Autonomous Water Sampler for Oil Spill Response

Abstract

A newly developed water sampling system enables autonomous detection and sampling of underwater oil plumes. The Midwater Oil Sampler collects multiple 1-L samples of seawater when preset criteria are met. The sampler has a hydrocarbon-free sample path and can be configured with several modules of six glass sample bottles. In August 2019, the sampler was deployed on an autonomous underwater vehicle and captured targeted water samples in natural oil seeps offshore Santa Barbara, CA, USA.

Full text



Citation: Gomez-Ibanez, D.;
Kukulya, A.L.; Belani, A.; Conmy,
R.N.; Sundaravadivelu, D.; DiPinto,
L. Autonomous Water Sampler for
Oil Spill Response. J. Mar. Sci. Eng.
2022,10, 526. https://doi.org/
10.3390/jmse10040526
Academic Editor: Shuo Pang
Received: 26 February 2022
Accepted: 7 April 2022
Published: 11 April 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Journal of
Marine Science
and Engineering
Communication
Autonomous Water Sampler for Oil Spill Response
Daniel Gomez-Ibanez 1, * , Amy L. Kukulya 1, Abhimanyu Belani 1, Robyn N. Conmy 2, Devi Sundaravadivelu 3
and Lisa DiPinto 4
1Woods Hole Oceanographic Institution, Woods Hole, MA 02543, USA; akukulya@whoi.edu (A.L.K.);
manyufacturing@gmail.com (A.B.)
2U.S. Environmental Protection Agency, Office of Research and Development, Cincinnati, OH 45268, USA;
conmy.robyn@epa.gov
3Pegasus Technical Services, Inc., Cincinnati, OH 45219, USA; sundaravadivelu.devi@epa.gov
4National Oceanic and Atmospheric Administration, Silver Spring, MD 20910, USA; lisa.dipinto@noaa.gov
*Correspondence: dgi@whoi.edu
Abstract:
A newly developed water sampling system enables autonomous detection and sampling
of underwater oil plumes. The Midwater Oil Sampler collects multiple 1-L samples of seawater
when preset criteria are met. The sampler has a hydrocarbon-free sample path and can be configured
with several modules of six glass sample bottles. In August 2019, the sampler was deployed on an
autonomous underwater vehicle and captured targeted water samples in natural oil seeps offshore
Santa Barbara, CA, USA.
Keywords:
oil spill; water sampling; autonomous underwater vehicle; autonomous survey; organic
analysis; environmental impact assessment; rapid response; plume tracking
1. Introduction and Background
Underwater oil spills can have long lasting impacts to ocean and coastal environments,
but are challenging to observe. A newly developed autonomous underwater vehicle (AUV)
water sampling system uses in situ sensing to detect dissolved oil, and then immediately
capture seawater in quantities sufficient to discriminate and measure its hydrocarbon
constituents while maintaining the independence and integrity of multiple samples.
This section reviews the need for midwater sample collection and previous work in
autonomous water sample collection. Section 2describes the concept and components of
the Midwater Oil Sampler (MOS). Section 3describes the qualification tests during design
development, and Section 4describes at-sea deployment of the Midwater Oil Sampler.
Discussion and conclusions are given in Sections 5and 6.
1.1. Obstacles to Underwater Oil Sampling
Underwater oil spill response relies on complementary observations from remote
sensing, shipboard measurements, and remotely operated and autonomous vehicles, all of
which have a role in understanding the flows of oil released into the ocean. However, only
laboratory chemical analysis of water samples can distinguish hydrocarbon classes and
accurately measure concentrations. This information is important to understanding the
fate and effects of oil in the ocean.
The Deepwater Horizon disaster of 2010 made clear that an underwater oil spill
is different from a surface spill, with a significant fraction of hydrocarbons retained in
midwater [
1
], and is therefore inaccessible to traditional methods of surface water collection.
However, even 12 years after Deepwater, it is challenging to access, locate, and sample a
midwater oil plume [2].
Obstacles to effective sampling of a midwater oil plume include, first of all, site access,
followed by the detection of the plume, and finally, the capture and shipping of water samples.
J. Mar. Sci. Eng. 2022,10, 526. https://doi.org/10.3390/jmse10040526 https://www.mdpi.com/journal/jmse

J. Mar. Sci. Eng. 2022,10, 526 2 of 13
1.1.1. Site Access
AUVs offer a uniquely capable platform for monitoring the extent of a midwater oil
plume, which may be difficult to reach due to depth, remoteness, competing marine traffic,
and hazards to human health and safety.
Towed or tethered remotely operated underwater vehicles (ROVs) and human occu-
pied vehicles (HOVs) can reach abyssal depths and offer immediate visual feedback to
operators. However, ROVs are limited to operations close to a surface vessel, and surface
vessels may be unavailable or not allowed to approach disaster sites due to ongoing re-
sponse efforts. Vessels that do approach may be exposed to hazards including crude oil
vapor, smoke, and explosions.
AUVs, in contrast, offer over-the-horizon access. They can be launched from surface
vessels beyond the disaster area, or from shore. AUVs can be mobilized quickly and work
independently from surface vessels, covering a large area to unlimited depths, discovering
midwater plumes that may not be connected to any visible surface expression. AUVs can be
programmed to travel in a lawnmower or sawtooth pattern, creating an XY grid map or YZ
section, both of which are impractical with an ROV. With each surfacing, an AUV may relay
data via satellite to a land-based server for human interpretation and situational awareness.
AUVs can be programmed to perform a broad scale search, first covering the search area
with widely spaced profiles, and later returning to areas with above-average fluorescence
detections or ‘hot-spots’ for a fine-scale survey using more focused sensors such as cameras
or water collection devices. This multi-scale survey can be automated and is adaptable in
real-time depending on changing mission objectives and situational awareness.
1.1.2. Oil Detection
In order to sample oil released underwater, it must first be detected using remote or
in situ sensors. Transported by ocean currents and its own buoyancy, the oil plume may
eventually reach the surface, forming a surface expression or “slick”, which can be detected
visually or with radar [
3
] from a ship, airplane, or satellite. However, some oil may be
neutrally buoyant at depth with no obvious connection to surface features [4].
In situ fluorometers are used to locate midwater oil plumes [
5
]. The Seabird SeaOWL
oil-in-water sensor detects fluorescent dissolved organic material (FDOM) with excitation
at 370 nm and emission at 460 nm [
6
]. SeaOWL also measures chlorophyll fluorescence,
with an excitation at 470 nm and emission at 690 nm, and backscatter at 700 nm, which can
help to disambiguate dissolved oil from other sources of FDOM. The SeaOWL is used in
real-time to target sample collections during an adaptive sampling mission.
1.1.3. Sample Capture and Shipping
Analysis of dissolved oil samples using gas chromatography mass spectroscopy (GC-
MS) can discriminate 91 well-known hydrocarbons, producing a fingerprint of the hy-
drocarbon source [
7
]. However, to deliver useful samples to a lab, the samples must be
collected and transported without altering hydrocarbon constituents.
Carryover contamination occurs when oil from one location is present in samples
intended to represent a different location. A surface slick may adhere to collection equip-
ment that is deployed to deeper water, causing oil to be detected at depths where there is
none. Decontamination of sampling equipment between deployments can be a challenge,
since clean water may be unavailable. If a common inlet tube is used to collect samples at
several locations, residual oil from the inlet may be distributed among all samples. To mini-
mize carryover contamination and simplify site decontamination, single-use disposable
sampling equipment should be used when feasible [8].
If water samples are exposed to heat or sunlight, volatile hydrocarbons may be lost
by evaporation. If water is collected in one container and then transferred to another,
sampled oil may remain stuck to the first bottle or transfer tubing, reducing the measured
concentration. When feasible, a water sampling system should avoid bottle-to-bottle
transfers and facilitate offload to cold storage immediately after collection.

