NOAA Technical Memorandum NMFS NE 172
Variability
of Temperature and Salinity
in the Middle Atlantic Bight and Gulf of Maine
Based on Data Collected as Part of the
MARMAP
Ships of Opportunity Program,
1978-2001
by Jack W. Jossi1,2 and Robert
L. Benway1,3
Postal Addresses: 1National Marine Fisheries Serv., Narragansett Lab.,
28 Tarzwell Dr., Narragansett, RI 02882
Print
publication date March 2003;
web version posted October 22, 2003
Citation: Jossi JW, Benway RL. 2003. Variability of temperature and salinity in the middle Atlantic bight and Gulf of Maine based on data collected as part of the MARMAP Ships of Opportunity Program, 1978-2001. NOAA Tech Memo NMFS NE 172; 92 p.
Download complete PDF/print version
ABSTRACT
Monitoring of the Middle Atlantic Bight (MAB) and Gulf of Maine (GOM)
has been conducted by the Marine Resources Monitoring, Assessment, and
Prediction Program’s (MARMAP’s) Ships of Opportunity Program
(SOOP) since the early 1970s. Portrayals of temporal and spatial patterns
of surface and bottom temperature and surface salinity for SOOP transects
crossing these regions during 1991-2001, and time plots of anomalous conditions
for spatially coherent sections of these transects during the period 1978-2001,
are presented. Annual transect averages and departures are presented for
both regions.
Lowest average annual surface temperature
in the MAB during 1978-2001 and in the GOM during 1978-97 (after which
coverage was insufficient) occurred in 1996, departing from baselines
in both regions by an average of 1.1°C. Highest average annual surface
temperature during the same respective periods occurred in 1995 for the
MAB, and in 1991 for the GOM, departing from baselines by 1.3°C for
the MAB, and by 0.4°C for the GOM. Three years of consistently low
average surface temperatures in the MAB ended in 1999, followed by two
more years of positive departures.
Surface salinities in the MAB had their lowest annual averages in 1998,
departing from baselines by 1.1 practical salinity units. Four years
of consistently low surface salinities in the MAB ended in 2000.
Bottom temperatures in the MAB were lowest in 1994, averaging 1.4°C
below the baseline. Bottom temperatures in the GOM were lowest in 1997,
averaging 0.5°C below the baseline. From 1999 through 2001, the MAB
bottom temperatures were consistently above the baseline, with the maximum
departure for the entire period of record of 1.8°C in 1999.
The annual transect average data and the time plots reveal that the
magnitude of departures from long-term means was greater for all features
and in all regions during the 1990s than during earlier years.
INTRODUCTION
The Marine Resources Monitoring, Assessment, and Prediction Program
(MARMAP) Ships of Opportunity Program (SOOP) has been consistently
collecting temperature and salinity information between Massachusetts
and Cape Sable, Nova Scotia, and from New York City across the continental
shelf and slope towards Bermuda, since 1978. A history of the program
and variations in surface and bottom temperature and surface salinity
through 1990 were summarized by Benway, Jossi, Thomas et al. (1993).
Included in the latter work were comparisons of time-space conditions
with 1978-90 baseline values. Since 1990, annual summaries, in similar
format, have appeared as NAFO [Northwest Atl. Fish. Organ.] SCR [Sci.
Counc. Res.] Documents (Benway et al. 1992; Benway,
Jossi, and Griswold 1993; Benway and Jossi 1994, 1995, 1996, 1997,
1998; Benway 1998). However, these post-1990 annual reports did not
utilize the consistent baseline of the original 1978-90 summary.
For that reason, it was decided to re-analyze the 1991-2001 data
and produce a report comparable with the first summary.
This report describes the common methods used during 1978-2001,
presents the 1978-90 baseline conditions of surface and bottom temperature
and surface salinity for the Middle Atlantic Bight (MAB) and Gulf
of Maine (GOM), portrays the 1991-2001 values and their departures
from baseline conditions, and presents 1978-2001 time plots of conditions
for several spatially coherent sections of the two transects.
MATERIALS
AND METHODS
SAMPLING LOCATIONS
The track lines of the ships
of opportunity varied on different monthly occupations of a route.
For consistency among sampling efforts, two route polygons were developed
for analysis and portrayal of the data collected in the two regions.
These polygons (Figure
1a and Figure
1b)
were based on composites of all sampling locations, such that oceanographic
features were assumed to vary along the polygons' long axes, but
to vary insignificantly normal to the long axes. Only data collected
within the polygons were included in analyses. A standard reference
position was chosen for each route from which the radial distance
to each sample location was calculated. These reference positions
were located at such distances beyond the narrow ends of the polygons
so that arcs passing through the sample location had little curvature
(see Figure
1a and Figure
1b). The calculation also produced a reference
distance which was positive from North America seaward.
The MAB sampling (Figure
1a) originated at Ambrose Tower (40°27.5'N,
73°49.6'W) and extended offshore approximately 500 km towards
Bermuda. This route is termed MARMAP Route MB, and the corners of
this polygon are defined by the following geographic positions: 40°34'N,
74°00'W; 40°20'N, 74°00'W; 38°30'N, 69°00'W;
and 36°44'N, 70°30'W. Sampling generally concluded within
the Gulf Stream. Because of the Gulf Stream's varying position, not
all sampling transects traversed the entire polygon length. A transect
along this route typically passed through shelf, slope, and Gulf
Stream water masses, and often crossed a portion of Deep Water Dumpsite
106. The major hydrographic features of this region were summarized
by Ingham (1982).
The GOM sampling extended from the Massachusetts coast to Cape Sable,
Nova Scotia (Figure
1b), for a distance of approximately 452 km.
This route is termed MARMAP Route MC, and the corners of this polygon
are defined by the following geographic positions: 43°30'N, 71°00'W;
43°30'N, 65°37'W; 43°00'N, 65°37'W; and 42°00'N,
71°00'W. This polygon included the following geographic regions
of the GOM: Massachusetts Bay, Wilkinson Basin, the central Gulf
ledges, southern Jordan Basin, Crowell Basin, and the western Scotian
Shelf.
DATA COLLECTION
The sampling regimen at every station included deployment of an
expendable bathythermograph (XBT) probe and concurrent collection
of a sea-surface bucket sample for temperature calibration and salinity
determination. In both polygons, XBT's were used to measure water-column
temperatures down to a maximum depth of 500 m. Stations were positioned
on the basis of time intervals rather than actual geographical locations.
Sampling was at least hourly in the MAB, and varied from every 1-2
hr in the GOM. Prior to 1981, all data were collected utilizing Sippican
analog recorders, and traces were digitized. Thereafter, data were
recorded digitally.
Ship position, time of XBT deployment, and water-column depth (taken
from a depth chart or fathometer) were recorded in the computer and
on a MARMAP log sheet. Full details on SOOP operational procedures
and theory are available in Benway and Jossi (in review).
TEMPORAL SCALES
OF SAMPLING
Monitoring of water temperature and salinity in the MAB and GOM
began in mid-year 1970. However, data collected before 1978 were
too incomplete for atlas portrayal. In the MAB, one or two (occasionally
three to four) sampling cruises were conducted per month. By comparison,
there was one sampling cruise per month carried out across the GOM.
Although it was desirable to sample during the first 2 wk of a month
for logistical reasons (i.e., being able to repeat sampling
in case of failure), sampling often occurred later in the month.
However, annual means of the variables measured on both routes reveal
no pattern or step change related to the variation of sampling times.
During 1986, no XBT's were deployed in the GOM, so no bottom temperature
data were collected. Gulf of Maine surface salinity sampling ceased
after 1993.
SPATIAL SCALES
OF SAMPLING
In the MAB, a typical transect began with the first station at
Ambrose Tower (40°27.5'N, 73°49.6'W; distance from reference
point = 17 km) (Figure
1a). Generally, subsequent stations occurred
about every hour (i.e., every 25-35 km depending on vessel speed)
across the continental shelf. Stations were spaced at 15-min (i.e.,
7-km) intervals near the continental shelf break. Thereafter stations
occurred each hour until the North Wall of the Gulf Stream was reached.
