For a long time, the traditional understanding of Chinese sedimentologists has been that continental sedimentary systems include alluvial fans at the mountain pass, rivers in the plain, lakes, deltas in transitional zones^[1,2,3]. In recent years, more and more foreign researchers have paid close attention to a type of fan-shaped sedimentary body formed predominantly by fluvial process, and the concept of fluvial fan has been proposed. Since 2000, the number of fluvial fan-related papers published has been rising significantly. The history of fluvial fan research can roughly be divided into four stages: (1) Before 2006, most researchers conducted preliminary research on fluvial fans on the basis of alluvial fan. (2) During 2007-2009, North et al.^[4] put forward a sedimentary model for fluvial fan, and defined fluvial fan, marking the first peak of fluvial fan research. (3) During 2010-2014, Hartley et al.^[5-6] introduced the concept of distributary fluvial systems (DFSs) based on satellite images and remote sensing observation, which attracted extensive attention from sedimentologists and made the classic controversy about fluvial fan and alluvial fan into a frontier research field, marking the second peak of fluvial fan research. (4) Since 2015, fluvial fan research has entered the third peak stage. In this stage, lots of researchers have delved into the origins, internal architecture, and seasonal factors, etc. of fluvial fans^[7], and the average number of papers published each year exceeded 130. Fluvial fans have remained a research hotspot in this stage.

In contrast to the progress of fluvial fan research abroad, only a few domestic researchers have set foot in this field. Zhang et al.^[8] gave an introduction to the concept of DFSs in 2017. Liu et al.^[9] described the study status of terminal fan. The study process abroad and in China should be introduced in details aiming at this important field possible to promote the sedimentology development. This paper gives a systematic elucidation on the concept, major research progress and controversy, identification marks of fluvial fan, and significance of fluvial fan research.

The concept of fluvial fan originated from alluvial fan research after a process of gradual evolution and separation. Denny^[10] pointed out that there were lobes formed by sudden current on alluvial fans, and the lobe size was a function of slope, radial length, and local geomorphology in 1967. Anstey^[11,12] thought alluvial fans formed through the expansion of lobes usually had a radius of less than 10 km, however, it is difficult to explain the large number of fans with a radius of more than 10 km in subaerial regions. In his book The Fluvial System, Schumm introduced fluvial process into the explanation of fan-shaped sedimentary bodies with stable fluid supply, which represents the basic rudiment of the concept of fluvial fan^[13]. Nemec et al.^[14] insisted on the presence of fluvial process in traditional alluvial fan system, and referred to alluvial fan with perennial channel flow on the surface as wet alluvial fan. Blair et al.^[15] held when a river flows from upland into a broad region, such as the Huaco or Jachal River in Argentina, the fan-shaped sedimentary bodies formed were neither alluvial fans nor deltas, but products of a fluvial environment, i.e., rivers or fluvial fans. According to the development degree of braided river, debris flow, and sheetflood, Gallway et al.^[16,17] classified alluvial fans into river-dominated, debris flow-dominated, and sheetflood-dominated fans. On the whole, before 2006, fluvial fans had attracted the attention of some researchers, but they were still considered a special type of alluvial fan.

In 2007, North et al.^[4] further defined fluvial fan, and pointed out that fluvial fans were products of frequent channel changes in the absence of horizontal restraints, and would turn out to be a regional diverging geometric morphology on a geological time scale, and single channels on fluvial fans were similar to fluvial sediments.

Hartley et al.^{[5, 18-19]} attributed fan-shaped sedimentary bodies with a length of 1-700 km dominated by fluvial and alluvial action as distributive fluvial system (DFS) on three scales: (1) large-scale fluvial megafans, with a radius of large than 100 km and an area of 1000-100 000 km²; (2) small-scale alluvial fans, with a radius of less than 30 km and an area of less than 100 km²; and (3) medium-scale fluvial fans with sizes in between the former two. Proposing the concept of DFS promoted fluvial fan research and also gave rise to fierce controversies. Through ongoing debates, some researchers began to adopt fluvial fan as the term for this type of fan after 2018.

With respect to the relationship between fluvial fan and alluvial fan, previous studies mainly proceeded from two aspects, i.e., origin and scale. In terms of origin, it is generally believed that fluvial fans are dominated inside by tractive current, while alluvial fans are dominated by gravity flow and sheet flow. In terms of scale, fluvial fans generally have a radius of more than 10 km, while alluvial fans are often less than 30 km in radius, and both of them would occur with a radius of 10-30 km. The differences between them are manifested in hydrology, geometric morphology, sediment transport mechanism, and stratigraphic architecture etc.^{[7, 17-18, 20]}. Due to differences in these aspects, compared to alluvial fans, fluvial fans are characterized by a lack of debris flow sediments, higher textural maturity of sediments, and larger scale.

Based on existing research results, fluvial fan can be defined as a type of fan-shaped sedimentary body developed at the mountain pass or in a plain, dominated by fluvial sediments diverging from the upstream apex to the downstream region on a geological time scale.

