­4 DISCUSSION
4.1 Net discharge and erosive potential
Net discharge (Qnet) from Araguari River towards Amazon River during the dry season reached 548 m3s-1; it increased to 3,688 m3s-1 during the rainy season. Therefore, river discharge was the main forcing agent between Araguari River and Amazon River mouth during the rainy season, which recorded magnitude 6.7 times greater than that of the dry season.
Urucurituba Channel is the one with the greatest erosive potential, both in lower Araguari River and in Amazon River mouth (Bailique). Maximum seasonal Qnet values were recorded in the two points of the channel (URU1 and URU2), which, consequently, recorded maximum current velocity values: VURU ≈1.01 ms-1 in the ebb tide and 0.67 ms-1in the flood tide – it is worth emphasizing velocity and depth are directly proportional to susceptibility to erosion (Novo, 2008). For example, expressive depths of approximately 28 m in almost the entire URU1 length, and of approximately 24 m in URU2 were notable (Santos et al., 2018). There was greater influence of tides spreading from Urucurituba and Gurijuba-Igarapé Novo channels and from branches in the floodplain to ARA1, during the dry season. Torres et al. (2018) have suggested that the recent clogging at Araguari River mouth, due to drainage of its water by Urucurituba Channel, has contributed to increase Qss and Qnet at Amazon River mouth, which has significantly accelerated erosion and accretion processes in Bailique archipelago region.
Despite the scarcity of information about the region, records show that mean discharge in June 2011 reached 2,540 m³s-1 (rainy season) and 316 m³s-1 in December 2011 (dry season) 60 km away from the old river mouth, just downstream the confluence with Igarapé Tabaco (Santos, 2012). However, Araguari River discharge has significantly decreased. In June 2013 (rainy season), it decreased to 34 m³s-1 and, subsequently, recorded negative discharge of –212 m³ s-1 (drought in September 2013), which indicated permanent flow direction inversion due to greater influence of the upstream tide, which has fully changed the hydrodynamic pattern of its mouth (Santos et al. 2018).
Water and sediment transport balances (Figures 4 and 7) have suggested high hydrological and hydrosedimentometric complexity in the region, as well as identified accelerated Urucurituba Channel evolution, which not only resulted in the expansion of a main drainage channel, but induced severe erosion processes, as well as uncontrolled branching of several waterbodies and of their connections. For example, Gurijuba-Igarapé Novo Channel stands out for its connection to Urucurituba Channel, given the presence of interconnected branches that increase the flow direction complexity, since these flows interact with each other, are chaotically distributed across the floodplain and get intensified by the tidal regime (Gallo and Vinzon 2015; Abreu et al., 2020). However, they remain intense even during the dry period, when the connection between floodable fields and the main system still influences the ecological dynamics in the region (Cunha and Sternberg, 2018).
Several environmental consequences arise from this new hydrological dynamics since it influences and worsens erosive processes and enables the development of new channels. This process intensifies the natural geological fragility of coastal estuarine areas and makes them apparently more attractive to agriculture and, consequently, to soil trampling by buffaloes due to riparian vegetation removal (Trimble 1994, Santos 2006).
4.2 Sediment transport: scenarios and implications
The hydrographic basins of Araguari and Amazonas rivers present different geological features, which result in white water in Amazon River and in clear water in Araguari and Igarapé Novo rivers. Thus, the main physical difference between both water types lies on the amount of suspended load carried by both drainages, which is even more prominent during the rainy season (Sioli, 1984), when the clear water in ARA1 and IGN stations presents SSC < 20 mgL-1 and the white water in other stations present values > 100 mgL-1 (Figure 5).
Urucurituba Channel recorded concentrations in the range of 130 mgL-1 (URU1) ≤ SSC ≤339 mgL-1 (URU2) in May 2018. Mesotidal intrusion through URU2 is one of the main forces transporting, and regulating the dynamics of, suspended solids. Similar situations, although under macrotidal regime, were observed by Carneiro et al. (2020) in Pará River estuary. According to Asp et al. (2018), Caeté River estuary presents wide tidal plains in an extensive range of macrotidal incursions, whose discharge and elevation lead to seasonal SSS variations. Gensac et al. (2016) reported impact on Sss dynamics at Amazon River mouth, mainly during the rainy season, when SSStends to decrease down the estuary, despite the flow peaks observed in Amazon River during this period (Valério et al, 2018).