The first station when traveling from Bermuda to New Jersey was at
the Gulf Stream North Wall, with sampling at the outbound intervals
until reaching Ambrose Tower. Surface water samples and water-column
temperature data were collected at each station.
In the GOM, sampling was initiated from either end of the Route
MC polygon until 1993. Thereafter, sampling took place from Nova
Scotia to Massachusetts (Figure
1b). Surface water samples were collected
every 1 or 2 hr (i.e., every 18 or 36 km), and water-column
temperature profiles were obtained no less than every 2 hr.
DATA PROCESSING
The depth of each XBT voltage reading was derived
as a function of time alone, using XBT descent equations shown in Table 1.
The XBT voltage readings were converted to temperatures in several
steps. First, a Sippican system MK-9 XBT Controller transmitted
hexadecimal voltage values (representing XBT thermistor measurements)
to a shipboard
personal computer. Second, the XBT program converted those hexadecimal
voltage equivalents to resistance, and then to temperature (°C),
in the following manner:
1. Convert hexadecimal value to decimal value.
2. Convert decimal value to (V) voltage (i.e., V = 10.0 × decimal
value / 4096).
3. Convert (V) voltage to (R) resistance as measured in ohms (i.e.,
R = (18094 1490.1) × V).
4. Convert (R) resistance to (T) temperature as measured in °C (i.e.,
T = -273.15 + {1 / [A + (B × ln R) + C × (ln R)3)]}, where:
A = 1.29502 × 10-3, B = 2.34546 × 10-4, and C =
9.9434 × 10-8).
This logarithmic equation was first presented by Steinhart and Hart
(1968). The constants were determined empirically from laboratory
tests of XBT thermistors (Georgi et al. 1980).
Quality control (QC) software was designed specifically for the
digital data format established in 1981. For each sampling station,
QC plots of temperature versus depth were produced, and all depth-temperature
pairs were listed. QC plots were used to determine water-column depth,
surface temperature, and bottom temperature, as well as to delete
inaccurate XBT drops. Water-column depth was determined by comparing
the log sheet depth with any bottom impact marks on the QC plot.
Surface bucket temperatures were used to calibrate probe values.
Bottom temperatures were chosen from corresponding probe bottom impact
marks located on the analog graph output.
Conductivity measurements of sea surface samples were generated
by use of a Guildline Model 8400a Autosal. Conductivities were converted
to salinity values using the UNESCO standard equations of state (UNESCO
1981).
Time-distance checks were performed on all recorded time and position
data to eliminate erroneous log entries. Analyses of the data revealed
considerable temporal and spatial variability for surface temperature,
surface salinity, and bottom temperature along both of the routes.
As a result, a time-space mapping technique was developed to portray
the data in a form that retained most of the original detail, to
show long-term means and variances, and to determine departures from
the means for individual years. Further, the technique permitted
the extraction of spatially coherent time series and the analysis
of relationships among measured values.
GRIDDING
To overcome problems associated with irregular sampling in both
space and time, the data were subjected to a gridding procedure.
The design of the gridding method was developed in part with other
research at NOAA Fisheries' Narragansett Laboratory (Jossi et
al. 1991; Thomas 1992). From each irregularly spaced raw data
matrix, the gridding technique calculated a curved planar surface
using interpolated data values at regular, spatial-temporal grid
points. Time-space grids of single-year values, base-period (1978-90)
mean values, standard deviation values, and single-year data algebraic
anomaly and standardized anomaly values were produced. The 1978-90
period was selected as the base period to compare the post-1990s'
results with those published for the 1970s and 1980s (Benway, Jossi,
Thomas et al. 1993).
Each grid was defined by route polygon reference
distance (range: 0-452 km) along the x-axis, time (range: 0-365.25
days) along the y-axis, and sample scalar values along the z-axis.
For MAB bottom temperature data, the grid values at greater than
the 210-km reference distance were converted to blanks due to the
limits of the XBT probe. All grids were constructed such that grid
points occurred at intervals of 17.38 km and 15.22 days, i.e.,
there were 27 grid columns (26 divisions in x) along each polygon,
and 25 grid rows (24 divisions in y) for every year or base period
(Table 2).
Grid dimensions
and techniques were chosen by considering, in descending order of
importance: 1) the closeness of fit between the actual data values
and the interpolated grid values as measured by the characteristics
of the residuals; 2) the average data coverage in time and space;
3) the portrayed rates of change of the scalar values; and 4) the
intended use of final products. The statistical characteristics of
the residuals are reported in Table 3.
Distributions of the residuals from the base-period grids, and of
a composite of the residuals from the single-year grids, are fairly
symmetrical (Figure 2 and Figure 3).
One exception is the residuals for route MB base-period bottom temperatures,
which although having a centered mean, exhibited extreme skewness
and kurtosis. This likely is due to the very rapid depth change at
the offshore limit of sampling, and the resulting significant difference
in depth sampled (i.e., water temperatures measured) with
small changes of reference distance. None of the residual sets satisfied
the Kolomogorov D statistic for normality. An inherent bias in this
gridding technique is the de-emphasis of extreme amplitudes in the
raw data values. However, the tails of the residual distributions
(Figure 2 and Figure 3) do not show that this is a severe problem.
The remainder of this section will detail the
sequence of gridding steps applied to all data types. First, a base-period
weighted mean grid, including all data collected during 1978-90,
was generated (Figures 4-9, panel A). Second,
the base-period grid surface, at the actual spatial-temporal location
of each observation, was subtracted from observed values to produce
a data set of base-period residuals (Figure 2). Calculation of these
values assumed a flat, rather than curved surface in each neighborhood.
Third, a weighted variance was found by squaring all residual values,
and then gridding these squared residuals. Finally, a standard deviation
grid (Figures 4-9, panel B) was created by taking the square root
of the variance grid (the calculation did not include the number
of observations, which exceeded 70 at all grid points). It should
be noted that all base-period mean and standard deviation grid values
were localized in time and space, because each grid value was the
weighted mean of all observations within the search ellipse surrounding
that grid point. The time-space distributions of the observations
used in these base-period calculations are shown in Figures 4-9,
panel C.
Maps of single-year conditions,
anomalies, and standardized anomalies were derived for each route
and data type (Figures 10-53 and Figures 55-66).
First, each single-year conditions map was created by gridding all
observed data from that year (panel A). Second, the base-period grid
surface, at the actual spatial-temporal location of each observation,
was subtracted from observed values, producing a set of algebraic
anomalies (i.e., residuals) for each year, which, in turn,
were used to produce annual anomaly maps (panel B). Finally, by dividing
the anomaly at each observation point by the corresponding interpolated
standard deviation (taken from the base-period standard deviation
grid surface), standardized anomalies were calculated for every observation
in a given year. Each annual standardized anomaly map was generated
by gridding these standardized anomaly values (panel C). For portrayal
of statistically significant events, standardized anomaly map contour
intervals were chosen by picking the 15.2% and 84.8% percentile levels,
as an approximation of one standard deviation, and the 2.3% and 97.7%
percentile levels, as an approximation of two standard deviations,
from the 13-yr data set of standardized anomalies.
The grid files calculated above also were used to produce base-period
means and annual means and anomalies for the transects as a whole.
Base-period means were calculated using all grid values from 1978
through 1990. Single-year means were obtained using the same process,
but for a single data year. Single-year anomaly means were calculated
by subtracting the base-period mean grid values from the single-year
mean grid values.
TIME PLOTS
Portrayals of the grid files indicate that neither transect exhibits
spatial coherence over its full extent. Further, these portrayals
are inconvenient for examining interannual variations.
To deal with these two issues in the GOM, cluster analyses (Thomas
1992) were used. Four clusters were established to represent spatially
distinct points along the transects for the variables involved. The
clusters selected (and the reference distances to the centers of
those clusters) were Massachusetts Bay (48 km), western GOM (165
km), eastern GOM (277 km), and Scotian Shelf (396 km). Slices at
these points were taken through all standardized anomaly grids for
the period of record to produce data for time plots. Grid surface
values were derived (again assuming a flat surface in the neighborhood)
each time the slice crossed a grid line (each 15.22 days except where
grids were blanked due to insufficient data), and are in the same
units (standard deviations) as the grids of origin.
