NOFO (Norwegian Clean Seas Association for Operating Companies) and NCA (Norwegian Coastal Administration) are cooperating closely to operationalize in situ burning as a response method in Norwegian waters. After introductory experiments during Oil on water 2016 together with S. L. Ross and SINTEF, NOFO and NCA have conducted in situ burn experiments in fire booms during Oil on water 2018 and 2019. These experiments included use of net to capture burn residue, igniters to start the fire when dropped from a multicopter drone, use of sensor packages from multicopter drones to monitor fire gases and soot in various parts of the smoke plume, and finally personnel in work boats taking surface samples and operating sensors to identify gases that might be hazardous to people. Both crude oil and fuel oils were burned, approximately 6 m3 for each burn. To our knowledge such field experiments have not been carried out before. The presentation will focus on operational aspects including photos and video footage, but also refer some findings from sensors and samples that might be hazardous to the environment as well as personnel.

Annual Norwegian oil on water (OOW) experiments are planned and conducted by NOFO (the Norwegian Clean Seas Association for Operating Companies) in close cooperation with NCA (the Norwegian Coastal Administration).

As a governmental agency and responder for the operating companies in Norway both NCA and NOFO, respectively, are in the process of evaluating in situ burning (ISB) as a response method to combat acute oil spills. ISB is first of all under consideration for oil spills in ice, but additional complexity and costs associated with field campaigns in ice determined initial experiments in open water. The first in situ burn experiment at OOW was done in 2016 where IOGP (The International Association of Oil & Gas Producers) was invited to conduct a herder experiment when burning uncontained fresh crude oil (Singsaas et al., 2017). This became a flying start to our ISB program, but in addition to this, it also revealed both differences and similarities regarding spill response in North America and Norway.

The experiment described in this paper was planned for the summer of 2018, but unfavorable weather conditions forced us to finalize the rest of the burns in 2019. The objectives with the ISB experiment reflect the need to document environmental impact, personnel hazard and to develop operational tactics for in situ burning:

  • verify ignition of various oil types (also emulsion) using an igniter operated from a drone

  • collect the residue from ISB and estimate burning efficiency

  • document release of exhaust gases and soot (Black Carbon) in the smoke plume including environmental toxicity measurements

  • document potential downfall from the smoke plume on the sea surface

  • monitor hazardous substances related to ISB in order to identify necessary safety precautions for personnel taking part in or being close to such activities

  • obtain necessary experience to operationalize ISB as an accepted spill response method

Test oils

One crude oil and three fuel oils were selected for a total of six burns:

  • Oseberg Blend (crude oil), ~ 30% evaporated, one out of three burns with emulsion

  • ULSFO (Ultra Low Sulphur Fuel Oil)

  • MGO (Marine Gas Oil)

  • IFO 180 (Intermediate Fuel Oil)

Fire booms with nets to recover burn residue

Two different types of fire boom were used, PyroBoom (2018, 2019) and American Fire Boom (2019). Both boom types were used during the Macondo oil spill in 2010. For the field experiment in 2018 a net was designed to collect the residue from each burn. This net had a mesh size of about 2 mm (heavy-duty mosquito net) and was attached to the inside of the PyroBoom. After some full-size testing without oil and followed by modifications, the final design of the net was fixed, where after six nets were prepared for the burn experiment.

The entire net had to be submerged to avoid being destroyed by the fire. Hence, the net was fixed to the fire boom with sections of heavy twine below the waterline. The front end of the net had weights to keep it submerged, and stainless-steel floats to keep it at a preselected depth when towing (Figure 1). Using the nets together with American Fireboom in 2019 required only minor modifications to the fire boom.

Figure 1.

Net for capturing burn residue attached to PyroBoom during preparations.

Figure 1.

Net for capturing burn residue attached to PyroBoom during preparations.

Close modal

Vessel and mode of operation

An anchor handling vessel, Strilborg, was used for all the burns. This vessel is 75 m long, has accommodation for about 30 persons, operates with 2 MOB/work boats, and has a main deck well suited for this type of operation. In 2018 the fire boom was spread out by using a flexible paravane (surface kite), which require a certain minimum speed through the water to get sufficient lifting force. In 2019 the paravane was skipped by having the ship tow the fire booms by sailing sideways through the water (Figure 2). This change made it possible to reduce the towing speed, and at the same time the fire booms were operating at the leeward side of the vessel, reducing both wind and waves for the boom and the fire.

