by John W. Day 1,*, H. C. Clark 2, Chandong Chang 3, Rachael Hunter 4,* and Charles R. Norman 5 1 Department of Oceanography and Coastal Sciences, Louisiana State University, Baton Rouge, LA 70803, USA 2 Department of Earth, Environmental and Planetary Science, Rice University, Houston, TX 77005, USA 3 Department of Geological Sciences, Chungnam National University, Daejeon 34134, Korea 4 Comite Resources, Inc., P.O. Box 66596, Baton Rouge, LA 70896, USA 5 Charles Norman & Associates, P.O. Box 5715, Lake Charles LA 70606, USA * Authors to whom correspondence should be addressed. Water 2020, 12(5), 1492; https://doi.org/10.3390/w12051492 Received: 20 April 2020 / Revised: 18 May 2020 / Accepted: 20 May 2020 / Published: 23 May 2020 (This article belongs to the Special Issue Sustainable Management, Conservation and Restoration in Deltaic Ecosystems with Special Emphasis on the Mississippi Delta) Download PDF Browse Figures Abstract Oil and gas (O&G) activity has been pervasive in the Mississippi River Delta (MRD). Here we review the life cycle of O&G fields in the MRD focusing on the production history and resulting environmental impacts and show how cumulative impacts affect coastal ecosystems. Individual fields can last 40–60 years and most wells are in the final stages of production. Production increased rapidly reaching a peak around 1970 and then declined. Produced water lagged O&G and was generally higher during declining O&G production, making up about 70% of total liquids. Much of the wetland loss in the delta is associated with O&G activities. These have contributed in three major ways to wetland loss including alteration of surface hydrology, induced subsidence due to fluids removal and fault activation, and toxic stress due to spilled oil and produced water. Changes in surface hydrology are related to canal dredging and spoil placement. As canal density increases, the density of natural channels decreases. Interconnected canal networks often lead to saltwater intrusion. Spoil banks block natural overland flow affecting exchange of water, sediments, chemicals, and organisms. Lower wetland productivity and reduced sediment input leads to enhanced surficial subsidence. Spoil banks are not permanent but subside and compact over time and many spoil banks no longer have subaerial expression. Fluid withdrawal from O&G formations leads to induced subsidence and fault activation. Formation pore pressure decreases, which lowers the lateral confining stress acting in the formation due to poroelastic coupling between pore pressure and stress. This promotes normal faulting in an extensional geological environment like the MRD, which causes surface subsidence in the vicinity of the faults. Induced reservoir compaction results in a reduction of reservoir thickness. Induced subsidence occurs in two phases especially when production rate is high. The first phase is compaction of the reservoir itself while the second phase is caused by a slow drainage of pore pressure in bounding shales that induces time-delayed subsidence associated with shale compaction. This second phase can continue for decades, even after most O&G has been produced, resulting in subsidence over much of an oil field that can be greater than surface subsidence due to altered hydrology. Produced water is water brought to the surface during O&G extraction and an estimated 2 million barrels per day were discharged into Louisiana coastal wetlands and waters from nearly 700 sites. This water is a mixture of either liquid or gaseous hydrocarbons, high salinity (up to 300 ppt) water, dissolved and suspended solids such as sand or silt, and injected fluids and additives associated with exploration and production activities and it is toxic to many estuarine organisms including vegetation and fauna. Spilled oil has lethal and sub-lethal effects on a wide range of estuarine organisms. The cumulative effect of alterations in surface hydrology, induced subsidence, and toxins interact such that overall impacts are enhanced. Restoration of coastal wetlands degraded by O&G activities should be informed by these impacts. Keywords: oil and gas production; produced water; oil and gas wetland impacts; induced subsidence; wetland restoration
1. Introduction
The first successful oil well was drilled in Louisiana in 1901 in southwestern Louisiana, in the Jennings field and marked the beginning of oil and gas (O&G) production in Louisiana that, by the mid-20th century, was the state’s primary industry [1]. Production of petroleum has been widespread in the Mississippi River Delta (MRD) and there is a dense network of pipelines both inshore and offshore (Figure 1). Hydrocarbon extraction in the MRD necessitated the dredging of canals through wetlands and waterbodies for navigation, pipelines, and O&G extraction. As canals were dredged, spoil material from canals was placed on the side of canals partially impounding wetlands; altering natural hydrology and salinity; decreasing nutrient, organic matter, and sediment exchange; changing vegetation composition and reducing vegetation productivity (e.g., [2,3,4,5]).
