team, consisting of representatives from US universities and government agencies, to exploit the Landsat 7 record for global change research. This team is addressing the technical and analytical means to process and analyze the core of this observation record, and for the first time in the history of the Landsat mission, the technical and operational aspects of the mission are being driven by the goals of the US science community. The expected outcome of these efforts is a rapid improvement in understanding the Earth system, as well as conceptual knowledge that will underpin significant advancements in the application of this technology for commercial, operational, educational, and research purposes. Pathways to achieve effective Landsat continuity in the early decades of the 21st century are also being given careful attention, and there is no question that the lessons learned from the Landsat 7 mission will strongly influence these next-generation sensor systems.

Remote Sensing of Environment↗

USGS2002

Impact of sensor's point spread function on land cover characterization: Assessment and deconvolution

Measured and modeled point spread functions (PSF) of sensor systems indicate that a significant portion of the recorded signal of each pixel of a satellite image originates from outside the area represented by that pixel. This hinders the ability to derive surface information from satellite images on a per-pixel basis. In this study, the impact of the PSF of the Moderate Resolution Imaging Spectroradiometer (MODIS) 250 m bands was assessed using four images representing different landscapes. Experimental results showed that though differences between pixels derived with and without PSF effects were small on the average, the PSF generally brightened dark objects and darkened bright objects. This impact of the PSF lowered the performance of a support vector machine (SVM) classifier by 5.4% in overall accuracy and increased the overall root mean square error (RMSE) by 2.4% in estimating subpixel percent land cover. An inversion method based on the known PSF model reduced the signals originating from surrounding areas by as much as 53%. This method differs from traditional PSF inversion deconvolution methods in that the PSF was adjusted with lower weighting factors for signals originating from neighboring pixels than those specified by the PSF model. By using this deconvolution method, the lost classification accuracy due to residual impact of PSF effects was reduced to only 1.66% in overall accuracy. The increase in the RMSE of estimated subpixel land cover proportions due to the residual impact of PSF effects was reduced to 0.64%. Spatial aggregation also effectively reduced the errors in estimated land cover proportion images. About 50% of the estimation errors were removed after applying the deconvolution method and aggregating derived proportion images to twice their dimensional pixel size. ?? 2002 Elsevier Science Inc. All rights reserved.

Remote Sensing of Environment↗

USGS2002

Seasonal comparisons of sea ice concentration estimates derived from SSM/I, OKEAN, and RADARSAT data

The Special Sensor Microwave Imager (SSM/I) microwave satellite radiometer and its predecessor SMMR are primary sources of information for global sea ice and climate studies. However, comparisons of SSM/I, Landsat, AVHRR, and ERS-1 synthetic aperture radar (SAR) have shown substantial seasonal and regional differences in their estimates of sea ice concentration. To evaluate these differences, we compared SSM/I estimates of sea ice coverage derived with the NASA Team and Bootstrap algorithms to estimates made using RADARSAT, and OKEAN-01 satellite sensor data. The study area included the Barents Sea, Kara Sea, Laptev Sea, and adjacent parts of the Arctic Ocean, during October 1995 through October 1999. Ice concentration estimates from spatially and temporally near-coincident imagery were calculated using independent algorithms for each sensor type. The OKEAN algorithm implemented the satellite's two-channel active (radar) and passive microwave data in a linear mixture model based on the measured values of brightness temperature and radar backscatter. The RADARSAT algorithm utilized a segmentation approach of the measured radar backscatter, and the SSM/I ice concentrations were derived at National Snow and Ice Data Center (NSIDC) using the NASA Team and Bootstrap algorithms. Seasonal and monthly differences between SSM/I, OKEAN, and RADARSAT ice concentrations were calculated and compared. Overall, total sea ice concentration estimates derived independently from near-coincident RADARSAT, OKEAN-01, and SSM/I satellite imagery demonstrated mean differences of less than 5.5% (S.D.<9.5%) during the winter period. Differences between the SSM/I NASA Team and the SSM/I Bootstrap concentrations were no more than 3.1% (S.D.<5.4%) during this period. RADARSAT and OKEAN-01 data both yielded higher total ice concentrations than the NASA Team and the Bootstrap algorithms. The Bootstrap algorithm yielded higher total ice concentrations than the NASA Team algorithm. Total ice concentrations derived from OKEAN-01 and SSM/I satellite imagery were highly correlated during winter, spring, and fall, with mean differences of less than 8.1% (S.D.<15%) for the NASA Team algorithm, and less than 2.8% (S.D.<13.8%) for the Bootstrap algorithm. Respective differences between SSM/I NASA Team and SSM/I Bootstrap total concentrations were less than 5.3% (S.D.<6.9%). Monthly mean differences between SSM/I and OKEAN differed annually by less than 6%, with smaller differences primarily in winter. The NASA Team and Bootstrap algorithms underestimated the total sea ice concentrations relative to the RADARSAT ScanSAR no more than 3.0% (S.D.<9%) and 1.2% (S.D.<7.5%) during cold months, and no more than 12% and 7% during summer, respectively. ScanSAR tended to estimate higher ice concentrations for ice concentrations greater than 50%, when compared to SSM/I during all months. ScanSAR underestimated total sea ice concentration by 2% compared to the OKEAN-01 algorithm during cold months, and gave an overestimation by 2% during spring and summer months. Total NASA Team and Bootstrap sea ice concentration estimates derived from coincident SSM/I and OKEAN-01 data demonstrated mean differences of no more than 5.3% (S.D.<7%), 3.1% (S.D.<5.5%), 2.0% (S.D.<5.5%), and 7.3% (S.D.<10%) for fall, winter, spring, and summer periods, respectively. Large disagreements were observed between the OKEAN and NASA Team results in spring and summer for estimates of the first-year (FY) and multiyear (MY) age classes. The OKEAN-01 algorithm and data tended to estimate, on average, lower concentrations of young or FY ice and higher concentrations of total and MY ice for all months and seasons. Our results contribute to the growing body of documentation about the levels of disparity obtained when seasonal sea ice concentrations are estimated using various types of satellite data and algorithms.

