CBP/TRS 66/91
September 1991
Dissolved Oxygen Trends
in the
Chesapeake Bay
(1984-1990)
f*-^m~
Chesapeake
Bay
Program
Printed on Recycled Paper
-------
Dissolved Oxygen Trends in the
Chesapeake Bay
(1984-1990)
September 1991
Produced by
Computer Sciences Corporation
under contract to the U.S. Environmental Protection Agency
Contract No. 68-WO-0043
-------
Table of Contents
Executive Summary—.....—.... ..............— .......i
Introduction[[[
Dissolved Oxygen in Estuarine Waters., 1
Determining Trends in Dissolved Oxygen :...2
The Data [[[2
Chesapeake Bay Mainstem Water Quality Monitoring
Program 2
Before the Monitoring Program 4
Data Analysis [[[5
The Interpolator 6
Results of the Analysis .—.... ..............—.......—....—.......9
Long-Term Trends in Dissolved Oxygen (1950-1990) 9
Dissolved Oxygen Trends (1984-1990) 11
Dissolved Oxygen Trends and Other Water Quality
Parameters •. 17
What The Results Mean.........—.....................—.....—...........23
Conclusions......................—........................—.....—.....—......26
-------
Executive Summary
Low dissolved oxygen is one of the most important problems facing
managers and scientists in the restoration of Chesapeake Bay. Oxygen is
critical to the health and survival of the Bay's aquatic life. Faced with
decreasing levels of dissolved oxygen associated with rising nutrient levels
(e.g., nitrogen, phosphorus), managers adopted a basin wide nutrient reduction
goal in 1987, to achieve at least a 40 percent reduction of nitrogen and
phosphorus entering the mainstem Chesapeake by the year 2000. To track
the progress toward this goal and to extablish long-term trends, an analysis
of current and historical dissolved oxygen data was conducted.
Since 1984, the Chesapeake Bay Program has overseen a comprehensive
water quality monitoring program in which the Bay's water is sampled
throughout the year at fixed stations. Water quality data collected prior to
1984 were not comparable to that of the current Monitoring Program and
were thus inadequate in determining historical baywide trends unequivocally.
Dissolved oxygen was examined in several ways to determine trends
through time. Using a volumetric "interpolator," dissolved oxygen was
analyzed over the entire Bay and regionally over segments. This interpolator
also has the capability of dealing with missing data and variable sampling
locations, and weights the influence of individual stations by the size region
they represent Long-term trends (1950-1990) in dissolved oxygen were
determined by examining changes in the volume of water containing under
5 mg/1 of dissolved oxygen and that under 1 mg/l as estimated by the
interpolator. Trends since 1984 were determined by examining changes in:
1) dissolved oxygen concentration;
2) the volume of water under certain target concentrations; and
3) oxygen deficit mass.
Dissolved oxygen in the upper Chesapeake Bay appears to be generally
related to water flow into the Bay from the Susquehanna River, the largest
source of freshwater to the Bay. The volume of hypoxic water decreased
during the drought years of the 1960s, then increased with the higher flows
of the early 1970s, and decreased again in the latter half of the decade
although flow was high.
Since 1984, a gradual downward trend in the average concentration of
dissolved oxygen was observed in the Bay with minima occuring in July
1986 and July 1989, and the volume of hypoxic/anoxic water fluctuated
widely. Under the hydrologic conditions of!984-1990, anoxic waters have
appeared each summer. This finding contrasts with the higher levels found
in the 1950s under similar hydrologic conditions. The volume of water
under 1 mg/1 increased in segments CB4 and CBS. The volume of water
under 5 mg/1 has generally decreased in the upper Bay segments through
CBS and has increased in lower Bay segments. The plot of dissolved oxygen
deficit for the entire Bay showed no definitive trend through time (Figure 10)
nor did the majority of plots for the individual segments. However, three
small segments of the Bay did show trends in dissolved oxygen through the
period of Monitoring Program data (1984-1990). The causes of the apparent
trends are not clear.
Although significant reductions in phosphorus levels have been achieved,
corresponding improvements in dissolved oxygen have not yet been seen.
-------
Introduction
Oxygen in the Bay's waters, like oxygen in the air, is critical for the survival
of its living resources. Thus, the levels of dissolved oxygen found in the Bay
are an important indicator of its water quality. The EPA Chesapeake Bay
Program (CBP), in its initial characterization of the Bay's water quality,
estimated that the amount of water with low or no dissolved oxygen had
increased between 1950 and 1980(CBP, 1983). The reductions in dissolved
oxygen were linked to nutrient enrichment from phosphorus arid nitrogen.