Syvitski et al.^{[21,22,23,24]} claimed that excess sediment load in the confluence area was the root cause of unstable channel belts. Most fluvial fans originate from tectonically active regions or broad, large-scale confluence areas. During the transport of overloaded clastics towards adjacent lowland, there is a persistent, strong trend of fluvial geomorphology construction over a time span of decades or centuries. Eventually, in the longitudinal direction, in the downstream direction with a gradual reduction of topographic gradient, fluvial fans dozens or even hundreds of kilometers in width could be formed. Back-arc foreland basins are usually favorable regions for the development of large-scale fluvial fans. For instance, an enormous number of fluvial fans occupy large areas of alluvial plains tilting down the Andes.

Mccarthy et al.^{[25,26,27,28]} discovered fluvial fans in rift, strike- slip, and back-arc basins too. Due to the topographic limits of basins of these types, fluvial fans developed in them are usually small in scale. Aslan et al.^[29-32] proposed that fluvial fans also existed in tectonically inactive regions under certain conditions, for example, a river flowing through upland or valley areas quickly entered an unlimited region.

Ielpi et al.^[33] studied the fluvial fan sedimentary model of the early Paleozoic Alderney Sandstone Formation, and suggested that enhancement of the interaction between river process and aeolian process was often accompanied by the decrease of channel and bar size. After investigating fluvial fans in the Kosi region of Northern India, Chakraborty et al.^[30] concluded that the majority of flood discharge was transmitted downwards via the distributive network of the fan area, providing an important drive for the formation of fluvial fans.

According to the viewpoints of all these researchers, fluvial fans are the products of certain conditions like tectonics, topography, climate, and sediment supply working together. First, the material basis of a fluvial fan is the large-scale confluence area at the upstream area of the river, which offers sufficient sediment supply. Second, the topographic basis is the adjacent lowland in the downstream direction, and the channel develops into a fan network after wedging in an endpoint. Affected by radial changes inside a channel, the fluvial fan has obviously different sedimentary characteristics at different positions^[34].

At present, there is still no widely accepted classification scheme of fluvial fans. However, researchers have conducted related studies from different perspectives. Some researchers, starting from the climatic environments in which fluvial fans develop, have identified fluvial fans of subtropical monsoon climate, arid climate, and extreme climate^{[5-6, 32, 35-37]}. Based on the types of basins in which fluvial fans develop, some researchers have classified fluvial fans into those of back-arc foreland basin, rift basin, and strike-slip basin^{[5-6, 28, 38-39]}. Other researchers suggested that according to the development positions of fluvial fans and their relationship with adjacent sedimentary environments, fluvial fans can be classified into fluvial fans developed on alluvial fan, in the middle of a large river, and at the end of a large river etc.^{[18, 40]}

Mikesell et al.^[41] studied fluvial fans formed by channel divaricating at the Lawrence Livermore National Laboratory (LLNL) in California, US, confirmed the existence of divaricating fluvial fans, and summarized their sedimentary model. According to the morphological characteristics of channels in fans, Hartley^{[5-6, 8]} classified the sedimentary types of DFSs (including fluvial fans) into six types, i.e., single-braided channel, divaricating braided channel, braided channel transformed to meandering channel, small-scale divaricating meandering channel, multi-meandering channel, and main meandering channel. After exploring the effects of channel wandering on the formation of Chacaito fluvial fan in Venezuela, Escalona et al.^[42] confirmed the existence of wandering fluvial fans.

By combining previous research results with existing geomorphological data on modern fluvial fans, starting from the origin and morphology of fluvial fans, considering the morphology of channels inside and external macro-morphology of the fluvial fan, the channels inside the fluvial fan are classified into four types, i.e., straight, bending, divaricating, and wandering. Next, according to evolutionary morphology, fluvial fans are classified into progressive, swinging, and secondary types. Progressive-type fan refers to the fans with overall forward stacked evolution and relatively complete fan morphology. A swing-type fan is formed through the transverse swinging of two or more subfans, and has asymmetric characteristics. Secondary-type fan refers to the fan mainly evolved from secondary fans formed by terminal channels, in which the primary fan has multiple secondary fans embedded in, showing nesting characteristics at the edges.

Correspondingly, the classification of fluvial fans should consider both morphology of internal channels and macroscopic fan types, and be a combination classification. Theoretically, there are 12 fluvial fan types and multiple transitional types. Fig. 1 shows several common fluvial fan types. The channels in a divaricating progressive-type fluvial fan are largely divaricating, and there may also be a few active main channels. The fan is formed by the gradual forward stacking evolution of fluvial sediments from internal channels (Fig. 1a). The channels of a straight progressive-type fluvial fan are straight, and the fan is dominated by overall forward stacking evolution (Fig. 1b). The channels in a divaricating secondary-type fluvial fan are divaricating, and the fan is formed by persistent secondary action at the channel terminal (Fig. 1c). The channels of a straight secondary-type fluvial fan are largely straight and rarely divaricating, and the fan also is dominated by secondary evolution at the channel terminal (Fig. 1d). The channels of a bending swinging-type fluvial fan are largely bending or meandering, and the fan is composed of several swinging subfans (Fig. 1e). The channels in a wandering swinging-type fluvial fan are basically wandering or scattered, and the fan is also composed of several swinging subfans (Fig. 1f).

