Atmospheric Remote Sensing: Instruments

FLOE Field Programs


Arctic Stratification 2014 phytoplankton

In July 2014, FLOE was installed in a NOAA Twin Otter to make the first synoptic measurements of subsurface phytoplankton layers associated with the retreating ice in the Arctic Ocean. We made a series of flights out of Barrow, Alaska over the Chukchi Sea to the west or the Beaufort Sea to the East. Our objectives were to see if thin plankton layers exist below the surface in the Arctic and how they might be affected by retreating ice. We have observed layers associated with stratification in other parts of the ocean [Churnside and Donaghay, 2009], and expected to find them in the Marginal Ice Zone as well. The lidar was able to measure the layer characteristics and the fractional ice cover simultaneously. These layers, which are not adequately captured in satellite data, will influence primary productivity, secondary productivity, fisheries recruitment, and carbon export to the benthos.

Churnside, J. H., and P. L. Donaghay, Thin scattering layers observed by airborne lidar, ICES Journal of Marine Science, 66(4), 778-789, doi:10.1093/icesjms/fsp029, 2009.

aboard NOAA Twin Otter
Jim Churnside monitoring lidar data in the front of the cabin in a NOAA Twin Otter. The lidar is visible in the rear of the cabin. The tank on the right is an auxiliary fuel tank to extend the aircraft range.
Arctic sea ice
Arctic sea ice out the window of the NOAA Twin Otter. Photos: Richard Marchbanks, NOAA / CIRES

Gulf of Mexico 2011 small pelagics and zooplankton

In late September and early October 2011, FLOE was installed on a small (King Air 90) twin-engine aircraft, deployed to Stennis International Airport, Mississippi. FLOE flew in coordination with the R/V McArthur II to survey epipelagic organisms (juvenile and adult small pelagic fish and gelatinous zooplankton) in the surface waters of the northern Gulf of Mexico potentially affected by the Deepwater Horizon oil spill and in surrounding areas. The oceanographic fish LIDAR was able to penetrate to > 30 m in offshore waters and 20 - 30 m on the shelf, except for the Mississippi River plume, where penetration was < 20 m. Few dense schools were detected, and those were generally found on the shelf. Several of these were positively identified as aggregations of moon jellies (Aurelia sp.). Large numbers of single targets were detected, especially off the shelf. Generally, more schools and single targets were detected at night than during the day, suggesting diurnal migration. Other features, including large layers and plumes were also observed. The layers are probably phytoplankton, but some of the plume structures might be oil from seeps. Comparison with the data from the surface vessel will be used in the final analysis to confirm the identity of the features detected by FLOE.

Chesapeake Bay menhaden

In a partnership with the Maryland Department of Natural Resources, we have been developing a technique to rapidly survey menhaden in Chespeake Bay. Airborne lidar and video techniques are both being considered. The lidar has much better depth penetration in the turbid waters of the bay, but High-Definition video has a much wider swath. As part of the lidar investigation, we measured the target strength of menhaden by comparison with standard calibration targets in a tank.

map of bay showing fish distribution
Map of menhaden distribution in Chespeake Bay.
fish school
Menhaden school.
calibration apparatus
Calibration targets and menhaden in tank.

Oregon and Washington sardine

Scientists use lasers and vessels to understand coastal pelagic ecosystem. How scientists are using advanced technologies such as lasers on planes as well as research vessels to determine abundance of commercially-important fish species in the Columbia River plume. Video: NOAA Northwest Fisheries Science Center

A National Ocean Partnership Program study has been investigating combinations of LIDAR, acoustics, and other data sources to determine the distribution of sardines and the relationships with environmental parameters. Partners include the NOAA Ocean Exploration and Research, Northwest Fisheries Science Center, the University of Washington, Oregon State University, and Flying Fish, LTD. The overall approach is to combine the data from instruments that have different space and time scales to get a more complete picture than any one a instrument alone can provide. The schematic below illustrates this concept. This effort builds on previous work in the area in collaboration with the NOAA Southwest Fisheries Science Center.

J. H. Churnside and J. J. Wilson, Power spectrum and fractal dimension of laser backscattering from the ocean, Journal of the Optical Society of America A, 23(11), 2829-2833, doi:10.1364/JOSAA.23.002829, 2006.