Based on the EPA report and other findings, a basinwide nutrient reduction
goal was adopted in 1987 to achieve at least a 40 percent reduction of
nitrogen and phosphorus entering the mainstem of the Bay by the year 2000.
It is expected that dissolved oxygen levels will rise as nutrient levels
decrease.
This summary of trends in dissolved oxygen is one in a series of reports on
the progress toward and reevaluation of the 40 percent nutrient reduction
goal. Reports describing trends in phosphorus and nitrogen in Chesapeake
Bay are companion documents to this report (CBP, 1991a, b).
Dissolved Oxygen in Estuarine Waters
The amount of oxygen in Bay waters is a function of several factors: 1)
surface aeration and oxygen production processes; 2) physical and chem-
ical processes affecting oxygen carrying capacity and transfer within the
water 3) biological and chemical processes which use oxygen. Atmospheric
oxygen enters the water through wind action and surface adsorption. When
sunlight is sufficient in surface and shallow waters, plant photosynthesis
generates oxygea The total amount of oxygen that water can hold depends
on both water temperature and salinity. The exchange of oxygenated surface
water with waters below is dependent on the degree and extent of density
stratification in the Bay, which, in turn, is a function of the temperature and
salinity of the different water layers. Plant and animal respiration, bacterial
decomposition, and abiotic chemical oxidation-reduction reactions all
consume oxygen.
The addition of nutrients to the water fuels photosynthesis, locally increasing
oxygen production initially, but also boosting plant production.
Decomposition of excess organic matter, which takes place primarily in the
Bay's bottom waters, remineralizes the nutrient constituents of the organic
matter. It is an oxygen-consuming process which delivers phosphorus and
nitrogen both to the sediments and the water.
The rates of these biological and chemical processes increase as temperature
increases. With a temperature rise, however, the amount of oxygen water
can hold (the saturation level) decreases. The saturation level is also
decreased by increasing salinity, but to a lesser extent. In the summer, water
-------
The Data
temperatures rise and salinity generally increases, causing a reduction in
dissolved oxygen potential. If the amount of unconsumed organic production
exported to the bottom is sufficiently large and the Bay waters sufficiently
stratified, then the transport of surface oxygen into lower depths is inhibited.
The bottom waters gradually become depleted of oxygen until they have
either minimal amounts of oxygen (hypoxia) or no oxygen (anoxia).
Determining Trends in Dissolved Oxygen
Determining long-term dissolved oxygen trends in the Bay is hindered by
the dynamic interactions of these many factors and by an historical record
that is inconsistent in space, time, and quality. These problems were
recognized in the research phase of the Chesapeake Bay Program, so that in
1984 the CBP initiated a long-term baywide water quality monitoring
program to evaluate the status and trends of a suite of water quality
parameters, including dissolved oxygen.
t
In this analysis, the historical (pre-1984) record was included to the extent
possible to provide a reference point to past conditions, but emphasis is on
trends since the beginning of the Monitoring Program, from June 1984
through September 1990.
Chesapeake Bay Mainstem Water Quality
Monitoring Program
The Chesapeake Bay Program, a cooperative effort between the federal
government and the states surrounding Chesapeake Bay, provides funds to
the states of Maryland and Virginia for the routine monitoring of 19 water
quality parameters at 49 stations in the mainstem of Chesapeake Bay
(Figure 1).
The Maryland Department of the Environment oversees the Program in the
Maryland portion of the Bay. In the lower Bay, Virginia Institute of Marine
Science and Old Dominion University (under contract to the Virginia Water
Control Board) share responsibility for sample collection and analysis.
At each station, measurements are taken for dissolved oxygen, water
temperature, conductivity, salinity, and pH. These measurements are
sampled from surface to bottom at 1 to 2 meter intervals using either a
Hydrolab probe or a Seabird Conductivity-Temperature-Depth (CTD)
metering system. Water samples for chlorophyll and nutrient analyses are
collected with a submersible pump and taken at surface and bottom at all
stations and at 1 meter above and below the pycnocline at stratified stations
(or at one-third and two-thirds the distance between surface and bottom if
no pycnocline exists).
