The U. S. National Aeronautics and Space Administration (NASA) has archived thousands of satellite images of density plumes in its online publishing outlet called 'Earth Observatory' since 1999. Although these images are in the public domain, there has not been any systematic compilation of configurations of density plumes associated with various sedimentary environments and processes. This article, based on 45 case studies covering 21 major rivers (e.g., Amazon, Betsiboka, Congo [Zaire], Copper, Hugli [Ganges], Mackenzie, Mississippi, Niger, Nile, Rhone, Rio de la Plata, Yellow, Yangtze, Zambezi, etc.) and six different depositional environments (i.e., marine, lacustrine, estuarine, lagoon, bay, and reef), is the first attempt in illustrating natural variability of configurations of density plumes in modern environments. There are, at least, 24 configurations of density plumes. An important finding of this study is that density plumes are controlled by a plethora of 18 oceanographic, meteorological, and other external factors. Examples are: 1) Yellow River in China by tidal shear front and by a change in river course; 2) Yangtze River in China by shelf currents and vertical mixing by tides in winter months; 3) Rio de la Plata Estuary in Argentina and Uruguay by Ocean currents; 4) San Francisco Bay in California by tidal currents; 5) Gulf of Manner in the Indian Ocean by monsoonal currents; 6) Egypt in Red Sea by Eolian dust; 7) U.S. Atlantic margin by cyclones; 8) Sri Lanka by tsunamis; 9) Copper River in Alaska by high-gradient braid delta; 10) Lake Erie by seiche; 11) continental margin off Namibia by upwelling; 12) Bering Sea by phytoplankton; 13) the Great Bahama Bank in the Atlantic Ocean by fish activity; 14) Indonesia by volcanic activity; 15) Greenland by glacial melt; 16) South Pacific Ocean by coral reef; 17) Carolina continental Rise by pockmarks; and 18) Otsuchi Bay in Japan by internal bore. The prevailing trend in promoting a single type of river-flood triggered hyperpycnal flow is flawed because there are 16 types of hyperpycnal flows. River-flood derived hyperpycnal flows are muddy in texture and they occur close to the shoreline in inner shelf environments. Hyperpycnal flows are not viable transport mechanisms of sand and gravel across the shelf into the deep sea. The available field observations suggest that they do not form meter-thick sand layers in deep water settings. For the above reasons, river-flood triggered hyperpycnites are considered unsuitable for serving as petroleum reservoirs in deep-water environments until proven otherwise.
Density plumes and their various configurations seen on satellite images have been a source of curiosity to the geologic community as well as to the general public. The National Aeronautics and Space Administration (NASA, 2018)[1] has archived satellite images on density plumes in its online publishing outlet "Earth Observatory" since 1999. NASA has used a variety of satellites, such as Aqua, Terra and Topex/Poseidon. However, there has not been a systematic attempt to compile variations in natural configurations of density plumes in the world’ s oceans and lakes. This article is the first attempt to compile natural variability of density plumes in 45 case studies worldwide (Table 1, Fig. 1[2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32]). Forel (1885)[33] first reported the phenomenon of density plumes in the Lake Geneva (Loc Lé man), Switzerland. Following Forel's pioneering work, Bates (1953)[34] suggested three types of density plumes in river-mouth deltaic environments: 1) hypopycnal plume for floating river water that has lower density than basin water (Fig. 2a); 2) homopycnal plume for mixing river water that has equal density as basin water (Fig. 2b); and 3) hyperpycnal plume for sinking river water that has higher density than basin water (Fig. 2c). Although river-mouth hyperpycnal plumes have received much attention[35, 36], plumes in other environments are equally important, but are generally ignored.
The primary purpose of this compendium of satellite images is to document the natural variability and complexity of density plumes not only at river mouths but also at other settings. The knowledge gained by this compilation would hopefully help in understanding the implications for external controls, hyperpycnal flows, submarine fans, and deep-water petroleum reservoirs. This article, which is rich in empirical data from a variety of environments, should serve as a guideline for students in their future research. Although the primary dataset for this study is from NASA's satellite images, other published photographic images are also used.
In practice, there is a tendency to use the two terms "flow" and "plume" synonymously. However, these two terms are not one and the same. In hydrodynamics, the term plume describes a condition when a column of one fluid moves through another fluid. To accommodate natural variability in plume types, a broad definition of plume is adopted in this article. Accordingly, a plume is defined as a fluid enriched in sedi- ment, ash, biological or chemical matter that enters another fluid. As it would be demonstrated later, there is a multitude of plume types in nature. Among them, the river plume is the most popular. National Oceanic and Atmospheric Administration (NOAA) Fisheries Glossary (2006, p. 42)[37] defines a River Plume as "Turbid freshwater flowing from land and generally in the distal part of a river (mouth) outside the bounds of an estuary or river channel."