Figure 2.

Strilborg towing the boom with paravane (left), and towing sideways (right).

Figure 2.

Strilborg towing the boom with paravane (left), and towing sideways (right).

Close modal

Spill permit and weather limitations

Like anywhere else there is a need for a spill permit when we conduct an oil-on-water experiment. Preferred wind speed is typically within 5 m/s for an ISB experiment, but conditions could be better or worse depending on the wind speed increasing, being constant over some time or decreasing. Consequently, we applied for and got a permit to conduct the burns at a maximum of 8 m/s wind. This gave us flexibility on site to consider whether conditions were acceptable for each burn. As a result, some burns were done at more than 5 m/s wind speed. Also, burns should not be carried out if it was raining. However, after igniting one of the burns, light rain started without being forecasted. Data collection from the drones were terminated due to the rain, giving less data than expected.

Fire boom deployment and operation

Prior to every deployment the fire boom(s) were spread out on the main deck to connect the net for residue, and floats were attached at the front of the net. This means that the middle part of the fire booms went overboard first (Figure 3), and the starboard and port side of the booms went out in parallel. In 2018 the paravane was connected to the starboard towline right in front of the boom, in 2019 the entire length of towlines was reeled out before the vessel turned 90 degree starboard where after the starboard towline was moved forward and the vessel could start towing sideways.

Figure 3.

Net attached to the PyroBoom being towed off the deck with a sea anchor (left). Storage tanks for different test oils are seen on both sides of the deck.

Figure 3.

Net attached to the PyroBoom being towed off the deck with a sea anchor (left). Storage tanks for different test oils are seen on both sides of the deck.

Close modal

After stabilizing the formation and speed of the vessel, the test oil was deployed through a floating hose and weir outlet in front of the fire boom. Keeping the weir outlet in right position sometimes proved to be a challenge, especially when wind and current had a different heading. When the test oil had been deployed, the overall request to the captain was to maintain the formation and speed through the water at a steady state during the burn.

Deploying test oil

The test oils were stored in containers on deck of the vessel. When preparing for a burn, the amount of oil to be used was transferred to a 10 m3 calibrated daytank connected to a pump and release equipment. A twin daytank was filled with seawater for flushing out oil from the release hose after the test oil was pumped. All the storage tanks, pumps and release equipment were connected before leaving harbor, hence all the transfer from storage tanks to sea surface could be done without connecting and disconnecting any hoses, minimizing unintended spills on board the vessel.

From the daytank the oil went through the transfer pump followed by a lightweight 3” diameter flat hose to a weir releasing the oil very gently in front of the boom (Figure 4). The weir outlet was kept in position by a workboat at all time during the release.

Figure 4.

Releasing test oil into the American Fireboom.

Figure 4.

Releasing test oil into the American Fireboom.

Close modal

Igniting oil in the fire boom

NOFO and NCA had sponsored the development of a new concept for igniting a burn, which included a new igniter containing 1 liter of gelled fuel as well as a PyroDrone (multicopter) to position and start the igniter in the oil. This field trial allowed us to verify the functionality in the field. With the onboard camera transferring live video to the drone pilot and a large screen, the PyroDrone flew from the vessel at 15–20 meters altitude, hoovered right above the oil in the boom for a vertical photo to estimate area coverage, where after it went down to 3–4 meters, armed the igniter and dropped it at the front of the slick (Figure 5). After this main task for the PyroDrone was completed, the drone was used extensively during all the burns to take photos and video footage from various positions.

Figure 5.

Igniter dropped from the PyroDrone to start the fire in the PyroBoom.

Figure 5.

Igniter dropped from the PyroDrone to start the fire in the PyroBoom.

