The U.S. Geological Survey collected low-altitude (typically 200-350 ft als) airborne thermal infrared, multispectral, and visual imagery data via a multirotor, small unoccupied aircraft system deployed along beaver-impacted sections of the East River and Coal Creek stream corridors, near the town of Crested Butte, CO. Visual imagery was collected in jpg format, and the images were compiled automatically into a larger stitched image (orthomosaic). Structure from Motion techniques were also applied to the visual imagery to derive time-specific digital surface models (DSM). Thermal infrared still images were collected in jpg and radiometric tiff formats, while multispectral data were collected in tif format. Although not done yet here, multispectral and thermal data can be compiled into orthomosaics and DSMs in a similar manner to visible light imagery.

The U.S. Geological Survey collected low-altitude (typically 200-350 ft als) airborne thermal infrared, multispectral, and visual imagery data via a multirotor, small unoccupied aircraft system deployed along beaver-impacted sections of the East River and Coal Creek stream corridors, near the town of Crested Butte, CO. Visual imagery was collected in jpg format, and the images were compiled automatically into a larger stitched image (orthomosaic). Structure from Motion techniques were also applied to the visual imagery to derive time-specific digital surface models (DSM). Thermal infrared still images were collected in jpg and radiometric tiff formats, while multispectral data were collected in tif format. Although not done yet here, multispectral and thermal data can be compiled into orthomosaics and DSMs in a similar manner to visible light imagery.

Small unoccupied aircraft systems (UAS) are now often used for collecting aerial visible image data and creating 3D digital surface models (DSM) that incorporate terrain and dense vegetation. Lightweight thermal sensors provide another sensor option for generation of sub meter resolution aerial thermal infrared orthophotos that can be used to infer hydrogeological processes. UAS-based sensors allow for the rapid and safe survey of groundwater discharge areas, often present in inaccessible, boggy, and/or dangerous terrain. Visible light and thermal infrared image data were collected March 2018 and March 2019, respectively, at Tidmarsh Farms, a former commercial cranberry bog located in coastal Massachusetts, USA (41°54'17.6"N 70°34'17.4"W), where a comprehensive stream and wetland restoration was performed. Wetland restoration actions at Tidmarsh Farms were made possible by a landowner decision to enroll in the U.S. Department of Agriculture (USDA) Natural Resources Conservation Service (NRCS) Wetland Reserve Easement Program. The Massachusetts Department of Fish and Game’s Division of Ecological Restoration (MDER) later became the lead manager for the design, permitting, and implementation of stream and wetland restoration actions on the site. In 2017, after the completion of the largest freshwater wetland restoration in Massachusetts to date, the property was purchased by the Massachusetts Audubon Society who in 2018 opened the Tidmarsh Wildlife Sanctuary to the public.

This dataset presents high-resolution thermal infrared (TIR) and true-color RGB (red, green, blue) imagery mosaics, longitudinal stream temperature profiles, thermal points of interest, and river centerlines from airborne surveys of the Quillayute, Dickey, Sol Duc, Calawah, South Fork Calawah, Sitkum, and Bogachiel Rivers (203 river kilometers; 126 river miles total). All datasets were produced and initially processed by NV5 Geospatial (NV5). The U.S. Geological Survey (USGS) performed additional processing of the longitudinal stream temperature profiles and thermal points of interest, as described below. TIR and RGB images were acquired by NV5 on August 29-31, 2022, using a FLIR SC6000 LWIR sensor and a Sony Alpha 7R III camera mounted in a fiberglass enclosure to a Bell 206 Long Range helicopter. Images were acquired during afternoon hours to maximize the thermal contrast between the river water and the banks. At a flying altitude of 350-450 m (1,148-1,476 ft) above ground level, the FLIR SC6000 and Sony Alpha 7R III achieved ground sampling distances of less than 50 cm (20 in.) and 10 cm (4 in.), respectively. TIR imagery mosaics (.tif) for individual surveys and a single RGB imagery mosaic (.sid) for the entire study area were developed by NV5, and river centerlines (.shp) were manually digitized by NV5 using the imagery mosaics as guides. Points were then generated by NV5 every 50 m (164 ft) along the centerlines to quantify the longitudinal stream temperature profiles (LTPs; .shp). Summary statistics, in degrees Celsius, were computed by NV5 for each point in the profile by sampling pixel values of water temperature along the centerline in the corresponding TIR mosaic within a 2-m (6.56 ft) radius buffer around each point. The statistical information was used by USGS to identify sampling points that fall on non-water features such as boulders or bridges, and then filter these points from the final dataset. LTPs assist in identifying the water temperature gradient in the river and changes in the gradient due to the potential influence of thermal exchange processes, such as water inflows (tributaries, lateral groundwater flow, hyporheic flow, etc.) or increased heating from a low percentage of effective riparian shading. These profiles are also an important component of models that estimate water temperature based on climate and land use scenarios. Thermal points of interest (POIs; .shp) were manually identified by NV5 and USGS across the channel, riparian zone, and floodplain. Such features include cold-water anomalies that may represent thermal refuges and serve as salmonid habitat. POIs were classified by USGS as one of four types: (1) tributary; (2) lateral groundwater / side channel / small tributary; (3) hyporheic / diffuse groundwater; or (4) point source effluent. Summary statistics were computed by USGS for each POI using a sample of water temperature values from pixels in the corresponding TIR mosaic within a 0.6-m (1.97 ft) radius buffer around each point. The automated sampling of the POIs included pixels that are not purely water, but instead mixed with other in-stream and riparian features, such as boulders, woody debris, and tree canopy. Therefore, the water temperatures reported for POIs where the 0.6-m radius sampling area contains mixed pixels are often skewed. The POI temperatures should thus serve as indicators where thermal heterogeneity requires additional investigation and potentially more precise quantification. All data is projected in UTM 10N and the horizontal datum is NAD83(2011).