In fluid mechanics, the term "flow" is used for a continuous, irreversible deformation of sediment-water mixture that occurs in response to applied shear stress, which is gravity in most cases[38]. In other words, not all plumes are flows. For example, floating hypopycnal plumes are not driven by gravity (Fig. 2a). However, both terms "flow" and "plume" are applicable to hyperpycnal type.
Because the original concept of hyperpycnal flows is closely tied to river floods[34], density plumes at river mouths are considered using 21 rivers in this study (Table 1). They are:
1)Amazon River, Brazil, Equatorial Atlantic; 2)Betsiboka River, NW Madagascar, Bombetoka Bay; 3)Chignik River, Alaska, USA, Pacific Ocean; 4)Congo (Zaire) River, West Africa, South Atlantic Ocean; 5)Connecticut River, New England region, USA, Long Island Sound; 6)Copper River, USA, Gulf of Alaska; 7)Ebro River, Iberian Peninsula, Mediterranean Sea; 8)Eel River, California, USA, Pacific Ocean; 9)Guadalquivir River, Southern Spain, Gulf of Cá diz; 10)Hugli River (a distributary of the Ganges River), India, Bay of Bengal; 11)Mackenzie River, Canada, Beaufort Sea; 12)Mississippi River, USA, Gulf of Mexico; 13)Niger River, West Africa, North Atlantic Ocean; 14)Nile River Delta, Egypt, Mediterranean Sea; 15)Onibe River, Eastern Madagascar, Indian Ocean; 16)Rhone River, France, Gulf of Lions, Mediterranean Sea; 17)Rio de la Plata Estuary, Argentina and Uruguay, South Atlantic Ocean; 18)Rupert River, Quebec, Canada, Rupert Bay; 19)Yangtze River, China, East China Sea; 20)Yellow River, China, Bohai Bay, Bohai Sea; 21)Zambezi River, Central Mozambique, Indian Ocean.
In this article, most satellite images used were taken by the Moderate Resolution Imaging Spectroradiometer (MODIS), an earth-observing instrument of the Aqua satellite.
The 45 case studies worldwide (Table 1) used in this article represent not only conventional river-mouth plumes, but also estuarine, meltwater, eolian, volcanic, cyclonic, upwelling, and other environments. In the following discussion, I have selected only certain case studies to emphasize specific details dealing with controlling factors. Although I tried to describe plume types based on shapes, it is not always possible because shapes are too complex in some cases. For this reason, I used their origin in classifying plumes in selected cases (e.g., volcanic plume, meltwater plume, dust plume, etc). Additional case studies would help resolve this problem.
The Mississippi River is the second-largest drainage system on the North American continent, second only to the Hudson Bay system. Historically, the Mississippi River delta has been the "classic" type locality for deltas[40]. Satellite images of the Mississippi River show one of the best developed deflecting density plumes (Fig. 2d). Walker and Rouse (1993) [2] attributed the deflection of plumes to shelf currents in the Gulf of Mexico.
The Yellow River (Huanghe), which is regarded as the world’ s largest contributor of fluvial sediment load to the ocean (Yu et al., 2011)[41]. Historically, it contributed a sediment load of nearly 100 million tons per year[42]. The Huanghe’ s average annual suspended-load concentration of 25 000 mg/L and flood stage concentration of 220 000 mg/L are, by 1983, the largest in the world[43]. Recent empirical data gathered from the Yellow River have been reviewed by Shanmugam (2018a)[44].
The Yellow River has revealed two distinctive types of plumes, namely simple lobate and horse's tail (Fig. 2e, 2f). The horse’ s tail shape is related to local deflection of plumes. The Yellow River is a river-dominated delta with unidirectional river currents, and it develops river-mouth muddy lobes (Fig. 2e). However, the development of a tidal shear front prevents the transport of sediment seaward (Fig. 3). Li et al. (2001)[45] showed that the shear front is an important dynamic factor in controlling rapid accretion at the Yellow River mouth. Suspended sediment converges and is deposited rapidly along the shear front zone. This is because a low-velocity zone is formed between two inverse flow bodies.
The Yangtze River (Changjiang) is the longest river in Asia (6 300 km). Satellite images show that the Yangtze River generates both hyperpycnal lobate and deflecting hypopycnal plumes (Fig. 4a). A distinction between hyperpycnal and hypopycnal plumes has been made on the basis of color. For example, hyperpycnal plumes are yellowish brown in color due to high sediment concentration, whereas hypopycnal plumes are bluish due to low sediment concentration (Fig. 4a).