J. H. Churnside, Polarization effects on oceanographic lidar, Optics Express, 16(2), 1196-1207, doi:10.1364/OE.16.001196, 2008.

survey drawing includes lidar on an aircraft, acoustics on a ship, and bottom-moored acoustics
Schematic of survey combining an aircraft, a ship, and fixed acoustic moorings.
power spectra of lidar return showing power law with exponent of 1.5
Power spectra of lidar return at 2 different depths. Except at the surface, the return had a power-law spectrum with an exponent of -1.5, suggesting a fractal process over a large reange of spatial scales.
depolarization of the lidar return as a function of attenuation
Relationship between lidar attenuation and depolarization for nearshore waters (green) and off shore (blue).

Bering Sea forage fish

The lidar was one of a number of instruments studying survey strategies to assess Bering Sea forage species. Lack of information on forage species composition, distribution, and movements hinders our understanding of their ecological role in the Bering Sea. Recognizing the need for development of forage species survey strategies, this study characterized forage species in the slope, shelf, and nearshore regions of the Bering Sea using direct (midwater trawl, MultiNet, beach seine, jig, ROV) and indirect (acoustics, LIght Detection And Ranging (LIDAR)) sampling technologies. Forage species distribution and quantity differed between shelf (6-100 m) and slope (6-100 m, 100-300 m, 300 m-bottom) regions. Acoustics suggest that shallow and deep layers contained dispersed backscatter while the middle layer contained patchy schools. LIDAR and visual measurements from aircraft documented a patchy distribution of surface plankton and densely shoaling fish that exhibited a high degree of temporal and spatial variability within the 10 d study period. In the nearshore, Pacific sand lance dominated catches and other commonly captured forage fish were YOY Pacific sandfish and gadids. Zooplankton density in the upper 100 m of the water column was significantly higher in nearshore waters. Though copepods were the most abundant taxa, euphausiids, second most abundant, provided more energy to predators due to their large size. We identified several potential candidate species/groups for assessment with acoustics and direct sampling. Other potential, near-surface species/groups could be surveyed with LIDAR and direct sampling. Our results suggest that shelf, slope, and nearshore regions should be surveyed separately and that additional work, in the form of species-or group-specific temporal studies, should be undertaken to refine survey designs.

M. F. Sigler, M. F., M. C. Benfield, E. D. Brown, J. H. Churnside, N. Hillgruber, J. K. Horne, S. Parker-Stetter, Survey Strategies for Assessment of Bering Sea Forage Species , North Pacific Research Board Final Report 401, 2006.

first image in evolution second image in evolution third image in evolution
Evolution of a biological hot spot, or intense foraging event in the SE Bering Sea.

Norwegian Sea mackerel

A broad-scale lidar survey was conducted in the Norwegian Sea in the summer of 2002. Since that survey, a number of data processing techniques have been developed, including manual identification of fish schools, multi-scale median filtering, and curve fitting of the lidar profiles. In the automated techniques, applying a threshold to the data has previously been shown to improve the correlation between lidar data and acoustic data. This is similar to applying a threshold to eliminate plankton scattering from acoustic data. We applied these techniques to the lidar data from the 2002 survey, and compared the results with the results of a mackerel survey performed by the ships FV "Endre Dyrøy" and FV "Trønderbas" during the same time period. Despite a high level of variability in both lidar and trawl data, we found that the broad-scale distribution of fish inferred from the lidar agreed reasonably well with the broad-scale distribution of mackerel caught by the FV "Endre Dyrøy" for manual processing of the lidar data and for manual processing of the lidar data using a median filter and a threshold level T > 1. We also measured the optical target strength of atlantic mackerel as a first step in a conversion to biomass.

E. Tenningen, J. H. Churnside, A. Slotte, and J. J. Wilson, Lidar target-strength measurements on northeast Atlantic mackerel (Scomber scombrus), ICES Journal of Marine Science, 63(4), 677-682, doi:10.1016/j.icesjms.2005.11.018, 2006.

relative catch per unit effort from the FV Endre Dyryi
Relative catch per unit effort from the FV "Endre Dyrøy."
distribution of mackerel inferred from manual processing of the lidar data
Distribution of mackerel inferred from manual processing of the lidar data. Schools in the top 3 m were neglected to more closely correspond with the depth of the headrope of the net on the ship. Correlation with ship data was 0.64.
distribution of mackerel inferred from automated processing of the lidar data
Distribution of mackerel inferred from automated processing of the lidar data. Correlation with ship data was 0.58.

