50 
Fishery Bulletin 1 12(1) 
Their results supported a hypothesis put forward 
by Joseph (1972) and concurred with the findings of 
Norcross and Austin. 2 Joseph (1972) analyzed the fluc- 
tuations of commercial landings for Atlantic Croaker 
from the Mid-Atlantic Bight (1890-1967) and discussed 
4 possible causes of their sudden decline after 1945 
(i.e., recruitment overfishing, habitat degradation by 
humans, multispecies interactions, and environmental 
forcing due to natural events). He dismissed the first 
three causes as potential driving forces of the declin- 
ing landings and attributed that decline to extremely 
low winter temperatures that had decimated overwin- 
tering age-0 Atlantic Croaker in estuarine nursery 
habitats. He documented evidence that large landings 
had been associated with warming of sea temperatures 
and that the historical declines in landings had always 
followed cooling trends. On the basis of this informa- 
tion, he originally proposed the overwintering mortal- 
ity hypothesis in juvenile Atlantic Croaker during cold 
winters, resulting in weak year-classes and future, low 
population sizes. 
This hypothesis has been repeatedly adopted in 
subsequent studies. In this respect, Norcross and Aus- 
tin 2 showed that the abundance of juvenile Atlantic 
Croaker in Chesapeake Bay (Virginia) during summer 
positively correlated with estuarine water tempera- 
ture during the previous winter. They associated the 
increase in catch in the mid-1970s with warmer winter 
temperatures and a decrease in catch in the late- 1970s 
with colder winter temperatures. 
With the exception of the study by Hare et al. 
(2010), studies of the population dynamics and man- 
agement of Atlantic Croaker have ignored environmen- 
tal effects on the processes modeled (Barbieri et al., 
1997; Lee, 2005; ASMFC 3 ’ 1 ). On the basis of a mecha- 
nistic recruitment-winter temperature hypothesis (de- 
scribed above), Hare et al. (2010) developed a coupled 
climate-population dynamics model. This model is an 
age-structured production model in which recruitment 
is generated through a stock-recruit relation, and the 
age composition is simulated to be conditional on the 
closest correspondence between predicted and observed 
harvests. The climate effects are log-linearly incorpo- 
rated into the model through a Ricker spawning-stock 
function with a temperature (i.e., MWET) variable. The 
coupled model indicates that both exploitation and cli- 
mate changes significantly affect Atlantic Croaker 
abundance. Importantly, Hare et al. (2010) found a 
significant correlation between the observed Atlantic 
Croaker recruitment and MWET, which thereby sup- 
ports the mechanistic recruitment hypothesis of Hare 
and Able (2007). 
The Atlantic Croaker stock in U.S. Atlantic waters 
can be considered data-moderate. In fact, this stock has 
3 ASMFC (Atlantic States Marine Fisheries Commission). 
2005. Atlantic Croaker stock assessment and peer review 
reports. ASMFC, Washington, D.C, 370 p. [Available from 
http://www.asmfc.org/, accessed January 2012.] 
been associated with many data sets, some of which 
were characterized by considerable uncertainty in their 
estimates and representativeness. For example, the 
ASMFC stock assessment subcommittees (ASMFC 3 ' 1 ) 
identified numerous small-scale (i.e., bay- or sound- 
specific) and 2 large-scale (i.e., spanning wide areas, 
many years, or both) survey indices of abundance, one 
coastwide or regional fishery-dependent index (i.e., the 
total catch per unit of effort [CPUE] from the National 
Marine Fisheries Service [NMFS] Marine Recreational 
Fisheries Statistics Surveys [MRFSS]), and various 
sources of fish kills and length data. 
Evaluation of these data sets and assessment pro- 
cedures revealed the following. The small-scale indices 
of abundance possibly reflected better local than coast- 
wide dynamics. The first stock assessment (ASMFC 3 ) 
lacked catch-at-age (CAA) data and dealt with conflict- 
ing trends in regional indices of abundance. The south- 
eastern (North Carolina-east Florida) shrimp trawl 
fishery (SESTF) bycatch, commercial fishery discards, 
and scrap (or bait) fishery landings are currently con- 
sidered significant but have been poorly characterized. 
The development of the MRFSS CPUE appeared unre- 
liable and raised concerns about its value as relative 
index of stock abundance (ASMFC 1 ). In this context, 
differing decisions and assessment choices have been 
adopted. Preference has been given to large-scale sur- 
vey indices and, in order to characterize recruitment, 
to a few small-scale indices developed from survey data 
collected in the so-called overwintering core area for ju- 
veniles. The first stock assessment of Atlantic Croaker 
(ASMFC 3 ) relied on an age-structured production mod- 
el (1973-2002). 
Because of the difficulties encountered in reconciling 
the conflicts between regional indices, regional models 
have been developed, thereby splitting the stock into 
the “northern” and “southern” management units. The 
model for the south Atlantic region, however, did not 
perform satisfactorily. Because that portion of the stock 
could not be assessed, emphasis was placed on the 
“northern” stock. In contrast, the 2010 assessment sub- 
committee (ASMFC 1 ) did not find evidence to support 
a north-south separation of the stock and conducted 
an assessment encompassing data for the coastwide 
stock. Moreover, this subcommittee developed matri- 
ces of CAA for 1988-2008 only and explored various 
assessment approaches, including continuity runs, but 
ultimately chose a statistical CAA model that uses the 
aforementioned CAA data. 
The results of that model form the basis for current 
management. Unfortunately, inadequate estimates of 
the SESTF bycatch and scrap fishery landings particu- 
larly hampered the determination of overfished status 
of the stock. Meanwhile, various ASMFC stock-assess- 
ment subcommittees and review panels documented 
information about climate effects on the population dy- 
namics of Atlantic Croaker. They consequently recom- 
mended that stock assessment models investigate envi- 
ronmental covariates to improve understanding of the 