The statistics of this study encompass modern fluvial fans developed in more than 280 regions worldwide. Among them, the western US, western South America, central Africa, the Middle East, and western China are major regions with fluvial fans. What needs to be particularly pointed is that fluvial fans are mostly in mid-latitude regions and rarely seen in high latitude regions, suggesting that their development is highly controlled by climatic factors.

China is a major region for the development of fluvial fans in the world, and in central and western China alone there are more than 50 regions with fluvial fans (hundreds in total). The peripheral regions of Tarim Basin, central and western regions of Inner Mongolia, and western regions of the Qinghai-Tibet Plateau are regions with concentrated development of modern fluvial fans. Among them, the peripheral regions of the Tarim Basin have several mega-fans more than 50 km in radius; the central regions of Inner Mongolia have fluvial mega-fans more than 150 km in radius of arid climate. In addition, by digging into literatures, we found major sedimentary basins, such as the Tarim Basin, Ordos Basin, and Bohai Bay Basin, all had favorable geological conditions for the development of ancient fluvial fans.

In 2006, Bennett et al.^[43] investigated three fluvial fans, Kings, Tuolumne, and Merced on the east side of the San Joaquin Valley in California, US, established a relationship between ground penetrating radar (GPR) signals and the activity of ancient soil and channels, and proposed that the enhancement of GPR signals corresponded to ancient channel active areas. They found that fluvial fans with active ancient channels were extensively developed with channel sediments, while those with inactive ancient channels had extensive development of fine-grained flood plain sediments.

In 2016, Escalona et al.^[42] explored the effect of channel swinging on the formation of Chacaito Fluvial Fan in Venezuela, and confirmed the effect of channel swinging on fan morphology and scale using a 2D numerical simulation method based on mixed water and sand flow. This study verified the existence of wandering fluvial fans by example and simulation.

In 2016, Galve et al.^[44] examined the effect of volcanism on the Santa Clara Fluvial Fan in northern Costa Rica. Relying on unmanned aerial vehicles (UAVs), GPR, satellite imaging, and other technical means, they discovered that volcanism and associated seismic activity greatly promoted the fluvial fan supply in this region and accelerated the expansion of the fluvial fan morphology and scale. The results of this study indicate that sediments produced by volcanoes, earthquakes, and other events directly affect the formation and scale of fluvial fans.

In 2016, Ielpi et al.^[33] investigated the Alderney Sandstone Formation in the Channel Islands of the UK, and overthrew the previous study that sediments in this region were the products of fluvial process alone. Instead, they argued that sediments there were multi-stage fluvial fan sediments affected by wind in arid climate. By comparing with modern fluvial fans in Mexico and Turkmenistan, they summarized the evaporation-dominated sedimentary characteristics at the end of fluvial fans in arid periods, and pointed out that intensified channel divarication and enhanced wind effect along the downstream direction were important features to identify fans of this type^[45]. Fans of this type are very similar with modern fluvial fan sediments in the desert environment in the peripheral regions of the Tarim Basin, so this study can provide reference for further study of fluvial fan sediments in China.

In 2016, Toorenenburg et al.^[46] studied the physical characteristics of fluvial fan sediments and reservoirs in the Ebro Basin of Spain. By combining digital outcrop technology with logging data, they analyzed the direct proportional evolutionary relationship between sand body connectivity and channel stability inside flood plains. They found that, under a humid climate, the distal area fine-grained sediments was highly developed, the channels were low in bending degree at the ends.

The modern fluvial fan of the Tabu River is located 200 km north of Siziwang Banner of Inner Mongolia, China, and formed by erosion and filling of the Tabu River from south to north. The Tabu River is an inland braided river under semi-arid climate, and its channel begins to divaricate from one apex, forming two-stage fluvial fan sediments 36 km and 43 km in radius respectively in an open and flat downstream region. On the west side of the first stage fan is a piedmont alluvial fan development area, where there are a large number of alluvial fans, with the largest radius of 5.6 km. The stage-I fan is an early-stage fluvial fan, with obvious abandoned channels and wind reworking signs. The second stage fan remains active today, and on the east-most side of the fan there are active fluvial systems. The staging of fans reflects the migration of the channel network from west to east (Fig. 2). The channel gradually dries up and terminates at the end of the fan. On the two stages of fans, 13 types of fluvial fan sedimentary architectures, including aeolian deposit, abandoned channel, channel, incised channel, gravity flow channel, basal gravel, terminal channel, lateral bar, downstream bar, embankment, flood fan, sheet flow, and flood plain fine-grained sediments were found. Among these, complex sedimentation-related architectures of channel account for 56%, suggesting that fan evolution is mainly controlled by fluvial system; flood plain complex architectures account for 34%, and aeolian sediments account for 10% (Fig. 3).