Gulf of Alaska capelin and herring

FLOE, along with a digital camera, was used in the waters off of southern Alaska in the summer of 2000 under funding from the North Pacific Marine Research Initiative. The Principal Investigator was Dr. Evelyn Brown of the Institute of Marine Science of the University of Alaska at Fairbanks. We evaluated airborne remote sensing, using lidar and color digital video, in the North Pacific in 2000. Specific objectives were (1) to determine lidar depth-penetration range, (2) to develop ocean color indices as a proxy for depth penetration and Chl a, (3) to compare lidar with acoustic and net-sampling data, (4) to define diurnal variability over large areas, and (5) to evaluate strengths and weaknesses. Depth penetration ranged from 18 to 50 m in non-silty water, with lowest values observed inshore by day and highest values on the continental shelf at night. A green index, derived from the three- band video data, was significantly related to depth penetration and was in general agreement with SeaWiFS satellite Chl a values. Significant correlations with acoustic data were obtained in an area with a high concentration of capelin, Mallotus villosus.

We returned in subsequent years in a collaboration with the Prince William Sound Science Center. We discovered a very high level of dissolved organic material in one of the inlets of Kodiak Island; the absorption increased dramatically in the inlet, without a corresponding increase in near-surface scattering. We demonstrated vessel avoidance by flying the lidar over one of the NOAA research vessels, the RV Miller Freeman. We demonstrated imaging of individual salmon using the laser as a light source for a gated intensified camera, which had an exposure time of 20 ns.

E. D. Brown, J. H. Churnside, R. L. Collins, T. Veenstra, J. J. Wilson, and K. Abnett, Remote sensing of capelin and other biological features in the North Pacific using lidar and video technology, ICES Journal of Marine Science, 59(5), 1120-1130, doi:10.1006/jmsc.2002.1282, 2002.

J. H. Churnside and J. J. Wilson, Airborne lidar imaging of salmon, Applied Optics, 43(6), 1416-1424, doi:10.1364/AO.43.001416, 2004.

lidar profiles plot
Lidar profiles in Uganic Bay showing very high absorption in the inner bay (red) relative to that in the outer bay (blue).
lidar signal vs position plot
Plot of lidar signal around NOAA RV Miller Freeman. Distance is measured from the center of the ship, with positive values in front. The red section shows where the signal is reduce several hundred m in front of and behind the ship. The increased return just behind the ship is scattering from bubbles in the wake.
school of salmon
Image of school of salmon taken with intensified camera under laser illumination. Camera shutter is opened when the laser pulse is illuminating the water at a depth of 5 m, so the fish at the surface appear as dark silhouettes.

Prince William Sound Alaska zooplankton

An aerial survey of Prince William Sound was performed on 14 May 2002 using FLOE. The flight was made just after the Prince William Sound Science Center performed a surface survey on 10-12 May that measured zooplankton concentrations in the Sound using acoustics and plankton net sampling. The plankton net sampling revealed that the large-bodied copepods of the genus Neocalanus a major constituent of zooplankton in the Sound. The zooplankton concentrations were measured in 8 areas of the Sound using a 420 kHz echosounder as the primary instrument. Lower frequencies were also used to correct for the return from fish in the beam. The largest concentrations of zooplankton were observed in the eastern part of the Sound, including the main basin and the Hinchenbrook Entrance. Using a technique common in acoustics, we applied a threshold to the LIDAR data to remove the return from low-level scatterers. Since the LIDAR target strength of plankton has yet to be measured directly, we varied the threshold level and compared the results with the acoustic results in the 8 areas. The best agreement was found with a threshold level of 2.75, relative to the background scattering level. With this threshold, the correlation between the LIDAR and echosounder results was 0.78.

