ORIGINAL PAPER

Reproductive dynamics of the sea urchin Paracentrotus lividuson the Galicia coast (NW Spain): effects of habitat and populationdensity

Rosana Ourens • Luis Fernandez • Marıa Fernandez-Boan •

Ines Naya • Juan Freire

Received: 19 September 2012 / Accepted: 3 April 2013 / Published online: 12 April 2013

� Springer-Verlag Berlin Heidelberg 2013

Abstract We studied the spatial variability in the size at

first maturity and the reproductive cycle of Paracentrotus

lividus in Galicia, contributing key information for the

exploitation and management of this resource. The size at

maturity varied between 20.4 (±1.2 SE) mm and

27.9 ± 1.2 mm and was smaller in areas of low population

density where sea urchins do not form patches. Using a

nonlinear model, we analysed the effect of depth, body

size, sex and population density on the temporal pattern of

the gonad index. The maximum and minimum indices were

obtained at 4 m depth in the months before and after the

spring spawning, respectively. The depth also affected the

cycle phase, and the sea urchins at 4 m depth spawned

9.4 ± 3.0 days later than the sea urchins at 8 m depth and

20.5 ± 3.0 days later than those at 12 m depth. Moreover,

the sea urchins living in patches showed a slight increase in

gonad size as a consequence of the better-quality habitat.

This shows that there is no intraspecific competition in this

area despite the high population densities reached

(18.5 kg m-2).

Introduction

Many echinoid species are traditionally eaten in various

regions of the world, and in Japan, the beginning of this

practice date backs to the ninth century (Lawrence 2007).

Asian and Mediterranean countries are the main markets

for this product, and the growing demand in the last dec-

ades has led to the expansion and development of new

fisheries and a staggering increase in the extraction rates

(Williams 2002). Consequently, the echinoid stocks have

decreased drastically on a world level and various cases of

overexploitation and collapse have been documented

(Andrew et al. 2002; Williams 2002; Micael et al. 2009).

This situation has made it urgent to revise the fisheries

management policies that govern the exploitation of these

resources searching new management strategies that guar-

antee the sustainability of these fisheries (Rogers-Bennett

et al. 2003). For this goal, both the fishing strategy and the

basic biological processes that determine the population

dynamics of the resource have to be understood. Repro-

duction is one of these processes, providing key informa-

tion for fisheries management such as size at first maturity

or spawning season. In addition, in species whose com-

mercial value resides in the gonads, such as the echinoids,

the study of the reproduction has another complementary

interest. In these cases, knowledge of the spatial and tem-

poral variations experienced by the gonad size provides

information on the expected commercial yield. This

information could be used to guide fishing strategy and

regulations, and it allows the planning of rotations and the

seasonality of the fishing operations, thus maximising

fishing profits.

The sea urchin fishery in Spain is focused on the species

Paracentrotus lividus. This echinoid is distributed along

the entire Mediterranean and NE Atlantic coasts, from

Communicated by M. Byrne.

Electronic supplementary material The online version of thisarticle (doi:10.1007/s00227-013-2236-2) contains supplementarymaterial, which is available to authorized users.

R. Ourens (&) � L. Fernandez � M. Fernandez-Boan � I. Naya �J. Freire

Grupo de Recursos Marinos y Pesquerıas, Universidad de A

Coruna, Rua da Fraga 10, 15008 A Coruna, Spain

e-mail: r.ourens@udc.es

Present Address:

J. Freire

Barrabes Next, C. Serrano 16-1, 28001 Madrid, Spain

123

Mar Biol (2013) 160:2413–2423

DOI 10.1007/s00227-013-2236-2

Ireland to Morocco, including the Canary and Azores

Islands (Boudouresque and Verlaque 2007). Several studies

on the species have investigated its reproductive cycle in a

wide range of regions, determining parameters such as the

spawning period and size at first maturity (e.g., Haya de la

Sierra 1990; Spirlet et al. 1998; Sanchez-Espana et al.

2004). Other studies, many carried out under culture con-

ditions, have also analysed the effect of certain environ-

mental variables on the mean gonad size (e.g., Guettaf and

San Martın 1995; Spirlet et al. 2000; Shpigel et al. 2004).

Ourens et al. (2011) carried out a statistical review of these

studies, and one of the most interesting results was the

latitudinal increase in the gonad index in Atlantic popula-

tions, as well as the higher values that these populations

showed in relation to the Mediterranean populations loca-

ted in similar habitats.

Despite all this information, little is known about the

variability in the reproductive cycle of this species on a

small scale. The habitat characteristics can lead to intra-

population variation in the spawning period or in the period

of maximum gonad production, and this information needs

to be taken into account when appropriate management and

exploitation strategies are determined for the local popu-

lation (the scale which is generally used in fisheries man-

agement, Orensanz and Jamieson 1998).

The aim of this work is to determine the variations in the

reproductive dynamics of the sea urchin P. lividus at the

fishing grounds, contributing information of interest for the

managers and fishers of this resource. At this spatial scale,

the population density and depth are potentially the most

influential parameters for the reproduction of the species.

The dense patches formed by sea urchins in certain areas

could result in intraspecific competition for food and space

and thus change the reproductive pattern. As an alternative

hypothesis, the higher quality of the habitat of the patches

(Tuya et al. 2007; Alvarado 2008) makes it possible to

maintain, or even increase, the growth and reproduction

rates despite the higher density, generating an Allee effect

(see Berec et al. 2007; Kramer et al. 2009). Similarly, depth

is the environmental variable that results in the largest

gradients in the distribution of the species. As in the pre-

vious case, this variable may indirectly affect the repro-

ductive rates given that it is linked to environmental factors

that influence the physiological state of the individuals,

such as the availability and quality of food, exposure to

waves, the temperature and the intensity of the light

(Garrabou et al. 2002). To test these hypotheses, we studied

the effect of population density on the size at first maturity

and quantified the variations in the temporal pattern of the

gonad index with certain environmental and population

parameters (population density, depth, sex and body size of

the sea urchins).

