Age and Growth Studies in Indian Oil Sardine (Sardinella longiceps) using Hard Part Microstructure, a Tool for Biological and Ecological Understanding

Age and growth characteristics of the Indian oil sardine (Sardinella longiceps) was studied by interpreting microstructures on hard parts. Microstructures were validated to the time scales by observing their frequency of formation on hard parts of the fishes reared in confinement. Among the hard parts, sagittal otolith alone was found suitable for ageing the species as frequency of microstructure formation on them followed definite time scale. The estimates of age at length data and growth parameters from otolith analysis indicated faster growth in the species than the earlier estimates by other methods. Despite an expected sexual divergence in their growth, results show identical growth in both sexes. It further highlighted the prevalence of very distinct temporal growth variation driven by habitat environment indicating significance of hard part studies on ecological understanding of fishes. It also aided in tracing the precise time of birth with high precision, identifying the cohorts that supported the fishery and possible inter-regional migration of the species. The data generated from hard part ageing would aid in better eco-biological understanding of species, precise stock assessment outputs and fishery forecasting.


Introduction
Indian oil sardine (IOS), Sardinella longiceps is the most dominant resource of the Indian coast, in terms of distributional range, abundance and contribution to the marine fish production. They support more than 10% of annual marine fish landings of India (Rohit et al. 2018). Several studies and reports are available on their fishery, biology, stock status and measures proposed during time to time for sustaining stock and fishery. Being one of the most decisive basic inputs, precise estimates of age/age parameter is a paramount requirement for assessing stock health (Bellido et al. 2000) and to decide on adequate measures for resource sustainability and fishery (Carbonara and Follesa 2019). Traditionally, growth and demography of fished populations in tropical waters were estimated following length-based methods (Bensam 1964;Banerji 1968;Sam Bennet 1965;Antony Raja 1969, 1970Palomares et al. 1987;Kurup et al. 1989;Annigeri et al. 1992;Yohannan et al. 1998;Bellido et al. 2000;Cubillos et al. 2001;Rohit and Bhat 2003;Ganga and Pillai 2006;Abdussamad et al. 2010;Jayabalan et al. 2014;Nair et al. 2016;Al-Anbouri et al. 2011;Rohit et al. 2018). However, the available estimates of population parameter by this method varied widely and lacked any proper scientific interpretations for it. Among the alternate techniques, interpretation of microstructures on hard parts are considered more accurate and reliable for estimating age and growth in fishes (Morales-Nin and Pertierra 1990). Despite suitability of this technique, only limited studies are available on ageing Sardinella species using hard parts for stock assessment (Dayaratne and Gjosaeter 1986). Earlier studies using hard part of sardines in India were limited to description of the rings on their otoliths (Nair 1949) and scales (Balan 1959(Balan , 1964. Recently, the age of S. longiceps from Oman waters (Al-Anbouri et al. 2011) and Sardinella sindensis from Persian Gulf (Dehghani et al. 2015) were studied using otolith microstructures. Other studies in sardine include ageing of Sardinella brasiliensis using growth increments in otoliths (Saccardo et al. 1988) and scales (Fontelles Filho et al. 2005).

3
The present study is aimed to identify suitable hard parts of the species which can provide reliable information on age with accuracy and precision to describe their growth characteristics. This study describes the validation of microstructures on scales, vertebrae and otoliths of S. longiceps, their precision and reliability in age determination and also demonstrates how effective it is for biological confirmation and understanding ecological interactions of the species.