The Yangtze River mouth is a complex setting in which both ocean currents and tidal currents are affecting sediment dispersal. Unlike the Yellow River that enters a protected Bohai Bay from major ocean currents, the Yangtze River enters the East China Sea affected by the warm, north-flowing, Kuroshio Current (Fig. 4b). As a consequence, muddy sediments brought by the Yangtze River are redistributed and deposited as a mud belt on the inner shelf[47].
Although both Yellow and Yangtze rivers develop hyperpycnal plumes at their river mouths, transport of hyperpycnal sediments from the river mouth to the deep sea has been either blocked or diverted by different external controls, such as tidal shear front and ocean currents (Fig. 5).
In a recent study, Luo et al. (2017)[5] recognized that extended density plumes tend to develop during winter months (Fig. 6a), which are absent during the summer months (Fig. 6b). In explaining the winter density plumes (Fig. 7), Luo et al. (2017)[5] make the case for vertical mixing of sediments by tides in the winter months using numerical modeling. Although the tides are active all year round, Luo et al. (2017)[5] suggest that their modeling shows that the sediment can only rise up to the surface in the winter, when temperatures and salinities at the sea surface and bottom are roughly the same. In the summer, however, an influx of fresh water from the Yangtze, combined with heating of the surface layers of the sea, prevents vertical mixing and keeps the resuspended sediment in the depths.
In the geologic and engineering literature, the focus of discussion on hyperpycnal flows is centered around fine-grained or common deltas. McPherson, Shanmugam, and Moiola (1987)[6] distinguished fine-grained deltas from coarse-grained deltas. The importance here is that braid deltas, a kind of coarse-grained deltas, are typical of high-gradient settings with high-velocity river flows. Because these braided rivers plunge into a standing body of water with multiple entry points, separated by braided bars, these rivers develop "linear" hyperpycnal plumes (Fig. 8). Distinguishing linear types is important because braid deltas are known to develop various types of sediment flows, including debris flows, in the subaqueous delta front[6].
Galloway (1976)[48] originally classified the Copper River delta as a "fan delta". According to McPherson, Shanmugam, and Moiola (1987)[6], the Copper River delta fits the characteristics of coarse-grained braid delta. The importance here is that it develops both coalescing riverine plume (Fig. 9a) and blanketing dust plume (Fig. 9b). In cases like this, it is extremely difficult to distinguish the complex interaction be- tween riverine and eolian input that occurs during sedimentation in the ancient sedimentary record.
The Golden Gate Bridge is located at the mouth of the San Francisco Bay connecting the Pacific Ocean (Fig. 12a). Barnard et al. (2006)[8] reported a field of giant sand waves of tidal origin beneath the Golden Gate Bridge at the mouth of San Francisco Bay in California (Fig. 12b). Repeated surveys demonstrated that the sand waves are active and dynamic features that move in response to tidally generated currents. The significance of the tidal sand waves at the mouth of the San Francisco Bay is that muddy lobate sediment plumes have been imaged by NASA at this area (Fig. 12c), mimicking many river-mouth sediment plumes, such as the one observed at the mouth of the Yellow River (Fig. 2e).
The Rí o de la Plata Estuary is located on the east coast of South America (Fig. 14a), bordering Argentina and Uruguay, is one of the largest estuaries in the world[50]. It is 280 km long and 220 km wide at its mouth, and its water depth does not exceed 10 m (Fig. 14b). It receives water and sediment from both the Paraná and Uruguay rivers with an annual mean discharge of 22 000 m3/s. The Rí o de la Plata Estuary shows dissipating plume with irregular front (Fig. 14c)[51, 52].
Gonzalez-Silvera et al. (2006)[9] studied Ocean color (OCTS, SeaWiFS) and sea surface temperature (AVHRR) images and evaluated spatial and temporal variability of the Brazil-Malvinas Confluence and La Plata Plume (20° to 45° S and 40° to 65° W). The data set covered the period from January 1997 to June 2003. Chlorophyll and SST data were compiled and analyzed. The results show a gradual increase of the northward intrusion of the La Plata Plume throughout the period lasting from summer to winter; the summer shape of the La Plata Plume showed a stronger penetration over the shelf on the Argentinean side of the estuary mouth; and the seasonal migration of the Brazil-Malvinas Confluence.
The Gulf of Cá diz is located in the north-eastern Atlantic Ocean (Fig. 16a). It is enclosed by the southern Iberian and northern Moroccan margins, west of Gibraltar Strait. Two major rivers, the Guadalquivir and the Guadiana, as well as smaller rivers, like the Odiel, the Tinto, and the Guadalete, reach the ocean here. In terms of ocean currents[12], it is one of the most complex oceanographic settings (Fig. 16b). Mimicking the current patterns, sediments that are emptied into the gulf by the Guadalquivir River exhibit a 'U-Turn' shape for the plume (Fig. 16c).
