Age, growth, mortality and reproductive seasonality of jolthead porgy, Calamus bajonado, from Florida waters

Sampling

Jolthead porgy (n = 630) were opportunistically sampled from fisheries landings by NMFS and state agencies’ port agents sampling the recreational and commercial fisheries along the SEUS coast during 2015 and 2016. Additionally we used eight (8) archived otolith samples collected by the SRHS in 2008 and 2011. Sampling occurred year round, All samples came from Florida, with southeast Florida (Ft. Pierce through Miami) accounting for 30% (n = 190) and the Florida Keys accounting for 70% (n = 448). The majority of samples were collected from fish caught by headboat anglers (n = 546), with private recreational anglers/charterboat anglers (n = 7) and the commercial fisheries sector (n = 85) accounting for the rest. All specimens were captured by either conventional vertical hook and line gear or divers with spears. Total lengths (TL) and/or fork lengths (FL) of specimens were recorded in millimeters (mm). Whole weight (W) in grams was recorded for fish landed in the recreational headboat fishery. Fish landed by commercial fisheries were eviscerated at sea and were excluded from the W–TL regression analysis. Sagittal otoliths were removed during dockside sampling and stored dry in coin envelopes.

Reproductive seasonality

Gonads were collected during 2015–2016 from fish sampled from headboats, as commercially caught fish were eviscerated at sea prior to sampling. Sex was assigned using macroscopic examination of the gonads (rounded cross section, females; triangular cross section, males) and recorded. Whole gonads were weighed to the nearest 0.01 grams, macroscopically staged using the general terminology of Brown-Peterson et al. (2011) and preserved in 10% buffered formalin for future histological processing and determination of timing of transition from female to male, which cannot be done reliably macroscopically. Reproductive seasonality was assessed for this study using a gonadosomatic index (GSI) for females, calculated as ${\displaystyle GSI=\left(gonadweight∕wholeweight\right)\ast 100.}$

Age determination and timing of opaque zone formation

Otoliths were sectioned on a low-speed saw, following the methods of Potts & Manooch III (1995). Three serial 0.5 mm sections were taken from the transverse plane near the otolith core. The sections were mounted on microscope slides with thermal cement and covered with mounting medium before analysis. The sections were viewed under a dissecting microscope at 12.5X using transmitted light.

Age determination was based on recording an opaque zone count and edge type code by an experienced reader (MLB) with extensive experience interpreting otolith sections (Burton, 2001; Potts, Burton & Myers, 2016) with no knowledge of date of capture or fish size for each sample. The edge codes refer to the type of zone, opaque or translucent, and in the case of translucent zones the amount of that zone between the last opaque zone formed and the otolith section edge. The codes used are outlined below:

1. 1 = opaque zone forming on the edge of the otolith section;

2. 2 = narrow translucent zone on the edge, generally <30% of the width of the previous translucent zone;

3. 3 = moderate translucent zone on the edge, generally 30%–60% of the width of the previous translucent zone;

4. 4 = wide translucent zone on the edge, generally >60% of the width of the previous translucent zone (Harris et al., 2007).

To ensure consistency in interpretation of the growth zones on the otolith, a subset of the otolith sections (n = 520; 82%) were then read by a second reader (JP), and an index of between-reader average percent error (APE) was calculated, following the methodology of Beamish & Fournier (1981). Where readings for a specimen disagreed, the sections were viewed again together. If consensus was reached the sample was retained; otherwise, it was excluded from further analysis.

Timing of opaque zone formation was assessed using edge analysis. The edge types were plotted by month of capture to determine if the opaque zones were deposited primarily in one season or month. Based upon edge frequency analysis, all samples were assigned a calendar age, obtained by increasing the opaque zone count by one if the fish was caught before that year’s opaque zone was formed and had an edge which was a moderate to wide translucent zone (type 3 or 4). Fish caught during the time of year of opaque zone formation with an edge type of 1 or 2, as well as fish caught after the time of opaque zone formation, were assigned a calendar age equivalent to the opaque zone count.

Growth

Growth of jolthead porgy was estimated using the Von Bertalanffy (1938) growth model:

L_t = L_inf(1 − e^{(−K(t−t₀))}) where L_inf = theoretical maximum length, K = Brody growth coefficient (rate at which maximum size is attained), and t₀ = theoretical age at size 0. Parameters were estimated from observed length at age using AD Model Builder estimation software (Otter Research Ltd., Sidney, B.C., Canada) [Mention of trade names or commercial companies is for identification purposes only and does not imply endorsement by the National Marine Fisheries Service, NOAA]. To account for growth of the fish throughout the year before or after its “birthday”, the calendar age of the fish (Age_c) was adjusted for the time of year caught (Mo_c) compared to month of peak spawning, or “birthdate”, (Mo_b), determined from the reproductive component of this study, thus creating a fractional, or monthly biological age, (Age_f): ${\displaystyle {Age}_{\mathrm{f}}={\mathrm{Age}}_{\mathrm{c}}+\left(\left({Mo}_{\mathrm{c}}-{Mo}_{\mathrm{b}}\right)∕12\right).}$

Body-size relationships

For weight–length relationships we regressed W on TL and FL (n = 577) using all fish with both lengths and whole weights sampled for this study, examining both a non-linear fit using nonlinear least squares estimation (version 9.4; SAS Institute, Cary, NC, USA) and a linearized fit of the log-transformed data. Residuals were examined to determine which regression provided the best fit. For length-length relationships, we regressed TL on FL and FL on TL (n = 622) using linear regression.