J. Mar. Sci. Eng. 2022,10, 526 3 of 13
Containers used to hold water samples may affect the measured hydrocarbon concen-
trations [
9
]. If samples are temporarily stored in plastic bottles, hydrocarbons can migrate
from the bottle to the water sample or vice versa. Volatile hydrocarbons may diffuse
through a plastic bottle and escape before analysis. To minimize these effects, typical
sampling protocols specify only glass bottles for oil sampling [8,10].
1.2. Existing Sampling Methods
Near-surface oil can be collected manually by direct dip or peristaltic pump [
8
], but
these methods are not feasible beyond a few meters of depth.
Deeper samples can be collected remotely using ROV-operated Niskin bottles, small-
diameter tubing [
11
], syringes [
12
], or gas-tight bottles [
13
]. However, these remote capture
devices are not suitable for AUV integration due to large size and weight.
A few water samplers have been integrated with long range AUVs. The Aqualab
sampler [
14
,
15
] used a multiport valve and 49 ethylene vinyl acetate (EVA) pouches to
collect 200 mL water samples beneath fjords in East Greenland in 2004 for isotope ratio
analysis, demonstrating the utility of using autonomous vehicles to retrieve samples
from otherwise inaccessible sites. The Suspended Particulate Rosette (SUPR) family of
samplers [
16
] also uses a multiport valve, but allows for different bottle sizes and materials
to be used for compatibility with various analytes. The Gulper [
17
,
18
] uses polymethyl
methacrylate (PMMA) plastic sample cylinders and silicone o-rings, materials selected for
consistent phytoplankton growth in primary productivity studies. Gulper was deployed to
collect water samples during the Deepwater Horizon oil spill response, although sampling
did not occur due to a flooded controller housing [4].
2. Midwater Oil Sampler Design, Materials, and Methods
This section describes the Midwater Oil Sampler and the other components of a
complete autonomous water sampling system including the AUV host platform, in situ
sensor suite, the Midwater Oil Sampler module, its operation sequence, and mission
programming. These parts work together to efficiently collect targeted water samples.
2.1. Host Platform
A REMUS 600 AUV [
19
] was chosen as a host platform for the Midwater Oil Sampler
because it is large enough to comfortably accommodate several 1-L bottles, along with
associated sensors, within its 32 cm diameter. The REMUS 600 operates to a 600 m depth,
with 3 m length, 300 kg mass, and expandable 5 kWh battery energy. Its aft section
implements core vehicle control, propulsion, energy storage, and communication functions,
while its forward section is a modular payload that can accommodate a variety of mission-
specific sensors using one of the supported payload electrical interfaces. The REMUS 600
is a member the REMUS family of AUVs, which includes the smaller REMUS 100 and
the larger REMUS 6000 autonomous underwater vehicles, developed at the Woods Hole
Oceanographic Institution and commercialized by Hydroid Inc. (Pocasset, MA, USA). The
Midwater Oil Sampler requires only DC power and RS-232 communication links to the
host, so it can also be used with other similar vehicles.
2.2. Integrated In Situ Sensors
For oil spill response, the REMUS 600 AUV was configured with several in situ sensors
along with the newly developed Midwater Oil Sampler. The sensor suite included a Licor LI-
192 photosynthetically active radiation (PAR) sensor, Seabird Sea-OWL fluorometer, Anderaa
4831F optode, Seascan Holocam, and a GoPro Hero 3 video camera. Only the fluorometer was
used for real-time triggering of water samples; other sensor data were analyzed post-mission
to serve as converging lines of evidence to confirm the presence of oil.

