FIGURE 1
The following sub-hypotheses were statistically tested: 1) Urucurituba Channel presents higher sediment transport rate than Gurijuba-Igarapé Novo Channel (> 50%), which causes the hydrological collapse of Araguari River mouth; 2) the rainy seasonal period, in combination to the tidal phase, significantly intensifies the solid discharge contribution coming from the Amazon River, which is later transported into Araguari River estuary. Therefore, this dynamics favors erosive events, as well as changes suspended sediment concentrations and water quality parameters; 3) the new configuration and complexity of the formed drainage system affects water balance and leads to variations in water quality parameters; and 4) factors such as location, distance from the Amazon River, seasonality and tidal phase can explain variations in suspended solid concentrations and in the quality of the water at Araguari River mouth.
2 MATERIALS AND METHODS
2.1 Study site
The study site lies between the right bank of lower Araguari River and the left bank of lower Amazon River, in the coastal-estuarine zone of Amapá State / Brazil. It is featured as a flat and low coast influenced by fluvial (North Channel of the Amazon River) and coastal (semidiurnal tides) processes (Torres et al., 2018) (Figure 1). It has hydromorphic soil associated with recent geological formations that originated quaternary plain areas that, in their turn, are naturally vulnerable to erosion processes (Rabelo, 2006).
The climate in the region is ruled by the Intertropical Convergence Zone (Schneider et al., 2014). The rainy season goes from December to July, whereas the dry season goes from August to November. Mean annual air temperature ranges from 27ºC to 27.5ºC, whereas mean annual rainfall rate ranges from 2,400 to 2,600 mm.year-1 (Oliveira et al., 2010).
Araguari River basin covers 1/3 of the total territory of Amapá State. It shelters important protected areas, from its source in Tumucumaque Mountains National Park to its low course near Piratuba Lake Biological Reserve (Dias et al., 2016).
Maximum flow of 3,415.8 m³s-1 (May 2014) and minimum flow of 155.5 m³s-1 (December 2013) were recorded between 2013 and 2015 in middle Araguari River, at Porto Platon station, which is located 270 km away from the original mouth (Cunha et al., 2014). However, the coastline of lower Araguari River, close to Amazonas, is influenced by tides whose amplitude is of approximately 3.4 m, in the North Channel of Amazon River (Figure 1).
Hydroelectric power generation (Coaracy Nunes series of plants - UHECN, Ferreira Gomes - UHEFG and Cachoeira Caldeirão - UHECC) is the main economic activity in the middle section of Araguari River basin (Cunha et al., 2011; Silva et al., 2017). Mining activity (gold, manganese, iron and pebble extraction) is carried out in its high section (Matta et al., 2008), whereas extensive buffalo breeding is performed in its low section, mainly in floodplain areas in the region (Silva et al., 2020).
2.2 Net discharge
Net discharge and current velocity were measured in acoustic Doppler profiler (Acoustic Doppler Current Profiler - ADCP, RD Instruments, Rio Grande model, 600 kHz), based on recommendations by Mueller and Wagner (2009).
Positive flow values (Q > 0) indicate flow downstream the Amazon River estuary (ebb), whereas negative flow values (Q < 0) indicate opposite flow direction – i.e., upstream the hydrographic basin (flood). Q = 0 indicates lack of net flow in the measured section (low or high tide). Monitoring stations are shown in Figure 1. Measurements were only taken during spring tide (12.5h cycle), in the dry (October 2017) and rainy (May 2018) seasons.
2.3 Suspended Sediments
One liter (1 L) of water-sediment mixture was collected on the surface of the channel (0.30 m deep) at 1-hour intervals, until the end of a tidal cycle, for suspended sediment (SSC) analysis purposes - 127 samples were collected in both seasonal campaigns.
SSC was obtained by filtering 250 mL of water samples in glass fiber membranes (0.45 µm porosity). It was used as dependent parameter and suspended solids discharge (Qss) estimator. Qss was used as dependent variable in order to test the sediment transport hypothesis based on environmental factors, hydrodynamic variables and water quality.
2.4 Suspended solids discharge
Qss was estimated by only taking into consideration the SSC load, which represented approximately 80-90% of the effective total solid discharge (Carvalho et al., 2014). The calculation was based on the simplified method described by the aforementioned authors and represented by the following equation:
\(Q_{\text{SS}}=0.0864.\ Q.\text{SSC}\) (1)
wherein: Qss represents the suspended solids discharge (t tidal cycle-1), Q is net discharge (m³s-1) and SSC is the suspended sediment concentration (mgL-1). Factor 0.0864 is the constant used for conversion into ton per day.
2.5 Water Quality Parameters
Water samples were collected in 2-L sampling bottle, in the middle of the channel section (surface), at 1-h intervals and throughout the semidiurnal tide cycle and in the dry (December) and rainy (May) periods.
Air temperature was measured in digital thermometer. Parameters such as water temperature, electrical conductivity, total dissolved solids and pH were measured in multiparameter probe (HANNA, model HI 9828). Turbidity was analyzed in AP 2000 IP turbidimeter (Policontrol Analytical Instruments) based on the nephelometric method. Secchi disk was used to measure water transparency. Instrutherm oximeter (model MO-900) was used to determine dissolved oxygen (DO) and DO saturation. Oakton Pcstestr 35 probe was used to measure salinity level.
2.6 Statistical Analysis
Hydrodynamic and water quality data were subjected to descriptive analysis and, subsequently, to uni- and multivariate analyses (R Development Core Team, 2020). After data non-normality were tested (Shapiro-Wilk test), Spearman correlation analyses were applied to test correlations between SSC and Qssversus water quality parameters, based on factors such as distance between collection points and the Amazon River, seasonality and tidal phase (flow and current velocity, Q and V, respectively).
Non-parametric Kruskal-Wallis test was used to investigate the spatial-seasonal effects of hydrodynamics and water quality on Ssc and Qss, based on the aforementioned factors (Crawley, 2007).
In addition, discriminating principal component analysis (DPCA) was applied to jointly test hydrodynamic and water quality variables (MacKinnon et al., 2016). After suitable transformation, DPCA was used to create normality and to equalize variances (Jombart and Collins, 2015). The goal was to enable the efficient discrimination of water quality and hydrodynamic groups based on few synthetic variables built as linear combinations of the original variables, which have the widest variation between and within groups (α ≤ 0.05) (Supplement).