Materials and methods

Study area and sampling design

The study was carried out in Galicia (NW Spain), in the

sites of Lira and Porto do Son (Fig. 1). Samples were

collected monthly between June 2006 and May 2008 at two

fishing grounds in each site: Ardeleiro and Os Forcados in

Lira, and Son and Queiruga in Porto do Son. These areas

are exposed to waves and the habitats are similar, so we

considered them as replicates. The substrate is formed by

large flat rocks alternating with vertical walls oriented in

different directions. There is abundant algal cover, mainly

in the shallow areas and during the spring and the summer

months, and the species Saccorhiza polyschides, Cystoseira

baccata and Laminaria spp. can form large forests. Hal-

idrys siliquosa, Codium spp., Desmarestia spp. and the

invasive Sargassum muticum are also abundant (Veiga

Villar 1999; Otero-Schmitt and Perez-Cirera 2002; Casal

et al. 2011).

Ardeleiro and Os Forcados were sampled at the depths of

4, 8 and 12 m, and at each depth, we sampled two sea urchin

distribution types: patches (high density) and dispersed (low

density). Due to the aggregation behaviour of the species,

sea urchins are often distributed in patches with population

densities that vary between 24 and 684 individuals m-2.

However, in some areas it is also common to find isolated

sea urchins. We called these areas low-population-density

areas, given that this variable ranged between 0.04 and 3.32

individuals m-2. At the fishing ground of Queiruga, speci-

mens were only obtained in low-density areas, maybe as a

consequence of the fishing activity, while at the fishing

ground of Son, there were only sea urchins at 4 m depth, but

in both distribution types (Table 1).

Using a 50 9 50 cm metal frame, we collected sea

urchins located in the central part of three patches at each

depth through diving. In the low-density areas a 3-m-long

rope with a weight at one end was used to describe a circle

with a 3-m radius, and the sea urchins were collected

within the circle. Like in the patches, three replicates were

obtained using this method.

In addition, a monthly record of the sea surface tem-

perature (SST) was obtained in the study area during the

sampling period. To do this, satellite images obtained with

the AVHRR (Advanced Very High Resolution Radiome-

ter) sensor with a spatial resolution of 1.1 km were used.

The images were processed by the Centro de Recepcion,

Proceso, Archivo y Distribucion de Imagenes de Obser-

vacion de la Tierra (CREPAD) using a multichannel

algorithm (Willis et al. 1985). Each sampling site was

covered by two satellite images, and the monthly SST was

estimated as the average of the two images.

2414 Mar Biol (2013) 160:2413–2423

123

Sample processing

All of the sea urchins in the samples were counted and

measured (diameter without spines, ±0.1 mm) to estimate

the population density and for later studies on size

frequencies.

For the study of the reproductive cycle, a mean of 340

sea urchins per month, and 20 per sampling site, depth and

distribution type were analysed.

The individuals were weighed (±0.01 g), and sex was

then identified with a macroscopic analysis of the gonads

and the gametes released. After removing excess water with

absorbent paper, the wet weight of the gonads (±0.001 g)

was recorded. The gonads were then placed in an oven at

60 �C for 48 h to obtain the dry weight (GDW, ±0.001 g).

Estimating the size at first maturity

To calculate the size at maturity, we chose the individuals

collected in the months of maximum gonadal development

(January to April, according to an exploratory analysis of

the data). We classified sea urchins as immature when their

gonads were not macroscopically visible or the gonad dry

weight was below the accuracy of the scales (±0.001 g).

The analysis was based on 1,386 individuals with sizes

between 5.6 and 83.7 mm, coming from the sampling

grounds of Os Forcados and Son. The other locations were

not included in the study because there were hardly any

immature individuals in these areas. Given that sea urchins

smaller than 40 mm were practically absent at 8 and 12 m,

we assume that recruitment of this species occurs in

Fig. 1 Locations of the

sampling sites

Table 1 Population density (mean ± SE) of P. lividus in each sampling area

Ardeleiro Os Forcados Queiruga Son

HD LD HD LD HD LD HD LD

4 m 76.44 ± 4.86 0.56 ± 0.03 132.70 ± 8.48 0.78 ± 0.04 – 0.69 ± 0.06 227.88 ± 15.55 0.61 ± 0.08

8 m 61.90 ± 3.74 0.46 ± 0.02 50.74 ± 2.04 0.42 ± 0.02 – 0.45 ± 0.05 – –

12 m 57.40 ± 3.33 0.42 ± 0.03 50.81 ± 2.09 0.49 ± 0.04 – 0.36 ± 0.04 – –

The patches were absent in Queiruga, and sea urchins were absent at 8 and 12 m in Son

HD high density, LD low density

Mar Biol (2013) 160:2413–2423 2415

123

shallow areas (unpubl. data) and that it was not possible to

estimate a size at maturity for each depth.

We used a logistic model with the logit function link

(Hardin and Hilbe 2007) to determine the percentage of

mature individuals in relation to size. We studied the effect

of the density (high and low) and study area on the slope of

the curve. The model was as follows:

P ¼ eaþb1�Dþb2�D�Densityþb3�D�Location

1þ eaþb1�Dþb2�D�Densityþb3�D�Location

where P is the probability of an individual being mature; a,

b1, b2 and b3 are parameters of the model; D is the diameter

of the individual; Density is the type of distribution (high or

low density); and Location is the sampling ground. The dot

symbol between two variables represents their interaction.

We defined the size at first maturity in two ways: 1)

diameter at which 50 % of the individuals in the population

are mature (D50) and 2) diameter at which 95 % of the

individuals of the population are mature (D95).

Estimating the gonad index

To avoid the disturbances produced by including juveniles

in the analysis, the study of the reproductive cycle was

based on individuals larger than the size at which 95 % of

the population was mature.

The gonadal growth of these individuals showed slight

negative allometry with respect to the somatic growth

(GDW = 1.3 9 10-5�D 2.8, n = 6,436). Therefore, to be

able to compare the gonad index of different samples, we

needed to use individuals from the same size class or use an

independent indicator of the diameter. We chose to use the

standardised gonad index (SGI) defined in Ourens et al.

(2012), which represents the residuals of the linear

regression between the logarithm of the gonad dry weight

and the logarithm of the sea urchin diameter (Fig. 2). This

indicator is centred at zero, and individuals in the phase

before spawning have larger gonads than the annual mean,

and thus their SGI will be high. However, the individuals in

the phase after spawning will have undeveloped gonads

that will result in negative SGI values. The main advantage

of this index with respect to the other indices most com-

monly used is that it allows comparisons between samples

with sea urchins of different diameters, collected at dif-

ferent locations or in different seasons of the year.