Materials and Methods
Fishery and biology of IOS along the southwest coast of India was monitored through weekly observations and laboratory studies during 2011-17. The population parameters monitored were size composition in the catch and spawning biology as described by Schreck and Moyle (1990). Population parameters were estimated from the length frequency data (LFD) through modal progression analysis (Pauly 1980;Pauly and Morgan 1987) using Electronic Length Frequency Analysis (ELEFAN I) programme (Gayanilo et al. 2005).
The microstructures on hard parts viz., scales, vertebrae and sagittal otoliths were subjected to image analyzing for extracting age and growth information following standard protocols (Brothers 1987;Green et al. 2009;Abdussamad et al. 2019). Microstructures were validated to time scale by ascertaining their frequency of formation in fishes reared in confinement by drawing random samples at regular intervals. Oil sardine juveniles trapped and reared in sea cages under natural habitat was used as study materials in this trial. The microstructure increments on scales and vertebrae were directly observed under transmission microscoped. Transverse section of sagittal otoliths at nucleus, polished to the required thickness were observed for microstructure formation at 10X, 40X, 100X magnifications. Microstructure increment was enumerated for each sampling interval. Hard parts of fishes collected from commercial catches were also subjected for similar analysis. Microstructure increments followed a regular time scale of one ring a day on sagittal otoliths, whereas on the vertebrae and scales it did not follow any consistent time scale. Therefore, otoliths alone were used for detailed analysis and interpretations. Growth parameters were estimated from age at length data and growth of the species was described using the von Bertalanffy growth formula (VBGF) (Von Bertalanffy 1938). Spawning seasonality was ascertained by direct biological monitoring and maturity analysis of the fishes and also by the interpretation of date of capture of the fishes and their age in days determined through otolith analysis. Since, hard part analysis indicated temporal growth variation in the species, correlation analysis was performed to test, the role if any of habitat ecology and biomass of the species on the growth as in Kendall and Sturat (1973). The upwelling, primary productivity, dissolved oxygen, sea surface temperature (SST) accessed from NASA's MODIS-Aqua sensor (https:// ocean color. gsfc. nasa. gov), Oceanic Nino index (ONI) to represent EL Nino Southern Oscillation (ENSO) from Climate Prediction Center (CPC) of National Oceanic and Atmospheric Administration (NOAA) are the ecological parameters tested.
A total of 36,791 IOS fishes were measured for length frequency and 13,084 were studied for the biology. Otoliths

Age and Growth
Age and growth were assessed by three approaches; direct measurement of growth in confinement by continuous monitoring, modal progression analysis of length frequency data (LFD) of fishes caught in commercial fishing and interpretation of microstructures on otoliths.

Direct Measurement
The oil sardines reared in open sea cages on natural food, exhibited almost uniform growth throughout the period and attained 166 mm size (TL) in 275 days (Table 1 and Fig. 1). Size of the juveniles at start of the trial was 74-77 mm with mean size of 75 mm, indicating all from a single brood. Considering 3 mm as size of the oil sardine larvae at hatching (Nair 1959), interpretation of growth curve developed by growth data from the rearing trial, juveniles were two-plus month old (67 days) at the start of trial. This suggested a total growth of 163 mm in 11 months under captive condition. The growth obtained in the rearing trial is much higher, compared to earlier reports of 100-160 mm growth in a year in natural condition (Hornell and Nayudu 1924;Balan 1964;Antony Raja 1969, 1970Annigeri et al. 1992;Yohannan et al. 1998;Nair et al. 2016).

Modal Progression Analysis
Growth parameters for age and growth estimations were derived from length frequency data by modal progression analysis. The von Bertalanffy growth curve fitted from these parameters provided a growth of 193 mm TL at the end of first years (Tables 2 and 3, Fig. 2). This is much higher compared to a growth of 166 mm in 11 months obtained in the present rearing trial and the growth ranges of 100-175 mm in one year reported by earlier researchers (Hornell and Nayudu 1924;Balan 1964;Antony Raja 1969, 1970Annigeri et al. 1992;Yohannan et al. 1998;Rohit and Bhat 2003;Nair et al. 2016) but is comparable to the growth reported for the species by Rohit et al. 2018.
In Oman waters, they are reported to attain 191 mm TL at the end of one year and 222 mm by second year (Jayabalan et al. 2014), which is almost same as the present estimate. The wide variation in the growth estimates by different researchers could be attributed to the individual bias in sampling, the complexities in the size frequency data used for analysis owing to their continuous spawning and recruitment and the estimation procedures employed. Partly it can be attributed to the variation in habitat environment, especially sea surface temperature (SST) which   persists along the regions where study was undertaken as opined by Ganga and Pillai (2006).

Microstructure Validation
The microstructures on hard parts appear as alternate dark and light bands under microscope in transmitted light (Fig. 3). Diurnal variations in growth of the animals are expected to produce similar variations in growth of hard part also and which appears as dark and light bands due to differential rates of minerals deposits. For ease of enumeration, generally dark bands were considered as daily rings. Campana and Nelson (1985) described the impact of day/ night cycles on the growth of the fishes and its resultant impact on hard parts as formation of growth rings. Al-Anbouri et al. (2011) also considered such microstructures on hard parts as daily rings, while ageing IOS from Oman waters. In the present study microstructure formation over time on hard parts viz, scales, vertebrae and sagittal otoliths of fishes reared in cages were monitored regularly. Microstructure increment was more frequent on scales and otoliths, whereas, it is very rare or in most cases no increments were observed on vertebrae (Table 1). Though microstructure formation on scales were more frequent during initial periods of growth, did not follow any definite time scale and that it almost ceases to form after certain period. However, their formation on sagittal otoliths were regular at one dark and light band a day throughout the period of growth, irrespective of size, sex and age of the fish. The uniformity of microstructure formation on otoliths established their suitability in age determination at day's precision in oil sardines and were considered most reliable. Therefore, otoliths alone were used for further studies. A variety of hard parts were in use across the globe for age and growth studies of fishes (Green et al. 2009;Abdussamad et al. 2019). Sardine (Sardinella Spp.) ageing in India and other areas were attempted by earlier researchers using otoliths (Nair 1949;Saccardo et al. 1988;Al-Anbouri et al. 2011;Dehghani et al. 2015) and scales (Balan 1959(Balan , 1964. Though types of hard parts were tested for their reliability in age determination of oil sardine, except otolith, others were found extremely unreliable as microstructure formation on them were found highly inconsistent.

