Good Practices for In-situ Burning of Marshes Based on Two Decades of Responses in Louisiana

Louisiana is rich in two types of resources: wetlands and oil exploration, production and transport facilities. Coastal wetlands cover over 3 million acres in Louisiana. The state has at least 87,764 miles of onshore pipelines (Louisiana Mid-Continent Oil and Gas Association, undated). Oil spills from pipelines that affect the interior of wetlands pose significant challenges for response, with the potential for removal efforts to cause addition damage to the habitat. In-situ burning is often considered to be the best response option; it can remove a large fraction of the oil while minimizing the amount of foot and vessel traffic on the oiled marsh surface. In this paper, we summarize the results of eight burns of oiled marshes in Louisiana over the period 2000 to 2019 for which some data on the impacts and recovery are available. The objective is to use the lessons learned from these past spills to develop best practices for future burns in wetlands. Figure 1 shows the locations of the burns, listed in Table 1, that are discussed.

The spill resulted from an explosive rupture in a pipeline carrying condensate; the amount of condensate released was not determined. The February 2000 burn was initiated by the landowner three days after the spill was discovered (the release could have occurred as much as five days prior to the burn). The soils were saturated, and there was 0.4–0.8 inches of water on the marsh surface. Fire breaks were created by laying down the vegetation using an airboat. The wind shifted during the burn, thus the burn extended both north and south of the oiled area. Although the oiled area covered 13 acres, the burned area extended to 135 acres. The burn removed all of the oil on the surface (there was no burn residue); however, it did not reduce the toxic effects of the condensate on the marsh vegetation for the 3–5 days prior to the burn. The vegetation in oiled/burned and unoiled/unburned areas was monitored for two years, using metrics of percent cover and stem density by species. Saltwort (Batis maritima) in oiled and burned areas recovered in 2 years; sea oxeye daisy (Borrichia frutesens), saltmeadow cordgrass (Spartina patens), and saltgrass (Distichlis spicata) had not recovered by year 2 but had started recovery (Figure 2).

This spill of at least 1,000 barrels (bbl) of condensate in the marsh interior was burned seven and eight days after the release in April 2001. It was surprising that the condensate did not evaporate over that period; there was still a layer of oil up to 1.5 inches thick in pools on the marsh surface. Water levels were 0–4 inches. The dense marsh canopy likely slowed evaporation. Two areas were burned. Once again, the burned area (98 acres) greatly exceeded the oiled area (12 acres), with the fire readily passing over the “firebreak” created by laying down the vegetation by airboat (Figure 3). There was no burn residue. However, some of the oil penetrated into crab burrows during low tide; the burn did not remove the oil in the burrows, so it was necessary to deploy sorbents to recover this oil as it was remobilized by the tides. The marsh was dominated by S. patens, smooth cordgrass (S. alterniflora), and D. spicata. The vegetation died completely in the areas of heaviest oil accumulations; thus, again, the burn did not reduce the toxic effects of a light condensate oil on the marsh surface for the 7–8 days prior to the burn. The site was surveyed in October 2001, six month after the burn. The lightly oil/burned vegetation had recovered, and the total petroleum hydrocarbons (TPH) and total polynuclear aromatic hydrocarbons (PAH) concentrations in the marsh soils were close to background except at the release site (Michel et al., 2003).

Hurricanes Katrina and Rita caused over 100 moderate to major oil spills, including the release of 100–200 bbl of South Louisiana crude oil from a Chevron tank farm on the left descending bank of the Mississippi River near Empire. The marsh was dominated by Olney's three-square bulrush (Schoenoplectus americanus) and S. patens. Merten et al. (2008) reported that the October 2005 burns were conducted one month after the release, with water levels of several inches. Airboats were used to create a firebreak, and fire teams wetted down the vegetation and had hoses to control the spread of the fire. The first burn was very intensive and did burn outside the firebreak in one area. However, the fire started to burn out once it burned beyond the area of pooled oil and heavily oiled vegetation. The burn on the second day was less intense because there was less oil and higher water levels. There was some burn residue on the surface (which was not enough to require removal), and some liquid oil seeping out of the soil after the burn. Therefore, sorbents were deployed to recover remobilized oil. Chevron funded a monitoring program. Baustian et al. (2010) documented excellent vegetative recovery after one full growing season (within nine months) using the following metrics: species composition, stem density, above- and belowground productivity, marsh resiliency, soil chemistry, soil residual oil, and organic matter decomposition. Figure 4 shows photographs of unburned and burned study plots.