Statistical analysis

The temporal pattern of the SGI shows the periodic nature

of the reproductive cycle (Fig. 3), and therefore to analyse,

it we designed a mathematical function that reflects this

periodicity (Fig. 4):

SGIi ¼ ai þ bi � senðp=6 � t þ xiÞ þ ei

where

• SGI is the standardised gonad index for an individual

ith,

• a is the SGI value at the inflection point, and we

consider it to be the mean annual value,

• b is the semi-amplitude of the cycle and can be

interpreted as the mean deviation of the SGI over time,

• g/6�t represents the annual periodicity of the cycle,

where t is the month,

• x is a phase parameter that shifts the sine wave left or

right, and

• e is the residual error.

Nonlinear mixed-effects models were fitted using the

nlme package in R (Pinheiro and Bates 2000). The vari-

ability produced by the sampling ground and year was

included in the random structure of the model. To do this,

we created a new variable called ground-year (eight lev-

els), which is the result of combining the variables ground

(four levels) and year (two levels). Initially, the random

effects were included in the estimation of the three

parameters (a, b and x).

The objective of this analysis was to evaluate the effect

of the depth, population density, sex and diameter of the

individuals on the temporal pattern of the SGI. The depth

was included as a categorical variable to estimate the three

parameters (a, b and x), while the variable sex was only

included in the estimation of the first two parameters (a and

b) because biologically a large phase lag between males

and females would not be possible in the spawning period.

Despite the SGI being independent of the body size by

definition, the diameter of the sea urchins may affect the

reproductive cycle, altering for example the spawning

season. Therefore, the diameter was introduced into the

model to estimate b and x.

1.6 1.7 1.8 1.9

−0.5

0.0

0.5

1.0

Log D (mm)

Lo

g G

DW

(g)

y = − 4.90 + 2.83·x

Fig. 2 Linear relationship between the logarithm of the diameter

(D) and that of the GDW for the set of individuals (n = 6,436)

2416 Mar Biol (2013) 160:2413–2423

123

The high density values reached in the patches (up to

684 ind m-2) could generate intraspecific competition for

food and space, leading to low gonadal development. To

test this hypothesis, we used the biomass density (kg m-2)

as an indicator of the competition given that it is a better

estimator of the intensity of food consumption than the

population density (ind m-2). Furthermore, we observed

that the highest population density occurs in patches with

large numbers of juveniles and recruits, which hardly

compete for food or space.

To determine the total weight of the samples, we esti-

mated the weight of the individuals (WW) by means of its

power-law relationship with the diameter (SM-I, WW =

2.1 9 10-3�D2.6, n = 8,917). The biomass density ranged

between 3.6 9 10-3 and 2.3 9 10-1 kg m-2 in low-den-

sity areas, and between 1.8 and 18.5 kg m-2 in patches.

This numeric variable was introduced into the equation to

estimate a, b and x. To estimate a, we also considered the

quadratic term in order to determine the existence of non-

linear relationships in the competition for food.

Following the protocol designed by Zuur et al. (2009),

modelling was initiated by using a complex model with all

explanatory variables as fixed components. In our case, the

fixed structure of the full model was the following:

ai ¼ a0 þ a1 � Depth8 þ a2 � Depth12 þ a3 � Biomass þ a4

� Biomassð Þ2þa5 � Sexi

bi ¼ b0 þ b1 � Depth8 þ b2 � Depth12 þ b3 � Biomass þ b4

� Sexi þ b5 � Diameteri

xi ¼ x0 þ x1 � Depth8 þ x2 � Depth12 þ x3 � Diameteri

þ x4 � Biomass

where a0, b0 y x0 represent the intercepts of the parameters

to be estimated for each individual ith; and the coefficients

a1,…, 5, b1,…,5 y x1,…,4 are the main effects of the covari-

ables. Depth8 and Depth12 are dummy variables that take

the value 1 when the depth is 8 and 12 m, respectively, and

0 otherwise. Similarly, Sex is the dummy variable that

takes the value 1 with males and 0 with females.

We then determined the best variance–covariance

structure for the random effects, and finally, we used a

backward elimination approach to select the variables that

affected the fixed structure. Starting values for modelling

−1.0

−0.5

0.0

0.5

1.0A

−1.0

−0.5

0.0

0.5

1.0B

−1.0

−0.5

0.0

0.5

1.0C

−1.0

−0.5

0.0

0.5

1.0D

J F M A M J J A S O N D J F M A M J J A S O N D

J F M A M J J A S O N D J F M A M J J A S O N DMonth

SG

I

Fig. 3 Temporal pattern of the SGI

at each sampling site

−0.2

−0.1

0.0

0.1

0.2

0.3

0.4

Month

SG

I

J F M A M J J A S O N D

α

β ω

− β

Fig. 4 Explanatory drawing of the parameters that define the

temporal pattern of the SGI

Mar Biol (2013) 160:2413–2423 2417

123

were estimated from numerical summaries and graphical

representations.

The minimum adequate model was chosen by compar-

isons of the Akaike information criterion (AIC, Akaike

1974), the Bayesian information criterion (BIC, Schwarz

1978), and, when models were nested, testing the

improvement in the likelihood ratio using a X2 test.

We used graphical methods (e.g., residual diagnostic

plots, and the observed trends augmented with the fitted

trends plots) to assess the adequacy of the final fitted

model.

Results

Size at sexual maturity

The sea urchins of Os Forcados matured before those of

Son (p \ 0.01), and maturity was reached earlier in the

low-population-density areas than in the patches

(p = 0.03). In accordance with this, the D50 varied a

maximum of 7.5 (±2.4 SE) mm between samples and was

20.4 ± 1.2 mm in low-density areas in Os Forcados and

27.9 ± 1.2 mm in the patches in Son (Fig. 5). Similarly,

the D95 also varied, and in these two extreme habitats,

95 % of the population matured with 29.6 ± 1.9 mm and

40.5 ± 2.1 mm, respectively.