J. Mar. Sci. Eng. 2022,10, 526 4 of 13
2.3. Sampler Design and Fabrication
The Midwater Oil Sampler was designed to minimize the contamination of samples
through the use of compatible materials and an optimized flow path. Both are described in
this section, along with the fabrication and assembly of the Midwater Oil Sampler.
2.3.1. Material Selection
Glass sample bottles were used to avoid hydrocarbon contamination that occurs with
plastic sample containers [
9
]. Single-use pre-cleaned wide-mouth amber bottles are readily
available from several vendors. Bottles were certified by the vendors to meet the U.S.
Environmental Protection Agency (EPA) standards [
20
], with negligible amounts of semi-
volatile organics and other contaminants. Glass bottles may optionally be baked in a muffle
furnace for decontamination and reuse. The sample volume of 1 L was chosen to maintain
typical detection limits while still allowing several samples to fit in an AUV payload. One
liter samples are consistent with the water sampling procedures developed in response to
the Deepwater Horizon disaster [
8
]. After recovery of the host AUV, sample bottles can be
quickly removed from the sampling system, capped, and shipped to a laboratory.
The only materials in the sample path other than the glass bottle were 316 stainless
steel and fluoropolymers (PFPE, FKM, and PTFE) to minimize hydrocarbon contamination.
2.3.2. Flow Path
The configuration of the water flow path was designed to minimize contamination.
Sample inlets for each bottle were independent to minimize carryover contamination. Water
entered each bottle through a 12 mm diameter by a 15 cm long upward-facing PTFE inlet
extension tube, which extended beyond the vehicle envelope to avoid collecting water that
has been in contact with the AUV skin. Inlet tubes were replaced after each mission to avoid
carryover contamination between missions and remove residue from surface slicks, which
is clearly visible post-mission in Figure 1. Past the inlet, water flowed through a check valve
and into a sample bottle via the bottle adapter. Check valves (Swagelok SS-4CP6-1/3-SC11)
trap water in the bottle and prevent water exchange with the environment when the pump
is not operating. The check valve and bottle adapter were made of 316 stainless steel
with fluoropolymer o-rings lubricated with Krytox fluoropolymer grease. Bottle adapter
assembly included a stainless-steel fill-tube inside the bottle to promote complete water
exchange. A second check valve was connected at the outlet port of the bottle adapter.
J. Mar. Sci. Eng. 2022, 10, x FOR PEER REVIEW 5 of 13
Figure 1. REMUS 600 AUV integrated with the Midwater Oil Sampler.
2.3.3. Fabrication and Assembly
The Midwater Oil Sampler uses off-the-shelf parts, rapid prototyping, and 3D-print-
ing to support efficient sampler fabrication and assembly.
A junction box subassembly (Figure 2) housed six pumps and supporting electronics,
which were sealed together in cast polyurethane within a thin plastic shell. The junction
box shell was formed by fused deposition modeling (FDM) of an acrylonitrile styrene
acrylate (ASA) filament. Before potting, pump housings were sandblasted to improve ad-
hesion to polyurethane, and then inserted into circular openings in the shell. Inside the
shell, uplink cable and motor wires were soldered to a control printed circuit board as-
sembly (PCBA). The control PCBA accepts a power supply of 18–36 V from the host AUV
and communicates with the host via RS-232. The control PCBA powers each pump for a
specified time. The PCBA also provides diagnostic information including power con-
sumed by each pump, temperature, and insulation resistance. Polyurethane (3M Scotch-
cast 2131) was poured into the shell, degassed under vacuum, and cured overnight.
Figure 2. Exploded view of the junction box sub-assembly. Parts are, clockwise, control PCBA, junc-
tion box shell, six pumps, and uplink umbilical cable.
Figure 1. REMUS 600 AUV integrated with the Midwater Oil Sampler.

J. Mar. Sci. Eng. 2022,10, 526 5 of 13
Because the pumps and other tube fittings were located downstream of the sample
bottle, beyond a check valve, they were not in the sample path and did not need to be
completely free of hydrocarbon contaminants. The outlet check valve was followed by
a brass 90
◦
elbow, followed by a silicone 90
◦
elbow, which was connected to a pump.
Each bottle had an independent water pump to pull pre-fill water out of the bottle during
sampling. Submersible impeller pumps (Shenzhen Century Zhongke Technology Co.,
Shenzhen, China, model DC40-1250) were potted without voids for pressure tolerance.
2.3.3. Fabrication and Assembly
The Midwater Oil Sampler uses off-the-shelf parts, rapid prototyping, and 3D-printing
to support efficient sampler fabrication and assembly.
A junction box subassembly (Figure 2) housed six pumps and supporting electronics,
which were sealed together in cast polyurethane within a thin plastic shell. The junction
box shell was formed by fused deposition modeling (FDM) of an acrylonitrile styrene
acrylate (ASA) filament. Before potting, pump housings were sandblasted to improve
adhesion to polyurethane, and then inserted into circular openings in the shell. Inside
the shell, uplink cable and motor wires were soldered to a control printed circuit board
assembly (PCBA). The control PCBA accepts a power supply of 18–36 V from the host AUV
and communicates with the host via RS-232. The control PCBA powers each pump for a
specified time. The PCBA also provides diagnostic information including power consumed
by each pump, temperature, and insulation resistance. Polyurethane (3M Scotchcast 2131)
was poured into the shell, degassed under vacuum, and cured overnight.
J. Mar. Sci. Eng. 2022, 10, x FOR PEER REVIEW 5 of 13
Figure 1. REMUS 600 AUV integrated with the Midwater Oil Sampler.
2.3.3. Fabrication and Assembly
The Midwater Oil Sampler uses off-the-shelf parts, rapid prototyping, and 3D-print-
ing to support efficient sampler fabrication and assembly.
A junction box subassembly (Figure 2) housed six pumps and supporting electronics,
which were sealed together in cast polyurethane within a thin plastic shell. The junction
box shell was formed by fused deposition modeling (FDM) of an acrylonitrile styrene
acrylate (ASA) filament. Before potting, pump housings were sandblasted to improve ad-
hesion to polyurethane, and then inserted into circular openings in the shell. Inside the
shell, uplink cable and motor wires were soldered to a control printed circuit board as-
sembly (PCBA). The control PCBA accepts a power supply of 18–36 V from the host AUV
and communicates with the host via RS-232. The control PCBA powers each pump for a
specified time. The PCBA also provides diagnostic information including power con-
sumed by each pump, temperature, and insulation resistance. Polyurethane (3M Scotch-
cast 2131) was poured into the shell, degassed under vacuum, and cured overnight.
Figure 2. Exploded view of the junction box sub-assembly. Parts are, clockwise, control PCBA, junc-
tion box shell, six pumps, and uplink umbilical cable.
Figure 2.
Exploded view of the junction box sub-assembly. Parts are, clockwise, control PCBA,
junction box shell, six pumps, and uplink umbilical cable.
The junction box subassembly and six bottle-adapter assemblies were all attached to an
anodized aluminum plate, with handles to facilitate installation of the 6-bottle module into
the host AUV (Figure 3). Each six-bottle sampling module was placed in a rigid aluminum
frame attached to the host AUV using 32 cm diameter ring joints. Yellow-painted syntactic
foam blocks were attached to the frame, surrounding the sample bottles. These foam
blocks protect the glass bottles, streamline the vehicle, and offset the weight of the sampler
and frame. More than one sampler frame may be mounted on an AUV at the same time,
allowing for 6, 12, or 18 samples to be collected during a single REMUS 600 mission.