The Strait of Gibraltar is a narrow strait that connects the Atlantic Ocean to the Mediterranean Sea and it separates Spain from Morocco. On satellite images, we can see the difference between massive sediment plume in Gulf of Cá diz to the west of the strait and swirly plankton plume in the Mediterranean Sea to the east of the strait (Fig. 17a). Conceivably, the massive plume may be related to increased sediment input from nearby during flood stages. The origin of swirly plumes could be attributed to the currents and internal waves in the area (Fig. 17b).
The ocean water moving east from the North Atlantic into the Mediterranean Sea has created turbulence in the form of internal waves (Fig. 17b). These internal waves— sometimes with amplitudes up to 80 m and wavelength up to 0.8 km[13] — occur primarily deep within the ocean, with just a mere crest emerging out through the surface. In cases of both plankton plumes and internal waves observed on the Mediterranean Sea side, the key controlling factor is the movement of ocean water through the strait. Aspects of origin of internal waves and tides along oceanic pycnoclines have been reviewed by Shanmugam (2013)[13] (Fig. 18). The link between internal waves and sediment plumes is not commonly discussed in the sedimentological literature. An exception is the study of Otsuchi Bay in Japan by (Masunaga et al. (2015)[14]. However, the exact configurations of plumes generated by internal waves are poorly understood.
Masunaga et al. (2015)[14] made observations of the run-up of internal bores in a shallow bay with a tow-yo instrument and mooring arrays with high spatial and temporal resolution. Their observations showed that internal bores propagate along the slope, accompanied by strong turbulent mixing and strong sediment resuspension in a shallow bay (Fig. 19). The isothermal displacement due to the bores reached 20 m vertically in a water depth of 40 m. Turbidity measurements showed suspended particles transported from the sloping bottom and offshore above the thermocline, forming an intermediate nepheloid layer (INL). At the head of the bore (dense water), a vortex accompanied by strong vertical motion induced strong vertical sediment resuspension and a steep isothermal displacement.
The Hugli River is a distributary of the Ganges River that empties its sediments into the Bay of Bengal. During monsoonal flood stages[53], the Hugli River develops anastomosing plumes at the mouth (Fig. 20). The Ganges River mouth is considered as a tide-dominated estuary[15]. The Bay of Bengal is known not only for severe monsoonal floods, but also for frequent cyclonic activity[21]. The origin of anastomosing plume type is not well understood. However, the Sagar Island and adjacent islands are rather permanent features that do control river flows and plume configurations.
The deep Gulf of Mannar (> 200 m bathymetry) and the shallow Palk Bay (< 7 m bathymetry), located between southeastern India and northern Sri Lanka, are in the northern Indian Ocean (Fig. 21a). This location provides an opportunity to understand the distribution of sediment plumes controlled by local bathymetry (Fig. 21d)[16], by reversal in monsoonal currents (Fig. 21b)[17], and by wave actions. Massive and swirly plumes are well developed in the shallow areas (Fig. 21b, 21c). Jagadeesan et al. (2013)[16] attributed the structuring of the mesozooplankton in this area to ocean currents. The significance of this case study is that plumes do not develop in the deep Gulf of Mannar.
The 0val-shaped island in the middle of the Palk Bay is called Nedunthivu (Tamil name meaning Long Island). The island’ s area is 50 km² , its length is 8 Km and its maximum width is about 6 km. Neduntivu, a flat island surrounded by shallow waters and beaches of coral chunks and sand, provides some of the sediment for the plume by wave actions.
The Zambezi River in Central Mozambique is a wave- dominated delta. It has developed a coalescing lobate plume due to multiple river mouths (Fig. 22b). The importance of longshore currents in modifying the delta is discussed by Mikhailov et al. (2015)[18].
Southern African margin is characterized by the south- flowing Agulhas Current on the Indian Ocean side and the north-flowing Benguela Current on the Atlantic Ocean side (Fig. 26a). The nutrient-rich waters cause the Benguela Upwelling and related high chlorophyll concentrations on the southwestern Africa (Shillington et al., 1992)[20]. As a consequence, a well-developed cloudy plankton plume is visible in the waters off Namibia in early November 2007 on a NASA- MODIS satellite image (Fig. 26b). In a chemical process related to the demise of planktons and their consumption by bacteria also results in the release of toxic hydrogen sulfide plume (Fig. 26c). It is unclear as to how these chemical plumes will be reflected in the deposit.