For the MAB, one point was selected at a reference distance of
101 km to represent continental shelf conditions usually not influenced
by either land mass or slope water conditions. This reference distance
also marks that point along the transect representing the mean center
of the Cold Pool (Cook 1985). A running-average fit was generated
by taking the average of all data collected within 15 mo before and
after a given time-plot point.
RESULTS
In the MAB, 294 cruises were conducted during the base period (1978-90).
An additional 183 cruises were conducted during 1991-2001, for a
total of 477. In the GOM, 209 cruises were conducted during the base
period. An additional 144 cruises were conducted during 1991-2001,
for a total of 353.
Contoured, 13-yr mean portrayals of surface temperature, bottom
temperature, and surface salinity values are shown for the MAB
(Figures
4-6) and GOM (Figures
7-9). Each figure contains three panels:
(A)
weighted means of the smoothed observations for the 13-yr time
frame; (B) estimated standard deviations of the weighted mean values;
and
(°C) data location within the 13-yr, 452-km base area.
Also presented are surface temperature, bottom temperature, and
surface salinity variations in time and space for each year, 1991-2001,
for the MAB and GOM (Figures 10-53 and Figures 55-66), with the exception
of 2000 for the GOM where sampling coverage was insufficient to
produce standard portrayals. Each single-year figure contains three panels:
(A) annual conditions; (B) departures of the annual conditions
from
the 13-yr mean values expressed as algebraic anomalies; and (°C)
departures of the annual conditions from the 13-yr mean values
expressed as standardized anomalies.
Figure 67 shows the average
relation between bottom depth and reference distance along the two
transects.
Figure 68 shows
the spatially coherent clusters from which GOM time plots were extracted. Figure 69 shows time plots of surface temperature, surface salinity,
and bottom temperature at mid-shelf of the MAB. Figure 70-73 show
time plots of these variables for Massachusetts Bay, western GOM,
eastern GOM, and Scotian Shelf.
In the summary of the first 13 yr of monitoring, Benway, Jossi,
Thomas et al. (1993) discussed 1978-90 baseline conditions
of surface and bottom temperature and surface salinity along the
MAB and GOM transects. This discussion is repeated below because
it and the base-period portrayals are central to understanding the
annual and inter annual variations of these features. Also, the variance
about these mean conditions is important when interpreting the significance
of single-year departures.
Annual conditions for 1991-98 have been presented as NAFO SCR
Documents, as cited in the "Introduction." Although one purpose
of this report was to present recalculated departures for these
years using consistent baselines, the recalculation showed no significant
differences from those resulting from the varying base periods
utilized in these former reports. Therefore, the 1991-98 recalculated
annual portrayals are presented, but only data collected for 1998-2001
are discussed in this report.
BASELINE CONDITIONS
DURING 1978-90
Middle Atlantic
Bight
Surface Temperature
Baseline conditions of surface temperature across the MAB (Figure 4A) show annual minimum values of less than 4°C in late February
very nearshore, of 4-8°C during mid-March on the shelf, and of
8-20°C during mid-March progressing southeastward through the
Slope and Gulf Stream water masses. The highest rate of vernal warming
takes place along the entire transect during late June, with peak
annual temperatures greater than 22°C over the shelf and greater
than 26°C offshore achieved by late August. From August to the
end of the year temperatures decline fairly uniformly over the entire
transect.
On the shelf, the periods of highest variability in these baseline
conditions (Figure 4B) occur in early June when standard deviations
about the 1978-90 means are in excess of 2°C. These variations
are largely due to interannual differences in the timing of the vernal
warming of the surface waters of the MAB. Variations in excess of
1°C occur over the inner shelf during the November and December
period. The November variations can be partly attributed to the interannual
differences of timing of the fall overturn. The December variations
include record-breaking, cold winter temperature values in 1978.
Offshore, the standard deviations about the baseline values generally
range from 2°C to more than 3°C, with the highest values
along the boundary between the Slope and Gulf Stream water masses.
The baseline values for this region are the most variable along the
transect due to the extensive migrations of both the Shelf/Slope
Front and the Gulf Stream North Wall.
Mean surface temperature for the entire transect
is 16.3°C (Table 4).
Surface Salinity
The dominant features of the time-space surface salinity field
across the shelf portion of the MAB transect (Figure 5A) are the
meltwater runoff from mid-March to late April, and the shorter duration
river discharge in mid-August, both concentrated within 30 km of
Ambrose Tower.
The Shelf-Slope Front, defined by 34.5 practical salinity units
(PSU), shows considerable spatial variation through the year, being
just seaward of the shelf break from January through April, then
migrating over 100 km further offshore by mid-June, and returning
to the area seaward of the shelf break by October. Even greater seasonal
excursion can be seen in the 35-PSU isohaline from about 240 km in
the winter to about 430 km in July and August.
Variations about these baseline conditions (Figure 5B) are highest
nearshore where standard deviations from February through April exceed
3 PSU, and values in excess of 1 PSU occur throughout the year, resulting
from variations in the timing and magnitude of runoff. Variations
in excess of 1 PSU occur across the shelf in June and offshore in
August and September. The influence of upstream conditions on these
values (particularly outflow from the GOM) is not easily determined
from this one transect, but is expected to be a contributing factor.
Mean surface salinity for the entire transect is 33.81 PSU (Table 4).
Bottom Temperature
Baseline conditions of bottom temperatures across the MAB shelf
(Figure 6A) show annual minimum values from less than 2°C nearshore
in mid-February to less than 5°C at the shelf break in late May.
The Cold Pool is a major feature of the bottom temperature regime
along this transect. It occurs during the period between the onset
of stratification (approximately from late March inshore to late
April offshore) and the fall overturn (approximately from early September
nearshore to early December offshore) (Cook 1985). Fall overturn
produces maximum bottom temperatures across the shelf, exceeding
16°C nearshore and 13°C offshore to the shelf-slope frontal
boundary.
Variations about baseline values (Figure 6B) show standard deviations
in excess of 4°C nearshore in late July and early August associated
with wind mixing of these shallow waters, and upwelling and downwelling
events common to that area. A large portion of the inner shelf during
September and October, and the outer shelf during October and November,
have standard deviations more than 2°C coinciding with the seaward
progression of the fall overturn. Beyond the 200-km reference distance,
the standard deviations exceed 2°C during most of the year, reflecting
the influence of the migrating Shelf-Slope Front on the bottom, as
well as the passage of warm-core rings. Midshelf deviations more
than 1°C occur during all but the winter months, and can be largely
attributed to interannual variability in Cold Pool temperature and
boundaries.
Mean bottom temperature for the entire transect is 8.8°C (Table 4).
Gulf of Maine
Surface Temperature
Baseline conditions of surface temperature across the GOM transect
(Figure 7A) range from a minimum of less than 3°C over the Scotian
Shelf in mid-March, and less than 4°C over Massachusetts Bay
in late-February, to a maximum of 19°C for Massachusetts Bay
and Wilkinson Basin, and over 14°C through Crowell Basin just
onto the Scotian Shelf in August. The highest rate of change due
to vernal warming occurs during late May through early July across
the entire transect. After August, surface temperatures, with the
exception of the mixed portions of the Scotian Shelf, decline rapidly.
Periods of highest variability occur during June to mid-October
over much of the transect (Figure 7B). Fall overturn generally occurs
during late October through early December, leading to further variability.
The high variations over the eastern end of the transect can be attributed
to variation in the Scotian Shelf Current.
Mean surface temperature for the entire transect
is 9.1°C (Table 5).