-------
Mainbay Monitoring Segments and Stations
CB8
Other
CB7
CB6
CB1
CB4
CBS
Relative Volume of
Water in Bay Segments
CB2
Station CB4.1C
Figure 1. Chesapeake Bay stations and segments
-------
The mainstem stations are usually sampled within the same 3-day interval,
with some exceptions due to extreme weather in winter and spring. Initially,
stations were sampled twice monthly from March through October and once
a month from November through February. However, in 1988, the once-
monthly winter schedule was extended to include October and March in
Virginia. At the same time, Maryland dropped the east and west lateral
stations from the winter sampling schedule. These changes in sampling for
Virginia and Maryland should have minimal effect on the dissolved oxygen
analysis, since the period of concern extends from April through September,
with dissolved oxygen levels usually becoming critically low between July
and September.
The EPA Chesapeake Bay Program Office (EPA-CBPO) Computer Center
in Annapolis, MD, compiles and maintains the Monitoring Program data.
Prior to submission to CBPO, the data-collecting organizations check the
data using a variety of quality assurance procedures. At the EPA-CBPO,
values are rechecked against acceptable ranges and questionable entries
are verified with the data originators. Some data are missing because of
equipment failure during data collection and because of the lack of
counterpart data. For example, data collected in Maryland during the
second cruises in March and October 1987-1990 (cruises 75,95,109, and
115), were not used in the analysis due to the fact that no counterpart
cruises took place in Virginia waters.
Before the Monitoring Program
Prior to 1984, the most extensive records of dissolved oxygen were collected
by the Johns Hopkins University Chesapeake Bay Institute (CBI), the
Virginia Institute of Marine Science, the U.S. Environmental Protection
Agency's Annapolis Field Office (AFO), and by Maryland and Virginia as
part of their state water quality monitoring programs. CBI and AFO data sets
sometimes included baywide surveys. These data, however, were usually
interrupted after one or more years once the objectives of the study were met.
The other data sets were generally locally focused.
The method for determining dissolved oxygen changed over the 33-year
record. Dissolved oxygen was originally determined using the Winkler
titration method. Later, oxygen measurements were made by meter, as in the
current monitoring program. The modem meters were usually calibrated
against a Winkler titfatioa Therefore, a 1:1 comparability of methodologies
was assumed.
The sampling designs of these studies differed according to individual
scientific objectives. Often, dissolved oxygen was measured only at the
surface or at several fixed depths: Vertical profiles at 1 to 2 meter intervals,
as in the current program, were rare. The historical data most similar to those
of the current program include once-monthly measurements for a year or
more. More frequently, measurements were made once or twice each
season. Some of the studies were designed to characterize oxygen levels in
the Bay. Other studies focused on Bay circulation and physical models,
nutrients and eutrophication, or were surveys of general water quality.
-------
The historical data posed a variety of quality assurance problems. To check
their quality, the original data were screened for outliers by comparing their
values to the range of values observed in the Monitoring Program for the
same season and segment. Any observation lying outside this range was
defined as an outlier. Outliers were checked against field sheets or data
reports and were omitted from the analysis if they could not be corrected or
confirmed.
The historical dissolved oxygen data from 1950 to 1983 were evaluated for
the suitability of their spatial and temporal coverage in this trend analysis.
Data for each year and month were mapped to determine if the location and
sample depth approximated that of the current Monitoring Program.
Dissolved oxygen levels are most critical for the Bay's aquatic life between
April and September, since this is when oxygen levels are at their lowest.
There were many years for which no data were available, and only a few
years where data were available for more than one or two of those months.
Data for the Virginia portion of the Bay were especially sparse.
The approach chosen was to compare data collected in Maryland Chesapeake
Bay for July of each year since it was the month most frequently sampled.
If necessary, these data were supplemented by data from June or August, if
July data were not available and the spatial coverage in June or August was
adequate. Therefore, analysis of the 40-year (1950-1990) trend in dissolved
oxygen was based on marginally sufficient July data (augmented by June or
August for some years) for 26 of 40 years in Maryland Chesapeake Bay.
Any trends found apply only to Maryland waters.
Data Analysis
There are several measures of dissolved oxygen which reveal various
aspects of the Bay's water quality. Concentrations of dissolved oxygen are
important since aquatic animals require minimum levels to survive. The
volume of water at particular concentrations and the amount of time that
oxygen deficient conditions persist are biologically significant and are also
useful indicators of oxygen conditions from one year to the next. A good
measure for analyzing trends is the total annual oxygen deficit, since it is an
indicator that accounts for differences in temperature and salinity. For these
reasons, several measures of dissolved oxygen were examined in this
analysis:
1) average concentration (milligrams per liter) of dissolved oxygen;
2) volume (cubic meters) of water at particular concentrations;
3) volume days (cubic meter days) or the volume of water at particular
concentrations on one sampling date multiplied by the number of days
between that sampling date and the next, summed over time;
4) oxygen deficit (kilograms) or the difference between the mass of
observed dissolved oxygen and the potential mass of dissolved oxygen
in solution at saturation. In computing oxygen deficit, if the mass of
oxygen in a cell was greater than saturation—which can occur in surface
-------
waters during active photosynthesis—the oxygen mass was denoted as
the saturation value for computation.