Close modal

Monitoring air quality, sampling, photo documentation

SINTEF, in cooperation with Maritime Robotics, performed an extensive monitoring of the smoke plume, using two dedicated drones (DJI Matrice 600 Pro, Figure 6) with sensors for emission gases (NOX, SO2, CO, and CO2), soot particle distribution (TSI DustTrak DRX Aerosol Monitor Model 8534) and sampling of soot particles. The DustTrak measured the particle fractions (PM) 1, 2.5, 4, and 10 μm. The soot particles were sampled in a closed-faced filter cassette with a PTFE filter (37 mm, 2.0 μm) connected to an MSA Escort Elf Pump with an airflow of 3 liters/min. In addition, samples of burned residue were collected from the sea surface. The University of Bergen measured the potential for human exposure by monitoring volatile organic compounds (Minirae 300 Photoionization detector (PID)), the particle distribution and concentration in the smoke fallout using sensors like the DustTrak also used in the drones (Szwangruber and Bråtveit, 2019). A heat flux sensor mounted on the cargo rail, the PyroDrone and two work boats were all operated from “Strilborg”. The drones monitoring the smoke were operated from the fishing vessel “Bøen”.

Figure 6.

Drone (DJI Matrice 600 Pro) equipped with monitoring instruments and sensors (two identically equipped drones were used)

Figure 6.

Drone (DJI Matrice 600 Pro) equipped with monitoring instruments and sensors (two identically equipped drones were used)

Close modal

Oil (1 liter) was sampled from the storage tanks when the oil was released on the sea surface, representing the unburned oil. Three residue samples were collected in the fire boom after the burn, located on the port side, in the middle, and on the starboard side in the boom, respectively. Viscosity, density and chemical composition including polyaromatic hydrocarbons (PAHs) were measured in the unburned oils and ISB residues.

As oil is burned, gas and smoke particulates are produced. The main objective of the drone sampling was to quantify the generation of gas and smoke particulates from the fire. To quantify the total flux of particulates and gas through a cross section of the smoke plume the following parameters were measured: The width and height of the plume, the velocity of gas and particles through the cross section (wind speed), and the concentration of particulates and gas. The drones flew in a pre-defined pattern in the smoke plume. A cross section of the plume 100 meters downwind of the fire boom was chosen as the primary target area for both drones. Another cross section 300 meters downwind was chosen as a secondary area for one of the drones. Transecting under the plume was set as a secondary priority for the other drone. The drones typically did three vertical and three horizontal transects to map the size and shape of the primary cross-sectional area.

One of the MOB-boats was used to monitor the potential for human exposure. The monitoring was performed up to 200 m from the fire.

Recovery of residue

After each burn, the net with residue was disconnected from the fire boom and transferred to the vessel as follows: Between the legs of a 6 m long U-shaped raft, the rim of a strong fishing net was attached to form a deep flexible cage which could be opened at the front end. After launching the raft from the main vessel and moving it into the fire boom, two (later three) persons on the raft disconnected the residue net section by section from the fire boom assisted by two MOB boats, one on each side of the boom. The freed parts of the net and residue were pulled into the cage, which was closed as soon as the entire net was positioned inside it. The raft was then pulled back to the ship where the flexible cage containing the fine meshed net with residue was transferred by crane to an open top container for weighing and sampling. Figure 7 illustrates this process.

Figure 7.

Disconnecting residue net from PyroBoom (left). Burn residue and residue net inside the flexible cage are lifted out of the water by the ship crane (right).

Figure 7.

Disconnecting residue net from PyroBoom (left). Burn residue and residue net inside the flexible cage are lifted out of the water by the ship crane (right).

Close modal

This section deals mainly with the operational aspects of the ISB experiment. Since analysis of the data from burn gases, soot and residue have not been completed, only an overview of some parameters is referred at the end of this section (Tables 1, 2 and 3). Annual summary reports from the oil on water experiments are available (NOFO, 2018 and NOFO, 2019).

Table 1

Experimental burns conducted during OOW in 2018 and 2019.

Experimental burns conducted during OOW in 2018 and 2019.
Experimental burns conducted during OOW in 2018 and 2019.
Table 2

Density (g/ml) and viscosity (cP) for unburned oils and their ISB residues. Density for Oseberg and IFO are measured at 80°C and recalculated to 15°C. Viscosity are at 10°C from the temperature-sweep (shear rate 10s−1 and 1°C/min).