In this research, 17 survey points were selected in the fluvial fan for detailed analysis (Fig. 2). The fluvial fan of the Tabu River complete in morphology and clear in apex, is dominated by conglomerate, sandstone, and argillaceous siltstone in lithology, and mainly composed of a channel complex, a flood plain, and aeolian sediments (Fig. 3). Typical profiles at proximal area, medial area, and distal area were selected for comparative analysis. At the proximal area, the sediments are mainly sandy gravel sediments 5-10 cm in grain size, which show characteristics of strong hydrodynamic forces, high-angle cross-beddings, obvious channel down-cutting, large single-stage channel scale, and small-scale of fine-grained backshore sediments (Fig. 4). At the medial area part, the sediments are mainly gravels and sands less than 5 cm in grain size. This part features weak hydrodynamic conditions, low-angle cross beddings, thinner single-stage channel sediments, poorer channel stability, enhanced overbank and flooding actions, and massive development of flood plain argillaceous and silty sediments (Fig. 5). The distal area is dominated by sandy and silty sediments, with a small amount of gravels less than 1 cm in grain size. Main characteristics of this part include extremely weak hydrodynamic conditions, non-development of cross beddings, small terminal scale, poor stability, superimposed development of multi-stage backshore argillaceous and silty sediments, and universal reworking of sediments at the channel terminal (Fig. 6).

After comparing the profiles of seven proximal survey points, six medial survey points, and four distal survey points in the two stages of fans, it was found that: (1) The proximal area has mainly channel sediments (42%), incised channel sediments (20%), and a small amount of gravity flow channel sediments (6%); and a low development degree (about 5%) of abandoned channels, suggesting that proximal area had sufficient provenance supply, strong hydrodynamic conditions, and stable channels. (2) At the medial part, the fluvial system expanded in transverse rapidly and dropped in the proportion of active channels significantly to 25%, a drop of 17% from that at proximal area; while channel bars, flood plains, and other flood plain sediments (38%) are commonly developed, there is also an increase in the number of abandoned channels (10%) and lateral bars (15%). All these imply that the medial area part had weak hydrodynamic conditions, poorer channel stability, frequent migration and abandonment of channels, and enhanced sedimentation. (3) At the distal area, channels gradually dried and reduced in loads, and the sediments are dominated by terminal channel (26%) and abandoned channel sediments (21%); all kinds of bars are low in development degree (6%). These observations suggest that the distal area has extremely low hydrodynamic force, small channel scales, low fluvial loads, and extremely poor stability (Fig. 3).

Aeolian sediments are found in all parts of the fan with difference. At the proximal area where sediment grains are coarse and the topography is narrow, aeolian sediments are small in scale. At the medial area part where the sediments are smaller in grain size and the fan expands in area, fine-grained backshore sediments are susceptible to be reworked by wind to form aeolian deposit of considerable scale. At distal area where the channels increase in number, but decrease in scale, and become poorer in stability, the sediments are small in grain size, dominated by fine sand and silt, and significantly affected by wind force, leading to rapid expansion of aeolian deposit scale.

Abandoned channels are widely developed in all parts of the fan. The proximal area has a small number of abandoned channels, mainly because of the strong hydrodynamic force and high stability of channel. At mid-fan part, the phenomena of channel migration and flooding become obvious, and the number of abandoned channels increases gradually. At the distal area, hydrodynamic conditions were extremely weak, and unstable terminal channels were the main channel systems; affected by factors such as wind force and aridity, a large number of abandoned channels come up.

In summary, the modern fluvial fan of the Tabu River is a typical fluvial fan formed by braided river. With large scale, complete sedimentary configuration, distinctively different sedimentary characteristics in various parts, and representative sedimentary phenomena, it can be taken as a typical profile for fluvial fan research.

Fluvial fan research is going deeper and deeper. As large- scale subaerial sedimentary bodies, fluvial fans are closely related to, and somewhat similar to, alluvial fans, rivers, lake deltas, and lake fan deltas, etc. Thus, the basic premise of deepening fluvial fan research is to sort out identification marks of fluvial fans to effectively identify them from similar sedimentary bodies.

In 2018, Ventra et al.^[18] claimed that alluvial and fluvial fans were ubiquitous sedimentary systems, with similar sedimentary backgrounds to some extent. With respect to the relationship between alluvial and fluvial fans, there have always been controversies. According to studies published in recent years, the two sedimentary systems are different in both basic morphology and sedimentary process, and have essentially different sedimentary facies associations and internal structure characteristics^[7].

On the morphological differences of them, in 1990, Gohain et al.^[47] for the first time introduced the term “megafan” to distinguish alluvial fan and fluvial fan, and defined fans with a radius of less than 30 km as alluvial fans, and those with a radius of more than 30 km as “megafans” (or fluvial fans). With understanding going deeper, Blair^[48] argued in 2003 that alluvial fans basically had a radius of less than 10 km (generally hundreds or thousands of meters), and that fans with a radius of more than 10 km were fluvial fans. Hartley et al.^{[5-6, 49-52]} further quantified this difference, and suggested that sedimentary fans with an area of more than 105 km² were fluvial fans, while those with an area of less than 105 km² were traditional alluvial fans.