Surface Salinity
The GOM surface salinity conditions vary less than those in the
MAB, but still show a wide range of values over the region (Figure 8A). Salinities range from a minimum of less than 30.5 PSU during
the peak runoff into Massachusetts Bay in early June, to a maximum
of near 33.5 PSU over the Scotian Shelf in early November. Unlike
the nearshore waters of the New York Bight during late summer - early
fall, there is no secondary pulse of river runoff
The standard deviation about the baseline conditions is less than
0.5 PSU for much of the time-space area (Figure 8B). However, deviations
reaching 1 PSU occur from April to June over portions of Massachusetts
Bay. Deviations between 0.5 and 0.7 PSU occur over portions of Crowell
Basin during January and February, and over scattered areas at the
eastern end of the transect during February through April. In addition,
standard deviations from 0.5 to 0.7 PSU occur from August through
December beginning over the Scotian Shelf and progressing westward
to include Crowell Basin.
Mean surface salinity for the entire transect is 32.43 PSU (Table 5).
Bottom Temperature
Conditions within the GOM range from a minimum of less than 3°C
over the Scotian Shelf from February through late April, and less
than 4°C over Massachusetts Bay during the same time period (Figure 9A). Central ledges bottom temperatures have average minima generally
between 6° and 7°C. Maximum temperatures greater than 10°C
occurred over the eastern end of the transect from late September
through early November. Maximum temperatures on the western end of
the transect occurred during the same time period but only reached
slightly over 9°C. These times of maximum bottom temperature
coincide with fall overturn for these shelf areas. These maxima are
absent for the basin areas which are commonly isolated from the overturn
event by the Maine Intermediate water, and where standard deviations
are generally less than 1°C.
The largest time-space standard deviations about the baseline conditions
appear over the Scotian Shelf during late fall and winter (>Figure 9B).
Mean bottom temperature for the entire transect is 6.7°C (Table 5).
ANNUAL CONDITIONS
DURING 1998-2001
Middle Atlantic
Bight -- 1998
Surface Temperature
There were above-average temperatures in the nearshore area from
January through April, while there were below-average temperatures
from the 175-km reference distance to near the end of the transect
during mid-February through mid-May (Figure 48). The above-average
nearshore temperatures may have been a consequence of the mild 1997/98
winter air temperatures as measured at Newark and JFK airports (National
Climatic Data Center 1998).
Annual minimum surface temperatures between 5° and 6°C
occurred inshore in February. From the outer shelf to about the 350-km
reference distance, the minimum of less than 6°C occurred in
March. This latter condition was from 4° to more than 8°C
below baseline values, and represents continuation of a negative
trend beginning in 1996. It also is the greatest negative departure
of surface temperature in the MAB since monitoring began, and exceeds
the previous record set in 1996 by over 2°C. These record low
offshore temperatures may be the result of increased transport of
water from the Labrador Shelf to the Slope Sea (Rossby and Benway
2000).
Annual maximum values occurred in August along the entire transect,
reaching over 24°C on the shelf, and over 26°C near the outer
end of the transect.
In September and early October, significantly positive anomalies
occurred at about the 40-km reference distance due, in part, to the
later-than-average occurrence of the fall overturn. A negative anomaly
in excess of 4°C was centered at the outer end of the transect
in November, and a significantly positive anomaly over 3°C occurred
at the 280-km reference distance in December.
Despite the record-breaking conditions in late winter and early
spring, the transect only averaged 0.6°C below normal in 1998
(Table 4).
Surface Salinity
Values ranged from less than 26.5 PSU in March and again in October
off Ambrose Light, to greater than 36 PSU at the outer end of the
transect in February and March (Figure 49).
The below-baseline salinities which have dominated this transect
since 1996 continued through 1998. One explanation for the early-in-the-year,
low inshore salinities was a combination of above-average precipitation
with mild 1997/98 winter conditions as measured at Newark and JFK
airports, both of which are compatible with heavy, early river discharge
in the apex of the MAB (National Climatic Data Center 1998). Low
offshore salinities may be the result of increased transport of water
from the Labrador Shelf region. A low North Atlantic Oscillation
(NAO) Index for the mid-to-late 1990s has been suggested by Rossby
and Benway (2000) as a possible indicator of increased transport
from the Labrador Shelf which, in turn, may cause a southerly shift
of the Gulf Stream and lead to lower offshore salinity values.
Most notable were: 1) the period from January into March when significantly
negative anomalies exceeded 2 PSU inshore near Ambrose Light, and
the period in April when significantly negative anomalies exceeded
3 PSU offshore from beyond the shelf break out to about the 425-km
reference distance; and 2) negative salinity conditions continuing
through most of the remainder of the year, with only minor exceptions
offshore in August and September. Salinities more than 1.5 PSU below
baselines occurred variously over the inner 250 km of the transect
after July. At Ambrose Light in October, salinities measured more
than 4 PSU below baselines. Another unusual pattern in the surface
salinity in 1998 was the shifting offshore of isohalines. For example,
the 30-PSU isopleth extended to the 140-km reference distance in
late July, some 100 km seaward of average, and occupied a good deal
more time-space area than average (Figure 5).
Salinity for the entire transect averaged 1.11 PSU below average
for 1998, the third consecutive year of lower-than-average salinities,
and the lowest mean annual salinity of the 1991-99 period (Table 4).
Bottom Temperature
Minimum bottom temperatures during 1998 were less than 6°C,
occurring during February nearshore, and during March and April over
the outer shelf (Figure 50). Maximum bottom temperatures on the inner
shelf greater than 12°C coincided with fall overturn in October.
Annual bottom temperature maxima greater than 13°C occurred over
the outer shelf in January and December.
Positive departures from baselines occurred generally over the
inner two thirds of the shelf from January to August, with departures
exceeding 3°C at Ambrose Tower in early May, and exceeding 2°C
on the mid-shelf in June. From mid-August through November, lower-than-average
conditions prevailed, with negative anomalies exceeding 4°C on
the mid-shelf in October. By December, conditions along most of the
transect were near baseline values.
The general pattern of anomalies during 1998 was somewhat similar
to those during the previous year, but for the transect as a whole,
bottom temperature values averaged 0.2°C below the baseline,
compared to 0.9°C above the baseline in 1997 (Table 4).
Gulf of Maine
-- 1998
Monthly coverage along the GOM transect has been difficult since
late 1993, and this difficulty has impacted the value of results
for all three data types in this report. Surface salinity sampling
became impossible at that time. Therefore, the following discussion
will be limited to those time-space areas where sufficient data exist.
Table 5 lists the annual means and anomalies for the transect as
a whole since 1991 only where sufficient data exist.
Surface Temperature
Adequate coverage began in April 1998 (Figure 51). By then temperatures
measured between 0° and more than 1°C warmer than average
from western Crowell Basin well onto the Scotian Shelf. Above-average
conditions spread westward to the central ledges in May and June,
and still further westward during July to August, when anomalies
of +2° to +3°C existed from Wilkinson Basin onto the western
Scotian Shelf, respectively. Higher-than-average air temperatures,
as measured at Boston and Providence airports during 7 out of the
first 8 mo of the year, may have contributed to these anomalous temperatures
(National Climatic Data Center 1999).
From October to the end of the year, the Crowell Basin and Scotian
Shelf portions of the transect were colder than average, with temperatures
more than 2°C below baseline. Offshore autumn winds led to
destratification of the water column, bringing below-average-temperature
bottom water
to the surface as disclosed by SOOP water-column temperature data.
During this same time period, the western GOM was close to baseline
values.
Annual minimum temperatures occur during the period of no coverage,
so they cannot be commented on. The annual maximum temperature
exceeded 21°C, and occurred during July in the Wilkinson Basin and
central ledges portion of the transect.
Bottom Temperature
Adequate coverage began in April (Figure 52). West of the central
ledges, the transect was slightly warmer than average from April
through December. The most interesting feature is the significantly-below-average
bottom temperatures in Crowell Basin in April, and the shifting of
that condition eastward across the Scotian Shelf by November, and
westward back onto the central ledges in December. These low bottom
temperatures may be a consequence of an increase in water transport
from the Labrador Shelf, as indicated by the NAO Index which decreased
during the mid-1990s (Rossby and Benway 2000).