Reproductive cycle

Description of the nonlinear model

To analyse the reproductive cycle, we used individuals

with a diameter larger than 40.5 mm, the size corre-

sponding to the D95 in the area where individuals reached

maturity latest (high-density areas in Son). Given that there

is no available information from Queiruga and Ardeleiro,

we had to assume that at least 95 % of the individuals from

these areas reached maturity at 40.5 mm.

The nonlinear mixed model fitted to these organisms

was simplified with respect to the full model, given that the

AIC of the two models was similar (-4,336.8 for the full

model and -4,337.5 for the final model) and there were no

significant differences in the likelihood ratio (p = 0.12).

The fixed structure of the final model was the following:

ai ¼ a0 þ a1 � Depth8 þ a2 � Depth12 þ a3 � Biomass þ a4

� Sexi

bi ¼ b0 þ b1 � Depth8 þ b2 � Depth12 þ b3 � Sexi

xi ¼ x0 þ x1 � Depth8 þ x2 � Depth12 þ x3 � Diameteri

The random effects were included in the three parameters

(a, b and x) and showed a general symmetrical positive-

definite variance–covariance matrix. A correction for het-

eroscedasticity was applied by allowing a variance struc-

ture with different spread per month and ground-year

(varIdent function in R). A correlation between the indi-

viduals belonging to the same sample was also observed,

and the error structures were assumed to be first-order auto-

regressive processes.

Diagnostic plots suggest that the residuals of the final

model meet all of the assumptions of normality and

homoscedasticity (SM-II). As the fitted values matched the

observed values closely, we concluded that the model fit is

satisfactory.

Spatial variability of the reproductive cycle

Although environmental and demographic conditions have

resulted in variations in the reproductive cycle, it is pos-

sible to observe a general temporal pattern with three dis-

tinct stages (Fig. 6). The largest SGI values were obtained

in the three or four first months of the year. The SGI then

started to decrease gradually with the spring, reaching the

lowest values in July or August, and then increasing again

at the start of autumn. This annual cycle was the opposite

0.0

0.2

0.4

0.6

0.8

1.0 A

0.0

0.2

0.4

0.6

0.8

1.0

0 10 20 30 40 50 60 70 80

0 10 20 30 40 50 60 70 80

Low density

High density

B

D (mm)

Pro

po

rtio

n o

f m

atu

re s

pec

imen

s

Fig. 5 Proportion of mature individuals for each size (± SE) in areas

of high and low population densities and at the Os Forcados (a) and

Son (b) fishing grounds. The lines on the top and bottom axes indicate

the number of mature and immature individuals observed,

respectively

2418 Mar Biol (2013) 160:2413–2423

123

of that described for the SST, which showed temperatures

higher than 15 �C from April to October (Fig. 6a).

The depth affected the three parameters estimated in the

reproductive cycle (Table 2). The semi-amplitude of the

cycle (b) was greatest at 4 m, so that the differences in the

mean monthly SGI were larger in this area, reaching a

value of 0.52 ± 0.03 g (the maximum differences in the

monthly SGI at 8 and 12 m were 0.43 ± 0.06 and

0.44 ± 0.06 g, respectively). However, the mean annual

SGI (a) was slightly higher in the 8-m area (p = 0.03) as

the result of a more stable temporal pattern. The parameters

a and b did not show significant differences between the

individuals located at 8 and 12 m depth (p = 0.10 and

p = 0.55, respectively). In parallel, the sea urchins located

in the deep areas showed an earlier reproductive cycle with

respect to the sea urchins in shallower areas (p \ 0.01).

Thus, the temporal pattern of the SGI at 4 m showed a

delay of 9.4 ± 3.0 days with respect to the reproductive

cycle of the sea urchins located at 8 m, and

20.5 ± 3.0 days with respect to those located at 12 m

depth (Fig. 6b).

The sea urchin diameter led to the opposite effect to that

of depth in the gonadal development cycle, so that there

was a delay of 5.1 ± 1.3 days for each 10 mm that the sea

urchin increased in size (Fig. 6c; Table 2). That is, the

difference in the reproductive cycle between the smallest

individuals (40.8 mm) and the largest (96.0 mm) was

28.3 ± 7.1 days.

The biomass density only affected the parameter a of the

model (p = 0.02, Table 2), increasing the mean annual

SGI 2.2 9 10-3 ± 9.2 9 10-4 for each kilogram of sea

urchin biomass (Fig. 6d). This positive linear effect of the

J F M A M J J A S O N D

12

16

20A

SS

T (

ºC)

1 2 3 4 5 6 7 8 9 10 11 12

−0.2

0.0

0.2

4 m8 m12 m

B

1 2 3 4 5 6 7 8 9 10 11 12

−0.2

0.0

0.2

45−50 mm65−70 mm85−90 mm

C

−0.2

0.0

0.2D

0−2 kg m−2

5−7 kg m−2

10−13 kg m−2

J F M A M J J A S O N D J F M A M J J A S O N D

−0.2

0.0

0.2

FemalesMales

E

Month

SG

I

Fig. 6 Seasonal SST (sea surface temperature) pattern and its

confidence interval of 95 % (a). Variations predicted by the model

for the temporal pattern of the SGI (standardised gonad index) with

the depth (b), sea urchin diameter (c), biomass density (d), and sex

(e). The curves were predicted by the fitted model, varying in each

plot a single factor. The reference values were: depth = 8 m,

diameter = 60 mm, density = 3 kg m-2, sex = female

Mar Biol (2013) 160:2413–2423 2419

123

density indicates that there is no intraspecific competition

between the individuals in patches, at least at the popula-

tion densities studied.

Finally, the mean SGI for males was 0.035 ± 0.004 g

higher than for females (p \ 0.01), perhaps as a conse-

quence of a more stable temporal pattern, which did not

reach minima as low as those reached for females in the

summer (Fig. 6e).

Discussion

We have studied the reproductive dynamics of P. lividus

through the size at first maturity and the reproductive cycle,

analysing the factors that cause spatial variations at local

scale.

The size at maturity of a population is frequently defined

as the size at which 50 % of the individuals are mature.

