124 
Fishery Bulletin 110(1) 
Ricker curve = 0.52) was related to a combination of 
broader spawning in time and space, a higher toler- 
ance to environmental factors, and stronger density- 
dependent mechanisms, especially intraspecific pre- 
dation (Fogarty et al., 2001). The higher recruitment 
variability of haddock (SD = 1.07) and more restricted 
spawning later in the season, coupled with different 
morphological and physiological features and feed- 
ing strategies indicate that activating factors (e.g., 
environmental effects) during the larval stage play a 
more important role in recruitment than constraining 
factors (e.g., density-dependent predation) during the 
juvenile period. 
The Northeast Fisheries Science Center (NEFSC) has 
been conducting groundfish and plankton surveys on 
the northeast continental shelf over the last 50 years 
(O’Brien et al. 1 ; Brodziak et al. 2 ). Ichthyoplankton data 
from two intensive sampling efforts, the Marine Re- 
source Monitoring, Assessment, and Prediction (MAR- 
MAP) surveys from 1977 to 1987, and the U.S. Global 
Ocean Ecosystems Dynamics (GLOBEC) surveys from 
1995 to 1999, have provided abundance estimates of 
eggs and larvae, their transport patterns in the cur- 
rents, growth and mortality rates, and prey and preda- 
tors. Process cruises were embedded within the stan- 
dard surveys to focus on specific mechanisms to learn 
more about the recruitment process in the first year of 
life, the period when the size of the year class is usu- 
ally established. A long-term goal was to develop readily 
obtainable environmental indices in future monitoring 
programs that could be used in models to provide a 
better estimate of recruitment. Because typical stock 
projections of 3—5 years require an assumption that 
recruitment will stay constant, any prior knowledge 
will be important for specifying annual catch limits 
between assessments. 
Our objective was to develop early-stage mortality 
relationships to hindcast recruitment of age-1 cod and 
haddock on Georges Bank. The primary hypotheses 
investigated were the following: 1) the abundance of 
eggs retained on the bank is a function of wind-driven 
transport; 2) larval-stage mortality is a function of 
wind-induced turbulence that promotes feeding, hence 
better survival; 3) the abundance of pelagic juveniles 
is related to the number of larvae that survive; and 4) 
the survival of demersal juveniles to age-1 recruits is 
a function of density-dependent predation. These and 
other recruitment hypotheses were also explored for 
Georges Bank haddock in Friedland et al. (2008). The 
ichthyoplankton data from the two intensive sampling 
1 O’Brien, L., N. Shepherd, and L. Col. 2006. Assessment 
of the Georges Bank Atlantic cod stock for 2005. NMFS 
NEFSC Ref. Doc. 06-10, 148 p. Available from Northeast 
Fisheries Science Center, 166 Water Street, Woods Hole, 
MA 02543-1097. 
2 Brodziak, J., M. Traver, L. Col, and S. Suther- 
land. 2006. Stock assessment of Georges Bank haddock, 
1931-2004. NMFS NEFSC Ref. Doc. 06-11. 114 p. Avail- 
able from Northeast Fisheries Science Center, 166 Water 
Street, Woods Hole, MA 02543-1097. 
periods, MARMAP and GLOBEC, were first used to 
hindcast recruitment, and were then compared with 
the virtual population analysis (VPA) recruitment es- 
timates (O’Brien et al. 1 ; Brodziak et al. 2 ) to determine 
the best life-stage mortality relationships. Then, by 
using the seasonal egg abundance estimated from the 
VPA-derived spawning stock biomass, proxies for egg, 
larval and juvenile mortality rates were applied to all 
years, 1977-2004, to estimate age-1 recruitment. 
Materials and methods 
Flistorically, peak cod spawning occurs on the north- 
eastern part of Georges Bank in February-April and 
in March-April for haddock (Fig. 1), and their eggs and 
larvae are transported south and west along the south- 
ern flank of Georges Bank but some part of the cohort 
is retained on the more shoal, central part of the bank. 
These commercially important species on Georges Bank 
have been monitored by the NEFSC annually by spring 
and fall bottom trawl and plankton surveys. Time series 
data are available since 1978 on spawning stock biomass 
(SSB) and recruitment (R) at age 1, as well as egg and 
larval abundances. The ichthyoplankton surveys have 
been used to produce estimates of the total seasonal 
production of eggs that then could be compared with the 
VPA-derived estimates of egg production given sufficient 
information (Lough et al., 2008). 
Life-stage recruitment model 
The model estimates the number of age-1 recruits from 
the initial abundance of eggs decreasing through various 
life-stage mortality rates to survivors at recruitment. In 
this study, the estimated mortality rates were initially 
used for each stage (without attempting to incorporate 
stochastically determined mortality processes) by using 
the following equation; 
R = E*e-(m 1 *t 1 + m 2 *t 2 + m g *£ 3 + m 4 *t 4 ), (1) 
where R = number of predicted recruits; 
E = initial number of eggs spawned or hatched; 
m 1 - the observed instantaneous mortality over 
the egg period f 1 ; 
m 2 = the larval observed mortality rate over the 
period t. 2 \ 
m 3 = the pelagic juvenile mortality over the period 
t 3 ; and 
m 4 = the demersal juvenile mortality over the 
period t 4 . 
The egg-stage duration (^ 4 ) of 19 days to hatching is 
used for both cod and haddock and is based on devel- 
opment time at a typical winter water temperature on 
Georges Bank of 5°C. The larval- and juvenile-stage 
durations were based on the length-at-age curves in 
Bolz and Lough (1988). Hatching of cod eggs to first lar- 
vae occurs in mid-March and for haddock in mid-April. 