Annual minimum bottom temperatures occur during the period of no
coverage, so they cannot be commented on. The annual maximum bottom
temperature exceeded 10°C, and was observed in Massachusetts
Bay in November.
Middle Atlantic
Bight -- 1999
Surface Temperature
Annual minimum surface temperatures between 4.2° and 6°C
occurred inshore from February through March. Annual maximum values
occurred in August along the entire transect, with temperatures reaching
24°-25°C on the shelf, and 28°-29.4°C offshore at
the outer end of the transect (Figure 53).
From January through April, above-average temperatures prevailed
in the nearshore area, much like the conditions in 1998. In January
and February, positive anomalies existed from the outer shelf to
the 320-km reference distance. A brief period of significantly
higher-than-average temperatures occurred at the very end of the transect
in January,
exceeding the mean by more than 4°C. Higher-than-normal air
temperatures at Newark and JFK airports during the 1998/99 winter
months may have
contributed to the continental shelf portion of this pattern (National
Climatic Data Center 1999), while the above-average temperatures
at the outer end of the transect were the consequence of the transect
encountering Gulf Stream water during January.
From late August through September, a significantly positive anomaly
with temperatures more than 2°C above average occurred at the
50-km reference distance. This inshore anomaly was most likely a
consequence of higher-than-average air temperatures (as measured
at the Newark and JFK airports) during the spring and summer months
(National Climatic Data Center 1999). With the exception of August
and October, above-average air temperatures prevailed through December
(National Climatic Data Center 2000) and possibly contributed to
water temperatures 2°C higher than average from mid-November
through December from mid-shelf to the shelf break.
Water temperatures 2°-4°C above average
beyond the shelf break during November and December coincided with
a large meander of the Gulf Stream bringing the North Wall over 100
km shoreward of its usual position (Figure 54).
In November, this feature coincided with an anomaly from 2° to
more than 4°C below average as slope water was displaced/entrained
along the feature’s eastern edge.
For 1999, the entire transect averaged 1.0°C above the baseline
surface temperature of 16.3°C, and appears to mark the end of
a 1996-98 cool period (Table 4).
Surface Salinity
From January into June, lower-than-average salinities
were generally observed along the shelf, with values of 2.4 PSU below
baseline at the 50-km reference distance in February, and 1.5 PSU
below baseline near the shelf break in May (Figure 55). A notable exception to this pattern was above-average salinity
occurring inshore of the 25-km reference distance in March. This
anomaly may have been a consequence of lower-than-normal winter and
spring precipitation (as measured at the Newark and JFK airports)
leading to reduced river runoff in the spring (National Climatic
Data Center 2000).
Beyond the shelf break, at the 300- to 450-km reference distance,
a low salinity anomaly was observed during February and March, with
values of 0.5-1.3 PSU below baseline. This anomaly coincided with
the passage of a warm-core ring which entrained a long streamer of
cooler, fresher shelf water through the outer part of the transect.
Salinity values for the transect were near average for the remainder
of the year, with the notable exceptions of a positive salinity anomaly
at the shelf break coinciding with downwelling observed in the SOOP
water-column temperature data during July and August, and a negative
salinity anomaly at the outer end of the transect in November. This
negative anomaly coincided with the offshore movement by the North
Wall of the Gulf Stream and the passage of a warm-core ring as mentioned
above.
Salinity values for the transect as a whole ranged from 27.5 PSU
in April near Ambrose Light to above 36.0 PSU at the outer end of
the transect in January, August, September, and December. For 1999,
the transect averaged 0.20 PSU below the baseline, continuing a 4-yr
period of below-average salinities (Table 4).
Bottom Temperature
From January through March 1999, bottom temperatures
across the shelf were up to 4°C above average, possibly due to
unusually mild winter air temperatures (as measured at Newark and
JFK airports) experienced during the 1998/99 winter months (National
Climatic Data Center 2000) (Figure 56). For
the transect as a whole, temperatures were generally above the baseline
throughout the year, with significantly positive anomalies of water
up to 8°C above average inshore during June and July, and 3°C
above average at the 140- to 175-km reference distance during May
and June.
Annual minimum bottom temperatures of less than 5°C were observed
in March nearshore. Annual maximum bottom temperatures greater than
19°C were observed during June to early August nearshore, and
bottom temperatures between 18° and 19°C occurred in October
on the inner shelf, coinciding with fall overturn.
Water temperatures between 2° and 3°C below normal were
observed beyond the shelf break during May-June and during October-December.
This latter period coincided with the passage of a large warm-core
ring with entrained streamers of cooler shelf water and a seaward
movement of the Gulf Stream North Wall mentioned above.
The transect as a whole averaged 10.6°C, the highest mean annual
temperature and departure from the baseline during the 1991-99
period (Table 4).
Gulf of Maine
-- 1999
Surface Temperature
Temperatures across the transect were near average
from January through April (Figure 57). There
was no coverage from May through the first half of July. When sampling
resumed, surface temperatures generally 1°-2°C above average
were observed in Massachusetts Bay and the Wilkinson Basin from July
through September. Surface temperatures 2°-3°C above average
were observed from eastern Wilkinson Basin to the eastern terminus
of the transect in August. By early October, the Crowell Basin had
surface temperatures more than 4°C above the baseline. Temperatures
across the Scotian Shelf were 2°-3°C higher than average
in September. With the exception of the Wilkinson Basin, temperatures
generally remained more than 1°C higher than average after late
October.
Annual minimum surface temperature for the transect was 1.4°C,
and occurred on the Scotian Shelf in March. Annual maximum values
for the transect reached 22°C, and occurred near the boundary
of Wilkinson Basin and the central ledges during July and August.
Bottom Temperature
The year began with average
temperatures across the Massachusetts Bay and Wilkinson Basin portions
of the transect. Colder-than-average bottom water was observed on
the central ledges and Crowell Basin in January (Figure 58). This cold water was a continuation of conditions in late
1998 (Figure 52). The cold water remained in
Crowell Basin through March when temperatures reached more than 2°C
below average.
Starting on the western Scotian Shelf in February, and spreading
across Crowell Basin in April, water more than 1°C warmer than
average occupied the time-space area. Adequate coverage was interrupted
from May through the first half of July. However, warmer water continued
to dominate the eastern GOM when sampling resumed. By late August,
this positive anomaly had spread westward onto the central ledges,
and by late September and early October, it had spread eastward across
the Scotian Shelf where temperatures more than 2°C above normal
were observed. These warmer-than-average bottom temperatures may
be a consequence of reduced coldwater transport from the Labrador
Shelf region, as suggested by a rising NAO Index during the late
1990s (Rossby and Benway 2000). Temperatures for the remainder of
the year were near average for Massachusetts Bay and western Wilkinson
Basin, and more than 1°C above average from the eastern Wilkinson
Basin to the Scotian Shelf.
Annual maximum transect bottom temperatures between 12° and
13°C occurred on the Scotian Shelf in late September and early
October. Annual minimum bottom temperatures between 1° and 2°C
were observed on the Scotian Shelf in March.
Middle Atlantic
Bight -- 2000
Surface Temperature
Annual minimum surface temperatures less than
4°C occurred nearshore in late February (Figure 59). Minimum temperatures offshore were between 8°C and 10°C
during the latter half of December. Annual maxima occurred in mid-August,
reaching greater than 22°C over the shelf, and greater than 28°C
near the end of the transect.
Anomalously warm water was found over the entire transect during
January and February where temperatures generally exceeded means
by 2°-4°C on the shelf, and by 6°-8°C offshore. This
anomalous feature was a continuation of the event of late 1999 resulting
from a large Gulf Stream meander (Figure 54). For the inner 300 km
of the transect, conditions were near normal from March through October.
From the shelf break to the 300-km reference distance in November
and December, temperatures were generally more than 2°C below
average. Offshore of the 320-km reference distance, temperatures
remained above average throughout the year, exceeding means by more
than 4°C in April, May, October, and December.
For 2000, the entire transect averaged 1.1°C above average
(Table 4).
Surface Salinity
Annual minimum surface salinity values occurred
from May through August at the inshore end of the transect (Figure 60). Annual maxima in excess of 36.5 PSU occurred beyond the
380-km reference distance from late October to the end of the year.