Figure 1. (a) Map of oil and gas pipelines in coastal Louisiana (http://www.dnr.louisiana.gov/assets/docs/oilgas/data/SLA_Pipelines.pdf) and (b) the distribution of oil and gas well permits issued between 1900 and 2017 that were ‘plugged’ or ‘abandoned’ as of 2017 in the 14 coastal parishes [18].
To understand the impacts of O&G activity on MRD wetlands, it is important to recognize that wetland degradation in O&G fields occurred much more rapidly (decades) than the natural deltaic cycle, which took place over centuries to millennia [5,6,7,8,9,10]. Naturally-occurring geologic faulting, sediment compaction, the delta lobe cycle, variability in river discharge, global sea-level change, tidal exchange, wave erosion, and storms such as hurricanes and frontal passages have shaped the MRD landscape for thousands of years [5,6,7,8,9,10,11,12,13,14,15]. However, in less than a century, more than 30,000 km of canals were dredged in the MRD [16,17,18], causing dramatic wetland loss (e.g., [9]) due to cumulative effects [4,19] of altered surface hydrology, induced subsidence and fault activation [20], and toxicity of produced water and spilled oil [21,22,23].
Our objective in this paper is to review the literature on the life cycle of O&G fields in the MRD by (1) describing the production history of hydrocarbons and produced water and (2) reviewing the impacts of surface alterations, induced subsidence, and toxic materials. Although, these impacts have been well documented, their interacting affects have been less well addressed. Thus, an important objective is to consider their cumulative and indirect impacts on coastal ecosystems, especially wetlands, in the MRD and to show how this information informs restoration. 2. Production History of Oil and Gas Fields
Hydrocarbon extraction has occurred in the MRD for over a century. Production increased rapidly and then declined as reserves were depleted and fluid output was dominated by produced water (Figure 2 and Figure 3). This is true for the overall production history and for individual fields that were active for 30–50 years or more.
Figure 2. Annual oil and gas production in southern Louisiana between 1945 and 2019. Data source: (http://www.dnr.louisiana.gov/assets/TAD/OGTables/Table03.pdf).
Figure 3. Composite histories of fluid production from oil and gas fields and wetland loss in south Louisiana. Production data from the Louisiana Department of Natural Resources and the PI/Dwights PLUS database [31]. Wetland loss values were determined by [32] and John Barras (unpublished data). These historical data, integrated across the delta plain, show close temporal and spatial correlations between rates of wetland loss and rates of fluid production [30]. Note that “water” in the legend refers to produced water.
Wetland loss in coastal Louisiana has been quantified [19,24,25]. Researchers have used different proxies of field development to correlate this land loss with specific oil and gas activities [26,27,28,29]. The authors of [30] reported the highest rates of wetland loss in the MRD correlated with peak hydrocarbon production (Figure 3). In [18] it was reported that the annual number of well drilling permits, a proxy for production, correlated with wetland loss rates (Figure 4). In this paper we investigate the factors contributing to these relationships.
Figure 4. The number of oil and gas permits issued annually and land loss rates [18].
Although individual fields show a similar pattern of production history and wetland loss, they differ in some details as illustrated for four different field complexes (Figure 5) that produced O&G over 4–6 decades. The Bully Camp and Madison Bay area had increased produced water development in the second half of their production history. Peak water production in the Bayou Rambio and DuLarge fields coincided with peak gas in this region that had relatively low oil capture. Water production was also high in the second half of the production history of this area. The pattern of water production in the Pointe au Chien study area was similar to gas production in the second half of the oil field life. In each example though, the production of oil, gas, and produced water varied over the field life-cycle, each field began, peaked, then gradually depleted over decades. All fields combine to yield the overall history of the coastal zone (Figure 2 and Figure 3). All four fields illustrate the concept of a development cycle that includes a run-up to peak production followed by a gradual decrease to total abandonment. Individual fields and the aggregate of all fields follow a similar life cycle. These patterns have important implications for the impacts of hydrocarbon extraction on coastal ecosystems, especially wetlands, for three main reasons. First, dredging of canals leads to severe alteration of surface hydrology that impacts coastal wetlands. Second, extraction of fluids (oil, gas, and water) leads to changes in pore pressure that result in induced subsidence and fault activation. Third, accidental spills and intentional releases of oil and produced water cause toxicity stresses that degrade coastal ecosystems. These factors are discussed in more detail below, after which we show how cumulative and interacting impacts multiply the individual impacts.