However, this is not the only definition (Williams and

Babcock 2005), and there are also various possible meth-

ods for calculating it. For example, classifying individuals

into mature or immature can be based on a macroscopic

analysis of the gonads, histological cuts or variations in the

gonad index (Haya de la Sierra 1990; Saborido-Rey and

Junquera 1998; Corgos and Freire 2006), and the statistical

procedures for estimating the size can also vary (Trippel

and Harvey 1991; Roa et al. 1999; Zhu et al. 2011). Due to

this and the spatial variability in the size at maturity

(Lozano et al. 1995), the studies on P. lividus show a

variation of up to 2 cm in the diameter at onset of sexual

maturity, which ranges between 20 and 40 mm (Semroud

and Kada 1987; Haya de la Sierra 1990; Sanchez-Espana

et al. 2004). According to this information, the size at

maturity estimated for P. lividus in this study is consistent

with the information previously published.

We have observed spatial variability in the size at

maturity related to the sea urchin distribution type. Various

studies on echinoids have documented the advantages

offered by patches to the individuals that constitute them,

such as protection against predators or wave action (Pearse

and Arch 1969; Freeman 2003; Vega-Suarez and Romero-

Kutzner 2011). In addition, patches are usually formed in

areas with good habitat quality, where food is abundant

(Vadas et al. 1986; Unger and Lott 1994; Alvarado 2008).

The individuals that live isolated from others do not have

these advantages and may initially invest more energy in

reproduction as an adaptation to high mortality rates,

leading to earlier maturity with respect to the sea urchins in

patches. This type of response to stressful situations has

been documented in other echinoids (Dix 1970; Kenner

and Lares 1991) as well as in P. lividus (Lozano et al.

1995).

Meanwhile, the reproductive cycle of P. lividus was

studied through the gonad index. Most published studies

analyse the spatial and temporal variability of this index

with linear statistical models, such as GLMs or ANOVAs

(e.g., Sellem and Guillou 2007; Gonzalez-Irusta 2009;

Hernandez et al. 2011). Given that the temporal pattern of

the gonad index is a cycle, studying it with these kinds of

models requires a categorical time variable with a certain

number of levels (e.g., the months). If the objective is also

to study the spatial variability of the temporal pattern (as in

our case), it is necessary to include a large amount of

interactions between the spatial variables and each of the

levels of the time variable. As a result, the output of the

model is difficult to interpret, and most studies limit

themselves to determining the effect of these variables on

the mean annual gonad index.

Using a nonlinear model is a good alternative for

studying the temporal pattern of the gonad index. Thus,

Ebert et al. (2012) developed a complex equation to rep-

resent mathematically the reproductive cycle of the sea

urchin S. purpuratus, using a gonad index different from

the one used in the present study. While the SGI corrects

potential biases produced by the allometric growth of the

gonad (Ourens et al. 2012), Ebert et al. (2012) incorporated

the effect of the allometry in the equation. In any case, both

models allow interpreting the effect of the independent

variables on the different components of the reproductive

Table 2 Parameters of the reproductive cycle of P. lividus estimated

with nonlinear mixed-effects models and their standard errors (SE)

Fixed effects Value SE t-value p value

a -0.014 0.013 -1.068 0.286

Depth 8 m 0.017 0.008 2.116 0.034

Depth 12 m 0.004 0.008 0.432 0.666

Males 0.035 0.004 8.996 \0.001

Biomass 0.002 0.001 2.385 0.017

b 0.258 0.017 15.171 \0.001

Depth 8 m -0.043 0.011 -3.861 \0.001

Depth 12 m -0.037 0.011 -3.190 0.001

Males -0.038 0.005 -6.884 \0.001

x 1.291 0.158 8.174 \0.001

Depth 8 m 0.162 0.052 3.129 0.002

Depth 12 m 0.353 0.052 6.842 \0.001

Diameter -0.009 0.002 -4.015 \0.001

Random effects SE Correlation

a 0.033 a b

b 0.043 0.053

x 0.203 -0.068 -0.463

Residual 0.220

The deviation associated with the random effects and the correlation

matrix are also shown

2420 Mar Biol (2013) 160:2413–2423

123

cycle, that is, the mean annual gonad index, amplitude and

the cycle phase.

The seasonality of the SGI in our study area reflects an

annual reproductive cycle in which the maximum gonadal

development occurs in winter before spawning takes place.

In the summer, the gonads have already released the

gametes, which results in low SGI values. The SGI begins to

increase again from September when the gonads start to store

new nutrients for the following gametogenesis. Similar

results have also been obtained in other regions of Galicia

(Urgorri et al. 1994; Catoira 1995; Montero-Torreiro and

Garcıa-Martınez 2003) and also in Europe and Africa (Byrne

1990; Gonzalez-Irusta 2009; Garmendia et al. 2010; Arafa

et al. 2012).

This general reproduction pattern can vary slightly

depending on the environmental and population conditions

in which the individuals live. Among the factors studied

here, depth may be one of the most influential in the

reproductive cycle. Although the mean annual SGI hardly

changed with the depth, individuals located in the shallow

areas seem to invest more energy in reproduction than

those sea urchins situated in deep areas, given the large

variation experienced by the size of their gonads along the

cycle. According to Walker et al. (2007), the gonads reach

their maximum size when most of the gonadal tissue is

constituted by phagocytic cells, which accumulate the

nutrients necessary for gametogenesis. Consequently,

gonads will be larger in those habitats where food is

plentiful and of high nutritional quality, as in the shallower

areas (Keats et al. 1984; Rogers-Bennett et al. 1995;

Guettaf et al. 2000; Ourens et al. 2011). Similarly, the

higher food availability in the 4 m depth area could lead to

the formation of a greater number of gametes or higher-

quality gametes that have large nutrient reserves. This

would explain the smaller gonad residuals observed at 4 m

depth after spawning. Following the same reasoning, the

different composition of the gametes (Unuma et al. 2003;

Walker et al. 2007; Hagen et al. 2008) and their different

contribution to the gonadal weight would justify the dif-

ferences that are observed between the SGI of males and

females after spawning.

The mathematical model that we used to represent the

reproductive cycle assumes symmetrical cycles, in which

the concave part of the curve is the convex mirror. In these

cases, the parameter x represents the phase shift between

cycles, as shown in Fig. 4. However, if the cycles were

asymmetrical, the parameter x may be indicative of the

changes that occur between cycles in the duration of the

period of accumulation of nutrients or of the spawning

period. This could be the cause of the variability produced

by the depth and the diameter in the parameter x in our

analysis. For example, the high availability of food in the

shallow zone could lead to a period of accumulation of

nutrients more prolonged than in deep areas, and therefore

a phase lag between cycles. Similarly, the metabolic rate of

echinoids decreases as the size of the individual increases

(Fuji 1967; McPherson 1968), which means that the

nutrients necessary for gametogenesis are accumulated

more slowly, resulting in a delayed spawning.