Remote Sensing of Environment↗

USGS2002

Achieving sub-pixel geolocation accuracy in support of MODIS land science

The Moderate Resolution Imaging Spectroradiometer (MODIS) was launched in December 1999 on the polar orbiting Terra spacecraft and since February 2000 has been acquiring daily global data in 36 spectral bands—29 with 1 km, five with 500 m, and two with 250 m nadir pixel dimensions. The Terra satellite has on-board exterior orientation (position and attitude) measurement systems designed to enable geolocation of MODIS data to approximately 150 m (1 σ ) at nadir. A global network of ground control points is being used to determine biases and trends in the sensor orientation. Biases have been removed by updating models of the spacecraft and instrument orientation in the MODIS geolocation software several times since launch and have improved the MODIS geolocation to approximately 50 m (1 σ ) at nadir. This paper overviews the geolocation approach, summarizes the first year of geolocation analysis, and overviews future work. The approach allows an operational characterization of the MODIS geolocation errors and enables individual MODIS observations to be geolocated to the sub-pixel accuracies required for terrestrial global change applications.

Remote Sensing of Environment↗

USGS2003

Effects of landscape characteristics on land-cover class accuracy

The effects of patch size and land-cover heterogeneity on classification accuracy were evaluated using reference data collected for the National Land-Cover Data (NLCD) set accuracy assessment. Logistic regression models quantified the relationship between classification accuracy and these landscape variables for each land-cover class at both the Anderson Levels I and II classification schemes employed in the NLCD. The general relationships were consistent, with the odds of correctly classifying a pixel increasing as patch size increased and decreasing as heterogeneity increased. Specific characteristics of these relationships, however, showed considerable diversity among the various classes. Odds ratios are reported to document these relationships. Interaction between the two landscape variables was not a significant influence on classification accuracy, indicating that the effect of heterogeneity was not impacted by the sample being in a small or large patch. Landscape variables remained significant predictors of class-specific accuracy even when adjusted for regional differences in the mapping and assessment processes or landscape characteristics. The land-cover class-specific analyses provide insight into sources of classification error and a capacity for predicting error based on a pixel's mapped land-cover class, patch size and surrounding land-cover heterogeneity.