Density (g/ml) and viscosity (cP) for unburned oils and their ISB residues. Density for Oseberg and IFO are measured at 80°C and recalculated to 15°C. Viscosity are at 10°C from the temperature-sweep (shear rate 10s−1 and 1°C/min).
Density (g/ml) and viscosity (cP) for unburned oils and their ISB residues. Density for Oseberg and IFO are measured at 80°C and recalculated to 15°C. Viscosity are at 10°C from the temperature-sweep (shear rate 10s−1 and 1°C/min).
Density (g/ml) and viscosity (cP) for unburned oils and their ISB residues. Density for Oseberg and IFO are measured at 80°C and recalculated to 15°C. Viscosity are at 10°C from the temperature-sweep (shear rate 10s−1 and 1°C/min).
Density (g/ml) and viscosity (cP) for unburned oils and their ISB residues. Density for Oseberg and IFO are measured at 80°C and recalculated to 15°C. Viscosity are at 10°C from the temperature-sweep (shear rate 10s−1 and 1°C/min).
Table 3

Estimated weight and BE after OOW 2019. Factors, such as boom leakage and residue sticking on the boom after ISB are not taken into account when calculating BE.

Estimated weight and BE after OOW 2019. Factors, such as boom leakage and residue sticking on the boom after ISB are not taken into account when calculating BE.
Estimated weight and BE after OOW 2019. Factors, such as boom leakage and residue sticking on the boom after ISB are not taken into account when calculating BE.

Oil on water 2018

With a two-week spill permit in June, it was expected to have a good chance to carry out all the five burns planned for. However, during this period there was only one single day that offered conditions within the weather window. We had two burns, the first one with Oseberg crude, the second one with ULSFO, one of the new low sulphur fuel oils on the market.

Burn #1 (2018, Oseberg crude, wind ~ 6–7 m/s)

Based on visual observations from the main vessel during and right after releasing 6.0 m3 of oil, it looked like the loss of oil from the PyroBoom (90 m long) was so great that ignition could be impossible. This made us drop the igniter as soon as possible. After such a start it was a surprise that the fire lasted for about 43 minutes, plenty of time to harvest data from the smoke plume and to take sample at the sea surface. After this first burn, the residue still in the net was all captured and recovered. However, this work was heavy and took long time. The parts of the PyroBoom that was in fire was seriously damaged. Patches of oil, some of this oil was burning, were observed from the PyroDrone behind the boom during the fire, which means that in total the oil loss was significant but could not be estimated based on visual observations only. The burning efficiency based on the ratio of weighed residue to weight of released oil would be significantly overestimated. Two sections (30 m) of the PyroBoom was disconnected and scrapped after this burn.

Burn #2 (2018, ULSFO, wind ~ 5 m/s)

With two sections gone, another residue net could not be attached to the rest of the PyroBoom. To carry out the second burn, the PyroBoom had to be used without a net, and a second vessel, OV Utvær (Figure 8), used their integrated sweeping arms during this burn to recover residue from the burn.

Figure 8.

OV Utvær with sweeping arms to recover residue from Burn #2.

Figure 8.

OV Utvær with sweeping arms to recover residue from Burn #2.

Close modal

This oil was kept heated in the storage tank to make it possible to pump, and 5.8 m3 was released into the PyroBoom. The oil covered about 50 m2 surface area when the igniter set it on fire. The oil has a high viscosity when the temperature is reduced, and there was hardly any oil loss seen from the vessel, both before and after the ignition. Only one igniter was necessary to start the fire but compared to the Oseberg crude oil it took a bit longer before the oil ignited properly. The fire lasted for about 48 minutes this time. Some residue was recovered by the vessel Utvær, but due to oil spread too much passed the mechanical recovery system to give a credible estimate of the burning efficiency. Again, the parts of the PyroBoom that was in the fire was so damaged that another two sections had to be scrapped.

In light of reports from ISB during the Macondo accident, we were surprised that the PyroBoom that had been in a fire for less than one hour was so seriously damaged. On the other hand, the manufacturer had warned us that deploying and recovering the PyroBoom several times for training purposes prior to the experiment, and in addition to that reeling it in and out during these operations, we might experience loss of integrity during the burns.

Oil on water 2019

On the positive side, having to postpone the rest of the burns meant tactics and equipment could be modified or replaced in the meantime. American Fireboom and more PyroBoom were purchased for the Oil on water 2019, and this time we only used brand new sections from both types of boom. Again, we got a spill permit for two weeks, and this time the weather conditions were favorable for three whole days, sufficient to do all five releases for in situ burns.