In the present case, Qss variations may have happened due to dilution, sedimentation and resuspension processes (Meade et al. 1985), and it suggests that Araguari River flow in the period of greatest fluvial discharge (rainy season) overlaps the hydrodynamics of the tides and substantially decreases the input of suspended sediments (unconsolidated and subconsolidated). On the other hand, the main factor controlling this difference in magnitude during the dry period lies on the dynamics of tides capable of overlapping Araguari River flow, since they transport white water with high Sss contents from the Amazon River into the basin, mainly via Urucurituba Channel.
The middle portion of Araguari River does not present high suspended sediment load because it was formed on Precambrian lands of Guiana Shield (Allison et al. 1996; Brito 2008; Bárbara et al. 2010; Santos & Cunha, 2015). However, ARA1 has shown expressive suspended solid discharge (1,354 t tidal cycle-1) during the rainy season, which was 3.4 times higher (4,586 t tidal cycle-1) than that during the dry period (Figure 6). Thus, in comparison to estimates (Qss = 575 t tidal cycle-1) by Santos and Cunha (2015), these solid discharge data suggest greater influence of flood tide currents through the channels interconnecting both basins.
Therefore, the highest Sss and Qssvalues observed for Urucurituba Channel did not only result from sediments transported by Amazon River, whose main source derives from erosive processes in the Andes (Gibbs 1967, Sioli 1984, Meade et al. 1985, Filizola and Guyot 2009), but it is also controlled by the action of local tidal currents that lead to significant erosion in the floodplain itself and in its sub-consolidated substrate. Santos et al. (2018) have observed mean erosion rate of 5 m month-1(60 m year-1) at URU1 site; they also reported that this section of the channel was only 55 m wide in 2011, but it had significantly increased to 321 m in 2016. In addition, maximum depth in this section of the channel was ≈35 m, similar to that of the Northern Channel of Amazon River (Abreu et al., 2020). If one takes into consideration that erosive potential increases as depth and flow speed increases (Novo, 2008), it is possible assuming that Urucurituba Channel tends to maintain and increase the rates of these processes due to its geomorphological features. Besides presenting the highest mean depth (28 m) in the measured cross section (URU1) and 34 m elsewhere (Santos et al. 2018), this channel also presents the highest mean current velocity among all channels (≈ 0.89 ms-1 during the dry season and ≈0.85 ms-1 during the rainy season).
The continuous and growing increase in suspended solids transport confirms the hypothesis about sedimentation evolution and worsening in the old Araguari River mouth, downstream ARA2 station and Urucurituba Channel, a process that has strong impact on the region because of the silting-up process taking place at the old river mouth. Impacts resulting from this process are significant; among them, one finds loss of environmental and ecosystem services, such as hydrological regime regulation, water quality, biodiversity maintenance (Junk et al. 2014), hydrological connectivity intensification/loss and, mainly, pororoca extinction (Cunha and Sternberg, 2018). Consequently, new deposition zones were identified, such as the one at the confluence of Urucurituba Channel with Araguari River, besides the formation of banks, islands on the left river bank and the clogging of Araguari channel, just upstream and downstream the confluence with the great tidal channel. Banks, which already have vegetation, are part of the new morphodynamics and depositional system along the river; they may be expanding in response to sediment transport adjustments observed in this stretch (Wang and Xu 2018). Currents upstream Araguari River present reduced velocity due to these processes. For example, Urucurituba Channel acts as suspended sediments’ exporter to floodplains during the dry season. On the other hand, during spring tides, it imports suspended sediments from floodplains, from Araguari River fluvial discharge and, finally, from the reflux discharge coming from Igarapé Tabaco region, during the rainy season (Figure 7).