J. H. Churnside and R. E. Thorne, Comparison of airborne lidar measurements with 420 kHz echo-sounder measurements of zooplankton, Applied Optics, 44(26), 5504-5511, doi:10.1364/AO.44.005504, 2005.

map of Prince William Sound showing ship and flight tracks
Map of Prince William Sound showing measurement lines for ship (red) and lidar (green) data.
copepod neocalanus
Copepod neocalanus.
depth profiles of echosounder and lidar returns
Depth profile measured by echosounder (red) and lidar (green).

Gulf of Mexico mullet and baitfish

In December 2000, FLOE went to Tampa, Florida to look for schooling fish in the shallow waters off the east coast of Florida in the Gulf of Mexico. On the left is a photograph of a school of mullet that was taken from the aircraft. We made several passes over this school. The echo grams are two consecutive passes over the same school from different directions. In the photograph, the school looks darker than the surrounding water. The lidar clearly shows the fish as a region of enhanced reflection.

In this experiment, the lidar measured a number of schools. A small boat equipped with a scientific echosounder was directed to each school and made acoustic measurements. In all, seven schools were captured both by the lidar and by the echosounder. More schools were targeted by the lidar, but it proved impossible to get the boat to most of these to obtain the acoustic measurements. The correlation between lidar and acoustics for the seven schools was almost perfect (99.6%). There was also a professional fish spotter on the plane with us. He was able to identify the species of the schools very reliably from the air.

J. H. Churnside, D. A. Demer, and B. Mahmoudi, A Comparison of Lidar and Echosounder Measurements of Fish Schools in the Gulf of Mexico, ICES Journal of Marine Science, 60(1), 147-154, doi:10.1006/jmsc.2002.1327, 2003.

menhaden school
Menhaden school.
lidar echogram of school
Lidar echogram of school.
second lidar echogram of the same school
Second lidar echogram of the same school.

Southern California Bight anchovy, sardine, squid

In January 1999, NOAA ESRL Chemical Sciences Division (formerly the Environmental Technology Laboratory) and the State of California Department of Fish and Game conducted an experiment with FLOE in California to detect squid. FLOE was reinstalled on the King Air and we flew around the southern California coast. The squid were plentiful as were the squid fishermen.

The following images are graphs of data. The X axis is time in seconds and the Y axis is depth in meters. The laser fires 30 shots per second. There are 60 seconds worth of data plotted, so these graphs represent 1800 shots from the laser. The aircraft we used was flying at about 75 m/s (meters per second), so 60 seconds represents about 4.5 Km in distance. To learn more about FLOE's hardware, refer to the instrumentation.

Select the image to view a slightly larger image, or select image links to the high resolution images showing the full detail of the data.

Image 1
Image 1: 11 Jan 99, 21:49 PST, 34° 03' N, 119° 00' W. The thick line at about 25 m is the bottom. The lighter areas above the bottom at 10-15 and at 50-55 s are squid.
Image 2
Image 2: 11 Jan 99, 21:07 PST, 33° 55' N, 119° 59' W. There is a group of squid at 10-12 s that is a few m from the bottom. A much larger group is at 30-35 s, and extends up to 10 m below the surface.
Image 3
Image 3: 12 Jan 99, 19:47 PST, 33° 54' N, 119° 58' W. The group here, at around 30 s, is clearly separated from the bottom. The bottom just under the school is at 50 m.
Image 4
Image 4: 12 Jan 99, 19:40 PST, 33° 53' N, 120° 00' W. The main group here is at 30-33 s, and extends from about 10 to 20 m above the bottom. There is a second group at about 37 s that is just a little higher. Above both groups is a plankton layer.

Atlantic off Spain, Portugal, France juvenile fish

In August and September of 1998 and 1999, FLOE flew on the Spanish Casa aircraft as part of the European project on Experimental Surveys for the Assessment of Juveniles (JUVESU). We found that the correspondence between acoustic surveys and the lidar surveys depended strongly on the time delay, consistent with the results of other studies. If the difference is more than about 2 days, the correlation is much lower. Significantly, repeat acoustic surveys and repeat lidar surveys also produced very different results if they were done a few days apart. The correspondence also depended on the type of fish aggregation. If most of the fish along a transect were in a single large school, the results varied greatly. On the other hand, small, scattered schools tended to be more consistent between techniques and between different passes over the same area with the same technique.