Age Determination of Wild Population
The microstructures on otoliths from wild caught fishes covering all size groups were analysed during 2012-2018 and the information were converted to length at age data. Annual growth curves constructed separately for males and females from these data indicated no distinct growth variations between sexes and thus rejected the nsorted hypothesis of sexual divergence in growth of the species. For all further analysis usorted data alone were used. The annual growth plots from unsorted data, shows considerable inter-annual variation in growth patterns (Fig. 4). The plots and curves produced higher clusters during 2012, 2013 and 2017 indicated comparatively faster growth and more flat and lower cluster for 2014, 2015 and 2016, slow growth during respective years. The plots and growth curve derived from the pooled age-length data for the 2012-2018 period produced two distinct clusters and growth curves corresponding to the fast and slow growth periods (Fig. 5). These findings suggested operation of temporal growth variation in the species with normal growth during 2012-13, 2017 and slow growth  (Table 4). Same magnitude of temporal growth variation was also evident for both males and females. The growth curve for 2011-13 is almost identical as that from modal progression analysis for entire the period (Table 3 and Fig. 5). However, the former shows much faster growth, 154, 201 and 221 mm respectively at the end of 6, 12, and 18 months compared to 144, 193 and 216 mm of latter. The estimate of annual growth constant (K) was larger (1.76 year-1 ) for the growth curve based on hard part analysis when compared with 1.57 year-1 from modal progression analysis. This indicated much higher growth potential in the species, which is not discernible from the routine modal progression analysis. The underestimates in the modal progression analysis can be attributed to the interference by different modes representing multiple cohorts in the population selectivity of the fishing gear and sampling bias.

Habitat Ecology
A perusal of environmental and oceanographic dynamics exhibited wide inter-annual fluctuations in magnitude and pattern of major variables during the period. The SST of the coastal waters varied between 27.8 and 28.7 °C (Fig. 6). The coastal waters were relatively cooler with small SST during 2011-13, which gradually increased thereafter and remained warmer during rest of the period.  Likewise, ONI was small during 2011-13 and high during the latter period. They were subjected to detailed statistical interpretation to assess the influence of ecological parameters on the temporal growth variation in the species. The prominent ecological variables, viz. sea surface temperature (SST), dissolved oxygen (DO), chlorophylla (Chl-a) coastal currents (CC), upwelling and El Niño southern oscillation (ENSO) and also species biomass were tested for significance, to have a measure their interaction on growth (Fig. 7). SST and ONI exerted significant (P < 0.05), but inverse impacts on growth. The inverse relationship observed in the analysis indicates small values of SST and ONI favored good growth, whereas large values have adverse impacts. The positive correlation between the growth and biomass (P < 0.01) is a natural indication of prevalence of ideal condition for growth, which always ensure successful recruitment and large biomass. Since the small SST and ONI values linked to La Nina and larger values with El Nino phases of ENSO, it can be concluded that it may the ENSO induced environmental modifications that produce growth variation in the species. Antony Raja (1970) reported growth variation in the species during 1961-65 and attributed it to population density. However, the alternating phases of El Nino and La Nina that prevailed during the period suggested that it might be the habitat environment that produced growth variation, as observed in the present study. Ganga and Pillai (2006) also discussed role of prevailing SST on growth rates in the species. The present findings explain how the age and growth information derived from hard part analysis aids in understanding growth and its temporal spread in the species, which is not distinguishable from other routine length frequency analysis and more so when data for several years are clubbed for analysis.

