In marine fish, stock complexity confers resilience to environmental variability and exploitation. Accurate stock identification is, therefore, a prerequisite to decipher the factors responsible for recruitment variation. We concatenated energetic reproductive investment with otolith microchemistry in order to delineate between stock spawning components in bonga shad Ethmalosa fimbriata in Senegal.
Satellite-derived (Moderate-resolution Imaging Spectroradiometer – MODIS Aqua, level 2, 0.1 degrees) sea surface temperatures were assessed at Joal (15 km radius) and Djifer once per sampling week. As no remote sensing data for inland waters are available, the Saloum River's surface water temperatures were recorded in situ once per sampling week with a digital thermometer (ama-digit ad 15 th; precision 0.4%; accuracy 0.4%). At Djifer and Foundiougne, salinity was determined once per sampling week according to the Practical Salinity Scale (PSS-78) with a handheld refractometer (Aqua Medic; precision 0.7%, accuracy 0.2%) using in situ water samples. For Joal, monthly means in MODIS satellite-derived (Aquarius, level 3, 0.5 degrees) sea surface salinities were used.
Monthly sampling was conducted at the Senegalese southern coast (SSC) and inside the Saloum River from February to October 2014, during Ethmalosa fimbriata's extended spawning season. Three environmentally contrasted study sites were chosen: Joal (SSC, 14°9.1' N; 16°51.7' W), Djifer (Saloum River's mouth, 13°57.8' N; 16°44.8' W), and Foundiougne (Saloum River's middle reaches, 14°8.1' N; 16°28.1' W). Fish were caught with gill nets (32 - 36 mm mesh size) by local fishermen and immediately stored on crushed ice after landing. Approx. 1000 fish per sampling site and month were examined randomly in order to find stage V females, i.e. mature individuals with ovaries containing fully hydrated oocytes. Sagittal otoliths were extracted, rinsed with ethanol (70%), and stored dry in Eppendorf caps. Females that spawned recently or lost part of their egg batch during handling were rejected.
Total wet mass (WM±0.01 g) and total length (LT, nearest mm) were obtained from individual stage V females. Ovaries were dissected, and oocytes were extracted out of one ovary lobe, rinsed with deionized water, and counted under a stereomicroscope. Around 80 hydrated oocytes per fish (ca. 70 for lipid analysis, ca. 10 for protein analysis) were transferred to a pre-weighed tin cap, stored in cryovials, and deep-frozen in liquid nitrogen. Dissected ovaries were transferred to a 4% borax buffered formaldehyde and freshwater liquid for fecundity analysis. Absolute batch fecundity (ABF) was estimated gravimetrically using the hydrated oocyte method for indeterminate spawners. The female's relative batch fecundity (RBF) was calculated by dividing ABF with the ovary free body weight (OFBM). The gonado-somatic index (GSI) of female spawners was calculated by dividing the ovary weight (OM, ±0.0001 g) by the ovary-free body weight (OFBM, ±0.1 g): GSI=OM ×[OFBM]^(-1) × 100. In order to access the nutritional status of female fish, a condition index (CI) was calculated for each individual using WM, LT, and b of the length-weight relationship: CI = WM x LT ^ -3.62 x 1000.
Oocytes in tin caps were freeze-dried (24 h) and weighed again to ascertain their dry mass (ODM±0.1 µg). The following equation was employed to calculate oocyte volumes (OV, mm3): OV =4/3 π(d_2/2)^2. Oocyte lipid content was determined by gas chromatography flame ionization detection (GC-FID). Three random samples were processed via GC mass spectrometry to ensure that all lipid classes were detected by the GC-FID. The total lipid content was assessed via summation of all individual lipid classes. For protein analyses, counted oocytes in tin caps were dried at 40°C for >24 h and weighted again for dry weight determination. Total organic carbon (C) and nitrogen (N) content was measured using a EuroVector EuroEA3000 Elemental Analyzer. From the total amount of N in the sample, the protein content was calculated according to Kjeldahl (Bradstreet, 1954), using a nitrogen-protein conversion factor of 6.25.
The oocyte gross energy content (J) was calculated on the basis of measured protein and lipid content, which were multiplied by corresponding energy values from literature: The amount of proteins per given oocyte (P, mg) was multiplied by a factor of 23.66 J mg-1 and subsequently added to the total amount of lipids per oocyte (L, mg) multiplied by 39.57 J mg-1 (Henken et al. 1986). Dividing the oocyte's energy content by the oocyte's dry weight allowed for calculation of the oocyte's calorific value (J mg-1). Further, the oocyte energy content of each individual E. fimbriata female was multiplied by its respective relative batch fecundity (RBF) to obtain a standardized estimate of the total amount of energy invested into a single spawning batch per unit body mass (SBEC, J g-1 OFBM): SBEC = [(P × 23.66 J [mg]^(-1) )+(L × 39.57 J [mg]^(-1) )]× RBF.
Dried sagittal otoliths were embedded in epoxy resin (Araldite 2020; Huntsman, USA) on glass slides. They were ground from the proximal side down to the nucleus using an MPS2 surface-grinding machine (GN, Nürnberg, Germany) and polished with a diamond-grinding wheel (grain size 15 µm). Concentrations of eight elements (Magnesium, Manganese, Copper, Zinc (Zn), Strontium (Sr), Yttrium, Barium (Ba), and Lead) were determined along transects of up to 2300 μm length along the rostrum's anterior edge on the otolith's proximal side. Analyses were carried out by Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS) using a NewWave UP193 solid-state laser coupled to a Thermo-Finnigan Element2 ICP-MS. The employed analytical procedure used a pulse rate of 10 Hz, an irradiance of ca. 1 GW cm−2, a spot size of 75 μm, and a traverse speed of 3 μm s−1. For external calibration the glass reference material NIST610 was analysed after every second ablation path, using reference values of Jochum et al. (2011). Data quality was validated by analysing a pressed pellet of NIES22 otolith with a calcium concentration of 38.8 wt. %, as well as through regular analyses of BCR-2G and BHVO-2G standard glasses. Precision was < 2% and the accuracy was < 13% for Zn. Recovery percentages were 100%, 101%, and 114% for Ba, Sr, and Zn, respectively. To account for the substitution of Calcium (Ca) by the divalent elements Ba, Sr, and Zn all element concentrations are given as element:Ca ratios. Mean otolith Ba:Ca, Sr:Ca, and Zn:Ca values throughout the entire female's lifetime are given in this dataset.
