Rabaoui et al.: Life history of Thenus orientalis in the Arabian Gulf 
129 
than 10 m (Roa-Ureta, 2015). Dhows may remain at sea 
for nearly a week, and tarads conduct 1-d operations. Both 
fleets are motorized. The dhow fleet is more important 
throughout the landings time series. It accounted for an 
average of 93% of the monthly catch of flathead lobster 
along the entire time series. 
We carried out a stock assessment of the flathead lobster 
with the method described in Roa-Ureta (2015), developed 
to assess data-poor fisheries. We fitted 2-fleet multiyear 
generalized depletion models to 14 years of monthly land- 
ings and fishing effort spanning the 168-month period 
from January 1995 through December 2008. Subsequent 
landings data were not available because landings of the 
flathead lobster were banned. The particular version of 
the model that was used is as follows: 
Cy = YC = LAE se =a8 
ea Ge i a] 
—M 
By 
j=P —M(t —-t: 
+e, Rixje ai e?, (6) 
where C, = the true catch in numbers in month ¢; 
l =a value of 1 or 2, and indexes the fleet; 
E = the observed effort in boat-days of fishing; 
N) = the initial abundance; 
M = the natural mortality rate; 
k = the scaling; 
a = the effort response; 
B = the abundance response; 
i = the index of the time step (=1—168); 
R; = the annual recruitment with index) (j=1,...,14 
for a total of 14 annual recruitments); and 
T, = the timing of annual recruitment with index j 
(in seasonal fisheries, the timing is always the 
first month of the season). 
In these models, parameter k is similar to catchability 
except that it is not affected by changes because of its 
relationship with abundance, and the effort response 
a and abundance response 8 accommodate nonlinear 
dynamics, namely the potential of effort to be saturated 
(a<1), effort synergy (a>1), hyper-stability (B<1), and 
hyper-depletion (B>1). 
We tried model fits with 4 alternative likelihood func- 
tions by using the data from each fleet (4 functions for 
each fleet data set) and 2 numerical methods to minimize 
the negative log-likelihood, leading to 32 model versions 
(4x4x2). The likelihood functions were exact normal, exact 
lognormal, and adjusted profile (i.e., with the dispersion 
parameter eliminated) approximations of the normal 
and lognormal. Precise formulas of these likelihoods are 
detailed in table 2 in Roa-Ureta (2019). Numerical proce- 
dures were the spectral projected gradient and the conju- 
gate gradient methods. These numerical techniques have 
worked well with generalized depletion models because 
they can handle moderately large optimization problems. 
With 14 years of data, the 2-fleet multiyear generalized 
depletion models include 36 (adjusted profile approxima- 
tions) or 38 (exact likelihoods) free parameters that can 
be estimated. The models were fit by using the package 
CatDyn (vers. 1.1-1; Roa-Ureta, 2019) in R. 
Results 
Spatial distribution of abundance and biomass 
The percentage of stations where flathead lobsters were 
caught varied between 16% and 25% of the total num- 
ber of stations sampled, during 2013-2015. The fre- 
quency of occurrence increased to 60% during the 2016 
trawl survey, mainly because of the increase in the fish- 
ing effort over the areas where flathead lobsters were 
caught during the previous surveys. The catch per unit of 
effort at the stations where flathead lobsters were caught 
ranged from 0.434 kg/h in 2013 to 0.593 kg/h in 2014. 
The overall catch per unit of effort of all stations was the 
lowest (0.067 kg/h) in 2013 and the highest (0.339 kg/h) 
in 2016 (Table 1). 
Results from the trawl surveys conducted from 2013 
through 2016 indicate that the flathead lobsters were 
patchily distributed in the Arabian Gulf, from the northern 
border between Saudi Arabia and Kuwait to the southern 
border with Bahrain. The abundance and biomass of the 
lobsters were both higher in the central and northern parts 
of the western Arabian Gulf (Fig. 1). The abundance at the 
stations where flathead lobsters were caught ranged from 
0.5 x 10°° to 4.8 x 10°? individuals/100 m? (Fig. 1A), and the 
biomass ranged from 0.001 to 0.800 g/100 m? (Fig. 1B). 
Reproductive biology 
The sex ratio of flathead lobsters captured in waters of 
Saudi Arabia was 1:1, with mature lobsters being dominant 
in the samples collected during November 2016. Few ber- 
ried and many spent flathead lobsters were also present in 
the samples. Few immature flathead lobsters were collected, 
and most of them were above 40 mm in carapace length 
and were in the preadult or adult stage. Maturity data sup- 
port a sex-separate maturity model for the flathead lobster 
with a high AIC difference of 6.01. Therefore, the maturity 
curve was fitted to data for females and males separately 
(Fig. 2). The carapace lengths at 50% and 95% maturity 
were 59.5 mm (standard error [SE] 0.6) and 64.6 mm (SE 
1.5) for females and 57.8 mm (SE 1.1) and 71.0 mm (SE 3.0) 
for males. Supplementary Figures 1 and 2 (online only) show 
the maturity stages of female flathead lobsters, as well as 
the variations in their percentages. Four berried flathead 
lobsters (Suppl. Fig. 3) (online only) were collected during 
the study period, and the eggs they carried were in the 
advanced developmental stage (Suppl. Fig. 4) (online only). 
The weight of an egg mass varied from 18.74 to 60.62 g, 
and the weight of a single egg ranged from 796 to 849 mg. 
The fecundity of the 4 berried lobsters ranged from 26,000 to 