The June 2014 burn was conducted 7 and 8 days after a 100 bbl release of South Louisiana crude oil into a Roseau cane (Phragmites australis, common reed) tidal freshwater marsh (water levels 1–2 feet) on the Mississippi River delta. The initial response involved skimming and sorbents, which required cutting pathways into the dense vegetation to access the oil. The fire burned out once it reached the end of the heavier oiling; of the 15 acres that were oiled, 5 acres burned. Follow-up operations after the burn included oiled vegetation cutting and debris removal, low-pressure flushing, herding, sorbent use, and skimming over approximately six weeks in burned and unburned areas. The Texas Petroleum Investment Company funded a three-year (2014–2016) study of the vegetation (percent cover, stem density, canopy height, and plant species composition), oil distribution, and sediment chemistry (Zengel et al. 2018, summarized below).

The burn was effective in rapidly removing ~80–90% of the floating and stranded oil from the marsh and reduced residual oiling on the marsh vegetation. No burn residues were observed. Oil concentrations in marsh soils were initially elevated in the oiled-and-burned sites; total polynuclear aromatic hydrocarbons (PAH) averaged 6 parts per million (ppm), likely due to the degree of oiling and other response activity rather than the burn alone, but were similar to reference and below background conditions by three months post-burn (<1 ppm). Oiling and response disturbance had little effect on the vegetation in sites that were not burned; however, heavier oiling in the burned area did not allow differentiating between oiling and burning effects. The combination of oiling and burning drastically reduced the vegetation as observed immediately after the burn. The originally dominant Phragmites, a non-native introduced variety, did not totally recover in terms of plant cover or stem density by two years post-spill (2016). However, the return of total plant cover (all species combined) was relatively rapid, reaching reference levels by three months. In addition, oiling and burning resulted in a mixed vegetation assemblage, including the native plant species broadleaf arrowhead (Sagittaria latifolia), delta arrowhead (Sagittaria platyphylla), pickerelweed (Pontederia cordata), and wild rice (Zizania aquatica) rather than an immediate return to dominance by Phragmites (based on relative plant cover in 2016, Figure 5). The mixed native marsh species in oiled and burned areas were viewed positively in terms of waterfowl habitat quality on the National Wildlife Refuge, compared to areas strongly dominated by introduced Phragmites. It was thought that Phragmites would continue to increase and eventually re-dominate the burned areas over several years. Overall, in-situ marsh burning was considered effective for this spill and resulted in enhanced vegetation conditions for at least several years after the burn.

The November 2017 spill of 30 bbl of South Louisiana crude from a pipeline spread to three acres in an intermediate marsh. Initial response operations involved skimming, manual removal, and use of sorbents, which resulted in extensive disturbance to the marsh surface, as shown in Figure 6. To prevent further damage, the site was burned 10 days after the release. Two of the three oiled acres burned, with an estimated 90–95% removal rate. Water levels were 0–6 inches. A fire team monitored the burn area throughout with two standby firefighting vessels. The fire team extinguished pockets of fire around the perimeter of the affected area and monitored small fires burning prior to extinguish all hot spots with the aid of a Forward Looking Infrared camera prior to sunset. Residual oil and oiled vegetation were removed mid-December, and sorbents were deployed for several months to recover any remobilized oil.

The heaviest residual oiling of sediments was in an area of about 3,000 square feet centered around the release site, where the sediments were oiled to 0.5–1 foot below the surface. The site was visited multiple times after the burn, where vegetation regrowth was observed in most areas, except around the release site, where the sediments contained an average oil content of 5,000 ppm TPH. Figure 7 shows the status of the vegetation around the release site as of 22 June 2018, seven months after the burn. There was little-to-no vegetative recovery at the release site (heavily oiled soils) and in areas heavily trampled prior to the burn.

The December 2017 spill of 50 bbl of South Louisiana crude oil from a pipeline below a floating intermediate marsh affected 0.5 acres, with oil trapped under the floating marsh. The marsh vegetation was dominated by spikerush (Eleocharis sp.), bulrush (Schoenoplectus spp.), and S. patens, with other common marsh species including bull tongue (Sagittaria lancifolia), cattails (Typha sp.), and needlerush (Juncus sp.), among others. The soils were saturated. The site was burned nine days after the release, with the burn extending to 3.3 acres (the burned area was 6–7x the oiled area). Figure 8 shows the extent of the oiled and burned areas. Post-burn operations included periodic hand raking of the heavily oiled marsh surface to break up, remove, and enhance weathering of remaining surface oil and oiled vegetation mats at the release site.