Therefore, the hypothesis of significant hydrodynamic and sedimentary changes was confirmed by intensified erosive events in Bailique Archipelago, Southern Urucurituba Channel (Figures 4 and 7), which receives expressive liquid and suspended solids contribution that did not exist until a few years ago (Torres et al. 2018). This outcome suggests that Urucurituba Channel indicates negative sedimentary mass balance (erosion) ranging from 14,675 t tidal cycle-1(12%), during the dry period, to 107,982 t tidal cycle-1 (88%), during the rainy season (Figures 6 and 7). However, the deposition zone (dry season) lies at Gurijuba River mouth, which presents intense accretion processes capable of forming islands, as well as muddy banks that progressively evolve to islands due to fast vegetation adaptation in this zone (Santana and Silveira 2005, Torres et al. 2018).
4.3 Water quality variability and tidal propagation influence
According to one of the research hypotheses, seasonal hydrological effect is one of the main factors influencing water quality dynamics in this estuarine region (Tables 1 and 3). Increased concentration of parameters associated with sediments (SSS, Turb, TDS and Secchi) and with saline intrusion (EC and salt) was also observed by Santos et al. (2014) in the region in 2011. Such an increase has been attributed to reduced dilution capacity observed in rivers during the lowest flow period, as well as to erosion processes taking place in rivers and channels.
According to Geyer (1995), fresh water coming from Amazon River mouth is partly maintained by the magnitude of its discharge, which is enough to produce strong current towards the ocean, even during low flow periods (Abreu et al., 2020). However, increased salinity level during the dry season, although incipient, can be attributed to lower river dilution in this period (hydrological effect of lower water volume and meteorological effect due to salt “precipitation” through the atmosphere) – this outcome suggests different oceanic influences on the Amazonian coast. Salinity level during the rainy season (0.01 ppt) (Tables 1, 2 and 3) was equal to salt levels observed before the influence of Urucurituba Channel (Santos et al. 2018). Salinity was a severe environmental issue in the region between 2013 and 2014, so much so that Araguari River flow has significantly decreased and enabled significant saline intrusion in the estuary through the original mouth (Figure 1) - ≈174 times greater than that of 2011 (ranging from 0.01 to 1.74) (Santos et al. 2018). However, salinity level has significantly decreased after 2013, due to simultaneous silting and blockage of flow at the river mouth - salinity value of 0.26 ppt was recorded in March 2015. Later, there was salinity peak reduction, such as the one observed in lakes connected to Araguari River (Piratuba Lake) - remaining salinity levels ranged from approximately 0.04 to 0.08 ppt (Cunha and Sternberg 2018).
The herein observed seasonal DO pattern has suggested opposite effect to that observed by Barbara et al. (2010) in Cutias do Araguari, 60 km upstream Gurijuba Channel, which recorded lower DO concentrations during the dry season (minimum of 6 mgL-1) and higher concentrations of it during the rainy season (≈8 mgL-1) (Tables 2 and 3). However, DO levels remained higher than 6 mgL-1. Santos et al. (2014) have found low OD values in lower Araguari River, downstream the study site (<3.0mgL-1), during the rainy season. Many of these values also remained low in subsequent campaigns, after the greatest influence of Urucurituba Channel (Santos et al. 2018). However, greater oxygen saturation, upstream Araguari River, may take place due to hydraulic accidents (dams) (> 7.0mgL-1), where differentiated turbulence favors atmospheric reaeration processes (Silva et al., 2020).