This dataset presents high-resolution thermal infrared (TIR) and true-color RGB (red, green, blue) imagery mosaics, longitudinal stream temperature profiles, thermal points of interest, and river centerlines from airborne surveys of the Quillayute, Dickey, Sol Duc, Calawah, South Fork Calawah, Sitkum, and Bogachiel Rivers (203 river kilometers; 126 river miles total). All datasets were produced and initially processed by NV5 Geospatial (NV5). The U.S. Geological Survey (USGS) performed additional processing of the longitudinal stream temperature profiles and thermal points of interest, as described below. TIR and RGB images were acquired by NV5 on August 29-31, 2022, using a FLIR SC6000 LWIR sensor and a Sony Alpha 7R III camera mounted in a fiberglass enclosure to a Bell 206 Long Range helicopter. Images were acquired during afternoon hours to maximize the thermal contrast between the river water and the banks. At a flying altitude of 350-450 m (1,148-1,476 ft) above ground level, the FLIR SC6000 and Sony Alpha 7R III achieved ground sampling distances of less than 50 cm (20 in.) and 10 cm (4 in.), respectively. TIR imagery mosaics (.tif) for individual surveys and a single RGB imagery mosaic (.sid) for the entire study area were developed by NV5, and river centerlines (.shp) were manually digitized by NV5 using the imagery mosaics as guides. Points were then generated by NV5 every 50 m (164 ft) along the centerlines to quantify the longitudinal stream temperature profiles (LTPs; .shp). Summary statistics, in degrees Celsius, were computed by NV5 for each point in the profile by sampling pixel values of water temperature along the centerline in the corresponding TIR mosaic within a 2-m (6.56 ft) radius buffer around each point. The statistical information was used by USGS to identify sampling points that fall on non-water features such as boulders or bridges, and then filter these points from the final dataset. LTPs assist in identifying the water temperature gradient in the river and changes in the gradient due to the potential influence of thermal exchange processes, such as water inflows (tributaries, lateral groundwater flow, hyporheic flow, etc.) or increased heating from a low percentage of effective riparian shading. These profiles are also an important component of models that estimate water temperature based on climate and land use scenarios. Thermal points of interest (POIs; .shp) were manually identified by NV5 and USGS across the channel, riparian zone, and floodplain. Such features include cold-water anomalies that may represent thermal refuges and serve as salmonid habitat. POIs were classified by USGS as one of four types: (1) tributary; (2) lateral groundwater / side channel / small tributary; (3) hyporheic / diffuse groundwater; or (4) point source effluent. Summary statistics were computed by USGS for each POI using a sample of water temperature values from pixels in the corresponding TIR mosaic within a 0.6-m (1.97 ft) radius buffer around each point. The automated sampling of the POIs included pixels that are not purely water, but instead mixed with other in-stream and riparian features, such as boulders, woody debris, and tree canopy. Therefore, the water temperatures reported for POIs where the 0.6-m radius sampling area contains mixed pixels are often skewed. The POI temperatures should thus serve as indicators where thermal heterogeneity requires additional investigation and potentially more precise quantification. All data is projected in UTM 10N and the horizontal datum is NAD83(2011).