Finally, the population density, which in this case is

related to the type of spatial distribution of the individuals

(aggregated vs. dispersed), increased the mean size of the

gonads. Although this increase was very slight, it shows

that there is not intraspecific competition for food or space

at high densities (up to 18.5 kg m-2). The particular

environmental characteristics that occur in the sea urchin

patches (e.g., higher food availability and protection from

predators and waves) may increase the carrying capacity of

the habitat. Thus, high densities do not promote the inverse

relationships that have been observed between the gonad

index and the population density in other studies where

densities were manipulated (Tomas et al. 2005). Further-

more, the dense aggregations of sea urchins not only

increase the gonad size, but also enhance the reproductive

success by decreasing the distance between individuals

(Pennington 1985; Levitan et al. 1992). In this sense,

Levitan (1991) found that fertilisation rates of Diadema

antillarum increased from 7 to 40 % when the density of

males was increased from 1 to 16 ind m-2. Estimating the

minimum aggregation level that would ensure the fertil-

isation of P. lividus is therefore a challenge we set for the

future.

In view of our results, the lower size at maturity of sea

urchins located in unfavourable habitats could be an

adaptation to high mortality rates. However, once indi-

viduals mature, they will develop greater gonads in areas

with higher habitat quality. We have also observed that sea

urchins located at 4 m invest more energy in reproduction

than individuals at greater depths, possibly as a response to

differences in food availability. This pattern contrasts to

that described by Ourens et al. (2013) for somatic growth in

the study area, who found larger individuals in deeper

areas. Inverse patterns between reproduction and growth of

P. lividus were also found by Turon et al. (1995), although

in that case the investment in reproduction was higher in a

habitat with food shortages and high hydrodynamics.

The information obtained in this study gains particular

interest for the managers and fishers of echinoids, as it

would allow them to adapt the harvesting strategy to the

spatial and seasonal variability in the reproductive

dynamics, thus optimising the productivity of the fishery.

Based on the results obtained, the minimum commercial

size (55 mm) and the closed season for P. lividus in our

study area (May–September) are two appropriate measures

that favour the reproduction of the species. The objective

of establishing a minimum harvesting size is to make sure

Mar Biol (2013) 160:2413–2423 2421

123

all individuals spawn at least once before becoming part of

the exploitable biomass. This objective is fulfilled if we

consider that more than 95 % of the population reaches

sexual maturity at 40.5 mm in all of the areas sampled.

Similarly, the species spawns at the end of spring, which

coincides with the end of the fishing season.

The maximum fishery productivity (in terms of gonad

harvest) could be obtained in the months before spawning

and at the shallow aggregations. However, probably a

management strategy that protects the shallow breeding

stock could be more adequate, due to high reproductive and

low growth rates of sea urchins in this habitat.

Acknowledgments This work was funded by the Spanish Ministry

of Education and Science and by the European Regional Develop-

ment Fund (ERDF). The authors wish to thank the other members of

the research team, especially G. Casal, N. Sanchez and M. J. Juan

Jorda, for their collaboration in processing the samples, estimating

SST data, and statistical advice.

References

Akaike H (1974) A new look at the statistical model identification.

IEEE Trans Autom Control 19(6):716–723

Alvarado J (2008) Seasonal occurrence and aggregation behavior of

the sea urchin Astropyga pulvinata (Echinodermata: Echinoidea)

in Bahıa Culebra, Costa Rica. Pac Sci 62(4):579–592

Andrew NL, Agatsuma Y, Ballesteros E, Bazhin AG, Creaser EP,

Barnes DKA, Botsford LW, Bradbury A, Campbell A, Dixon JD,

Einarsson S, Gerring PK, Hebert K, Hunter M, Hur SB, Johnson

CR, Juinio-Menez MA, Kalvass P, Miller RJ, Moreno CA,

Palleiro JS, Rivas D, Robinson SML, Schroeter SC, Steneck RS,

Vadas RL, Woodby DA, Xiaoqi Z (2002) Status and manage-

ment of world sea urchin fisheries. Oceanogr Mar Biol Annu Rev

40:343–425

Arafa S, Chouaibi M, Sadok S, El Abed A (2012) The Influence of

season on the gonad index and biochemical composition of the

sea urchin Paracentrotus lividus from the Golf of Tunis. Sci

World J 815935:8. doi:10.1100/2012/815935

Berec L, Angulo E, Courchamp F (2007) Multiple Allee effects and

population management. Trends Ecol Evol 22(4):185–191

Boudouresque CF, Verlaque M (2007) Ecology of Paracentrotus

lividus. In: Lawrence JM (Ed), Edible sea urchins: biology and

ecology, 2nd edition, Elsevier. Dev Aquac Fish Sci 37:243–285

Byrne M (1990) Annual reproductive cycles of the commercial sea

urchin Paracentrotus lividus from an exposed intertidal and a

sheltered subtidal habitat on the west coast of Ireland. Mar Biol

104:275–289

Casal G, Kutser T, Domınguez-Gomez JA, Sanchez-Carnero N,

Freire J (2011) Mapping benthic macroalgal communities in the

coastal zone using CHRIS-Proba mode 2 images. Estuar Coast

Shelf Sci 94(3):281–290

Catoira JL (1995) Spatial and temporal evolution of the gonad index

of the sea urchin Paracentrotus lividus (Lamarck) in Galicia,

Spain. In: Emson R, Smith A, Campbell A (eds.), Echinoderm

research 1995. A.A. Balkema, Rotterdam, pp 295–298

Corgos A, Freire J (2006) Morphometric and gonad maturity in the

spider crab Maja brachydactyla: a comparison of methods for

estimating size at maturity in species with determinate growth.

ICES J Mar Sci 63:851–859

Dix TG (1970) Biology of Evechinus chloroticus (Echinoidea:

Echlnometridae) from different localities. 3. Reproduction.