Monitoring of the site was conducted in January and September 2018 and November 2019 to date. As of 2019, the monitoring results indicate that the burn did not effectively remove all the released oil near the immediate source, probably due to liquid oil trapped under the floating marsh, as well as more weathered oil mixed with sediments in and under the matted vegetation. Oiling conditions were improving in the heavily oiled sites across the monitoring periods, likely aided by periodic raking and the colonizing vegetation. Burning greatly reduced plant cover and height, particularly in the heavily oiled sites, but also in the unoiled and burned sites. Nine months post-burn, plant cover and height had recovered in the unoiled and burned sites, with greater evenness of dominant species compared to the reference sites. Nine and 23 months post-burn, vegetation was recovering in the oiled and burned sites with new growth coming in from the margins of the most heavily oiled area around the release site, colonizing all the oiled monitoring plots by 23 months though the central part of the release site was still unvegetated. In the heavily oiled and burned area, Schoenoplectus spp. was usually the single dominant plant species, though by 23 months most other species typical of the marsh were also found in at least some of the oiled stations. In contrast, the reference and unoiled/burned sites were typically dominated by Eleocharis or Spartina patens, with Schoenoplectus being co-dominant or secondarily dominant. The reference and unoiled burned sites also generally had greater plant species richness than the oiled and burned sites (total number of species observed). Burned areas with lesser oiling further from the release site appeared qualitatively recovered in terms of vegetation structure, based on visual examination. On-going monitoring through 2020 will continue to track vegetation recovery and oiling conditions for this site.

The November 2018 spill of 40 bbl of condensate from a pipeline below a floating intermediate marsh affected 0.3 acres. The marsh was saturated with the oil. The burn was conducted four days later and remained within the footprint of the oiled area (Figure 9). The fire burned for approximately three hours. A post-burn site assessment reported that all the crude condensate had burnt off (no residues).

This location consisted of two separate but adjacent sites: 1) a recent site of oil with 0.5–2 inches of water and oil on the marsh surface and thick vegetation where South Louisiana Crude oil was released from a flow line; and 2) an older release (+1 month prior) of South Louisiana crude oil and produced water with about 1 inch of water and oil from a produced water disposal line, where the vegetation had died back from the high salinity of the produced water (the dead vegetation at the produced water release site is visible in Figure 10). Water levels were 0.75 inches over the marsh surface.

Both sites were burned on 21 August 2019, six days after the release was reported. The burn at Site 1 extended only to the oiled perimeter, and there was some burn residue, thus raking was recommended to break up the remaining oil and oiled vegetation (Figure 11). The burn at Site 2 removed only the floating oil on the water surface. However, the heavily oiled soils continued to be a source of remobilized oil. Further assessments were required to identify the extent and degree of subsurface oil to determine appropriate treatment options.

Lessons learned from the use of in-situ burning of oiled marshes are summarized in Table 2 and cross-walked with the eight case studies discussed above. These lessons learned reinforce the results of other in-situ burning in marsh case studies (Michel et al., 2002, 2003; Zengel et al., 2003; Michel and Rutherford, 2013).

Responders should consider use of in-situ burning of oiled marshes under these conditions:

• When rapid removal of oil on the marsh or water surface (not just on the vegetation) is required to:

  • Reduce exposure of wildlife to oiling

  • Keep oil from spreading more widely or to other sensitive areas

  • Avoid trampling of soils/mixing oil into soils during manual or mechanical removal

• For difficult-to-access areas where logistics and waste generation would be challenging under typical response methods

• When the spill site is remote, sparsely populated, and distant from human infrastructure

• When winds are less than 12 miles per hour, and preferably lower (API, 2015)

• For light to medium oils (e.g., light refined products other than gasoline; light to medium crude oils)

• When the oil is relatively fresh and has not formed a stable emulsion

• When the vegetation is mostly herbaceous (grasses, sedges, rushes), rather than woody

• When vegetation is dormant (in fall and winter)

• When marsh soils are covered by a layer of water or are water-saturated

• When the burned vegetation will not be flooded post-burn or re-oiled

When use of in-situ burning of oiled marshes may not be appropriate:

• Where air quality and/or public health and safety concerns are high due to smoke and or burns that could escape control (consider sensitive populations, roads, waterways, etc.)

• Where there is a high risk of an uncontrolled burn

  • Strong winds

  • Ineffective fire break/fire control

  • Other risks of severe wildfire, including in adjacent habitats (drought, dry conditions, high fuel loads)

• When the oil will not burn

  • Oil is emulsified and will not sustain ignition

  • Oil has penetrated into the soils, with minimal surface oil

• If the oil type/conditions will leave a large amount of burn residue

• If there is a risk of impact on protected species in the marsh (e.g., endangered birds mammals).

We want to thank the many responders and agencies who shared their photographs, videos, and on-site observations for the in-situ burns included in this paper: Brandi Todd and Ololade Ajilore, Office of Response and Restoration, National Oceanic and Atmospheric Administration; Louisiana Oil Spill Coordination Office; U.S. Coast Guard; Chris Sanfilippo, Texas Petroleum Investment Company; Jim Elliott, Teichman Group, LLC; Forefront Emergency Management; and Delta National Wildlife Refuge.