A wealth of satellite information is available on cyclone- induced sediment transport across the United States Atlantic shelf[21]. Category 4 Hurricane Floyd (September 7-17, 1999), for example, was nearly 933 km across. It affected the East Coast of the United States from Florida to Maine, with maximum damage to North Carolina. A comparison of satellite images of shelf waters shows a striking contrast in resuspended sediment between fair-weather and storm-weather periods (Fig. 27). During fair-weather periods, for example, shelf waters were calm and free of resuspended sediment along the Florida-Georgia-South Carolina-North Carolina coast (Fig. 27a). In contrast, during storm-weather periods, shelf waters contained abundant resuspended sediment as the tropical Hurricane Floyd passed over the East Coast on September 16, 1999. The power of Floyd was so intense that it stirred up the entire water column across the United States Atlantic continental shelf (Fig. 27b). In this area, the shelf is 100 km wide and 200 m deep at the shelf edge. These satellite images are the most convincing visual evidence for shelf-wide resuspension of sediment. These responded sediments that reach the shelf-slope break and cascade down the slope. Such plumes are labeled "cascading plume" (Fig. 27b).
Analogous to the 1999 Hurricane Floyd, which resuspended muddy sediments over the entire width of nearly 100 km shelf on the U.S. Atlantic margin (Fig. 27b), the 2009 Tropical Storm resuspended muddy sediments and caused a swirly cyclonic plume that is nearly 150 km in maximum width in the northern Gulf of Mexico (Fig. 28b).
Tsunamis play a major role in both erosion and deposition of sediments. Shanmugam (2006)[22] explained tsunami-related deposition in deep-water environments in four progressive steps: (1) triggering stage, (2) tsunami stage, (3) transformation stage, and (4) depositional stage. The transformation stage is evident in sediment-rich backwash flows during the 2004 Indian Ocean tsunami at Kalutara Beach, southwestern Sri Lanka (Fig. 29). The incoming ocean waters are clearly blue in color (implying sediment free), but these waters transform into brown in color near the coast because of their incorporation of sediment. The transformation to brown color is the result of the wave breaking, These returning sediments are, both suspended and bedload components, are called "backwash plumes".
Meltwater sediment plumes from ice sheets of Greenland are common in the Labrador Sea (Fig. 30a). Chu (2014)[23] reviewed the elements of the Greenland ice sheet hydrologic system and proposed depositional models for the buoyant sediment plumes and turbidity currents (Fig. 30c). Mankoff et al (2016)[54] discussed structure and dynamics of a subglacial discharge plume in a Greenlandic fjord.
The most striking feature of the Great Bahama Bank is the swath of whitish blue waters west of the Andros Island (Fig. 32a). This swath of waters is nearly 350 km in length in the northwest-southeast direction and about 100 km in width, just west of the Andros Island. This swath is labeled as "whitings plume" (Fig. 32a). The reason is that the region west of Andros Island on the Great Bahama Bank is dominated by fine lime mud and pellet mud sediments[57]. The seas overlying these muds are well known for episodic, highly turbid events that produce milky white waters, historically referred to as “ whitings” [58]. The origin of whitings has been attributed to both direct precipitation of lime mud[58, 59] and by resuspension by fish activities[25], by wind[26], and by the Florida Current[27].
In the South Pacific Ocean, coral reef environments have developed a ring plume (Fig. 32b). In the South Pacific, the dark blue waters of the Coral and Solomon Seas, the coral reefs and forested islands of the Louisiade Archipelago stretch southeastward from the tip of Papua New Guinea for over 350 kilometers. The lower left corner of the satellite image shows part of the northwestern coast of the largest island in the archipelago, Tagula Island (also called Vanatinai Island). The Louisiade Archipelago has reefs and islands that are in various stages of “ evolution” from fringing reef to coral atolls, which result in ring plumes (Fig. 32b). Khanna and Yadav (2008)[28] discussed reef development near the Tagula Island.
An astronaut photograph showing a dust plume surging out over the Red Sea and reaching most of the way to Saudi Arabia (Fig. 33a). The point source of this plume is the delta of the southern Egyptian river Khor Baraka. Invariably, dust plumes are highly variable in shapes. This is because at the time of contact of dust from air to water, local oceanographic factors, such as ocean currents coupled with waves, would determine plume shapes.
Satellite images off the Yucatan Peninsula show feathery plume (Fig. 33b). The plume is a complex mix of both sediment and plankton. Both winds and waves are considered responsible for the generation of plumes. The blue-green cloud in this image roughly matches the extent of the shallow continental shelf west of the Yucatan Peninsula. A few days before the image was taken by NASA, strong winds churned the Gulf. The sediment was probably brought to the surface in shallow waters by strong waves.