Figure 5. Production history of selected oil and gas fields in the Mississippi delta. Upper left—Bully Camp; upper right—Bayou Rambio and DeLarge fields; lower left—Madison Bay field area; lower right—Pointe au Chien study area [30]. Note that “water” in the legends refers to produced water.3. Patterns of Wetland Loss and Oil and Gas Activities
From 1930 to 2010 about a quarter (about 5000 km2) of MRD coastal marshes were lost, mostly by conversion to open water [33,34] with two general patterns (Figure 6). Wetland loss has been pervasive across the coast with high loss near the mouth of the Mississippi River, in the Barataria and Terrebonne basins, and in the Chenier Plain. Two areas stand out with much less loss. One is the central coast that receives input from the Atchafalaya River, which carries about a third of total Mississippi River discharge, that flows into shallow bays and wetlands over a wide arc along the central Louisiana coast. The other zone is on the northeastern flank of the delta in the seaward reaches of the Pontchartrain estuary.
Figure 6. Wetland loss in coastal Louisiana from 1932 to 2016. Red and yellow areas have high land loss rates. Note that land loss is low in the central coast and in the northeastern flank of the delta. (Source: [33,34] The map can be downloaded at https://pubs.usgs.gov/sim/3381/sim3381.pdf for detailed examination of specific areas of change. See also https://pubs.er.usgs.gov/publication/sim3381).
A number of studies have discussed the causes of wetland loss and addressed the role of O&G activities in this loss (Table 1). O&G activities reduce forces leading to wetland sustainability (e.g., sedimentation, accretion, vegetation biomass production) and enhance forces leading to deterioration (e.g., increased subsidence, saltwater intrusion, vegetation decline, toxicity stresses). O&G impacts on wetlands are both direct and indirect. Direct loss is due both to canal dredging and spoil placement, while indirect impacts include alteration of surface hydrology, induced subsidence, and introduction of toxic substances. Direct impacts are reported to have caused between 6% and 70% or more land loss while indirect impacts have caused between 20% and 80% or more of wetland loss (Table 1).
Table 1. Estimates of the percentage of wetland loss caused by all oil and gas activities (overall) or by direct or indirect impacts of oil and gas activities.
Wetland loss patterns are generally similar to temporal patterns of production (Figure 7). The different studies were done using different methods and represent different conceptualizations of the delta and sub-areas sampled. Regardless of the proxy used to describe the oil and gas life-cycle, the relation to land loss is unmistakable. The rates of land loss differ because they are for different areas (e.g., total coast vs. deltaic and Chenier plains) and for different time periods and the study methods, though similar, are not the same [32,34,44]. Peak loss occurred generally between 1960 and 1980. The curve for total O&G production is sharper than that of land loss likely because of delayed impacts of O&G as well as other causes of land loss (e.g., isolation from riverine input, edge erosion, hurricane impacts).