Significantly high salinities were centered at the 90- and 180-km
reference distances in February, and at the 440-km reference distance
in January. The inner 350 km of the transect were generally near
average from late March through August. From mid-September to late
October, salinities dropped to more than 1 PSU below average over
the inner 75 km of the transect. From the shelf break to the 300-km
reference distance, salinities more than 1.5 PSU below average occurred
during November and December.
The outer end of the transect was generally saltier than average
from May to the end of the year, with values exceeding means by as
much as 1-1.5 PSU.
The entire transect averaged 0.29 PSU above average during 2000
(Table 4).
Bottom Temperature
Annual minimum bottom temperatures less than
5°C occurred over the inner 50 km of the transect in early
March (Figure 61). Annual maxima greater
than 22°C
occurred at about the 20-km reference distance in early August.
Like 1999, bottom temperatures over much of the shelf were above
average; the only exception being at the 150-km reference distance
in April and again in November. Values exceeded means by 1°C
over much of the outer shelf from January through October, and inshore
by more than 8°C in early August.
Ordinarily, maximum bottom temperature is reached in late September,
and coincides with the fall overturn (Figure 6). In 2000, the maximum
was reached in early August due to recurring downwelling in the New
York Bight apex region (Pacific Fisheries Environmental Laboratory
2000), and a maximum 65-km shoreward excursion of the Gulf Stream
North Wall from early June through August (Johns Hopkins University
Applied Physics Laboratory 2000).
Gulf of Maine
-- 2000
As noted earlier, sampling coverage of surface and bottom temperatures
and surface salinity in the GOM during 2000 was insufficient to produce
the standard portrayals.
Middle Atlantic
Bight -- 2001
Surface Temperature
Annual minimum temperatures of less than 4°C
occurred over the inner shelf in mid-February (Figure 62). Minimum temperatures at the 450-km reference distance were
just over 20°C, and occurred in early April. An annual maximum
greater than 22°C was reached over the entire shelf by mid-July,
with a progressive persistence from inshore to offshore. The end
of the transect reached maximum temperatures greater than 28°C
in September.
The warmer-than-average water beyond the shelf break in 2000 generally
continued to be present through March 2001, exceeding means by
more than 4°C. Brief and spatially isolated anomalous events occurred
through the middle third of the year (e.g., cold water between
the 200- and 275-km reference distances in May, and hot water beyond
the 400-km reference distance in August and September). This latter
feature resulted from the passage of a warm-core ring across the
transect (NOAA CoastWatch Northeast Regional Node 2002). Other
significantly positive anomalies began to appear over the outer shelf
in October,
resulting from a belated fall overturn in that region. Significantly
high surface temperatures extended over the first 340 km of the
transect by December, coinciding with the second-warmest November-January
air temperatures in several bordering states since weather records
began in 1895 (National Oceanic and Atmospheric Administration
2002).
For 2001, the entire transect averaged 0.8°C above average
(Table 4).
Surface Salinity
Annual minimum salinities of less than 27.5 PSU
were centered very nearshore in early April (Figure 63). Highest values occurred from January through mid-March at
the outer end of the transect.
Salinities more than 1.5 PSU below average briefly occurred between
the 230- and 310-km references distances in January, a remnant of
a feature with an inshore-undulating time-space margin during 2000.
Salinities more than 2.0 PSU below average also occurred seaward
of the spring runoff plume during April and May. An area centered
on the shelf break in between April and June reached significantly
low values (by more than 2.0 PSU) during the middle of this time
period. Moderately-higher-than-average salinities occurred offshore
during the first 3 mo of the year, and again at isolated locations
inshore and offshore in August and September, and offshore in December.
These offshore anomalies correspond well with positive anomalies
of surface temperature.
For 2001, the entire transect averaged 0.09 PSU above average (Table 4).
Bottom Temperature
Typical annual minimum bottom temperatures of
less than 4°C occurred in February at about the 50-km reference
distance (Figure 64). Annual maximum temperatures
in excess of 18°C occurred nearshore in July.
In terms of departures from the 1978-90 baselines, the first half
of the year generally remained within normal variability. By early
July, significantly-warmer-than-average water was present over more
than 50 km of the outer shelf, and anomalously cooler water occupied
the slope area beyond 200 km. The 10-11°C water of the upper
100 m of a warm-core ring was found about 65 km shoreward of the
average location of water with these temperatures, leading to the
positive anomalies. Likewise, the ordinarily deeper and colder water
of the ring was present at shallower depths in August and September
due to the ring’s shoreward position.
Gulf of Maine
-- 2001
Surface Temperature
Continued poor coverage occurred for the transect
in 2001 (Figure 65). Near-normal annual minimum
temperatures less than 4°C occurred in Massachusetts Bay during
February and early March. Annual minimum values of 2°C were recorded
on the eastern end of the transect from February to late March.
Temperatures over the western margins of the Scotian Shelf were
more than 1°C below average during the same February-to-late-March
period. Measurements taken in June over the western 300 km of the
transect indicate that that portion of the GOM was headed for an
anomalously warm summer, perhaps exceeding means by nearly 3°C;
lack of coverage prevents verification. However, from June through
September, the western Scotian Shelf had significantly low temperatures,
falling below the baselines by more than 3°C. From late November
to the end of the year, temperatures were near average over the entire
transect.
No annual means for the entire transect were calculated.
Bottom Temperature
Annual minimum bottom temperatures on the ends
of the transects were similar to those for surface temperatures in
magnitude (i.e., less than 4°C in the west; less than 2°C
in the east), although the minima persisted longer at the bottom
than they did at the surface (Figure 66). No sampling occurred during the usual annual maxima on the bottom.
Significantly warm water appeared from the 150- to 230-km reference
distances in January. By June, this feature had reached westward
to about the 90-km reference distance. This westward movement of
warm water is consistent with intrusions of slope water from the
Northeast Channel that follows the deeper channels of the GOM in
areas of the sampling transect. Figure 66, Panel B, shows continuity
of the features during the first 6 mo of the year, while panel C
does not. This difference is due to the temperature departures in
April not reaching significant levels.
The significantly cold water between the 240- and 320-km reference
distances during mid-March to mid-April is a westward extension
of the anomalously cold water which occurred over the Scotian Shelf
during at least the first 9 mo of the year. Here, and in the vicinity
of Cape Sable in June, bottom temperatures were more than 2°C
below average.
Coverage was not sufficient to allow calculation of a transect-wide
average of bottom temperature for the year.
TIME PLOTS OF
ANOMALIES FOR THE 1978-2001 TIME SERIES
The five time plots are the result of taking a multiyear slice
through the standardized anomaly grids at one reference distance
in the MAB, and four in the GOM. An adequate analysis of these time
plots is beyond the scope of this paper. However, the overlay of
a 15-mo running average reveals features of the signal that are felt
to be significant enough to warrant comment.
Middle Atlantic
Bight
The location chosen to represent the MAB was the 101-km reference
distance, or approximately 40°N, 73°W, along the transect
(Figure 1a).
Surface Temperature
Isolated months during 1978-2001 departed significantly
from the 1978-90 means (Figure 69). Departures
in excess of two standard deviations were more numerous in the 1990s
than in the previous years, even after adjustments to account for
the 1990s not being included in the base period. Sequential, monthly
positive or negative departures were more numerous in the 1990s than
in previous years. The December 2001 and January 2002 anomalies both
exceeded the previous series record of September 1990.
Finally, anomalies averaged higher after 1990 (+0.39) than through
1990 (-0.07).
Surface Salinity
Isolated months during 1978-2001 departed significantly from the
1978-90 means, and are especially more prevalent in the late 1990s
than in earlier years (Figure 69). There is more month-to-month consistency
in the surface salinity departures than in the surface temperature
departures. Uninterrupted positive departures of 2 yr (i.e.,
1980-81 and 1985-86), and uninterrupted negative departures of 2-3
yr (i.e., 1996-98 and 1998-99), occurred.