Volcanic ash plumes were reported from Bogoslof Island, Bering Sea in June 2017 (Fig. 34b). The challenge in studying volcanic ash plumes is that they travel hundreds to thousands of kilometers depending on the magnitude of eruptions, global atmospheric circulation patterns, and bulk density of ash[29]. Furthermore, analogous to dust plume, volcanic ash plumes are highly variable in shapes.
Paull et al. (1995)[30] reported that seafloor venting of microbial gases occurs at 2 167 m water depth over the Blake Ridge diapir on the Carolina Continental Rise. Gas-rich plumes were identified acoustically in the water column up to 320 m above a pockmarked sea floor associated with active chemosynthetic biological communities. In this area, plumes and venting fluids emanate from near a small fault that extends downward toward a dome in the bottom-simulating reflector (BSR), indicating that fluid and/or gas migration is associated with gas hydrate-bearing sediment below. These plumes might be caused by gas bubbles or buoyant clumps of gas hydrate that float upward from the seafloor. Distribution of gas hydrates has been mapped (Fig. 37).
Ruppel and Kessler (2017)[61] discussed aspects of subglacial model gas hydrates in high latitude regions.
Satellite image shows tendril configuration of plumes in Lake Michigan (Fig. 38a). In this example, suspended sediments transformed the southern end of Lake Michigan. Ranging in color from brown to green, the sediment filled the surface waters along the southern coastline and formed a long, curving tendril extending toward the middle of the lake, induced by wind activity.
Swirly plumes in Lake Erie (Fig. 38b), USA were attributed to seiche[1]. Seiche is a large standing wave that occurs when strong winds and a quick change in atmospheric pressure push water from one end of a body of water to the other. de Jong and Battjes (2004)[31] discussed the atmospheric origin of seiche.
In this first attempt, 24 different configurations of density plumes have been recognized. They do not constitute a systematic classification based on a single unifying concept. For example, although most configurations are named after shape (e.g., lobate), other factors are also used, such as color (e.g., whitings), process (e.g., eolian), etc. The 24 types are broadly grouped into 14 general categories (Fig. 39). Given this complex variability, any attempt to advocate facies models for deposits of hyperpycnal flows, under the false assumption that river- mouths are the only type of hyperpycnal flows, is misleading.
This and other recent studies have revealed that there are major implications for understanding 1) external controls, 2) hyperpycnal flows, 3) submarine fans, and 4) deep-water petroleum reservoirs.
External controls are allogenic in nature, which are external to the depositional system, such as uplift, subsidence, climate, eustacy, etc. However, external controls of density plumes are much more variable and include some common depositional processes (e.g., tidal currents). At least, 18 external controls of plumes have been recognized in this review (Fig. 40): 1) Tidal shear front (Fig. 3): The Yellow River[3]; 2) Winter season (Fig. 6): The Yangtze River[5]; 3) Ocean currents (Fig. 14): Rio de la Plata Estuary[4, 51, 52]; 4) Tidal currents (Fig. 12): San Francisco Bay[1, 8]; 5) Monsoonal current (Fig. 21): Indian Ocean[16]; 6) Eolian (Fig. 33): Red Sea; 7) Cyclones (Fig. 27): Gulf of Mexico, U. S. Atlantic shelf[21]; 8) Tsunamis (Fig. 29): Sri Lanka, Arabian[22]; 9) Braid delta and related high gradients and coarse sediments (Fig. 8): Alaska, Pacific Ocean[6]; 10) Seiche in lakes (Fig. 38b): Lake Erie, USA[1]. Seiche is a large standing wave that occurs when strong winds and a quick change in atmospheric pressure push water from one end of a body of water to the other. de Jong and Battjes (2004)[31] discussed the atmospheric origin of seiche. 11) Upwelling (Fig. 26): Off Namibia[20]; 12) Phyoplankton (Fig. 26): Off Namibia; 13) Fish activity (Fig. 32): The Great Bahama Bank[25]; 14) Volcanic eruptions (Fig. 34b): Bering Sea)[1]; 15) Glacial melt (Fig. 30): Greenland[23]; 16) Coral reef (Fig. 32b): South Pacific Ocean[1]; 17) Pockmarks: Carolina Continental Rise (Fig. 37), North Atlantic Ocean[30]; 18) Internal waves and tides[13] and related internal bores (Fig. 19).
I have discussed the importance of external controls, such as tidal shear front and ocean currents earlier. Future studies should consider external controls in developing meaningful depositional models.