Several time scales were used: sampling date, monthly mean, and annual
total. Water year (instead of calendar year) was used to define a year-long
period. A water year begins on October 1 of one year and extends through
September 30 of the following year. For example, wateryear 1985 extended
from October 1, 1984 through September 30, 1985. Water year is often
appropriate in environmental and biological contexts, as it best represents
seasonal and annual hydrology. To compare complete water years, the
partial water year 1984 was excluded since the Monitoring Program did not
begin until late June.
For analytical purposes, the Chesapeake Bay basin is divided into segments.
Salinity regimes, land use, and management areas all factor into the
segmentation scheme (CBP, 1983). The main Bay is divided into eight
segments (CB1 through CBS, Figure 1); the eastern and western embay-
ments and areas at the mouths of tributaries are in separate segments. In
this analysis, regional trends were examined in the eight CB segments.
Baywide trends were derived for the entire region, including peripheral
areas, as well as the main Bay segments. The piechart in Figure 1 shows
the relative contributions of the segments to the total.
The Interpolator
The volumetric "interpolator" developed by Computer Sciences Corporation
(Reynolds and Banner, 1989), is an analytical tool that provides a way to
look at the volume and mass of dissolved oxygen over a region rather than
at particular locations. The interpolator is capable of handling missing data
and variable sampling locations—a major consideration in the historical
data—and also weights the influence of individual stations by the size region
they represent.
Figure 2 illustrates how the interpolator works. The study area is divided
into a three dimensional grid. The bathymetry or depth profile of the Bay
determines the number of cell layers at any location. Actual sampling
locations lie within the grid, and the values at the center of each cell are
estimated by weighting measurements from nearby stations according to
their distance from the center point. Mass or concentration within each cell
is computed and the cells are averaged, totaled or grouped depending on the
desired measurement (e.g., average concentration, total mass, or volume of
water).
The Chesapeake Bay was divided into 57,871 cells, each cell measuring 1
kilometer long by 1 kilometer wide by 1 meterdeep. At each station, vertical
dissolved oxygen measurements were available at 1 or 2 meter intervals.
Where measurements were missing, values were linearly interpolated from
the closest measurements above and below the missing data point A
weighting method commonly applied in geostatistics—the inverse distance
squared of the four nearest neighbors (Monmonier, 1982)—was used to
interpolate cell values horizontally between stations (Figure 3).
-------
Inputs to the Interpolator
cC9*
w •«. ' .i!!!!;;;;;;;;!:!^!!'
•'ilS:i;i.
\
Figure 2. Interpolator three-dimentional grid.
-------
How the Interpolator Works
Station 4
Figure 3. Inverse-Distance-Squared Estimation Method.
-------
Using the interpolator, the concentration of dissolved oxygen was estimated
for each cell in the grid. Volume and mass were computed by summing the
volume or mass of all the cells. Monthly means were obtained by calculating
the average dissolved oxygen concentration, mass, or volume for each
sampling date, then averaging all dates for each month. Oxygen deficit was
calculated for each date, then summed within the water year. Cubic meter
days were also summed within each water year.
The usefulness of the estimates generated by the interpolator were evaluated
by a technique called jackknifing (Clark, 1980). In this approach,
measurements at monitored stations were compared to estimates for that
location. To do this, observed data for the station in question were
eliminated and values were estimated for the station using interpolated
values from the closest cells. The correlation of estimated with observed
values was calculated separately for shallow Qess than 10 meters) and deep
locations. The correlation coefficient for shallow locations was 0.93
(N=42494); the correlation coefficient for deep locations was 0.96 (N=24184).
The high correlation coefficients provide confidence in the interpolator
estimates for dissolved oxygen.
Results of the Analysis
Long-Term Trends in Dissolved Oxygen
(1950-1990)
Long-term (1950-1990) trends in dissolved oxygen were determined by
examining changes in the volume of water containing under 5 mg/1 and
under 1 mg/1 dissolved oxygen in the Maryland portion of Chesapeake Bay.