J. Mar. Sci. Eng. 2022,10, 526 6 of 13
J. Mar. Sci. Eng. 2022, 10, x FOR PEER REVIEW 6 of 13
The junction box subassembly and six bottle-adapter assemblies were all attached to
an anodized aluminum plate, with handles to facilitate installation of the 6-bottle module
into the host AUV (Figure 3). Each six-bottle sampling module was placed in a rigid alu-
minum frame attached to the host AUV using 32 cm diameter ring joints. Yellow-painted
syntactic foam blocks were attached to the frame, surrounding the sample bottles. These
foam blocks protect the glass bottles, streamline the vehicle, and offset the weight of the
sampler and frame. More than one sampler frame may be mounted on an AUV at the
same time, allowing for 6, 12, or 18 samples to be collected during a single REMUS 600
mission.
Figure 3. Exploded view of the Midwater Oil Sampler module. From top to bottom are the top foam
block, sampler module, and sampler frame.
2.4. Operation Sequence
For convenient bottle preparation and fast access to samples, modules of six bottles
are easily removable from the AUV. The following steps were used to prepare and operate
the Midwater Oil Sampler:
1. Clean bottles are installed by screwing them in from below. The sampler’s plywood
service cradle provides clearance below the sample bottles (Figure 4).
2. Distilled water is pumped into each bottle by temporarily connecting a flexible tube
from a distilled water supply to each bottle inlet.
3. One or more modules are loaded into the host AUV. The sampler umbilical cable is
mated, and the top foam block fastened over the top of the sampler. New inlet exten-
sions are attached to each inlet barb. The vehicle is now ready to launch.
Figure 3.
Exploded view of the Midwater Oil Sampler module. From top to bottom are the top foam
block, sampler module, and sampler frame.
2.4. Operation Sequence
For convenient bottle preparation and fast access to samples, modules of six bottles
are easily removable from the AUV. The following steps were used to prepare and operate
the Midwater Oil Sampler:
1.
Clean bottles are installed by screwing them in from below. The sampler’s plywood
service cradle provides clearance below the sample bottles (Figure 4).
2.
Distilled water is pumped into each bottle by temporarily connecting a flexible tube
from a distilled water supply to each bottle inlet.
3.
One or more modules are loaded into the host AUV. The sampler umbilical cable
is mated, and the top foam block fastened over the top of the sampler. New inlet
extensions are attached to each inlet barb. The vehicle is now ready to launch.
4.
During an underwater mission, when a pump is activated by the host AUV, ambient
seawater flows into one bottle, passing through the inlet extension, inlet check valve,
and fill tube.
5.
After the host AUV is recovered, the top foam block is removed, and the umbilical
cable is disconnected. The sampler is removed and placed in its service cradle.
6.
Sample bottles, attached to the sampler only by their threaded necks, are unscrewed
and capped with PTFE-lined screw caps. No bottle-to-bottle water transfer is necessary.
7. Bottles are labeled and placed in coolers for transport to the laboratory.

J. Mar. Sci. Eng. 2022,10, 526 7 of 13
J. Mar. Sci. Eng. 2022, 10, x FOR PEER REVIEW 7 of 13
4. During an underwater mission, when a pump is activated by the host AUV, ambient
seawater flows into one bottle, passing through the inlet extension, inlet check valve,
and fill tube.
5. After the host AUV is recovered, the top foam block is removed, and the umbilical
cable is disconnected. The sampler is removed and placed in its service cradle.
6. Sample bottles, attached to the sampler only by their threaded necks, are unscrewed
and capped with PTFE-lined screw caps. No bottle-to-bottle water transfer is neces-
sary.
7. Bottles are labeled and placed in coolers for transport to the laboratory.
Figure 4. One sampler module in its plywood service cradle.
2.5. Mission Programming and Sample Triggering
The host REMUS 600 AUV control system is responsible for dynamically positioning
the vehicle within an oil plume and triggering the sampler to fill a bottle. The AUV oper-
ator specifies intent in the form of a mission script, composed of several sequential objec-
tives. Sampling can be accomplished with either of two different mission objectives de-
veloped for plume sampling. To sample at a specified position, a Point and Gulp objective
may be used. To sample when certain fluorometer (FDOM) criteria are met, an Adaptive
Sampling objective may be used. The Point and Gulp objective is used when desired sam-
pling locations are known in advance, based on previous AUV missions or other observa-
tions. During the Point and Gulp objective, the vehicle transits to a specified latitude, lon-
gitude, and depth, and then navigates, at constant depth, in a 20 m radius circle around
the specified point, at 2 knots speed, through two revolutions. This circle pattern allows
enough time for the sampling pump to run for 90 s, exchanging pre-fill water with ambi-
ent seawater.
The Adaptive Sampling objective is intended to provide additional observations when
elevated fluorescence is detected and may optionally be configured to take a water sam-
ple. Typically, an AUV survey consists of a series of grid lines covering an area of interest.
The Adaptive Sampling objective allows the survey path to adapt to observations. When
the triggering criteria are met, the vehicle navigates in a spiral around the point of highest
FDOM, gathering additional observations, before resuming the previously specified
wide-area grid survey. The Adaptive Sampling objective allows the triggering criteria for
water sampling to be set independently from the triggering criteria for additional survey
passes.
Figure 4. One sampler module in its plywood service cradle.
2.5. Mission Programming and Sample Triggering
The host REMUS 600 AUV control system is responsible for dynamically positioning
the vehicle within an oil plume and triggering the sampler to fill a bottle. The AUV operator
specifies intent in the form of a mission script, composed of several sequential objectives.
Sampling can be accomplished with either of two different mission objectives developed for
plume sampling. To sample at a specified position, a Point and Gulp objective may be used.
To sample when certain fluorometer (FDOM) criteria are met, an Adaptive Sampling objective
may be used. The Point and Gulp objective is used when desired sampling locations are
known in advance, based on previous AUV missions or other observations. During the
Point and Gulp objective, the vehicle transits to a specified latitude, longitude, and depth,
and then navigates, at constant depth, in a 20 m radius circle around the specified point,
at 2 knots speed, through two revolutions. This circle pattern allows enough time for the
sampling pump to run for 90 s, exchanging pre-fill water with ambient seawater.
The Adaptive Sampling objective is intended to provide additional observations when
elevated fluorescence is detected and may optionally be configured to take a water sample.
Typically, an AUV survey consists of a series of grid lines covering an area of interest. The
Adaptive Sampling objective allows the survey path to adapt to observations. When the
triggering criteria are met, the vehicle navigates in a spiral around the point of highest
FDOM, gathering additional observations, before resuming the previously specified wide-
area grid survey. The Adaptive Sampling objective allows the triggering criteria for water
sampling to be set independently from the triggering criteria for additional survey passes.
Triggering criteria include “thresh_min”, which specifies a minimum FDOM count,
and “thresh_inflate” (i.e., how much higher the background must be for the FDOM sensor
measurement to trigger optional observation or sampling). For example, a 10% increase
above background fluorescence could trigger additional observations, while a 20% increase
could trigger a water sample. Water sampling occurs only after additional observations are
completed, and the water sampling circle is centered at the point of highest FDOM.
3. Design Verification
Several important functions of the Midwater Oil Sampler were tested before its first
deployment in a midwater oil plume. Verification tests are outlined below.
3.1. Pressure Tolerance
Pumps were qualified for use at ambient pressure by placing one pump in a hydrostatic
pressure test chamber filled with mineral oil, with electrical penetrations to supply 12 V DC