P. Carrera, J. H. Churnside, G. Boyra, V. Marques,C. Scalabrin, and A. Uriarte, Comparison of airborne lidar with echosounders - a case study in the coastal Atlantic waters of southern Europe, ICES Journal of Marine Science, 63(9), 1736-1750, doi:10.1016/j.icesjms.2006.07.004, 2006.

Ghost Nets

Ghostnet refers to lost or abandoned fishing gear that drifts in the ocean continuing to catch fish and entangle marine mammals, turtles, and sea birds. The synthetic materials currently used in fishing nets decay extremely slowly, so these nets can continue to drift for years. Many of these end up trapped on the coral reefs, where entanglement rates are even higher than in the open ocean and where they damage the fragile coral. To remove these nets from reefs, divers must cut the nets off with knives and load them into inflatable boats. It is extremely laborious and dangerous work. During 2001, a multi-agency effort consisting of 3 ships and 18 divers removed nearly 70 tons of debris during 270 ship days at sea, clearing only two atolls in the 1200-mile Hawaiian Archipelago.

Given the magnitude of the problem and hazards associated with cleaning the reefs, a multi-agency effort is under way to locate ghostnets in the open ocean and collect them before they reach reefs. (Visit High Seas GhostNet for more information.) The ghostnet team has shown that many of these nets pass through the Sub-Tropical Convergence Zone north of the Hawaiian Islands in the spring of each year when the convergence is particularly strong. Oceanic debris of all types is expected to accumulate here and in other convergence zones. For this reason, searching for ghostnets and other debris in convergence zones was expected to be more efficient than searching the entire ocean. The effectiveness of this approach has been demonstrated in the North Pacific.

W. G. Pichel, J. H. Churnside, T. S. Veenstra, D. G. Foley, K. S. Friedman, R. E. Brainard, J. B. Nicoll, Q. Zheng, and P. Clemente-Colon, Marine debris collects within the north Pacific subtropical convergence zone, Marine Pollution Bulletin, 54(8), 1207-1211, doi:10.1016/j.marpolbul.2007.04.010, 2007.

satellite image of an eddy in the Gulf of Alaska
Satellite image of an eddy in the Gulf of Alaska.
imaging lidar image of an active net
Imaging lidar image of an active net.
Ghostnet 2001 logo
GhostNet 2001

The Ghostnet team used satellite imagery to locate convergence zones in the North Pacific. An aircraft equipped with a suite of remote sensors was directed to these zones to see if there was, in fact, a higher concentration of debris in these areas. The sensors included an imaging lidar, visible imagers, and an infrared imager. The image of an eddy in the Gulf of Alaska is located in the sea-surface height image from a satellite. This type of eddy indicates convergence of the surface currents. The locations of detections are presented as circles on the flight track across the eddy. It is clear that there are more detections across the eddy than on either side, so this type of search is much more efficient than a random search of the ocean. The next step is to tag nets during an aerial search, so they can be located and retrieved by surface ships.

East Sound plankton

An interest in thin plankton layers led us to examine lidar data from a number of the deployments to see if we could see thin layers. Thin layers of phytoplankton or zooplankton may contain densities of organisms ranging up to 1000 times those found just above, or below the structure. These extraordinary concentrations of living material must have important implications for many aspects of marine ecology (e.g. phytoplankton growth dynamics, micro- and macrozooplankton grazing, behaviour, life histories, predation, harmful algal blooms). For our purposes, a thin layer was defined as a layer in the data less than 3 m thick. In all, we detected over 2000 km of such layers in a total of 80,000 km of flight lines. One interesting feature was that thin layers often followed dynamical structures like internal waves.

J. H. Churnside, Spatial characteristics of thin scattering layers observed by airborne LIDAR , Final Report to Office of Naval Research, 2005.

J. H. Churnside and L. A. Ostrovsky, Lidar observations of a strongly nonlinear internal wave train in the Gulf of Alaska, International Journal of Remote Sensing, 26(1), 167-177, doi:10.1080/01431160410001735076, 2005.

chart with all flight tracks analyzed for thin layers in red
Chart with all flight tracks analyzed for thin layers in red.
thin layer showing abrupt change in depth
Thin layer showing abrupt change in depth.
strong internal wave train
Strong internal wave train.