Natural mortality

We estimated the instantaneous rate of natural mortality (M) of jolthead porgy using several methods:

(1) Hewitt & Hoenig’s (2005) longevity mortality relationship ${\displaystyle M=4.22∕t_{max}}$

where t_max is the maximum age of the fish in the sample;

(2) Charnov, Gislason & Pope’s (2013) method using life history parameters ${\displaystyle M={\left(L∕L_{\infty }\right)}^{-1.5}K}$

where L_∞ and K are the von Bertalanffy growth equation parameters and L is fish length at age;

(3) Then et al.’s (2015) t_max-based estimator ${\displaystyle M=4.899t_{max}^{-0.916};}$

and (4) Then et al.’s (2015) growth-based estimator ${\displaystyle M=1.118K^{0.73}{L}_{\infty }^{-0.33}.}$

The equation of Hewitt & Hoenig (2005) and the two equations from Then et al. (2015) use either maximum age or von Bertalanffy growth parameters to generate a single point estimate. The Charnov, Gislason & Pope (2013) method, which incorporates life history information via the growth parameters, is based upon evidence suggesting that M decreases as a power function of body size. This method generates age-specific rates of M and has recently been used in Southeast Data Assessment and Review (SEDAR) stock assessments (SEDAR, 2017a).

The estimated percent of the population to survive to the oldest observed age in the population was calculated with the natural mortality rates applied to the fully-recruited ages in the fishery (modal age + 1). The following equation was used: ${\displaystyle \text{\%}survivorship=100\ast \left(exp\left(-\sum \left(M{age}_{\mathrm{f}}-M{age}_{\mathrm{o}}\right)\right)\right),}$

where Mage_f = natural mortality at the first age of full recruitment and Mage_o = natural mortality at the oldest age in the population.

Total and fishing mortality

The total instantaneous mortality rate Z was estimated using a catch curve analysis of the age-frequency of the samples (Beverton & Holt, 1957). We used only fully recruited ages (modal age plus one), since the age group at the top of the catch curve may not yet be fully vulnerable to the fishing gear (Everhart, Eipper & Youngs, 1975). The instantaneous rate of fishing mortality, F, was estimated by subtracting M from Z.

Reproductive seasonality

Gonads were collected from 178 individuals (22 males, 156 females). Due to lack of samples we did not analyze male reproductive seasonality. Jolthead porgy exhibit increasing ovary development from December to March, with peak spawning in March, as indicated by gonadosomatic index (Fig. 1A). This trend is supported by a plot of relative proportion of gonad stage by month (Fig. 1B), showing fish with primarily hydrated oocytes from December–March, followed in subsequent months by a preponderance of fish in a regressing or resting gonad stage. The lack of gonad samples from June through September is likely due to a combination of two factors. First, the species was not encountered as frequently by samplers during the months of June through August. The reason for this temporal absence in the landings is unknown, but it is apparent in the otolith collections as well (Fig. 2). Secondly, it is probable that any jolthead porgy that were encountered were in a resting/regressing stage and the gonads were overlooked. This is a plausible explanation for why no gonad samples were taken in September, when sample sizes rebounded.

Age determination and timing of opaque zone formation

We sectioned 638 jolthead porgy sagittae collected. Opaque zones were counted on 635 (99.5%) jolthead porgy sections. Three samples were determined to be illegible and excluded from the analyses.

We were able to assign an edge type to 99% of our samples (n = 633) for our analysis of opaque zone formation timing. Jolthead porgy otoliths exhibited opaque zones on the margin March–June, with a peak in April (Fig. 2). Opaque zones on the edge were absent from July through February. A shift to narrow translucent edge was noted from May to September. Moderate to wide translucent edges were found October through January, and the widest translucent edge was found during February and March, prior to peak opaque zone formation in April. We assumed that opaque zones on jolthead porgy otoliths were annuli.

Calendar ages were assigned as follows: for fish caught January through June and having an edge type of 3 or 4, the annuli count was increased by one; for fish caught in that same time period with an edge type of 1 or 2, calendar age was equivalent to annuli count; for fish caught from July to December, the calendar age was equivalent also to the annuli count.