Figure 7. Oil and gas production and land loss studies for coastal Louisiana. Loss values are from [32,34,44] where rate values represent an average over the time interval shown. Production data are from Louisiana Department of Natural Resources. These historical data show close temporal and spatial correlations between rates of wetland loss and rates of hydrocarbon production.4. Indirect Impacts of Oil and Gas Exploration and Production on Coastal Louisiana Wetlands
4.1. Alteration of Surface Hydrology Due to Canal Dredging and Spoil Placement
MRD wetlands are dependent on sheet flow for exchange of water, sediments, and nutrients to sustain wetland health. Unimpeded wetland hydrology facilitates alternating flooding and draining of wetlands. Canals and spoil banks are linear and intersecting, and spoil banks are higher in elevation than surrounding wetlands and normal high tides [4,9,15,45]. Placement of dredge spoil impounds wetlands, reducing or eliminating surface water exchange and tidal influence, reducing sediment deposition onto wetlands, impeding the exchange of materials (e.g., nutrients, sediments, organisms) between semi-impounded marshes and the surrounding marsh, and increasing inundation duration while decreasing inundation frequency [2,4,16,46,47,48,49,50]. Spoil banks trap water, increasing water logging and decreasing drainage, sediment accretion, and vegetation productivity (Figure 8) [4,15,16,19,46,48,51,52,53,54,55,56,57,58,59,60,61,62,63]. As surface flow is minimized by spoil banks, water may only be introduced into impounded wetlands when water levels are elevated during frontal passages or major storms.
Figure 8. Total hours flooded, for one month, at a reference marsh with a natural berm and a partially impounded area where about 75% of the natural berm had been replaced by a dredged canal spoil bank, Golden Meadow oil and gas field south of New Orleans, LA [52]. The dashed line shows what the relationship would be if there were no partial impoundment.
Due to spoil banks, accretion in un-impounded marshes may be up to five times higher than that in impounded marshes [3,57,61,64,65,66,67]. Canals also promote saltwater intrusion into wetlands previously isolated from direct exchange with higher salinity waters [2,4,27,47,48,68,69,70,71]. Increases in salinity cause changes in vegetation composition and reduced productivity and/or death of fresh and low salinity marsh species and lead to formation of open water [53,71,72,73]. Introduction of saltwater increases sulfate concentrations, which can be reduced to sulfides in anaerobic soils, that stresses and causes mortality to low salinity wetland vegetation [73] and rapid decomposition and collapse of soil organic matter and soil structure peats [74]. Reduced vegetation productivity and vegetation death exacerbate land loss because plant roots bind soils and increase soil strength. When roots die, the wetland rapidly loses elevation and is more vulnerable to erosion [74,75].
Over time, canals widen as a result of spoil bank undercutting, erosion, and collapse causing additional wetland loss [45,46]. Localized subsidence along pipeline canals occurs along the flanks of spoil banks when the weight of the spoil depresses the surface of the marsh and leads to the formation of linear ponds behind the spoil banks due to subsidence from the weight of the spoil, trapping of water at the base of spoil banks and blocking of sediment input [76]. The authors of [77] documented soil compaction beneath spoil banks created more impenetrable soils that reduced ground water movement. These processes isolate the wetland behind the spoil bank from both above- and belowground water exchange.
O&G canals are deep, straight channels while natural waterways are primarily shallow and sinuous tidal channels [4]. As canal density increases in an area, the density of natural channels decreases because canals preferentially capture water flow from natural channels (Figure 9) [37,52] in a process termed ‘channel theft’, because deep, straight canals transport water more efficiently than natural shallow and sinuous channels [50].
Figure 9. The relationship between canal density and the density of natural channels. The data are averages of replicate 1-km2 grids (numbers shown by symbol) in the region of Leeville, LA, a saline marsh area [37].
Once dredged, spoil banks are not permanent features and have a life cycle of their own. They disappear over time due to compaction, subsidence, sea-level rise, and erosion. Spoil banks protect remnant marsh from wave erosion and as spoil banks disappear remnant marsh can be lost due to wave attack, leading to further wetlands loss. In Figure 10, subaerial land that disappeared between the two mapping dates (center panel in Figure 10) in the Leeville Field includes both spoil banks and marsh.
Figure 10. Aerial imagery of a portion of the Leeville oil and gas field in southern Louisiana showing the disappearance of spoil banks and marsh between 1998 and 2017. Bayou Lafourche is in the upper left in each of the images [5]. Areas depicted in red in image B disappeared between the two dates. Red linear strips are spoil banks that disappeared. Green shows subaerial land, both spoil banks and marsh, that was still present in 2017. The Southwest Louisiana Canal at the top of image A was dredged in the late 19th century. In 1998, some spoil banks along the canal were still present, but by 2017 they had disappeared. Louisiana highway 1 is shown at the upper left in image A. By 2017, a new, elevated highway was constructed (white lines in Image C). The width of images is approximately 2.8 km.