J. Mar. Sci. Eng. 2022,10, 526 8 of 13
power to the pump inside. The chamber was pressurized to 69 MPa, equivalent to 7000 m
depth, with the pump operating. The pump was run for 24 h while pressurized, with no
change in power consumption. Pump flow rate was found to be unchanged after the test.
After the assembly of pumps into the junction box, each complete multi-pump assem-
bly was placed underwater and electrical resistance from umbilical power ground pin to
seawater was measured with a multimeter. No ground faults were detected in any of the
assembled junction boxes.
3.2. Flow Rate
To measure the pump flow rate, each complete sampler was filled with fresh water, and
the inlet tube was immersed in a bucket of water. A single pump was activated continuously.
Water exiting the pump was collected in empty 1 L bottles. With each sampler tested,
bottles filled in 20–30 s. Variability was mainly due to air trapped in the impeller housing,
something that occurs during bench testing, but not when submerged. In this application
with check valves and other restrictions, flow rate was conservatively estimated to be greater
than 2 L per minute. To ensure complete displacement of pre-fill water by ambient (sample)
seawater, pumps were run for 90 s, enough for at least three water changes.
3.3. Exchange Fraction
To promote rapid and complete exchange of pre-fill with ambient seawater, an inlet
tube extended down from the bottle adapter into the bottle to guide newly introduced
seawater to the bottom of the sample bottle, while displaced water exits out of the top. Tests
using dye-colored water to track the exchange of water in the bottle confirmed that this tube
promotes a more complete exchange of water for each liter pumped. A density gradient
between buoyant distilled pre-fill and heavier saline sample water also helps keep the newly
introduced seawater below and separate from the less dense pre-fill, so that the bottle is
filled from bottom to top with little mixing. To verify the exchange of water, a sampler was
pre-filled with distilled water, placed in seawater, and each pump was commanded to pump
for 90 s. After pumping, the salinity in the bottle was within
±
3% of ambient seawater,
indicating that little of the original pre-fill water remained in the bottle after sampling.
3.4. Leak Testing
The water sampler secured water in the sample bottle by means of two check valves.
These valves were sealed closed by springs when the pumps were powered off. To test the
integrity of the seals, six bottles were pre-filled with fresh tap water, measuring less than
1 ppt salinity. The entire sampler was suspended in seawater overnight. After removing
the bottles, salinity was measured again and remained less than 1 ppt.
Performance of the check valves can also be verified during operation of the sampler
by using distilled water for pre-fill and reserving a “trip blank”, a bottle that is never
pumped. By measuring the salinity of the trip blank, the integrity of bottle seals can be
confirmed during actual sampling operations.
4. At-Sea Testing Results
In August 2019, a multi-agency cooperative exercise located, mapped, and sampled natu-
rally occurring oil seeps near Santa Barbara, California, USA. The exercise simulated a rapid
response to an oil spill, and employed U.S. Coast Guard buoy tender George Cobb, several
underwater vehicles, an aerial drone, and satellite remote sensing. Participants from several
organizations gained familiarity with the AUV operations, developed processes to deliver data
products in near real-time to the U.S. National Oceanic and Atmospheric Administration’s
Environmental Response Management Application (ERMA) geodatabase, and demonstrated
the integration of AUV operations as part of a coordinated oil spill response [21].
This section describes the autonomous survey missions, water sampler operation, and
sample analysis.