Burn #3 (2019, Oseberg Blend, American Fireboom, wind ~ 4 m/s)

This was the first burn in 2019, with the same type and amount of oil as the year before, but with American Fireboom and the same residue net that was designed for the PyroBoom. A representative from the manufacturer of the fire boom was actively taking part in the operations of the boom, both training prior to the experiment and during burning. Attaching the net to another fire boom proved to work well, so did the operation. 6.0 m3 of Oseberg Blend was released into the boom. From the vertical photos, the surface area of the oil just before igniting was approximately 55 m2, making an average thickness of nearly 110 mm. The burn lasted for 63 minutes. By towing the fire boom sideways, reducing the length of the towlines and having a third person on the raft, the process of disconnecting the net with residue and bringing it on board the main vessel was improved considerably. Towing sideways also allowed us to tow at a lower speed than in 2018 without having the boom formation collapse, giving us better control and reduced oil leakage. The parts of the American Fireboom that had been in the fire were brittle and damaged. After bringing the boom on board the vessel to attach a new net for the next burn, it was obvious that these parts had to be replaced.

Burn #4 (2019, IFO 180, American Fireboom, wind ~ 4 m/s)

The oil for the next burn was an Intermediate Fuel Oil 180. No significant oil loss was observed. The volume released was 4.2 m3 resulting in a surface area of about 25 m2 and an average thickness of nearly 170 mm. The burn was successful with a 37 minutes burn time. The residue was very viscous, had high density and was extremely sticky. Disconnecting the net with residue from the fire boom was difficult and time consuming, and it was evident that lots of water became encapsulated in the residue. Since the encapsulated water was impossible to remove, the weight of the residue was too high, making the burning efficiency underestimated. On the other hand, some loss of residue due to difficulties in collecting and keeping the residue within the net, will increase the estimated burn efficiency. In total we believe the burn efficiency is underestimated.

Burn #5 (2019, Oseberg Blend, PyroBoom, wind ~ 4 m/s)

Burn #5 was done to repeat Burn #1 (2018) with a PyroBoom that had not been used, everything else being the same, with the expectation that the PyroBoom this time would suffer less damage in the fire. 5.6 m3 of Oseberg Blend was released, the surface area just before the burn was about 110 m2, and the average thickness about 50 mm. The oil was easily ignited, and the burn was good. With such area to burn, the burn time was 44 minutes, and the resulting burning efficiency was estimated to 91 % without accounting for any oil losses. The parts of the PyroBoom that had been in the fire had less damage than in Burn #1, but still these parts had to be scrapped after the burn.

Burn #6 (2019, MGO, American Fireboom, wind ~ 6–7 m/s)

This burn was with Marine Gas Oil (MGO), which has low density and very low viscosity. With such properties we expected to see a significant oil loss as soon as the oil was released. The surface area of the oil prior to ignition was approximately 115 m2, giving slightly more than 50 mm average thickness since the volume of oil was 6.0 m3. The oil was effectively held into place by the American Fireboom, it was easy to ignite, and the burn lasted for about 28 minutes. For all the burns this year, the distance between the main vessel and the fire was about 100 m, and the MGO was the only oil that gave heat radiation that could be felt on the face when watching from the bridge wing of Strilborg. A handheld infrared sensor indicated burn temperatures above 1000 degrees Celsius. Only minor quantities of residue were left after the burn, estimating the burning efficiency above 95%.

Burn #7 (2019, Oseberg Blend emulsion, American Fireboom, wind ~ 6 m/s)

This release was the last one, and the oil was a water in oil emulsion from Oseberg Blend with 52% water content. A 4-liter sample of this emulsion was easy to burn at ideal conditions outdoors at the local fire department prior to the offshore experiment. 6.0 m3 emulsion was released in the American Fireboom, and covering a 50 m2 surface area the average thickness was about 120 mm. With similar weather conditions as in the previous burns, we did not succeed in maintaining the fire for more than some minutes, even with manual igniters and supplying additional igniter fuel directly onto the emulsion. The burn was unsuccessful, and the emulsion was combatted by dispersant application from nozzles at the bow of Strilborg, guided by the PyroDrone and finally by the Norwegian surveillance aircraft.

Net to recover residue

To be able to obtain a spill permit for an ISB experiment it was necessary to recover the burn residue. A lot of effort was used to develop a method to recover the residue from the burns. The resulting net had to be tested and modified several times, and several days of waiting for better weather conditions was spent training in protected waters. Based on visual observation, practically all the residue collected in the nets during the fires was recovered by disconnecting the net and bringing it on board the ship. Collecting residue that was released after the burn without a net proved less successful.