The amount of water with dissolved oxygen concentrations less than 5
mg/1 was relatively small in 1950 and 1957 (Figure 4) and increased
dramatically in 1958. Note that in Figure 4, the years for which adequate
data did not exist are not shown. The smallest volume of water under 5 mg/
1 was in 1965. Highs in the mid-1970's were about two thirds of the 1958
maximum. Peak volumes after the beginning ofthe Monitoring Program (in
1986,1989) were also about this magnitude.
Figure4. Volume of low dissolved oxygen
in Maryland Chesapeake Bay in July.
ffl Volume < 5 mg/1
• Volume < 1 mg/l
Long-Term Trends
20000.
10000.
5000-
195Q
1060
1070
108O
1000
-------
Variation in the volume of water with dissolved oxygen less than 1 mg/1
showed a similar pattern through time (Figure 4) with some differences in
magnitude. In 1958, the volume of water containing less than 1 mg/1 was
large, however, it constituted a smaller proportion of the water volume under
5 mg/1 than in subsequent "bad" years. The volumes less than 1 mg/1 in 1986
and 1989 were larger than any previous year for which there were data.
As freshwater flows into the Bay from the tributaries, it affects the depth and
degree of water column stratification. This stratification, in turn, controls
the amount of oxygen deprivation below the pycnocline. In some parts of
the watershed, rainfall and snowmelt carry nutrients from the land to the
streams, tributaries, and ultimately to the Bay. The amount of freshwater
inflow at a given time may then indicate the quantity of water-borne
nutrients. To provide the hydrologic background for trends in dissolved
oxygen, the Susquehanna River was used since it is the largest source of
freshwater to the Bay.
Figure 5 depicts the annual flow in the Susquehanna from 1950-1990
(averaged from July of one year through June of the next) compared to the
90-year(1900-1990) mean flow. Flowduringthe 1950s varied about the 90-
year mean. From 1961 to 1969, the freshwater flow was considerably
reduced; from 1971 to 1979, the flow was well above the mean. Another
period of reduced flow followed in 1979 and continued until 1982. Since
1984 (when the CBP Monitoring Program began), flow has been variable
but generally below the 90-year mean. Flow decreased from 1984 through
1988 and increased in 1989 and 1990. The highest flow of the decade
occurred in the June 1983-July 1984 period, just preceding the start of the
Monitoring Program.
Figure 5. Low dissolved oxygen in
Maryland Bay compared to annual
Susquehanna flow.
o Volume of D.O. < 1 mg/1
^— Susquehanna Flow
— — 90-Year Mean
60000
? soooo.
u 20000.
10000-
Susquehanna Flow and
Low Dissolved Oxygen
1950
o o
1960
4000
-6000
-4000
-2000
1970
1980
1990
10
esc &«•.*>•
-------
Dissolved oxygen in the Maryland Chesapeake Bay appears to be generally
related to the Susquehanna's flow. (Rgure 5). The volume of hypoxic water
decreased during the drought years of the 1960s, then increased with the
higher flows of the early 1970s. It decreased again in the latter half of the
decade, although flow was high. Since 1984, the volume of hypoxic/anoxic
water has fluctuated widely. The very high volumes in July 1986 and July
1989 are particularly notable. These volumes of low dissolved oxygen are
much higher than expected compared to equivalent flows in previous years.
As mentioned above, the selected July data from 1950 to 1983 are marginal,
sporadic, and limited to the upper half of the Bay. Conditions at a single
point in time do not adequately represent an entire season or year, and
dissolved oxygen may have been much higher or lower in other months and
years for which data were not available.
Dissolved Oxygen Trends (1984-1990)
Since oxygen is crucial to the Bay's aquatic life, the onset, spatial extent, and
duration of oxygen deficiency is particularly important to their well-being.
Dissolved oxygen goals recently drafted by CBP to protect and restore
Chesapeake Bay's living resources identify 1,3, and 5 mg/1 as key threshold
levels for assessing the health of Bay waters (CBP, 1991c). Another critical
concentration is around 0.2 mg/1, below which anaerobic chemical reactions
occur that release nutrients and toxic sulfldes from the sediments. The
Monitoring Program data provide an opportunity to examine both seasonal
and annual trends as well as identify differences among Bay segments.
A gradual downward trend in the average concentration of dissolved oxygen
was observed in the Bay from 1984 to 1990 ( see total Bay, Figure 6), with
minima occurring in July 1986 and July 1989. This trend was not evident
in all mainbay segments. Segment CB1 showed a slight upward trend in
concentration, while segment CB2 showed no trend (Figure 6). Segments
CB3 through CBS exhibited downward trends of varying degree (Figure 6).
