Abstract

A 38-year time series (1959 to 1996) of weekly
observations of abundance of the marine
diatom Skeletonema costatum and related
plankton habitat variables in lower
Narragansett Bay, Rhode Island was
analyzed. A phytoplankton growth model was
then developed for investigation of
mechanisms regulating Skeletonema annual
abundance pattern. A ca. 45% decline in
Skeletonema abundance was detected, from
2,292 cells per ml (prior to September 1974)
to 1,263 cells per ml (after August 1980).
Winter-spring decreases were greatest, with
declines in March Skeletonema abundance
from near 3,400 cells per ml before a 1977
change-point to mean abundance of ca. 1,000
cells per ml after 1977. Three types of
Skeletonema annual abundance patterns
(winter-spring [w-s], summer, or autumn
bloom-dominated) were characterized, with
the frequency of w-s dominated annual cycles
decreasing in the later portion of the time
series.

Winter-spring Skeletonema bloom years were
bright, windy and cold, with reduced
zooplankton (primarily Acartia hudsonica)
abundance in the first quarter (January-March)
of the year. Summer and fall Skeletonema
bloom years were dark, warm, and had calm
winds and elevated A. hudsonica abundance
in the first quarter of the year. Years in which
an index of winter weather pattern, the North
Atlantic Oscillation Index (NAOI), was low had
colder winter water temperatures and w-s
bloom-dominated Skeletonema annual
cycles. In elevated NAOI years (mainly during
the 1980s and 1990s), warmer winters
occurred and summer-fall blooms dominated.
Linear regression models suggested that
Narragansett Bay Skeletonema annual
patterns are a function of both nearfield and
farfield control. Initial (week 1 of the year)
Skeletonema abundance, position of the Gulf
Stream north wall and first quarter
zooplankton abundance differentiated
between high and low abundance
Skeletonema years. Two variables, initial
(week 1) Skeletonema abundance and state
of the NAOI, were able to distinguish w-s from
summer bloom dominated years.

A growth model was used to identify
mechanisms for the observed long-term
changes in Narragansett Bay Skeletonema
abundance and bloom pattern. The growth
model simulated major features of
Skeletonema annual cycles, replicating bloom
timing and magnitude. Model results indicated
light-limited growth in winter, a brief period of
silicon-limited growth in early spring, followed
by nitrogen-regulated growth in late spring,
summer and early autumn. Grazing
(Acartia-based zooplankton grazing was
dominant) and flushing losses offset
Skeletonema growth. In response to warmer
winter water temperatures characteristic of the
1980s and 1990s, simulated winter and
spring zooplankton grazing rate and quahog
grazing rate increased 25-50 percent during
1959-96. Simultaneous with the increased
grazing, model-inferred immigration of
Skeletonema cells from upper Narragansett
Bay declined. Model results implied that
temperature-dependent increases in
winter-spring grazing and reduced
immigration were responsible for the
observed long-term decline in Narragansett
Bay Skeletonema costatum.

Email: dborkman@gsosun1.gso.uri.edu