Alluvial and fluvial fans also differ fundamentally in structure. Fluvial fans have typical fluvial system on the surface, and different types of sediments in the broad areas of channel and backshore; better grain size and sorting of sediments than alluvial fans; diverse sedimentary facies, and complex transverse and longitudinal changes and associations of sedimentary facies. Channel gradient changes make fluvial fans prograde forward and aggrade laterally at the same time. Chakraborty et al.^[53] proposed that channel would gradually decrease in width and depth and braided river would transit to meandering river frequently towards the downstream direction.

Based on the previous viewpoints, it can be concluded that the differences between fluvial and alluvial fans mainly manifest in the following aspects:

(1) There are significant differences between them in scale. According to the statistics on more than 100 fluvial fans worldwide using satellite imaging, most fluvial fans have a radius of 10-50 km, and the largest one has a radius of more than 160 km. In contrast, the majority of alluvial fans have a radius of less than 10 km. In summary, fans with a radius of more than 30 km are largely fluvial fans, while those with a radius of less than 10 km are mostly alluvial fans; fans with a radius of 10-30 km can be either of them.

(2) Alluvial fans are formed in piedmont open regions through the rapid accumulation of sediments carried by seasonal floods in mountainous areas, while fluvial fans are formed through the radial accumulation of sediments when a river flows from adjacent upland areas into an unlimited region of an open depression.

(3) In terms of sediment transport mechanisms, in alluvial fans, the main transport mechanisms are gravity flow and sheet flow, while in fluvial fans, the major transport mechanism is fluvial action with obvious channel abandonment effect, and sedimentation is mainly controlled by tractive currents.

(4) In terms of overall sediment characteristics, sediments of alluvial fans are characterized by debris flow chaotic accumulation, and low textural maturity and ordering degree; while fluvial fans are characterized by tractive current sediments with higher textural maturity and stronger regularity in transverse and longitudinal directions.

(5) In terms of the internal structural stability of sedimentary bodies, alluvial fans have lower subfacies stability. As alluvial fans are steep in slope, adjacent to provenance and small in area, sediments of debris flow, channel, sheet flood etc. respond intensely to provenance change, and have frequent changes in facies^[54]. In contrast, fluvial fans appear in areas with gentle slopes and larger area, and have channels, flood plains, peat bogs, and other microfacies; so the associations of various subfacies and microfacies are in a more orderly manner.

Friend^[20] was the first to point out the differences between ancient fluvial fans and modern fluvial systems. According to him, the sedimentary characteristics of fluvial fans include a gradual decrease in sandstone grain size and stratum thickness, and the increase in the proportions of fine-grained sediments, small-scale cross beddings, and horizontal beddings towards the downstream direction. Compared with typical fluvial facies, subaerial fluvial fans terminate at distributive channel network, and do not reach the sea or lake. Kelly et al.^[55] used the fluvial fan model to describe a same-level distributive network. According to their description, with the divaricating of the river, the fan gradually progrades forward and laterally towards the downstream direction. Although the two models fail to describe the characteristics of typical and complete fluvial fans, it can be seen that the typical morphological differences between the fluvial fan and ordinary fluvial systems lie in the divarication at the end of the fluvial fan, and the transverse expansion of channels.

Fluvial fans are complex sedimentary bodies formed by divaricating river nodes under frequent channel shifts in the absence of horizontal restraints. They are obviously different from small-scale flood fans formed under flood action in fluvial facies. Fluvial fans have multiple channels on the surface, and the sedimentary systems formed by these channels on a long geological time scale show regional radial distribution. This is the most significant morphological difference that distinguishes fluvial fans from other non-fan-shaped fluvial systems, especially fluvial systems limited by a valley^[4].

Transformation between fluvial fan and fluvial system may occur under certain conditions. According to Leier et al.^[50], given a sufficiently large horizontal space, a fluvial fan could take shape at any part of a river, and the same river could give rise to multiple fluvial fans along its course. The part where a fluvial fan emerges must meet two conditions: (1) The river transits from narrow to open topography. (2) The environment is suitable for expansion and divarication of river. Given that periodic flood events increase channel instability, large-scale fluvial fans are mostly developed in regions where rivers have seasonal changes in discharge significantly.

After reviewing the viewpoints of different researchers, we found that fluvial fans differ from ordinary fluvial facies in development scale and proportions of typical sedimentary units: (1) Fluvial fans have large fan parts, high migration and evolution degree of channels inside, many abandoned channels, and reworked sediments larger in amount and scale than ordinary fluvial facies. (2) In terms of position, fluvial fans are sedimentary bodies developed by fluvial systems under specific conditions. That is, it's a special type of fluvial system. When a river flows from one endpoint along the downstream direction into an adjacent depression, the fluvial system extends rapidly in the transverse direction, and the sediments it carries accumulate and are reworked within a fan area, thus forming fluvial fan. (3) From the perspective of distribution characteristics of sedimentary facies, fluvial systems are more significantly controlled by topography, and confined within a narrow environment in most cases. Due to the limited transverse development area, the sedimentary facies are smaller in scale^[56]. Fluvial fan is a kind of fluvial system without horizontal restraints, so various sedimentary facies can evolve within a wider scope over longer time, ending in larger scale.