N Z J Mar Freshw Res 4:385–405

Ebert TA, Hernandez JC, Russell MP (2012) Ocean conditions and

bottom-up modifications of gonad development in the sea urchin

Strongylocentrotus purpuratus over space and time. Mar Ecol

Prog Ser 467:147–166

Freeman SM (2003) Size-dependent distribution, abundance and

diurnal rhythmicity patterns in the short-spined sea urchin

Anthocidaris crassispina. Est Coast Shelf Sci 58:703–713

Fuji A (1967) Ecological studies on the growth and food consumption

of Japanese common littoral sea urchin, Strongylocentrotus

intermedius (A. Agassiz). Memoirs of the Faculty of Fisheries,

Hokkaido University, 15:1–160

Garmendia JM, Menchaca I, Belzunce MJ, Franco J, Revilla M

(2010) Seasonal variability in gonad development in the sea

urchin (Paracentrotus lividus) on the Basque coast (Southeastern

Bay of Biscay). Mar Pollut Bull 61:259–266

Garrabou J, Ballesteros E, Zabala M (2002) Structure and dynamics

of North-western Mediterranean rocky benthic communities

along a depth gradient. Est Coast Shelf Sci 55:493–508

Gonzalez-Irusta J (2009) Contribucion al conocimiento del erizo de

mar Paracentrotus lividus (Lamarck, 1816) en el Mar Cantab-

rico: ciclo gonadal y dinamica de poblaciones. Dissertation,

Universidad de Cantabria

Guettaf M, San Martın GA (1995) Etude de la variabilite de l0indice

gonadique de l0oursin comestible Paracentrotus lividus en

Mediterranee nord-occidentale. Vie Milieu 45(2):129–137

Guettaf M, San Martın GA, Francour P (2000) Interpopulation

variability of the reproductive cycle of Paracentrotus lividus

(Echinodermata: Echinoidea) in the South-Western Mediterra-

nean. Mar Biol 80:899–907

Hagen NT, Jørgensen I, Egeland ES (2008) Sex-specific seasonal

variation in the carotenoid content of sea urchin gonads. Aquat

Biol 3:227–235

Hardin JW, Hilbe JM (2007) Generalized linear models and

extensions, 2nd edition. Stata Press, 387 pp

Haya de la Sierra D (1990) Biologıa y ecologıa de Paracentrotus

lividus en la zona intermareal. Dissertation, Universidad de

Oviedo, Spain

Hernandez JC, Clemente S, Brito A (2011) Effects of seasonality on

the reproductive cycle of Diadema aff. antillarum in two

contrasting habitats: implications for the establishment of a sea

urchin fishery. Mar Biol 158:2603–2615

Keats DW, Steele DH, South GR (1984) Depth-dependent reproduc-

tive output of the green sea urchin, Strongylocentrotus droeba-

chiensis (O.F.Muller), in relation to the nature and availability of

food. J Exp Mar Biol Ecol 80:71–91

Kenner MC, Lares MT (1991) Size at first reproduction of the sea

urchin Strongylocentrotus purpuratus in a central California kelp

forest. Mar Ecol Prog Ser 76:303–306

Kramer A, Dennis B, Liebhold AM, Drake JM (2009) The evidence

for Allee effects. Popul Ecol 51:341–354

Lawrence JM (2007) Edible sea urchins: use and life-history

strategies. In: Lawrence JM (ed) Edible sea urchins: biology

and ecology, 2nd edition, Elsevier. Dev Aquac Fish Sci 37: 1–9

Levitan DR (1991) Influence of body size and population density on

fertilization success and reproductive output in a free-spawning

invertebrate. Biol Bull 181:261–268

Levitan DR, Sewell MA, Chia FS (1992) How distribution and

abundance influence fertilization success in the sea urchin

Strongylocentotus franciscanus. Ecol 73(1):248–254

Lozano J, Galera J, Lopez S, Turon X, Palacın C, Morera G (1995)

Biological cycles and recruitment of Paracentrotus lividus

(Echinodermata: Echinoidea) in two contrasting habitats. Mar

Ecol Prog Ser 122:179–191

2422 Mar Biol (2013) 160:2413–2423

123

McPherson BF (1968) Feeding and oxygen uptake of the tropical sea

urchin Eucidaris tribuloides (Lamarck). Biol Bull 135:308–321

Micael J, Alves M, Costa A, Jones M (2009) Exploitation and

conservation of echinoderms. Oceanogr Mar Biol Annu Rev

47:191–208

Montero-Torreiro MF, Garcıa-Martınez P (2003) Seasonal changes in

the biochemical composition of body components of the sea

urchin, Paracentrotus lividus, in Lorbe (Galicia, north-western

Spain). J Mar Biol Ass UK 83:575–581

Orensanz JM, Jamieson GS (1998) The assessment and management

of spatially structured stocks: an overview of the North Pacific

Symposium on Invertebrate Stock assessment and management.

In: Jamieson GS, Campbell A (eds) Proceedings of the North

Pacific Symposium on Invertebrate stock assessment and

management. Can Spec Publ Fish Aquat Sci 125: 441–459

Otero-Schmitt J, Perez-Cirera J (2002) Infralittoral benthic biocoe-

nosis from Northern Rıa de Muros, Atlantic Coast of Northwest

Spain. Bot Mar 45:93–122

Ourens R, Fernandez L, Freire J (2011) Geographic, population, and

seasonal patterns in the reproductive parameters of the sea urchin

Paracentrotus lividus. Mar Biol 158:793–804

Ourens R, Freire J, Fernandez L (2012) Definition of a new unbiased

gonad index for aquatic invertebrates and fish: its application to

the sea urchin Paracentrotus lividus. Aquat Biol 17:145–152

Ourens R, Flores L, Fernandez L, Freire J (2013) Habitat and density-

dependent growth of the sea urchin Paracentrotus lividus in

Galicia (NW Spain). J Sea Res 76:50–60

Pearse JS, Arch SW (1969) The aggregation behavior of Diadema

(Echinodermata, Echinoidea). Micronesica 5:165–171

Pennington JT (1985) The ecology of fertilization of echinoid eggs:

the consequences of sperm dilution, adult aggregation, and

synchronous spawning. Biol Bull 169:417–430

Pinheiro JC, Bates DM (2000) Mixed effects models in S and S-Plus.