PyroDrone with igniter

Using the PyroDrone with the igniter was highly successful during these experimental trials. It is fast, precise in locating the ignitable oil, and can improve the safety of personnel. The advantage will be even better in ice covered waters where a workboat may have difficulty in accessing the oil. With the present size and shape of the igniter, the amount of fuel could be increased to three times as much (3 liters) by making a minor modification. With a somewhat reduced flying time due to the extra weight, the PyroDrone will still be able to operate such an igniter.

Burn efficiency (BE)

All estimates of burning efficiency (BE) above are probably overestimated since the loss of oil (prior to and during burning) have not been included. There is a clear understanding that BE depends on many factors, implying that BE could differ from one burn to another even if everything seems to be the same. Whether the residue will sink or not also depends on the BE: a higher burning efficiency leads to higher density of the residue, which means that the residue could sink or float depending on how successful the burn in question is, along with environmental conditions and oil type.

Evaluating fire booms

During all the burns conducted spreading out the fire boom using either paravane or towing sideways, both types of fire boom suffered too much damage to be used a second time after being brought back on deck for attaching a new residue net. For our mode of operation, this experience indicated that although the robustness of these fire booms is not very different, the American Fireboom might last longer. This is a challenge both regarding costs as well as resupplying fire boom during an acute spill if you have multiple contained burns. A water-cooled fire boom is therefore an option that should be investigated. Such booms are inflated and could be stored on a reel, making it a possible solution for our vessels and our mode of operation.

Safety precautions for personnel

Some preliminary results related to safety hazards for personnel are available based on measurements on board two MOB boats operating downwind from the fire and the main vessel Strilborg about 150 meters upwind. Measured particle fractions PM<1, PM1, PM2.5, PM4 and PM10 indicate that the fraction with the ultra-fine and most dangerous particles represented the major part of the soot, independent of oil type. Certain types of filter half masks can offer acceptable protection to personnel operating near the downfall during shorter periods, provided no air leakage between the skin and the mask. Volatile organic compounds (VOC) were measured with Photoionization detector (PID) and showed concentrations far below thresholds for human concern.

Since analysis and reporting of the data gathered during the in situ burns are still underway, we only have some preliminary conclusions to offer from the experiment conducted in 2018 and 2019:

  • ✓ The igniter operated by the PyroDrone has been verified as a quick and effective tool to start an in situ burn. Especially in ice conditions, the method is an improvement over current methodology. Overall safety of personnel will also be improved by increasing the distance to the oil/burn area.

  • ✓ Burning efficiency was estimated for most of the burns, but in general the numbers are overestimated since loss of oil was only visually observed and not included.

  • ✓ This experiment has demonstrated that it is possible to recover most of the residue after contained burning using fire boom with a fine meshed net attached sub surface.

  • ✓ Both PyroBoom and American Fireboom suffered too much damage to be used for a second burn after being brought on deck for attaching a new residue net.

  • ✓ With a spill permit for a total of four weeks (2018 and 2019), only four days had acceptable wind conditions for the in situ burn experiment. This indicates that ISB probably has limited weather applicability in Norwegian offshore waters.

  • ✓ The ultra-fine and most dangerous particles (PM<1) represented the major part of the soot, independent of oil type, but the downfall was mainly concentrated to areas with visible smoke. Filter half masks give acceptable protection to personnel near the downfall during shorter periods.

  • ✓ Measurements showed extremely low concentrations of VOC.

1.
Singsaas,
I.,
Cooper
D.,
Potter
S.,
Lewis
A.,
Daling
P.S.,
Bråtveit
M.
(
2017
).
Field experiment to validate herder and in situ burning use in open water
.
HISB (Herder and In-Situ Burning) project. SINTEF report OC2017 A-034, ISBN 978-82-7174-279-9.
2.
NOFO
(
2018
):
Summary report from the oil-on-water experiments 2018 (Norwegian).
3.
NOFO
(
2019
):
Summary report from the oil-on-water experiments 2019 (Norwegian).
4.
Szwangruber,
M.,
and
Bråtveit
M.
(
2019
).
Oil on water 2019: Monitoring of particle downfall from the smoke plume during in situ burning
.
Draft report from Department of Global Public Health and Primary Care
,
The University of Bergen
,
Norway (in Norwegian)
.