Based on flume experiment results, North et al.^{[4, 57]} suggested that, different from channel branching in ordinary river-dominated deltas, the extension of a fluvial fan is the result of frequent silting-up and migration of multiple channels. In fluvial fan, channel divarication is not the division of one channel into two, but a “pseudo-divarication” phenomenon in the profile in which channels of two stages (or more) change courses from one point under the influence of siltation, discharge change, and other factors in different historical periods. This theory fundamentally distinguishes the sedimentary mode of fluvial fans from that of traditional deltas^[58].

By observing and investigating modern fluvial sedimentary systems, Olariu et al.^[58-59] found that typical deltas and fan deltas mainly developed in the vicinity of the base level of stable water bodies, and were significantly affected by fluctuations in the base level of lakes or seas. The diffusion of sediments is obviously controlled by the water body, and the accumulation and preservation degree of sediments depend on rise and fall of base level^[60]. Furthermore, lake fan deltas usually rely on the provenance supplied by alluvial fans, and their fan-delta plains have typical alluvial fan characteristics and weak fluvial action. Near shore underwater fans, sublacustrine fans, and other underwater fans are mainly developed below the base level of stable water bodies. In terms of sedimentary mechanism, they are obviously affected by gravity flow. The changes in accommodation space caused by base level fluctuations also exert a vital effect on the scale of various underwater fans. In contrast, modern fluvial fans can emerge in broad continental facies regions such as mountainous areas and plains. Additionally, fluvial fans can transport sediments to lakes or seas along any continental position, and can be found in many regions, from isolated inland basins unaffected by base level^{[4, 61]} to lakeshores and coasts jointly controlled by rivers and lakes (seas)^{[29, 49, 62]}. Under this mode, the fluvial fans developed in isolation in mountainous areas or on the plains are naturally different from lake fans, mainly due to their different development positions. Fluvial fans developing near lakeshores usually enter water bodies, forming fluvial fan deltas. A major difference that distinguishes fluvial fan deltas from various lake fans is that fluvial fan deltas have obvious fan areas formed by fluvial action, and small-scale deltas and fan deltas at the end of the fluvial fan. This characteristic has been observed in the Ganjiang region of Poyang Lake and the Gangcha region of Qinghai Lake in China.

In brief, fluvial fans differ from lake deltas, fan deltas, and various underwater fans in the following four aspects: (1) From the perspective of development position, fluvial fans have a wider development position scope, and can be found in mountainous regions, on plains, and in the vicinity of lakes. (2) In terms of fan composition, fluvial fans are mainly affected by fluvial action, and limitedly affected by lake action. Even the fluvial fan delta has a large-scale fan formed by mainly fluvial action; under lacustrine condition, small-scale regular delta sediments would emerge at the edges of the primary fan. (3) From the perspective of control factors, fluvial fans are deposited and reworked predominantly under fluvial action, and its primary fan is either far away from the lake or prograding by a limited distance towards the lake, so it is less affected by the stable water body. (4) In terms of internal channel distribution patterns, the fluvial systems inside fluvial fans are dominated by branch channels, and are not necessarily active in the same stage; they may represent spatial-temporal migration of several wandering rivers. In comparison, the channel systems inside lake deltas and fan deltas are dominated by distributive channel networks with most distributive channels active in the same stage. In various kinds of underwater fans, gravity flow channels take the majority.

Weissmann et al.^[6] have summed up four general characteristics of DFSs (including fluvial fans): (1) The alluvial system creates sedimentary action towards the direction of a basin center in unlimited region. (2) The channels appear in a radial pattern from the apex to the downstream direction. (3) In normal circumstances, a sedimentary body with a transverse convex pattern and a longitudinal concave pattern would be formed. (4) There is a node, above which the fluvial system is located in an incised valley, and under which the fluvial system is distributed on the active sedimentary lobe. Moscariello^[7] has summarized four main characteristics of fluvial fans: (1) They have typical fluvial sedimentary process, and continuous sedimentary zones in channels and on banks. (2) Changes in the topographic gradient and patterns of channels in downstream and lateral directions are consistent. (3) They have regular changes of sedimentary structures of clasts towards the downstream direction. (4) They have facies associations evolving in regular patterns in the transverse and longitudinal directions. Ventra et al.^[18] have sorted out seven general characteristics of fluvial fans: (1) Starting from the apex of the fluvial fan, the fluvial system and residual channel axes on the fan show radial distribution. (2) The primary fan is frequently superposed with secondary fans formed by isolated channels. (3) From the apex to the toe of the fluvial fan, channel patterns transit from braided, to wandering, to highly bending, and the channels terminate or converge into a single stable channel at the end of the fan. (4) Under arid conditions, channels would decrease in width and depth towards the downstream direction of the fan, forming so-called terminal fans. (5) Towards the downstream direction, typical sedimentary structures of channel and overbank sediments decrease gradually. (6) Non-channel area increases gradually towards the downstream direction. (7) Perennial and seasonal water levels also decline towards the downstream direction.