Springer, New-York, p 528

Roa R, Ernst B, Tapia F (1999) Estimation of size at sexual maturity:

an evaluation of analytical and resampling procedures. Fish Bull

97:570–580

Rogers-Bennett L, Bennett WA, Fastenau HC, Dewees CM (1995)

Spatial variation in sea red urchin reproduction and morphol-

ogy: implications for harvest refugia. Ecol Appl 5(4):1171–

1180

Rogers-Bennett L, Rogers DW, Bennett WA, Ebert TA (2003)

Modeling red sea urchin (Strongylocentrotus franciscanus)

growth using six growth functions. Fish Bull 101(3):614–626

Saborido-Rey F, Junquera S (1998) Histological assessment of

variations in sexual maturity of cod (Gadus morhua L.) at the

Flemish Cap (north-west Atlantic). ICES J Mar Sci 55:515–521

Sanchez-Espana AI, Martınez-Pita I, Garcıa JF (2004) Gonadal

growth and reproduction in the commercial sea urchin Paracen-

trotus lividus (Lamarck, 1816) (Echinodermata: Echinoidea)

from southern Spain. Hydrobiologia 519:61–72

Schwarz G (1978) Estimating the dimension of a model. Ann Statist

6(2):461–464

Sellem F, Guillou M (2007) Reproductive biology of Paracentrotus

lividus Echinodermata: Echinoidea) in two contrasting habitats

of northernTunisia (south-east Mediterranean). J Mar Biol Ass

UK 87:763–767

Semroud R, Kada H (1987) Contribution a l0etude de l0oursin

Paracentrotus lividus dans la region d0Alger: indice de repletion

et indice gonadique. In: Boudouresque CF (ed) Colloque

international sur Paracentrotus lividus et les oursins comestibles.

GIS Posidonie, Marseilles, pp 117–124

Shpigel M, McBride SC, Marciano S, Lupatsch I (2004) The effect of

photoperiod and temperature on the reproduction of the

European sea urchin Paracentrotus lividus. Aquaculture 232:

343–355

Spirlet C, Grosjean P, Jangoux M (1998) Reproductive cycle of the

echinoid Paracentrotus lividus: analysis by means of the

maturity index. Invertebr Reprod Dev 34(1):69–81

Spirlet C, Grosjean P, Jangoux M (2000) Optimization of gonad

growth by manipulation of temperature and photoperiod in

cultivated sea urchins, Paracentrotus lividus (Lamarck) (Echi-

nodermata). Aquaculture 185:85–99

Tomas F, Romero J, Turon X (2005) Experimental evidence that

intra-specific competition in seagrass meadows reduces repro-

ductive potential in the sea urchin Paracentrotus lividus. Sci Mar

69(4):475–484

Trippel EA, Harvey HH (1991) Comparison of methods used to

estimate age and length of fishes at sexual maturity using

populations of white sucker (Catostomus commersoni). Can J

Fish Aquat Sci 48:1446–1459

Turon X, Giribet G, Lopez S, Palacın C (1995) Growth and

population structure of Paracentrotus lividus (Echinodermata:

Echinoidea) in two contrasting habitats. Mar Ecol Prog Ser

122:193–204

Tuya F, Cisneros-Aguirre J, Ortega-Borges L, Haroun RJ (2007)

Bathymetric segregation of sea urchins on reefs of the Canarian

Archipelago: role of flow-induced forces. Est Coast Shelf Sci

73:481–488

Unger B, Lott C (1994) In-situ studies on the aggregation behaviour

of the sea urchin Sphaerechinus granularis Lam. (Echinider-

mata: Echinoidea). In: David B, Guille A, Feral JP, Roux M

(Eds) Echinoderms through time. Proceedings of the Eighth

International Echinoderm Conference, Dijon, France, 1993.

Balkema, Rotterdam, pp 913–919

Unuma T, Yamamoto T, Akiyama T, Shiraishi M, Ohta H (2003)

Quantitative changes in yolk protein and other components in the

ovary and testis of the sea urchin Pseudocentrotus depressus.

J Exp Biol 206:365–372

Urgorri V, Reboreda P, Troncoso J (1994) Dispersion, demografıa y

produccion gonadal de una poblacion de Paracentrotus lividus.

Universidade de Santiago de Compostela. 172 pp

Vadas RL, Elner RW, Garwood PE, Babb IG (1986) Experimental

evaluation of aggregation behavior in the sea urchin Strongylocen-

trotus droebachiensis.A reinterpretation. Mar Biol 90:433–488

Vega-Suarez W, Romero-Kutzner V (2011) Patron de distribucion

espacial de Paracentrotus lividus. An Univ Etol 5:21–30

Veiga Villar AJ (1999) Caracterizacion de la flora y vegetacion

bentonica marina intermareal y de su riqueza en recursos

explotables en las Rıas Baixas Gallegas (NO. Penınsula Iberica).

Dissertation, Universidade de A Coruna, Spain

Walker CW, Unuma T, Lesser MP (2007) Gametogenesis and

reproduction of sea urchins. In: Lawrence JM (ed) Edible sea

urchins: biology and ecology, 2nd edition, Elsevier. Dev Aquac

Fish Sci 37: 11–30

Williams H (2002) Sea urchin fisheries of the world: a review of their

status, management strategies and biology of the principal

species. Department of Primary Industries, Water and Environ-

ment, Government of Tasmania, p 27

Williams JR, Babcock RC (2005) Assessment of size at maturity and

gonad index methods for the scallop Pecten novaezelandiae. N Z

J Mar Freshw Res 39:851–864

Willis JK, McClain EP, Pichel WG, Walton CC (1985) Comparative

performance of multichannel sea surface temperatures. J Geo-

phys Res C 6:11587–11601

Zhu GP, Dai XJ, Song LM, Xu LX (2011) Size at sexual maturity of

bigeye tuna Thunnus obesus (Perciformes: scombridae) in the

tropical waters: a comparative analysis. Turk J Fish Aquat Sci

11:149–156

Zuur AF, Ieno EN, Walker NJ, Saveliev AA, Smith GM (2009)

Mixed effects models and extensions in ecology with R.

Springer, Berlin, p 574

Mar Biol (2013) 160:2413–2423 2423

123