Opaque zones on jolthead porgy otoliths were clear and easy to interpret (Fig. 3) and edge types were assigned consistently between readers. Between-reader APE was 0.06% (n = 520), meeting Campana’s (2001) acceptable value for APE (5% for species of moderate longevity and reading complexity). In fact, the readings only disagreed on two of 520 samples and only by ± one year in both cases. Both readers noted that those two otolith samples were of marginal quality. Consensus on between-reader counts was reached on all samples, and no samples were excluded from further analyses. The opaque zones were easy to trace from the sulcal groove out into the lateral plane on the dorsal side of the sections. The opaque zones were also matched with those on the ventral side, but due to refraction of the crystalline matrix on the ventral portion, the most consistent counts were made on the dorsal portion.

Growth

Growth of jolthead porgy was modeled using fractional ages based on the birth month, or month of peak spawning, of March determined from the reproductive samples collected. Jolthead porgy in this study ranged from 235–680 mm TL and ages 1–13, but only five fish were older than age-8 (Table 1). The standard deviation of length-at-age increased with age, but the coefficient of variation (CV) remained relatively constant. Thus, we assumed constant CV in our model estimation. Growth is described by the following equation: ${\displaystyle L_t=737\left(1-e^{-0.14\left(t+2.02\right)}\right)\left(n=635;\text{Fig. 4}\right).}$

While our data included large numbers of aged 2–4 fish, we had just one fish smaller than 250 mm TL. This lack of smaller fish is likely explained by gear selectivity, as our samples were all fishery-dependent and fish smaller than this size are likely unable to recruit to the hooks used in the fishery, or by fishers not retaining small fish in their catch, especially in light of the bag-limit regulations. Therefore we re-estimated the growth model using the method of McGarvey & Fowler (2002), which uses a left-truncated normal probability density function of length to adjust for the bias imposed by minimum size limits (or some other selectivity imposed by the fisheries) by assuming a zero probability of capture below the minimum size limit. We assumed a de facto minimum size limit for this analysis of 250 mm TL and estimated the growth model parameters with t₀ freely estimated by the model. This procedure has the effect of pulling the growth curve down to simulate smaller fish length-at-age for the youngest ages. We applied the full, untruncated normal likelihood to specimens in the study not subject to the minimum size limit. Parameters were estimated by minimizing the negative sum of log-likelihoods using AD Model Builder estimation software. The resulting growth model is: ${\displaystyle L_t=724\left(1-e^{-0.14\left(t+1.9\right)}\right)\left(n=634,\text{Fig. 4}\right).}$

Body-size relationships

Statistical analyses revealed an additive error term (variance not increasing with size) in the residuals of the W – TL and W – FL relationships for jolthead porgy, indicating that direct non-linear fits of the data were appropriate. The relationships are described by the following equations: ${\displaystyle W=1.1\times 10^{-5}{TL}^{3.05}\left(n=577\right)\text{and}W=4.2\times 10^{-5}{FL}^{2.90}\left(n=577\right).}$

The relationships between TL and FL are described by the equations

${\displaystyle TL=1.09\times FL+20.44\left(n=622;r^2=0.99;p<0.0001\right)\text{and}FL=0.90\times TL-14.26\left(n=622;r^2=0.99;p<0.0001\right).}$

Natural mortality

Natural mortality (M) was estimated to be 0.32 y⁻¹ for jolthead porgy using Hewitt & Hoenig’s (2005) method integrating all ages into a single point estimate, using the maximum age from our study of 13 yrs. M was estimated to be 0.47 y⁻¹ using Then et al.’s (2015) t_max—based method and 0.03 y⁻¹ using their growth-based equation. Age-specific estimates of M using Charnov, Gislason & Pope (2013) are presented in Table 1. We used the midpoint of each age (e.g., 0.5, 1.5, 2.5, etc.) to calculate age-specific M, because the Charnov, Gislason & Pope (2013) method cannot mathematically calculate M for age-0. Also, for stock assessment purposes where the integer age is used to describe the entire year of the fish’s life, the mid-point gives the median value of M for that age.

When considering the cumulative estimate of survivorship on the fully recruited age to the oldest age, the Hewitt & Hoenig (2005) method estimates 2.8% survivorship, while the Charnov, Gislason & Pope (2013) estimate is 8.3% and Then et al.’s (2015) aged-based estimate is 0.6%. When estimating survivorship only on the fully recruited ages, this estimate omits a large amount of mortality that occurs on younger fish, thus affecting the survivorship estimate. If we use all ages (0–13), survivorship is estimated to be 1.1% using Hewitt & Hoenig (2005) constant M, 1.3% using the Charnov, Gislason & Pope (2013) age-specific M, and 0.1% using Then et al.’s (2015) age based estimate.

Total and fishing mortality

Jolthead porgy were fully recruited to the fishery by age-3. Catch curve analysis of the sample age frequency plot results in an estimate of Z = 0.70. If we use the Hewitt & Hoenig (2005) point estimate of M, this results in a value of fishing mortality F = 0.38. If we use the age-varying method of Charnov, Gislason & Pope (2013), then our estimates of F range from 0.35 –0.50 for ages 3–9.