By combining existing research results with the case study of fluvial fan of the Tabu River, we have sorted out the differences between fluvial fans and related sedimentary bodies (Table 1) and ten common characteristics of typical fluvial fans: (1) They generally have an average radius of more than 10 km. (2) They have primary fluvial systems entering an open region with a low gradient from a confluence area or system. (3) They have a single apex, are convex in transverse and concave in longitudinal directions, and are distributed radially from the apex downward. (4) They have typical fluvial sedimentary characteristics, and are unaffected by stable water bodies; they can be formed either by distributive channel systems, or by swinging of several rivers. (5) Due to the topographic gradient, the channels spread out transversely and branch radially at the end; so toward the downstream direction, the number of abandoned or active channels increase, down-cutting action of the channel becomes weaker, the proportion of non-channel sediments increases, and the sediments become smaller in grain size and higher in textural maturity. (6) The development scale of fluvial fan is significantly affected by climate and geomorphology, and the primary fan is usually superposed with secondary fans or lobes formed by multiple isolated channels. (7) Inside fans, channel and overbank sediments develop contiguously, peat bogs, flood plains, and other sedimentary micro-facies also emerge in quite large scale; and the channel-backshore ratio shows regular changes inside the fans. (8) The fluvial fans have diverse types of ends, some converging into an axial river, some terminating at the end of branch channel, some terminating at a sea, lake, or other stable water bodies. (9) With channels frequently changing course and in abandonment, fluvial fans have abandoned channels and aeolian sediments well-developed. (10) The sedimentary sequence of fluvial fan is controlled by the interactions between channel construction and channel abandonment, and may have positive and reverse rhythms alternately.

Research on fluvial fan sediments has undergone several stages, and many related or similar terms have been put forward, such as wet alluvial fan^{[13, 63]}, alluvial megafan^[63,64], terminal fans^[65], fluvial megafan/large fluvial fan, and distributive Fluvial System (DFS) etc.

When fluvial fan is presented as a sedimentary system with equivalent status to alluvial fan and fluvial facies, the terms related to alluvial fan, such as wet alluvial fan, alluvial megafan, and the terms related to fluvial facies such as terminal fan, and fluvial megafan, can all be replaced by the concept of fluvial fan. In addition, the generalized DFS includes sedimentary types on three scales, i.e., small-scale alluvial fan, medium-scale typical fluvial fan, and large-scale fluvial megafan. Thus, it is necessary to analyze the relationship between DFS and fluvial fan.

DFS describes the sedimentary pattern of a type of channel and flood plain. It refers to the fan-shaped sediments formed towards the downstream direction by clasts carried by channels originating from a confluence area after they enter a basin and begin to diverge from an apex.

Some researchers even maintain that DFSs have dominated the main geomorphological elements of modern sedimentary basins. Around 88% of modern fluvial sediments belong to DFS, while sediments of other distributive systems only account for 1%-12%^[19]. According to this viewpoint, the alluvial fan, fluvial fan, and megafan described above all belong to DFS^{[5-6, 8]}.

Weissmann et al. pointed out that the scale of a DFS mainly depended on the size of channels developed within the basin, and most basins contain several DFSs. General characteristics of DFSs include: (1) The alluvial system produces sediment towards the direction of the basin center in an unlimited region. (2) The channels are in a radial pattern from the apex to the downstream direction. (3) A sedimentary body convex in transverse and concave in longitudinal directions would be formed. (4) There is a node, above which the fluvial system is located in an incised valley, and under which the fluvial system is distributed on sedimentary lobe. According to the viewpoint of Wessiman et al., at the macroscopic level, the DFS is characterized by fan sediments, dominated by fluvial sediments inside, and similar to those of rivers in internal sedimentary characteristics. Thus, at a macroscopic level, aside from the above four general characteristics, DFS also has the combination characteristics of different types of rivers. Hartley et al.^{[5-6, 8]} classified DFSs into six types, single braided channel, branched braided channel, braided transited to meandering channel, branched meandering channel, multiple meandering channel, and single meandering channel. Sambrook et al.^[6] even claimed that DFSs dominated the sedimentary zones of all modern sedimentary basins. Since 2010, many researchers have studied DFSs from different perspectives on the basis of the concept of DFS^{[24, 35]}. What needs to be especially highlighted is that there have always been doubts about DFS since the concept was first proposed. In a review published in 2012, Fielding et al.^[36] raised a series of questions on the scientific validity and necessity of DFS term, the integrity of evidence, and data accuracy from two aspects, i.e., modern sedimentation and ancient sedimentation, and pointed out that sediments of trunk rivers equivalent to modern large rivers were preserved in large quantities in stratigraphic records, and research on modern tributary systems showed that modern tributary systems could substitute DFSs.