J. Mar. Sci. Eng. 2022,10, 526 9 of 13
4.1. Autonomous Survey Summary
A REMUS 600 AUV, with in situ sensing and Midwater Oil Sampler payloads, con-
ducted gridded surveys of areas known to host underwater oil seeps. The REMUS 600
executed 19 missions in five days, covering a total distance of 126 km. Total survey time
was 23 h and 55 min, while individual missions lasted between 0.5 and 3 h.
Discharges from several natural gas and oil seeps were variable during the plume
mapping exercise. To locate and sample active oil plumes, it was valuable to have multiple
aerial and underwater vehicles active simultaneously, providing real-time situational
awareness, which was used to plan short 1 to 2 h underwater sampling missions based
on the most recent observations. Thirteen of 19 missions did not detect FDOM above the
threshold for sampling. Over five days, the AUV collected a total of 13 seawater samples.
4.2. Water Sampler Operation
In preparation for the oil plume mapping exercise, eight water sampler modules
were staged on the Cobb. Distilled water was purchased on shore and loaded on to the
Cobb in 20 L polycarbonate carboys. This water was used to pre-fill empty sample bottles
before loading the sampler modules into the REMUS 600 AUV. Before each day of the AUV
operations, a total of twelve 1 L bottles were loaded into two water sample modules.
At the end of each day’s operations, newly filled bottles were removed from the samplers.
Two 40 mL subsamples were poured from each sample bottle into glass vials required for
benzene, toluene, ethylbenzene, and xylene (BTEX) analysis. The original sample bottles with
remaining water were capped and labelled with the date, sampler serial number, and pump
number. Bottles were packed in ice for overnight shipping to a lab for analysis.
After removing the sample bottles, the samplers were decontaminated by first scrub-
bing the sampler in a tub of Sunshine Makers Simple Green All-Purpose Cleaner diluted
10:1 with water, and then pumping a solution of 10:1 diluted Simple Green from a reservoir
through each inlet and pump channel, with a temporary bottle installed, followed by a
freshwater rinse.
4.3. Laboratory Water Analysis
A total of 19 water sample bottles were sent to a U.S. EPA lab in Cincinnati and
analyzed within 14 days of collection. Thirteen bottles contained seawater samples and five
bottles were blanks, filled with distilled water, and carried onboard the AUV during one or
more missions but never pumped. These bottles were known as trip blanks. One bottle
was pre-filled with distilled water, but was never loaded onto a sampler, never placed into
the ocean, and known as a DI (deionized water) blank.
In the laboratory, salinity was measured using a portable refractometer at 22
◦
C. All
thirteen seawater samples exhibited salinity above 30 ppt, providing confidence that the
pumps effectively exchanged the blank freshwater with ambient seawater.
The 40 mL subsamples were used for volatile organic compound (VOC) analysis using
Method 524.3 [
22
] modified to use dynamic headspace extraction instead of purge and trap.
Other hydrocarbons were extracted from the 1 L bottles with dichloromethane and
concentrated to 1 mL following Method 3500C [
23
]. Alkanes and polycyclic aromatic hydro-
carbon (PAH) concentrations were quantified by gas chromatography/mass spectrometry
(GC/MS) following Method 8270D [
7
]. Concentrations of individual alkanes and PAHs
above the reporting limits were summed to compute total alkane and PAH concentrations.
The water sample analysis results are listed in Table A1.
5. Discussion
In the coastal waters of Santa Barbara, there are limited freshwater sources of fluo-
rescent natural organic matter entering the ocean. Thus, salinity is high and background
FDOM fluorescence is predictably stable at ~2 ppb quinine sulfate equivalent (QSE) with
slightly higher concentrations in the surface waters. Deviations from these concentrations
indicate sources of new fluorescent material such as petroleum hydrocarbons. Mission 7,

J. Mar. Sci. Eng. 2022,10, 526 10 of 13
completed on 27 August 2019, provides an example of a successful Adaptive Sampling objec-
tive. At six locations along a transect, as shown in Figure 5, the AUV encountered FDOM
significantly above the background. After a helical search to locate the depth with highest
FDOM, the AUV then collected a water sample at each location. The first four samples
were collected at a 11–13 m depth and two more were collected 20 m below the surface.
J. Mar. Sci. Eng. 2022, 10, x FOR PEER REVIEW 10 of 13
bottles were blanks, filled with distilled water, and carried onboard the AUV during one
or more missions but never pumped. These bottles were known as trip blanks. One bottle
was pre-filled with distilled water, but was never loaded onto a sampler, never placed
into the ocean, and known as a DI (deionized water) blank.
In the laboratory, salinity was measured using a portable refractometer at 22 °C. All
thirteen seawater samples exhibited salinity above 30 ppt, providing confidence that the
pumps effectively exchanged the blank freshwater with ambient seawater.
The 40 mL subsamples were used for volatile organic compound (VOC) analysis us-
ing Method 524.3 [22] modified to use dynamic headspace extraction instead of purge and
trap.
Other hydrocarbons were extracted from the 1 L bottles with dichloromethane and
concentrated to 1 mL following Method 3500C [23]. Alkanes and polycyclic aromatic hy-
drocarbon (PAH) concentrations were quantified by gas chromatography/mass spectrom-
etry (GC/MS) following Method 8270D [7]. Concentrations of individual alkanes and
PAHs above the reporting limits were summed to compute total alkane and PAH concen-
trations.
The water sample analysis results are listed in Table A1.
5. Discussion
In the coastal waters of Santa Barbara, there are limited freshwater sources of fluo-
rescent natural organic matter entering the ocean. Thus, salinity is high and background
FDOM fluorescence is predictably stable at ~2 ppb quinine sulfate equivalent (QSE) with
slightly higher concentrations in the surface waters. Deviations from these concentrations
indicate sources of new fluorescent material such as petroleum hydrocarbons. Mission 7,
completed on 27 August 2019, provides an example of a successful Adaptive Sampling ob-
jective. At six locations along a transect, as shown in Figure 5, the AUV encountered
FDOM significantly above the background. After a helical search to locate the depth with
highest FDOM, the AUV then collected a water sample at each location. The first four
samples were collected at a 11-13 m depth and two more were collected 20 m below the
surface.
Figure 5. Vehicle track during REMUS 600 Mission 7, 27 August 2019. The color of the
track indicates the measured fluorescence.
Vertical depth profiles showed that SeaOWL FDOM and backscatter maxima oc-
curred between 10 and 15 m for waters in Mission 7, which was not coincident with the
chlorophyll maximum at 7 m depth. This suggests that the elevated FDOM and
Figure 5.
Vehicle track during REMUS 600 Mission 7, 27 August 2019. The color of the track indicates
the measured fluorescence.
Vertical depth profiles showed that SeaOWL FDOM and backscatter maxima occurred
between 10 and 15 m for waters in Mission 7, which was not coincident with the chlorophyll
maximum at 7 m depth. This suggests that the elevated FDOM and backscatter observed at
these locations were not from biological activity. The presence of oil was also supported by the
AUV’s holographic camera, which detected oil droplets at these locations [
20
], even though
elevated levels of hydrocarbons were not detected in seawater samples from Mission 7.
Alkanes, PAHs, and BTEX measured by GC/MS in all seawater samples collected
during the spill response exercise were uniformly low. Concentrations of individual hydro-
carbons in the seawater samples were similar to concentrations found in trip blanks, the DI
blank, and lab reagent blank. These measurements were consistent with uncontaminated
samples from a site with very little oil present.
Sampler 4 was reused on 29 August after Simple Green site decontamination on
27 August. Samples and blank offloaded from the site-decontaminated sampler showed no
increase in hydrocarbons compared to other samples and blanks, suggesting that Simple
Green site decontamination does not introduce additional hydrocarbons into the samples.
The results of sea testing and sample analysis support the absence of contamination
from the Midwater Oil Sampler itself. However, further trials in midwater plumes with
higher concentrations of oil are necessary to confirm the intended operation.
6. Conclusions
AUV-based oil sampling complements surface- and ROV-based sampling methods.
AUV fluorometer surveys alone provide situational awareness during an underwater oil
spill. With the addition of a water sampler to the AUV sensor payload, spill investigations
can return an accurate inventory of oil constituents in a shorter time and with less effort
compared to other midwater sampling methods.
The Midwater Oil Sampler was designed to minimize contamination and expedite the
handling of seawater samples. The sampler was integrated with an AUV and deployed in
open-ocean natural oil and gas seeps near Santa Barbara, USA. By collecting targeted seawater
samples only when oil was detected via fluorescence, the time and expense of analysis was
incurred only when samples contained valuable information about the oil plume.