The U.S. Geological Survey collected low-altitude airborne thermal infrared data and visual imagery via a multirotor, small unoccupied aircraft system deployed from the northern bank of Oh-be-joyful Creek and adjacent to the Peeler fault, approximately 6 kilometers northwest of the town of Crested Butte, in Gunnison National Forest, Colorado, on August 17, 2017. Thermal infrared still images were collected in jpg and radiometric tiff formats, and non-radiometric thermal infrared video was collected. The radiometric thermal infrared still images were compiled automatically into a larger stitched image (orthomosaic). Visual imagery was collected in jpg format, and the images were compiled automatically into a larger stitched image (orthomosaic). Structure from Motion techniques were applied to the visual imagery to derive a time-specific digital elevation model (DEM).

Small unoccupied aircraft systems (UAS) are now often used for collecting aerial visible image data and creating 3D digital surface models (DSM) that incorporate terrain and dense vegetation. Lightweight thermal sensors provide another sensor option for generation of sub meter resolution aerial thermal infrared orthophotos that can be used to infer hydrogeological processes. UAS-based sensors allow for the rapid and safe survey of groundwater discharge areas, often present in inaccessible, boggy, and/or dangerous terrain. Visible light and thermal infrared image data were collected March 2018 and March 2019, respectively, at Tidmarsh Farms, a former commercial cranberry bog located in coastal Massachusetts, USA (41°54'17.6"N 70°34'17.4"W), where a comprehensive stream and wetland restoration was performed. Wetland restoration actions at Tidmarsh Farms were made possible by a landowner decision to enroll in the U.S. Department of Agriculture (USDA) Natural Resources Conservation Service (NRCS) Wetland Reserve Easement Program. The Massachusetts Department of Fish and Game’s Division of Ecological Restoration (MDER) later became the lead manager for the design, permitting, and implementation of stream and wetland restoration actions on the site. In 2017, after the completion of the largest freshwater wetland restoration in Massachusetts to date, the property was purchased by the Massachusetts Audubon Society who in 2018 opened the Tidmarsh Wildlife Sanctuary to the public.

The U.S. Geological Survey collected low-altitude airborne thermal infrared data and visual imagery via a multirotor, small unoccupied aircraft system deployed from the northern bank of Oh-be-joyful Creek and adjacent to the Peeler fault, approximately 6 kilometers northwest of the town of Crested Butte, in Gunnison National Forest, Colorado, on August 17, 2017. Thermal infrared still images were collected in jpg and radiometric tiff formats, and non-radiometric thermal infrared video was collected. The radiometric thermal infrared still images were compiled automatically into a larger stitched image (orthomosaic). Visual imagery was collected in jpg format, and the images were compiled automatically into a larger stitched image (orthomosaic). Structure from Motion techniques were applied to the visual imagery to derive a time-specific digital elevation model (DEM).

This page provides access to infrared, multispectral, visual image data, and derivative products collected along a beaver-impacted section of the East River from August 12-17, 2017 and July 28-August 2, 2018. This page may be updated in the future with additional data and analysis related to drone-based imaging.

The University of Connecticut and the U.S. Geological Survey (USGS) collected low-altitude (30-50 m above ground level) airborne visible-light imagery data via a quadcopter, small unoccupied aircraft system (UAS or ‘drone’) deployed along two tributary confluence locations within the Housatonic River: Mill Brook (latitude: 42°52’18” N, longitude: 73°21’48” W) and Furnace Brook (latitude: 41°49’16” N, longitude: 73°22’17” W). Both tributary confluence sites serve as critical summer thermal refuge for cold water-adapted poikilotherms. The objectives for this data collection included the creation of high-resolution orthomosaic images of the two tributary confluences to infer bank and instream structures and mixing processes at the tributary confluences. Detailed site-scale maps such as these are important tools for managers and researchers aiming to protect and conserve populations at risk. The UAS (Mavic 2 Zoom, DJI Enterprises) was flown several times per day, at wind speeds below 10 mph, capturing RGB imagery from March 24-25, 2021. The UAV flights collected single RGB JPG images at 30-50m above ground level using the double-grid flight pattern on the third-party app Pix4D Capture (https://www.pix4d.com/product/pix4dcapture). The images were stitched automatically into several orthomosaic images using Agisoft Metashape (Agisoft LLC, St. Petersburg, Russia) software as described in the ‘processed_data’ subfolders of this data release. Structure from Motion techniques were also applied to the visual imagery to derive time-specific, digital surface models (DSM) of the exposed banks and some subsurface features.