As O&G canals are so pervasive, it has been suggested that canals are responsible for practically all wetlands loss. For example, the authors of [18,27] plotted land loss from 15-min quadrangles against canal density and concluded that land loss was directly related to the percent of canals in each map, indicating that almost all land loss was related to canals (Figure 11). However, in [5] and [9] the authors showed that this approach is flawed because it statistically relates all land loss in 15-min quadrangles (which cover about 66,000 ha) to canal density in the quadrangle even when it is neither spatially or functionally related to land loss patterns. Additionally, as we discuss below, induced subsidence causes land loss but is not functionally related to surface alterations in hydrology. In addition, toxic stress due to spilled oil and produced water cause vegetation stress and mortality. However, altered hydrology, induced subsidence, and toxic compounds interact synergistically and cause wetland loss as indicated by the interactions demonstrated in the Leeville field (Figure 10).
Figure 11. The land loss rate from the 1930s to 1990 and canal density in 15-min quadrangle maps. The authors of [18] concluded that this demonstrates that most wetlands loss is due to canals. This is incorrect, however, because of the large size of 15-min quadrangle maps and other causes of land loss due to induced subsidence and toxic stress.4.2. Oil and Gas Production Induced Subsidence
Oil, gas, and brine extraction depletes the hydrocarbon reservoir, often precipitously, resulting in pressure drop, compaction, and fault activation (usually reactivation); and this change at depth translates upward, manifesting at the surface as subsidence and faulting. That is, without pressure support the depleted reservoir collapses, and the pressure difference at the nearby associated fault plane nucleates slippage there [4,20,28,30], leading to a subtle, and sometimes dramatic, overprint integrated with other O&G related processes. Observations generally confirm the relationship between fluid production and subsidence in coastal Louisiana: for example, wetland loss is typically higher in the vicinity of oil and gas fields [9,20,29,78,79,80,81,82,83]. On the coastal Chenier Plain, the authors of [83] suggested that paleo-sea level elevations, vertically offset by 0.5–1 m on a transect near the area of maximum oil and gas production, were influenced by this production. Local rates of measured subsidence in oil and gas fields in south central Louisiana (often more than 20 mm/year) were much higher than regional rates in the Mississippi River Delta (about 10 mm/year) [84]. In [30] the authors analyzed releveling surveys, remote images, subsurface maps, stratigraphic sections, and hydrocarbon production data in relation to wetland loss for the Terrebonne-Lafourche basins and found that the highest rates of subsidence coincided with the location of oil and gas fields (Figure 12) [30]. This study also showed that subsidence rates were greater in the later epoch (1982–1993) than the earlier (1965–1982); that is, subsidence accelerated late in the cycle of O&G production. In [20,30,84] it was concluded that these rapid changes were likely caused by induced subsidence and fault reactivation due to oil and gas activity.
Figure 12. Map showing average subsidence rates between 1965 and 1993 in south Louisiana. Areas of highest subsidence rates (>12 mm/year; hatched pattern) correlate closely with locations of oil and gas fields. Lowest average subsidence rates are located between major producing fields [30].
Elevation change related to faulting in the MRD is often marked by an arcuate scarp separating marsh and water [85], while subsidence related to a reservoir’s compaction is spread over and beyond the production area (up to kilometers away from the producing wells) [20]. Subsidence involved with O&G fields typically begins around the time of peak production [84], and continues for an extended time as the field is depleted, often decades after peak production [29,79,86], though it too reaches an ultimate limit [79,84].