J. Mar. Sci. Eng. 2022,10, 526 11 of 13
Author Contributions:
Conceptualization, D.G.-I. and A.L.K.; Methodology, D.G.-I., A.L.K., A.B., R.N.C.
and D.S.; Software, D.G.-I.; Validation, D.G.-I., A.B. and D.S.; Supervision, A.L.K. and D.G.-I.; Writing—
original draft, D.G.-I.; Writing—review, A.L.K., D.G.-I., D.S., L.D. and R.N.C.; Funding Acquisition,
A.L.K., L.D. and R.N.C. All authors have read and agreed to the published version of the manuscript.
Funding:
This work was supported by the United States Bureau of Safety and Environmental
Enforcement under contract number E18PG00001.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: The data presented in this study are available in article.
Acknowledgments:
We thank Liam Cross for help with the assembly, testing, and operation of
samplers; Kevin DuCharme and Erin Fischell for the REMUS software support; Sean Whelan and
Noa Yoder for help with the AUV operations; Captain and crew of the USCGC George Cobb for
at-sea support. D.G.-I. acknowledges support during writing and editing from the U.S. Norway
Fulbright Foundation and Oslo Metropolitan University.
Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the
design of the water sampler; in the collection, analyses, or interpretation of data; in the writing of the
manuscript, or in the decision to publish the results. The views expressed in this article are those
of the authors and do not necessarily represent the views or the policies of the U.S. Environmental
Protection Agency, National Oceanographic and Atmospheric Administration, or the Bureau of Safety
and Environmental Enforcement. Mention of the trade names of commercial products does not
constitute endorsement or recommendation for use.
Appendix A
Results of the sample water chemical analysis are included in Table A1.
Bottle ID number 190826-01 was the deionized water (DI) blank, which was filled
from the Santa Barbara distilled water supply but never loaded into a sampler. Bottles
190827-006-005, 190827-006-006, and 190829-004-005 were trip blanks; these were loaded
into a sampling module and travelled with the AUV on several missions before being
offloaded and analyzed with the other samples.
Table A1. Results of water hydrocarbon analysis.
Collection Date Sample
Identifier
Collection
Condition Sampler/Bottle Depth
(m)
∑BTEX
(ng/L)
∑Alkanes
(µg/L)
∑PAHs
(µg/L)
26 August 2019 190826-01 DI Blank n/a n/a 385.53 1.11 0.32
27 August 2019 190827-004-001 Mission 5b 4/1 11.2 255.91 1.22 0.29
27 August 2019 190827-004-002 Mission 6 4/2 9.7 192.80 1.36 0.52
27 August 2019 190827-004-003 Mission 6 4/3 14.1 223.89 1.33 0.35
27 August 2019 190827-004-004 Mission 7 4/4 11.1 92.36 1.25 0.21
27 August 2019 190827-004-005 Mission 7 4/5 11.8 96.65 1.93 0.33
27 August 2019 190827-004-006 Mission 7 4/6 12.5 147.74 2.49 0.26
27 August 2019 190827-006-001 Mission 7 6/1 11.4 45.33 2.49 0.23
27 August 2019 190827-006-002 Mission 7 6/2 20.2 147.86 2.01 0.26
27 August 2019 190827-006-003 Mission 7 6/3 20.6 123.76 1.90 0.16
27 August 2019 190827-006-004 Mission 8 6/4 13.8 136.85 0.66 0.14
27 August 2019 190827-006-005 Trip Blank 6/5 n/a 259.31 2.46 0.33
27 August 2019 190827-006-006 Trip Blank 6/6 n/a 243.27 0.98 0.39