Araguari River, as well as Gurijuba-Igarapé Novo and Urucurituba channels, develop in a floodplain, without hydraulic accidents, in downstream direction, which may lead to reduced reaeration rates (Santos et al. 2014). On the other hand, Amazon River naturally presents low DO values very often (lower than 5 mgL-1), mainly due to high respiration rates, as well as to high particulate and dissolved organic matter concentrations in water (Ward et al. 2013, 2015, Sawakuchi et al. 2017). Thus, DO concentration decrease can be easily attributed to the influence of Amazon River on Araguari River, due to oxygen demand and nutrient supply such as N, S and P (Ward et al., 2015). The influence on overall water quality starts at Urucurituba Channel, which enables higher DO concentrations in IGN and ARA1 stations (> 6 mgL-1 in the two seasons) and lower DO concentrations in the remaining monitoring stations (Tables 2 and 3). However, low oxygen concentration can also be associated with high turbidity, which reduces light penetration (essential to photosynthesis), mainly in the region of lakes (Brito 2008; Cunha and Sternberg, 2018). On the other hand, Damasceno et al. (2015) have found high DO concentration values (7.18 mgL-1 during the dry season and 6.52 mgL-1 during the rainy season) in a section upstream Amazon River mouth (Macapá) - these values were higher than the ones often observed for these channels.
The observed pH corresponds to acidic water, due to high organic matter concentration and decomposition by microorganisms (Ward et al., 2013), which increases during the rainy season. However, Barbara et al. (2010) observed trend of seasonal pH variation in the middle course of Araguari River, which recorded very low pH value = 4.7 (confluence of the main Amapari River tributary) and pH = 5.2 at IGN, which is 54 km away from the study site. This seasonal pattern was observed in large Amazonian rivers (Jari River) and in lakes connected to these rivers (Ajuruxi Lake); besides, it is seen as typical feature of Amazonian waters (Da Silva et al. 2013, Damasceno et al. 2015, Abreu and Cunha 2017). It happens because semidiurnal tides (Gallo and Vinzon 2015) take place on a daily basis in the Amazon River Delta and used to influence almost all water quality parameters in lower Araguari River until 60-80 km away from its original mouth, at the time it flowed into the Atlantic Ocean (Santos et al. 2014). However, nowadays the tidal wave spreads up to 75 km from the new river mouth, as indicated by the negative flow at IGN station (Figure 4A).
The lower influence of flood and ebb tides on parameters such as DO, DOsat and SSC differ from that observed in other studies focused on investigating water quality under tidal influence in Amazonian estuaries. Moura and Nunes (2016) recorded significant tidal effects for Turbidity, pH and DO – Turbidity was mostly evident at low tide, whereas pH and DO were mostly evident at high tide. Alves et al. (2009) have also found positive correlation between DO and tidal currents. Araújo (2018) has found positive relationship between flood tide and parameters such as EC, TDS, salinity and turbidity in Guajará Bay (Pará), as well as negative correlation to water temperature and lack of significant effects on pH and DO.
The differentiation caused by tides suggests the interaction between two water types coming from Araguari and Amazonas rivers. For example, pH values ranging from 5 to 8 are typical of Amazon rivers with clear water (Junk et al. 2011). Thus, the smallest spatial variations indicate stretches presenting intense mix of white water coming from Amazon River with clear water coming from Araguari River. Both rivers present slightly higher acidic pH at low tide and values closer to neutral pH at high tide. On the other hand, increased turbidity, TDS, EC and decreased water transparency during high tide are highly significant due to increased transport of suspended load from the Amazon River - these parameters are highly correlated to QSS.
Differences in tidal amplitudes also affect water quality response and vary depending on spatial drainage configuration. Tidal amplitude under syzygy condition is greater downstream the estuary. URU2 and GUR recorded tidal amplitude in the order of 3 m, whereas URU1 recorded 1.9 m; however, this amplitude decreased to just 1.2 m towards the core of the hydrographic basin (ARA1) (Figure 1). Thus, the spatial factor (distance from the Amazon River) reflects on several water quality parameters and is more intense in Urucurituba Channel than in Gurijuba-Igarapé Novo Channel. In addition, the morphological drainage pattern of Gurijuba River tends to hinder water flow from Amazon River to Araguari River, due to greater distance between the two basins through this channel. On the other hand, Urucurituba Channel presents significantly straight drainage channel, hydrodynamic magnitudes increased by tides and greater influence of water velocity and discharge, which favor different water quality responses.
A detailed, although summarized, analysis of all water quality and hydrodynamic variables is available in the supplementary file of this manuscript (Supplementary Figures 1 and 2).