Longevity
Maximum life span of the species, the age of the largest fish (221 mm TL) observed was estimated by VBGF growth equation using growth parameters from modal progression analysis of LF data and age at length data from otolith microstructure interpretation. It is 1.7 years (20.6 months) Fig. 7 The result of correlation analysis depicting the role of ecology on the growth of the species (ONI-oceanic nino index, SST-sea surface temperature DO-dissolved oxygen, Chla-chlorophyl-a) (*** P ≤ 0.001, ** P ≤ 0.01 and * P ≤ 0.05) Fig. 8 The annual length frequency plot of IOS (2011-17) by the former method and 1.5 years (18 months) by latter. Earlier estimates of their life span from the region were 2 years (Hornell and Nayudu 1924) and 2.5-2.6 years (Nair et al 2016;Rohit and Bhat 2003;Ganga and Pillai 2006;Abdussamad et al. 2010), following modal progression analysis of LF data. Devanesan (1943), estimated it as 14 years from the microstructures on scales, whereas Nair (1949Nair ( , 1960 estimated it as 3-4 years from otolith analysis. Jayabalan et al. (2014) reported their longevity as 2.5 years in the Oman waters. The present study highlighted a much faster growth in the species and consequently had a much shorter life span than earlier estimates.
The estimate of their size obtained from otolith analysis is 201 mm at the end of one year (Table 3 and Fig. 8). It is evident from the pooled length frequency plot that nearly 99.5% of the stock was succumbed to fishing or natural causes during their first year of life. Despite this, their rapid growth, early maturity, high fecundity and moderate to high resilience, would continue to support good fishery year after year under normal conditions (Rohit et al. 2018). However, being a short lived, prolonged unfavorable conditions may elicit undesirable changes like spawning and recruitment failures and may result in possible setback to fishery during the subsequent years even under normal fishing. This finding can be used to explain the frequent inter-annual fluctuations in their landings.

Spawning Periodicity
Spawning periodicity was monitored directly by biological observations and back calculations from the date of fishing and age at length data from hard part analysis. The biological observation suggested prolonged spawning during April-September along the coast with peak in May and June. Earlier reports also point out spread of the peak spawning activity during pre-monsoon and monsoon months (Hornell and Nayudu 1924;Nair and Chidambaram 1951;Bensam 1964;Prabhu and Dhulkhed 1967;Sarang and Sundaram 2010). Irrespective of the wide spread in spawning activity, earlier reports suggested peak spawning during monsoon months along the region (Devanesan 1943;Nair 1959;Balan 1959Balan , 1964Antony Raja 1970;Kumar and Balasubramanian 1987;Shah et al. 2018).
The spawning periodicity portrayed by back tracing age and date of fishing data indicates year-round spawning in the species with two peaks, major during premonsoon (April-June) and minor during post-monsoon (August-November) period (Fig. 9). This is in contrary to the present biological observations and earlier reports of their spawning periodicity. This indicates that major part of their landing was supported by pre-monsoon/monsoon recruits of the region itself. The other part of the landing which was by post-monsoon recruits might have migrated in from adjacent regions, highlighting the occurrence of interregional movement in the stock. Such inter-regional movement of the species and the possible drivers were described by earlier workers (Krishnakumar et al. 2006;Jayaprakash 2007) and according to them there is no physical barriers which restrict such movements.

Conclusions
The present study was designed to ascertain the suitability of various hard parts in age and growth determination in IOS and to address the issue of wide variation in their population biomass reported. Among the several hard parts, the study established sagittal otoliths alone as suitable for ageing oil sardines. The age information obtained from otolith were found to be highly reliable and can be used as an alternative method to assess age and growth characteristics of fishes and also in proper understanding biology and ecology of the species, which would complement to interpret the stock assessment outcomes. Though conventional methods employed provided estimates of age and growth, selecting a method that provides a reliable information was continuing as a challenge. The present study made through a combination of conventional length-based methods and hard part microstructures validated through direct observation on growth of the fish in confinement has confirmed that hard parts, though time consuming, provide precise age and growth estimates. The findings of the present study clearly indicated the beneficial role of hard part analysis in understanding the ecological influences on fish biology including growth and spawning/recruitment success. Such information would enable researchers to interpret and trace the cause for population fluctuations, aberrations in spawning patterns and influence of sudden and prolonged changes in environment on the growth and physiology of the species. Further, analysis of otolith microstructure enabled to precisely determine the birth Fig. 9 Spawning periodicity chart back traced from age and date of catch data of the fishes landed along the region time and age of the fish which in turn helped to trace and identify the cohorts which contributed to the fishery. Such comprehensive studies, combination of ecological and biological understanding of the fishery would go a long way in providing timely forecast of the fishery.