Insights into the representativeness of biodiversity assessment in large reservoir through eDNA metabarcoding

Nova Avanhandava (NA) HPP upstream of the reservoir, and secondarily by the flows across the Pereira Barreto channel which vary according to the operational requirements of the two hydroelectric power plants. The TI reservoir area has multiple uses, such as fish farming, pro- fessional and recreational fishery, tourism associated with boat trips, and cattle breeding, mak- ing it an anthropic area [25, 26]. Hydrodynamic modeling The modeling domain, used for hydrodynamic modeling, covered the entire water mirror of the reservoir. For the water renewal ratio and stream speed computations, the hydrodynamic simulations utilized the Environmental Hydrodynamics Base System (SisBAHIA1, http:// www.sisbahia.coppe.ufrj.br/), a model used for natural water bodies under different scenarios [27]. The two-dimensional module of the FIST3D (Filtered in Space and Time 3D) hydrody- namic model was used to simulate water flow velocity and rate of water renewal during drainy and dry seasons. The FIST optimized spatial discretization allows for an optimal presentation of jagged contours and complex bathymetries. Bathymetric data was interpolated in the Sur- face Mapping System software (Surfer) using the Kriging regression method, with a spacing of 100 meters to obtain the Digital Terrain Model (DTM) for hydrodynamic modeling (S1 Fig). For the reservoir contour and spatial delimitation, the spatial discretization of the model was performed via biquadratic quadrangular finite elements. The water velocity was measured for

both rainy (April 1 to June 30, 2017) and dry (October 23, 2020 to January 21, 2021) seasons for the entire area of the reservoir. The mesh has a grid resolution ranging from around 300 meters in the tributary rivers to approximately 2 km in the Três Irmãos Reservoir main body (S1 Fig). Daily inflows from the NA reservoir and the arms where the Mato Grosso brook and Baix- ote stream are, and daily outflows of the TI dam and Pereira Barreto channel, in addition to the fluvial inflow of the tributary and the winds, were used as variables or constraints in the model. Due to the reservoir’s large size (~132 km in length), a compartmentalized analysis of the rate of water renewal during a simulated three-months period was applied. The region was divided into 14 compartments (Fig 1B). The conservation of mass principle was applied to a generic substance diluted in the water mass with concentration C. Pilot sampling Pilot sampling was conducted in September 2020 at one site in the reservoir (Fig 1A). Water sampling was conducted in accordance with ethical guidelines and Brazilian legal regulations. No permits were required for the described study, which complied with all relevant regula- tions. To evaluate the best depth for sampling, three 1L water samples were collected at the sur- face, middle, and bottom of the water column (1-, 13-, and 25-meter depths, referred to as P01, P13 and P25, respectively) (S1 Table). Three independent samples (R1, R2 and R3) were collected at each depth, resulting in nine samples. For each sample, one liter of water was col-

lected utilizing a Van Dorn bottle, then stored in sterilized polypropylene bottles (Nalgon, Cat. No. 2330) at 8º C until filtered using 0.22 μm Sterivex filters (Merck-Millipore, Cat. no. SVGP01050) in 60 mL increments using disposable syringes. After filtration, the remaining liquid was expelled by injecting air, and the filters were stored at 4º C in sterile Longmire’s buffer (Tris 100 mM, EDTA 100 mM, NaCl 10 mM, SDS 0.5%) until the DNA extraction. Non-disposable equipment was cleaned with sodium hypochlorite solution (10% v/v) and rinsed with autoclaved distilled water to avoid cross-sampling contamination. PLOS ONE eDNA metabarcoding reservoir PLOS ONE | https://doi.org/10.1371/journal.pone.0314210 January 24, 2025 4 / 22