No trend during the time period was apparent, although anomalies
averaged lower after 1990 (-0.34) than through 1990 (+0.17).
Bottom Temperature
Greater departures in the 1990s also occurred in the bottom temperature
data (Figure 69). The phase changes of the smoothed values are quite
similar through the time period for the bottom temperature and the
surface temperature. Departures from 1994 through 2000 were considerably
larger than during the earlier part of the series.
Anomalies averaged slightly higher after 1990 (+0.10) than through
1990 (-0.07).
Massachusetts
Bay
The location chosen to represent Massachusetts Bay was the 48-km
reference distance, or approximately 70°20'W, along the transect
(Figure 1b).
Surface Temperature
With the exception of isolated monthly departures
near, or in excess of, two standard deviations, the 1978-88 period
exhibited no enduring anomalous surface temperatures (Figure 70). During 1989 to early 1990 and during 1992 to mid-1994, mostly
colder-than-average conditions prevailed. Positive anomalies were
briefly present from mid-1990 through 1991. After 2000, most departures
were positive.
The smoothed values depart from the mean considerably less than
those for surface temperature in the MAB.
No trend during the time period was apparent.
Surface Salinity
Salinity anomalies generally shifted from negative values at the
beginning of 1978 to a positive peak by mid-1980, declined generally
through early 1984 to a period minimum, rose sharply to a positive
peak in mid-1985, declined generally to a negative point in late
1987, and after generally climbing to a positive peak in 1990, again
declined generally to a negative point at the end of the sampling
period in 1993 (Figure 70). The longest sustained period was that
of negative anomalies in 1983 and 1984.
Bottom Temperature
From 1978 to 1981, bottom temperatures were near normal (Figure 70). Positive departures occurred during 1982 and 1983, followed
by near-average values in the mid-1980s. From 1987 through early
1990, and from early 1992 to mid-1994, values were generally negative,
after which departures became inconsistent, with several significantly
warm months. Departures in the late 1990s were less excessive than
in the earlier period, and might result in a warming trend for these
data -- anomalies averaged slightly higher after 1990 (+0.10) than
through 1990 (-0.11).
Western Gulf of
Maine
The location chosen to represent the western GOM was the 165-km
reference distance, or approximately 68°55'W, along the transect
(Figure 1b).
Surface Temperature
Variations from 1978 through 1990 followed a
similar pattern to those for surface temperature anomalies in Massachusetts
Bay, except that they were of slightly larger magnitude (Figure 71). Positive anomalies generally occurred from 1983 to early
1985; negative anomalies occurred in 1982, generally for a fairly
prolonged period from mid-1986 to 1991, and generally again from
mid-1991 to 1994. This pattern was followed in 1996 by the lowest
annual average temperature anomaly of the period, from which values
began increasing to reach their highest of the period by 2000. This
latter pattern was not seen in Massachusetts Bay.
No trend was apparent, although the last 5 yr of the period exhibited
consistent and numerous significant positive departures.
Surface Salinity
The western GOM surface salinity pattern follows that of Massachusetts
Bay very closely (Figure 71). The only major exception was that in
the western GOM, the 1985 positive anomalies persisted to the beginning
of 1987. The maximum positive anomaly of the time series occurred
in mid-1985, and the maximum negative anomaly occurred in mid-1984.
No trend was apparent during the time period.
Bottom Temperature
Patterns here were also very similar to those for bottom temperature
in Massachusetts Bay, although the departures were of less magnitude
(Figure 71). The maximum negative anomaly of the time series occurred
in late 1996. Positive anomalies generally dominated from 1991 to
near the end of 2001, with the maximum positive anomaly of the time
series occurring in mid-1991. Variations were larger in the late
1990s than earlier in the period.
Anomalies averaged higher after 1990 (+0.43) than they did through
1990 (-0.29).
Eastern Gulf of
Maine
The location chosen to represent the eastern GOM was the 277-km
reference distance, or approximately 67°37'W, along the transect
(Figure 1b).
Surface Temperature
Patterns in the eastern GOM followed those to
the west very closely (Figure 72). The maximum
negative anomaly of the time series occurred in early 1988, and the
maximum positive anomaly of the time series occurred in late 1999.
The only major difference from the pattern to the west occurred in
1998 when a brief but major shift to cold water interrupted a general
4-yr uptrend. The larger anomalies of the late 1990s -- compared
to the earlier years -- are again present.
Anomalies averaged higher after 1990 (+0.22) than they did through
1990 (-0.07), largely as a result of values from 1998 through 2000.
Surface Salinity
Variations in this portion of the GOM are quite similar in pattern
and magnitude to those to the west (Figure 72). The maximum positive
anomaly of the time series occurred in early 1990, and the maximum
negative anomalies of the time series occurred in 1984 and 1987.
Negative anomalies dominated after 1986.
Bottom Temperature
Patterns and magnitudes of bottom temperature departures in the
eastern GOM are quite similar to those in the western GOM from 1978
to 1990 (Figure 72). Thereafter, the patterns begin to go out of
phase, and the magnitudes, especially in the late 1990s, were much
more negative in the eastern GOM, exhibiting significantly-lower-than-average
temperatures in 1998 and early 1999. The maximum positive and negative
anomalies of the time series occurred in 1998 and 1997, respectively,
and were part of the usual late 1990s relatively extreme anomalies.
Anomalies averaged higher after 1990 (+0.08) than through 1990
(+0.32).
Scotian Shelf
The location chosen to represent the Scotian Shelf was the 396-km
reference distance, or approximately 66°15'W, along the transect
(Figure 1b).
Surface Temperature
From 1978 through mid-1993, surface temperature
anomaly patterns on the Scotian Shelf generally followed those in
the eastern GOM (Figure 73). However, the magnitudes
of departure in the former were considerably less than those in the
latter. From 1993 to 1998, temperatures varied in opposite phase
to those to the west. By 1999, both areas exhibited warmer-than-average
temperatures, with the departures on the Scotian Shelf of lesser
magnitude. Values returned to near average at the end of 2000, became
negative during much of 2001, and returned to a positive phase at
the end of the time series.
Larger anomalies were apparent during the 1990s than during earlier
years, and no trend was seen over the entire time period.
Surface Salinity
Patterns of variation in surface salinity deviations were generally
in phase with those in the eastern and western GOM, and to a lesser
extent with those in Massachusetts Bay, during the entire period
of sampling. However, like surface temperature, the salinity departures
from means were less than those in the western sections.
No trend was apparent during the time period.
Bottom Temperature
Patterns of variation in bottom temperature deviations as determined
by the smoothed data seemed to be in synchrony -- except for 1988-92
-- with the patterns in the eastern GOM (Figure 73). However, these
shifts often resulted in Scotian Shelf temperatures being in a positive
phase while those to the west were in a negative phase, and vice
versa.
Bottom temperature measurements from the eastern GOM are obtained
from considerably greater depths (Figure 72) than the measurements
from the Scotian Shelf, and would be expected to represent different
water masses. Unfortunately, an investigation of a connection between
these patterns is beyond the scope of this paper.
DISCUSSION
This paper presents time-space portrayals of surface temperature
and salinity and bottom temperature data from the MAB and the GOM
during 1991-2001. The style of the portrayals was chosen to aid in
comparison of these data with those published earlier (Benway, Jossi,
Thomas et al. 1993). Also presented are time plots of anomalous conditions
of these data for the 1978-2001 period.
Record high and low values for all three data types, and for both
geographical regions, all occurred during the 1990s. Surface temperatures
along both transects during 1996 averaged the lowest during the
entire 1978-2001 period, departing from baselines by 1.1°C in both
regions. The highest annual average surface temperatures in the MAB occurred
during 1994 and 1995, and in the GOM during 1991. Three years of
consistently low surface temperatures in the MAB ended in 1999.
Surface salinities in the MAB had their lowest annual averages
in 1998, departing from baselines by 1.1 PSU. Four years of consistently
low surface salinities in the MAB ended in 2000.