J. Mar. Sci. Eng. 2022,10, 526 12 of 13
Table A1. Cont.
Collection Date Sample
Identifier
Collection
Condition Sampler/Bottle Depth
(m)
∑BTEX
(ng/L)
∑Alkanes
(µg/L)
∑PAHs
(µg/L)
29 August 2019 190829-004-001 Mission 16 4/1 9.0 176.36 0.70 0.20
29 August 2019 190829-004-003 Mission 17 4/3 6.2 222.97 0.77 0.28
29 August 2019 190829-004-005 Trip Blank 4/5 n/a 256.20 1.28 0.28
29 August 2019 190829-011-001 Mission 18 11/1 11.4 163.93 0.85 0.15
6 September 2019
190906-LRB Lab Reagent Blank n/a n/a 32.96 0.53 0.33
References
1.
Reddy, C.M.; Arey, J.S.; Seewald, J.S.; Sylva, S.P.; Lemkau, K.L.; Nelson, R.K.; Carmichael, C.A.; McIntyre, C.P.; Fenwick, J.;
Ventura, G.T.; et al. Composition and Fate of Gas and Oil Released to the Water Column during the Deepwater Horizon Oil Spill.
Proc. Natl. Acad. Sci. USA 2012,109, 20229–20234. [CrossRef] [PubMed]
2.
White, H.K.; Conmy, R.N.; MacDonald, I.R.; Reddy, C.M. Methods of Oil Detection in Response to the Deepwater Horizon Oil
Spill. Oceanography 2016,29, 76–87. [CrossRef]
3.
Solberg, A.H.S.; Storvik, G.; Solberg, R.; Volden, E. Automatic Detection of Oil Spills in ERS SAR Images. IEEE Trans. Geosci.
Remote Sens. 1999,37, 1916–1924. [CrossRef]
4.
Zhang, Y.; McEwen, R.S.; Ryan, J.P.; Bellingham, J.G.; Thomas, H.; Thompson, C.H.; Rienecker, E. A Peak-Capture Algorithm
Used on an Autonomous Underwater Vehicle in the 2010 Gulf of Mexico Oil Spill Response Scientific Survey. J. Field Robot.
2011
,
28, 484–496. [CrossRef]
5.
Fuller, C.B.; Bonner, J.S.; Kelly, F.; Page, C.A.; Ojo, T. Realtime geo-referenced detection of dispersed oil plumes. Int. Oil Spill Conf.
Proc. 2005,2005, 693–696. [CrossRef]
6.
Walsh, I.D.; Koegler, J.; Koch, C.; Mitchell, T.O. Robust Tools to Detect Subsurface Oil: Fluorescence Sensors. In Proceedings of the
OCEANS 2015—TS/IEEE Washington, Washington, DC, USA, 19–22 October 2015; pp. 1–6.
7.
U.S. Environmental Protection Agency. Test Method 8270D: Semivolatile Organic Compounds by Gas Chromatography/Mass Spectrome-
try (GC-MS); Update V to the Third Edition of the Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, EPA
Publication SW-846; U.S. Environmental Protection Agency: Washington, DC, USA, 2014.
8.
Wilde, F.D.; Skrobialowski, S.C. US Geological Survey Protocol for Sample Collection in Response to the Deepwater Horizon Oil Spill,
Gulf of Mexico, 2010; U.S. Department of the Interior; U.S. Geological Survey: Reston, VA, USA, 2011; p. 178.
9.
U.S. Environmental Protection Agency. Test Method 8000D: Determinative Chromatographic Separations; Update VI to the Third
Edition of the Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, EPA Publication SW-846; U.S. Environmental
Protection Agency: Washington, DC, USA, 2018.
10.
U. S. Environmental Protection Agency EPA Region 9 Sample Container and Preservation List. Available online: https://www.
epa.gov/aboutepa/epa-region-9-sample-container-and-preservation-list (accessed on 10 October 2021).
11.
Klump, J.V.; Paddock, R.; Lovalvo, D. A Multiple-Loop, ROV Controlled, In-Situ Water Sampler. J. Great Lakes Res.
1992
,18,
309–315. [CrossRef]
12.
Di Meo, C.A.; Wakefield, J.R.; Cary, S.C. A New Device for Sampling Small Volumes of Water from Marine Micro-Environments.
Deep. Sea Res. Part I Oceanogr. Res. Papers 1999,46, 1279–1287. [CrossRef]
13.
Seewald, J.S.; Doherty, K.W.; Hammar, T.R.; Liberatore, S.P. A New Gas-Tight Isobaric Sampler for Hydrothermal Fluids. Deep.
Sea Res. Part I Oceanogr. Res. Pap. 2002,49, 189–196. [CrossRef]
14.
Dodd, P.A.; Price, M.R.; Heywood, K.J.; Pebody, M. Collection of Water Samples from an Autonomous Underwater Vehicle for
Tracer Analysis. J. Atmos. Ocean. Technol. 2006,23, 1759–1767. [CrossRef]
15.
Dowdeswell, J.A.; Evans, J.; Mugford, R.; Griffiths, G.; McPhail, S.D.; Millard, N.; Stevenson, P.; Brandon, M.A.; Banks, C.;
Heywood, K.J.; et al. Autonomous Underwater Vehicles (AUVs) and Investigations of the Ice–Ocean Interface in Antarctic and
Arctic Waters. J. Glaciol. 2008,54, 661–672. [CrossRef]
16.
Breier, J.A.; Rauch, C.G.; McCartney, K.; Toner, B.M.; Fakra, S.C.; White, S.N.; German, C.R. A Suspended-Particle Rosette
Multi-Sampler for Discrete Biogeochemical Sampling in Low-Particle-Density Waters. Deep. Sea Res. Part I Oceanogr. Res. Pap.
2009,56, 1579–1589. [CrossRef]
17.
Bird, L.E.; Sherman, A.; Ryan, J. Development of an Active, Large Volume, Discrete Seawater Sampler for Autonomous
Underwater Vehicles. In Proceedings of the OCEANS 2007, Vancouver, BC, Canada, 29 September 2007; pp. 1–5.
18.
Pennington, J.T.; Blum, M.; Chavez, F.P. Seawater Sampling by an Autonomous Underwater Vehicle: “Gulper” Sample Validation
for Nitrate, Chlorophyll, Phytoplankton, and Primary Production. Limnol. Oceanogr. Methods 2016,14, 14–23. [CrossRef]
19.
Fiester, C.; Gomez-Ibanez, D.; Grund, M.; Purcell, M.; Jaffre, F.; Forrester, N.; Austin, T.; Stokey, R. A Modular, Compact, and
Efficient Next Generation REMUS 600 AUV. In Proceedings of the OCEANS 2019—Marseille, Marseille, France, 17–20 June 2019;
pp. 1–6.

J. Mar. Sci. Eng. 2022,10, 526 13 of 13
20.
U.S. Environmental Protection Agency, Office of Emergency and Remedial Response. Specifications and Guidance for Contaminant-
Free Sample Containers, Directive 9240.0-05A; U.S. Environmental Protection Agency: Washington, DC, USA, 1992; p. 28.
21.
DiPinto, L. Three-Dimensional Mapping of Dissolved Hydrocarbons and Oil Droplets Using a REMUS-600 Autonomous
Underwater Vehicle. Available online: https://www.bsee.gov/sites/bsee.gov/files/research-reports//1100aa.pdf (accessed on
25 February 2022).
22.
Prakash, B.; Zaffiro, A.D.; Zimmerman, M.; Munch, D.J.; Pepich, B.V. Method 524.3: Measurement of Purgeable Organic Compounds in
Water by Capillary Column Gas Chromatography/Mass Spectrometry, Version 1.0; U.S. Environmental Protection Agency, Office of
Ground Water and Drinking Water: Cincinnati, OH, USA, 2009.
23.
U.S. Environmental Protection Agency. Test Method 3500C: Organic Extraction and Sample Preparation, Final Update IV to the Third
Edition of the Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, EPA Publication SW-846; U.S. Environmental
Protection Agency: Washington, DC, USA, 2007.