In discussing river-mouth processes, geologists, geophysicists, and hydraulic engineers use process terms to represent hyperpycnal flows that are not consistent in meaning with each other, such as single-layer and multi-layer hyperpycnal flows (Fig. 41). For example, the following 16 concepts and terms are used in the literature[3, 34, 35, 54, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75]: 1) density flow[55], 2) underflow[34], 3) hyperpycnal flow[33, 56], 4) hyperpycnal underflow[34], 5) hyperconcentrated flow[57], 6) low-density hyperpycnal plume[34], 7) high-density hyperpycnal plume[34], 8) high-turbid mass flow[58], 9) supercritical hyperpycnal flow[59], 10) tide-modulated hyperpycnal flow (Fig. 41D)[3], 11) cyclone-induced hyperpycnal turbidity current[60], 12) buoyancy-dominated flow[61], 13) hyperpycnal turbidity current[62], 14) turbidity front[47], 15) turbidity current[34, 63, 64, 65, 66], 16) multi-layer hyperpycnal flows[67].
In this context, Gao et al. (2015)[76] suggested that the Yellow River has undergone a regime shift in response to resuspension induced by tidal currents and waves. This shift has presumably resulted in the replacement of hyperpycnal flows by bottom-turbid layers. The difference between the two is that hyperpycnal flows behave as a single layer without vertical stratification in density or velocity (Fig. 41a), whereas bottom-turbid layers reveal vertical stratification in density and velocity (Fig. 41b). Clearly, there is no clarity on the fluid mechanics of hyperpycnal lows among various researchers. Amid this uncertainty, it is premature to apply genetic facies models of hyperpycnites to petroleum reservoirs (e.g., Yang et al., 2017b)[77, 78].
Conventionally, submarine fans were related to deposition from turbidity currents[80]. Recently, origin of submarine fans have been attributed to hyperpycnal flows[74, 82]). Zavala and Arcuri (2016)[74] proposed two types of hypothetical hyperpycnal flows, namely, sandy and muddy types. In this classification, the lofting plume (i.e., positively buoyant) in sandy hyperpycnal flows is of significance. Similar lofting models were also proposed by Steel et al. (2016)[82]. Zavala and Ar-curi (2016)[74] also used the classification of turbidity currents by Mutti et al. (1994)[83] into low-density turbidity currents and high-density turbidity currents. Despite our poor understanding[84, 85] of the behavior of high-density turbidity currents and hyperpycnal flows in deep-water environments (Shanmugam, 2016; 2018b), Zavala and Arcuri (2016)[74] proposed two types of hyperpycnal flows, namely, sandy and muddy types. Importantly, they proposed two types of turbidites, namely "intrabasinal turbidites" and "extrabasinal turbidites" (Fig. 42). Intrabasinal turbidites are those with sediments derived locally from adjacent shelf and got transported into the basin by “ classic” turbidity currents. In contrast, extrabasinal turbidites are those with sediments derived from distant land and delta and got transported into the basin by “ flood- triggered" turbidity currents or hyperpycnal flows (Fig. 42). However, there are major problems with these concepts (Shanmugam, 2018a)[44]:
1)The concept of lofting hyperpycnal flow is problematic because it defies basic principles of buoyancy. In discussing buoyancy effects in fluids, Turner (1980)[86] explained that positively buoyant plumes cannot be hyperpycnal (i.e., plume with excess density cannot loft). By definition, hyperpycnal flows are negatively buoyant due to their excess density.
2) Zavala and Arcuri's (2016)[74] classified turbidity currents and fans based on provenance (internal source vs. external source), which is in conflict with the conventional definition of turbidity currents based on Newtonian rheology and turbulent state[87, 88].
3) The hypothetical model of extrabasinal turbidites and related hyperpycnite fans is untenable for two reasons. First, the concept of high-density turbidity currents, which serves as the basis, is not only theoretically flawed[89], but also empirically undocumented in the world's oceans[90]). Second, the model fails to take into account the most fundamental factor on developing submarine fans, which is submarine canyons[44].
4) Modern examples do not show development of submarine fans by hyperpycnal flows in modern settings. For example, hyperpycnal flows and related sediment plumes are totally absent in the deep Gulf of Mannar (Fig. 21b).
1)The river-flow triggered turbidity currents (i.e., hyperpycnal flows)[72] in experiments are purely muddy flows (Fig. 43).