Bottom temperatures in the MAB were lowest in 1994, averaging 1.4°C
below the base period. Lowest bottom temperatures in the GOM occurred
in 1997, averaging 0.5°C below the baseline. In the MAB, 1999
had the highest average bottom temperatures for the period of record
(+1.8). In the GOM, 1994 had the highest average bottom temperatures
for the period of record (+0.5).
As the monitoring program continues, special attention will be
paid to the relatively extreme behavior of the ecosystem that occurred
in the 1990s.
ACKNOWLEDGMENTS
For their generous cooperation for over a decade, we extend our
appreciation to the captains, officers, and crews of the: Oleander,
Bermuda Container Lines; Yankee Clipper, Claus Spect, Hamburg,
Germany; and C/V Skogafoss and C/V Godafoss, Skogaline
Ltd., St. John, Antigua. Appreciation also is proffered to all individuals
who have volunteered to ride these ships, in particular those riding
the Oleander. Special thanks are extended to the staff of
the National Ocean Service's Office of Ocean Observations for their
continued support. Lastly, for all employees who have passed through
this office and for all other participating vessels over the last
22 yr too numerous to name, thanks for a job well done.
REFERENCES
CITED
Benway, R.L. 1998. Surface and bottom temperatures and surface
salinities: New York to the Gulf Stream, Massachusetts to Cape Sable
N.S., 1997. NAFO [Northwest Atl. Fish. Organ.] SCR [Sci.
Counc. Res.] Doc. 98/17; 12 p.
Benway, R.L.; Jossi, J.W. 1994. Surface and bottom temperatures,
and surface salinities: New York to the Gulf Stream, Massachusetts
to Cape Sable, N.S., 1993. NAFO [Northwest Atl. Fish. Organ.] SCR [Sci.
Counc. Res.] Doc. 94/15; 16 p.
Benway, R.L.; Jossi, J.W. 1995. Surface and bottom temperatures,
and surface salinities: New York to the Gulf Stream, Massachusetts
to Cape Sable, N.S., 1994. NAFO [Northwest Atl. Fish. Organ.] SCR [Sci.
Counc. Res.] Doc. 95/15; 16 p.
Benway, R.L.; Jossi, J.W. 1996. Surface and bottom temperatures,
and surface salinities: New York to the Gulf Stream, Massachusetts
to Cape Sable, N.S., 1995. NAFO [Northwest Atl. Fish. Organ.] SCR [Sci.
Counc. Res.] Doc. 96/11; 11 p.
Benway, R.L.; Jossi, J.W. 1997. Surface and bottom temperatures,
and surface salinities: New York to the Gulf Stream, Massachusetts
to Cape Sable, N.S., 1996. NAFO [Northwest Atl. Fish. Organ.] SCR [Sci.
Counc. Res.] Doc. 97/06; 11 p.
Benway, R.L.; Jossi, J.W. 1998. Departures of 1996 temperatures
and salinities in the Middle Atlantic Bight and Gulf of Maine from
historical means. J. Northwest Atl. Fish. Sci. 24:61-86.
Benway, R.L.; Jossi, J.W. In review. Ships of opportunity (SOOP)
survey sampling. In: Griswold, C.A., ed. MARMAP ecosystem surveys:
operations manual. Available from: National Marine Fisheries Service,
28 Tarzwell Dr., Narragansett, RI 02882.
Benway, R.L.; Jossi, J.W.; Griswold, C.A. 1993. Surface and bottom
temperatures, and surface salinities: New York to the Gulf Stream,
Massachusetts to Cape Sable, N.S., 1992. NAFO [Northwest
Atl. Fish. Organ.] SCR [Sci. Counc. Res.] Doc. 93/27;
15 p.
Benway, R.L.; Jossi, J.W.; Prezioso, J. 1992. Surface and bottom
temperatures, and surface salinities: New York to the Gulf Stream,
Massachusetts to Cape Sable, N.S., 1991. NAFO [Northwest
Atl. Fish. Organ.] SCR [Sci. Counc. Res.] Doc. 92/16;
19 p.
Benway, R.L.; Jossi, J.W.; Thomas, K.P.; Goulet, J.R. 1993. Variability
of temperature and salinity in the Middle Atlantic Bight and Gulf
of Maine. NOAA [Natl. Ocean. Atmos. Adm.] Tech. Rep.
NMFS [Natl. Mar. Fish. Serv.] 112; 108 p.
Cook, S.K. 1985. Temperature conditions in the Cold Pool 1977-1981:
a comparison between Southern New England and New York transects. NOAA [Natl.
Ocean. Atmos. Adm.] Tech. Rep. NMFS [Natl. Mar. Fish. Serv.]
24; 22 p.
Georgi, D.T.; Dean, J.P.; Chase, J.A. 1980. Temperature calibration
of expendable bathythermographs. Ocean Eng. 7:491-499.
Ingham, M.C., editor; Armstrong, R.S.; Chamberlin, J.L.; Cook,
S.K.; Mountain, D.G.; Schlitz, R.G.; Thomas, J.P.; Bisagni, J.J.;
Paul, J.F.; Warsh, C.E. 1982. Summary of the physical oceanographic
processes and features pertinent to pollution distribution in the
coastal and offshore waters of the northeastern United States, Virginia
to Maine. NOAA [Natl. Ocean. Atmos. Adm.] Tech. Memo.
NMFS [Natl. Mar. Fish. Serv.]-F/NEC-17; 166 p.
Johns Hopkins University Applied Physics Laboratory. 2000. Gulf
Stream, multi-day merged data. <http://fermi.jhuapl.edu:80/avhrr/gs/averages/index.htm>.
Accessed Aug. 30, 2001.
Jossi, J.W.; Goulet, J.R.; Green, J.R. 1992. 30 years of zooplankton
variability in the Gulf of Maine. In: Wiggins, J.; Moores, C.N.K.,
eds. Proceedings of the Gulf of Maine Scientific Workshop, Woods
Hole, MA, 8-10 January, 1991. Boston, MA: Univ. of Massachusetts
Urban Harbors Inst.: 255-258.
National Climatic Data Center. 1998. Local climatological data,
edited. <http://www.ncdc.noaa.gov/pdfs/lcd/lcd.html>.
Accessed Oct. 2, 2000.
National Climatic Data Center. 1999. Local climatological data,
edited. <http://www.ncdc.noaa.gov/pdfs/lcd/lcd.html>.
Accessed Oct. 16, 2000.
National Climatic Data Center. 2000. Local climatological data,
edited. <http://www.ncdc.noaa.gov/pdfs/lcd/lcd.html>.
Accessed Oct. 16, 2000.
National Oceanic and Atmospheric Administration. 2002. November
2001 - January 2002 warmest on record in U.S.; global temperatures
warmest on record in January. <http://www.noaanews.noaa.gov/stories/s869.htm>.
NOAA CoastWatch Northeast Regional Node. 2002. GOES sea surface
temperature. <http://newton.gso.uri.edu/thermo6/ocean_goes_result/html_dir/thumb_page1.html>.
Accessed Mar. 15, 2002.
Pacific Fisheries Environmental Laboratory. 2000. Arrowmark graphic
data product, May 2000. <http://www.pfeg.noaa.gov>.
Accessed Aug. 30, 2001.
Rossby, T.,; Benway, R.L. 2000. Slow variations in mean path of
the Gulf Stream east of Cape Hatteras. Geophys. Res. Lett. 27(1):117-120.
Sippican Corporation. 1970. XBT system linearity equation: ocean
engineering bulletin No. 1. Available from: Sippican Corp., Marion,
MA.
Steinhart, J.S.; Hart, S.R. 1968. Calibration curves for thermistors. Deep-Sea
Res. 15:497-503.
Thomas, K.P. 1992. Initiation and decline of Ceratium tripos blooms
in the Gulf of Maine: dependence upon water column structure. [M.S.
thesis.] Kingston, RI: Univ. of Rhode Island; 127 p.
UNESCO [U.N. Ed. Sci. Cult. Organ.]. 1981. Background papers and
supporting data on the practical salinity scale 1978. UNESCO
Tech. Pap. Mar. Sci. 37; 144 p.