The Mississippi Delta is particularly susceptible to subsidence and faulting since rapid deposition of sands and clays has created a weak, metastable situation that responded from early in its geologic history with “landslide-like” faulting (down to the coast listric normal faulting) parallel to the coast, and that faulting has progressed upward as deposition continued—that is, growth faulting [14,87,88,89]. The sedimentary section developed so rapidly that there was little time for consolidation, cementation, or in many cases, normal pressure equilibration. At the same time, these faults and related rollover anticlines on their downthrown side, formed hydrocarbon traps, the basis for many of the present-day O&G fields on the delta (Figure 13). To add complexity, the low density, easily deformed Louann salt layer that began near the base of the geologic section flowed upward in various geometries, creating salt domes dragging up steeply tilted beds, faults, and anticlinal features that became hydrocarbon traps as well. Growth faults in the delta move episodically over their lifetime, and along segments a few kilometers in length (e.g., [90]). Over the cycle of MRD petroleum development, a number of these growth faults related to oil and gas fields have been reactivated, with consequent displacement and surface subsidence on the fault downthrown side. The mechanism of reactivation along these growth faults is expedited by a poroelastic reduction of horizontal confining stress, which occurs as a result of the fluid withdrawal and consequent pore pressure decrease [91,92]. That is, faults related to MRD O&G fields that break the surface today are mostly reactivated growth faults involved with the reservoirs below—reservoirs that these faults created in the first place.
Figure 13. Typical Mississippi River Delta oil and gas reservoir, a rollover anticline on the downthrown side of a down-to-the-coast normal fault. Oil, gas, and water production affects the subsurface in and around the producing reservoir. The depletion process leads to production induced fault activation, reservoir compaction with additional compaction of bounding shales as shown in the subsurface. The surface manifestation of these subsurface changes is shown as a composite of land loss on the downthrown side the activated fault plus subsidence over the compacted oil, gas, and water producing reservoir.
The mechanism of O&G subsidence is involved with the collapse of the reservoir itself. Subsidence happens when production of oil, gas, and water (and sometimes sand) reduces reservoir sand pore pressure to the point where it can no longer support the overburden [93]. Thus, the delta’s young, often poorly consolidated reservoir sands, sandwiched between shales, often undergo compaction with production drainage and this deformation is transferred to the surface as a subsidence bowl form. The authors of [94] modeled this reservoir compaction and surface deformation process using the concept of a nucleus-of-strain (the impacted reservoir), together with an elastically deforming half-space (the geologic section above), spreading the effect and reducing the displacement to produce a halo-like effect in and around the O&G field. In [79] the authors used this concept along with 1982–1993 epoch releveling at the Leeville, Golden Meadow, Cut Off, and Valentine fields and the authors of [95] modeled the Lapeyrouse field (Figure 12). At Lapeyrouse, modeling matched the measured releveling when a reactivated down-to-the-coast growth fault was added to the deformation of the compacted disc reservoir sand. This reservoir compaction subsidence process is found in fields along the Louisiana coast and often acts in concert with O&G field related faulting.
In [29] the authors considered the extended timeframe of subsidence where displacement related to depleted O&G fields did not stop as the production cycle ended, but accelerated significantly for decades. As shown in Figure 14, this study indicates (a) greater subsidence over traversed O&G fields for each epoch (panels A and B), (b) increased subsidence rates (accelerated subsidence) in the later epoch (panel C), and (c) areas of greatest subsidence and rates likely related to fault displacement.
Figure 14. Elevation changes during epoch 1 (1965–1982; A) and epoch 2 (1983–1993; B), and the rates of subsidence in the two epochs (C). Over the entire transect, subsidence rates were greater in epoch 2 than in epoch 1. The yellow squares in (A) and (B) indicate an arbitrarily selected reference station approximately 8 km south and outside of the projected Leeville (see Figure 10) to estimate the magnitudes of local production-related subsidence signals [29].
While post production time-dependent subsidence is intuitively expected due to factors such as slow dissipation of pore pressure in the reservoir (e.g., [96]), and time-dependent creep in the reservoir [97], this subsidence typically happens over only a few years, not decades after and not with acceleration. The authors of [29] considered the role of the bounding shale above and below the sand reservoir itself (Figure 15), and modeled compaction in the Valentine field (Figure 12 and Figure 14) in two stages, the first being poroelastic compaction of the reservoir sand during active production, followed by the time-delayed compaction resulting from slow poroelastic and viscoplastic compaction of the low-permeable bounding shale as included pore water slowly drained into the depleted sand.