Remote Sensing of Environment↗

USGS2003

Lithologic mapping in the Mountain Pass, California area using Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) data

Evaluation of an Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) image of the Mountain Pass, California area indicates that several important lithologic groups can be mapped in areas with good exposure by using spectral-matching techniques. The three visible and six near-infrared bands, which have 15-m and 30-m resolution, respectively, were calibrated by using in situ measurements of spectral reflectance. Calcitic rocks were distinguished from dolomitic rocks by using matched-filter processing in which image spectra were used as references for selected spectral categories. Skarn deposits and associated bright coarse marble were mapped in contact metamorphic zones related to intrusion of Mesozoic and Tertiary granodioritic rocks. Fe-muscovite, which is common in these intrusive rocks, was distinguished from Al-muscovite present in granitic gneisses and Mesozoic granite. Quartzose rocks were readily discriminated, and carbonate rocks were mapped as a single broad unit through analysis of the 90-m resolution, five-band surface emissivity data, which is produced as a standard product at the EROS Data Center. Three additional classes resulting from spectral-angle mapper processing ranged from (1) a broad granitic rock class (2) to predominately granodioritic rocks and (3) a more mafic class consisting mainly of mafic gneiss, amphibolite and variable mixtures of carbonate rocks and silicate rocks. ?? 2002 Elsevier Science Inc. All rights reserved.

Remote Sensing of Environment↗

USGS2003

Mapping mine wastes and analyzing areas affected by selenium-rich water runoff in southeast Idaho using AVIRIS imagery and digital elevation data

Remotely sensed hyperspectral and digital elevation data from southeastern Idaho are combined in a new method to assess mine waste contamination. Waste rock from phosphorite mining in the area contains selenium, cadmium, vanadium, and other metals. Toxic concentrations of selenium have been found in plants and soils near some mine waste dumps. Eighteen mine waste dumps and five vegetation cover types in the southeast Idaho phosphate district were mapped by using Airborne Visible-Infrared Imaging Spectrometer (AVIRIS) imagery and field data. The interaction of surface water runoff with mine waste was assessed by registering the AVIRIS results to digital elevation data, enabling determinations of (1) mine dump morphologies, (2) catchment watershed areas above each mine dump, (3) flow directions from the dumps, (4) stream gradients, and (5) the extent of downstream wetlands available for selenium absorption. Watersheds with the most severe selenium contamination, such as the South Maybe Canyon watershed, are associated with mine dumps that have large catchment watershed areas, high stream gradients, a paucity of downstream wetlands, and dump forms that tend to obstruct stream flow. Watersheds associated with low concentrations of dissolved selenium, such as Angus Creek, have mine dumps with small catchment watershed areas, low stream gradients, abundant wetlands vegetation, and less obstructing dump morphologies. ?? 2002 Elsevier Science Inc. All rights reserved.

Remote Sensing of Environment↗

USGS2003

Mapping vegetation in Yellowstone National Park using spectral feature analysis of AVIRIS data

Knowledge of the distribution of vegetation on the landscape can be used to investigate ecosystem functioning. The sizes and movements of animal populations can be linked to resources provided by different plant species. This paper demonstrates the application of imaging spectroscopy to the study of vegetation in Yellowstone National Park (Yellowstone) using spectral feature analysis of data from the Airborne Visible/Infrared Imaging Spectrometer (AVIRIS). AVIRIS data, acquired on August 7, 1996, were calibrated to surface reflectance using a radiative transfer model and field reflectance measurements of a ground calibration site. A spectral library of canopy reflectance signatures was created by averaging pixels of the calibrated AVIRIS data over areas of known forest and nonforest vegetation cover types in Yellowstone. Using continuum removal and least squares fitting algorithms in the US Geological Survey's Tetracorder expert system, the distributions of these vegetation types were determined by comparing the absorption features of vegetation in the spectral library with the spectra from the AVIRIS data. The 0.68 μm chlorophyll absorption feature and leaf water absorption features, centered near 0.98 and 1.20 μm, were analyzed. Nonforest cover types of sagebrush, grasslands, willows, sedges, and other wetland vegetation were mapped in the Lamar Valley of Yellowstone. Conifer cover types of lodgepole pine, whitebark pine, Douglas fir, and mixed Engelmann spruce/subalpine fir forests were spectrally discriminated and their distributions mapped in the AVIRIS images. In the Mount Washburn area of Yellowstone, a comparison of the AVIRIS map of forest cover types to a map derived from air photos resulted in an overall agreement of 74.1% (kappa statistic=0.62).

Wyoming↗