2) Published data on river-flood derived hypepycnal flows are muddy in texture and they occur close to the shoreline in inner shelf environments. There has not been any study documenting transport of sand and gravel by river-flood triggered hyperpycnal lows across the entire continental shelf and into the deep-water basin in modern settings (double headed arrow, Fig. 43a). Mississippi River generates some spectacular density plumes at its mouth in shallow-water shelf environments (Fig. 44). But we do not know whether these plumes reach the deep sea. Available data show that density plumes tend to dissipate and die out as they travel away from the shoreline within a few kilometers (Fig. 44b). However, there are examples of plumes that reach the shelf edge and beyond. Such plumes are triggered by cyclones, tsunamis, eolian dust, volcanic explosions, upwelling, etc., but they are unrelated to river floods.
3) An important lesson learned from this study is that most river-mouth plumes are deflected or blocked by external controls. Therefore, hypypycnal flows are not viable transport mechanisms of sand and gravel across the shelf into the deep sea.
4) In a comprehensive review of hyperpycnal flows, Talling (2014, p. 179)[92] concluded that "Weak and dilute flows generated by plunging hyperpycnal flood discharges most likely deposits thin (mm to < 10 cm) and fine grained sediment layers, similar to those documented for hyperpycnal flows in lakes and reservoirs (Fig. 6d, 6e). The available field observations suggest that they do not form meter-thick sand layers in deep water settings, as has been previously proposed (Mulder et al., 2003)."
For the above reasons, river-flood triggered hyperpycnites are considered unsuitable for serving as petroleum reservoirs in deep-water environments until proven otherwise.
A global evaluation of density plumes suggests (Fig. 40): six different environments, six compositional material, 11 sources, 18 external controls, and 24 configurations. Despite their wide natural variability, individual density plumes have not been investigated to understand their depositional characteristics. The challenge in studying density plumes is that a single type (e.g., swirly) can be generated by different mechanisms (e.g., cyclone, seiche, upwelling, etc.). To date, no one has investigated how these different types of density plumes are preserved in the sedimentary record.
The results of this study raise some fundamental questions:
1) Are configurations of density plumes ephemeral or sustained for a period of time?
2) Does each configuration result in a unique depositional layer?
3) What is the relationship between depositional environment and plume configuration?
4) How external factors control plume configuration?
5) Can we distinguish deposit of a particular plume configuration from associated turbidites, contourites, or hemipelagites?
Amid these obstacles, opportunities do exist for future research projects.
I thank Dr. Huaixian Xu, Executive Chief Editor of Petroleum Exploration and Development (PED) for inviting me to contribute this article for the Special Issue of PED in celebrating the 60th anniversary of RIPED. 2018 also marks the 10th anniversary of my association with RIPED. I am grateful to NASA for their excellent collection of satellite and other images of density plumes triggered by rivers, tides, glaciers, volcanic eruptions, cyclones, etc. I also thank NASA, Jet Propulsion Laboratory, NOAA, Wikipedia, Clastic Detritus, SEOS, AAPG Bulletin, GSA Bulletin, Geology, Journal of Palaeogeography, Oceanography, Progress in Oceanography, Journal of Sedimentary Research, American Geophysical Union, Elsevier, Wiley, Butterworth-Heinemann, DigitalGlobe, and R. D. Kreisa for their images used in this article. As always, I am thankful to my wife Jean for her general comments.
The authors have declared that no competing interests exist.
[1] |
|
[2] |
|
[3] |
|
[4] |
|
[5] |
|
[6] |
|
[7] |
|
[8] |
|
[9] |
|
[10] |
|
[11] |
|
[12] |
|
[13] |
|
[14] |
|
[15] |
|
[16] |
|
[17] |
|
[18] |
|
[19] |
|
[20] |
|
[21] |
|
[22] |
|
[23] |
|
[24] |
|
[25] |
|
[26] |
|
[27] |
|
[28] |
|
[29] |
|
[30] |
|
[31] |
|
[32] |
|
[33] |
|
[34] |
|
[35] |
|
[36] |
|
[37] |
|
[38] |
|
[39] |
|
[40] |
|
[41] |
|
[42] |
|
[43] |
|
[44] |
|
[45] |
|
[46] |
|
[47] |
|
[48] |
|
[49] |
|
[50] |
|
[51] |
|
[52] |
|
[53] |
|
[54] |
|
[55] |
|
[56] |
|
[57] |
|
[58] |
|
[59] |
|
[60] |
|
[61] |
|
[62] |
|
[63] |
|
[64] |
|
[65] |
|
[66] |
|
[67] |
|
[68] |
|
[69] |
|
[70] |
|
[71] |
|
[72] |
|
[73] |
|
[74] |
|
[75] |
|
[76] |
|
[77] |
|
[78] |
|
[79] |
|
[80] |
|
[81] |
|
[82] |
|
[83] |
|
[84] |
|
[85] |
|
[86] |
|
[87] |
|
[88] |
|
[89] |
|
[90] |
|
[91] |
|
[92] |
|