However, if all sedimentary bodies of this type in continental sedimentary systems were incorporated into DFS, it isn’t able to highlight the dominant role of fluvial action in the formation processes of these fans. Meanwhile, the DFS stresses on the geomorphological characteristics of sedimentary fans, and this classification scheme cannot be used to satisfactorily study fans at different parts from the perspective of origin in sedimentological applications.

5.2.1. Sedimentological significance

Research on fan-shaped sedimentary bodies is an old yet vibrant field of sedimentological research. In the traditional sedimentological system, alluvial fan, fan delta and delta, and deep-water fan are the main fan types of subaerial region, transitional region, and sea (lake) region, respectively^[1,2,3]. Proposing the concept of fluvial fan is another revision and supplementation to traditional sedimentology by modern sedimentation research. As continental sedimentary fan, fluvial fan can interact with alluvial fan, fan delta, and delta, forming complex sedimentary bodies. By further exploring the relationships between fluvial fan and alluvial fan, river, delta, and fan delta, we can get a better understanding on the role of river in continental sedimentary processes and promote the advance of sedimentological research.

As a new sedimentary mode, fluvial fan is triggering a rediscovery of modern and underground sedimentary systems worldwide. Hartley et al.^[5,6] analyzed the characteristics of sedimentary systems in more than 700 basins all over the world, and found more than 400 large-scale fan-shaped sedimentary bodies differing from typical alluvial fans and deltas. From satellite imaging, many fluvial fans were also found in Xinjiang and Inner Mongolia, China. In addition, there are many large-scale continental petroliferous basins in China. Fluvial fans, as ubiquitous continental sedimentary bodies in modern sediments, must have existed in ancient petroliferous basins. Therefore, investigating fluvial fans as independent sedimentary systems can give us a deeper understanding on the evolution-inheritance relationship between different continental sedimentary systems from the perspective of sedimentation origin. The new understandings in these aspects will have a profound influence on sedimentological research.

5.2.2. Significance for oil-gas exploration

Fluvial fans commonly have channel sand bodies inside^[4] which can be potential high-quality reservoirs^[7]. Some fluvial fans have a thickness up to 1 km. Through comparison, both alluvial fans and deltas can develop high-quality oil-gas reservoirs. Fluvial fans, as continental sedimentary bodies developed between alluvial fans and deltas, are larger than alluvial fans and closer to provenance than deltas^[18]. There are wide flood plain sedimentary zones inside fluvial fans, which can form fine reservoir-cap associations with channels and overbank sand bodies.

So far, fluvial fan reservoirs have been discovered in several oilfields both in China and abroad. The fluvial fan developed under arid climate in the Neuquen Basin of central Argentina has formed a high-quality reservoir of 10-12 m thick and dozens of kilometers wide, which has made a stable yield of oil and gas^[66]. In the fluvial fan sediments developed under humid climate and volcanism in the San Jorge Basin of southeastern Argentina, the reservoirs are 300-1500 m thick, making this area an important oil-producing area^[7]. Many regions in Southern North Sea also have fluvial fan sediments and produce petroleum industrially^[7]. In China, potential fluvial fan reservoirs have been discovered in the upper Es2 submember of the Paleogene Shahejie Formation of Pucheng Oilfield, the Triassic Baikouquan Formation of the Mahusag, the Fuyu Reservoir of the Sanzhao sag in the Songliao Basin etc.^[9].

Based on the above analysis, it is inferred that Songliao Basin, Bohai Bay Basin, Junggar Basin, Ordos Basin, Qaidam Basin etc. of continental petroliferous basins in China probably have fluvial fan sediments of a considerable scale^[67,68]. Investigating fluvial fan sediments distributed in these areas may change the understandings on sedimentary systems of continental petroliferous basins in China, and have a major impact on oil-gas exploration^[58].

The concept of fluvial fan has been evolving continuously. The proposal of DFS has greatly increased the depth and breadth of fluvial fan research; however, compared with DFS, the “fluvial fan” concept independent from alluvial fan, can better meet demands of sedimentological research.

Fluvial fans have typical identification marks, and are significantly different from alluvial fans, fluvial facies, and lake deltas in development position, hydrodynamic condition, sedimentary system morphology, and sedimentary facies associations, etc.

The development of fluvial fan is controlled by multiple factors. Climate, tectonics, provenance, wind field, volcanic events, and other external factors control the development and scale of fluvial fan; and the activity of channels inside a fluvial fan controls the facies development and lithological associations inside the fan.

Fluvial fan research will enrich traditional sedimentological understandings, and is currently triggering a rediscovery of modern and ancient continental sedimentary systems worldwide. Investigating modern fluvial fan systems is of vital significance for hydropower construction, environmental impact assessments, and geological hazard prevention, and exploring ancient fluvial fan sediments will offer new ideas for research on oil-gas reservoirs.