Not only is the concentration of dissolved oxygen important to the Bay's
aquatic life, but also the extent and duration of low dissolved oxygen. Figure
7 shows each segment's volume of dissolved oxygen at 5,3,1, and 0.2 mg/
1 for each sampling date. (Note that plots for CB1, CB2, CB3, and CBS are
not shown since the volume of low dissolved oxygen was negligible in these
segments). The time when oxygen levels dropped most rapidly was between
April and May or May and June. Hypoxic/anoxic conditions occurred
quickly thereafter. In segment CB4, for example, water with less than 0.2
mg/1 had appeared by May in 1985,1988, and 1990 (Figure 7). In 1988,
these anoxic waters remained through the summer and into October.
Station CB4.1C in segment CB4 (Figure 1) is a typical station in the
Chesapeake Bay deep trench. The dissolved oxygen profiles at this station
(Figure 8) show chronic severe oxygen depletion in this part of the Bay.
Notice that anoxic or near anoxic water .extended over 75 percent of the
water column. In 1984 and 1989, for example, anoxia extended from the
bottom of the Bay (at 35 meters deep) up to 7 meters below the surface.
11
-------
Dissolved Oxygen
CB8
Other
CB7
CB6
CB4
CBS
Relative Volume of
Water in Bay Segments
ie
14.
> 12.
10.
i
' a.
a.
4.
2.
0.
CB1
OCIM octas octee octa? octaa owes octw
CB4
14.
1> 12.
I"'
I'"
i 4.
2.
0.
o
o
0 CD
°° o 0° »0 "f •>
o o oo «•
-------
Concentration
CB2
CBS
14.
§ 12.
3 8'
1 '1
•5
l;
0.
16
14.
1> 12.
i" 10-
i4-
2.
0.
16,
14.
1>12.
• 10-
2.
0.
o° o
o» /o »o »„ 0%o ^
0 . ° . ° o „
oo o.o o o
000°0
ee°
°>k> % ™ o
14.
jf 10-
8 o-
!••
2-
OctB4 OcJBS OctBS Oct 87 Oct 88 Oct89 Oct90
CBS
is
o
o «_ .f f -
"I o 0°0 «* o""
o-o o ,,-
„ °o° Oo° 00° oo «0
e*o f ° of o o CD
o ^ o o
c 10.
S
|8.
OetB4 OctSS OctSB Oc»87 OctBB OctBS OctSO
Total Bay
o
°. o., - *
'°0 % °o o- * --o
0 „ 0 00 00
" - o - . - - o —
*• v- •/ v V V \-
o
o o
0 000
° 0° „ 0°
e»°o o»«8°
°<» iff 0° o
oct 84 OCIBS octae octe? octaa octas OMM
CB6
0 0
oc» V oco o8° « *.
o * «/> V «b «*><»>
OctB4 OctBS Oct 86 Oct 87 OctBB OctBS Oct 90
Figure 6. Average monthly concentration of
dissolved oxygen in Chesapeake Bay Segments
(Trend lines werecomputedusingdata from October
1984 through September 1990).
Out84 Oct85 OctSS Oct87 OctSS OctBS OctSO
13
-------
CB4
Volume of Low Dissolved
9000
Oct84 Oct85 Oct86 Oct87 Oct88 Oct89 Oct90
9000
CBS
CB6
Oct84 Oct85 Oct86 Oct87 Oct88 Oct89 Oct90
9000
^ 7500-
6000-
4500-
I
o
2
Oct84 Oct85 Oct86 Oct87 Oct88 Oct89 Oct90
14
-------
Oxygen Over Time
CB7
30000
^ 25000-^
CO
5>
1> 20000-
15000-
1 10000-
5000
9000
£
£
i
6000-
.•§ 4500
o
3000-i
1500 H
— I 1 1 1 1
Oct84 Oct85 Oct86 Oct87 Oct88 Oct89 Oct90
Total Bay
Oct84 Oct85 Oct86 Oct87 Oct88 Oct89 Oct90
Figure 7. Volume of water with low dissolved oxygen concentrations.
|<.2; ^<1.0; D<3.0; il<5.0
15
-------
Anoxia
Surface o
-5
-10
CO
-15
M
-20
-25
-30
Bottom'35
J l M M J S N J M M "j S Ki J" M M J"S N J M M J S Ki J M M j" S Ki J M M J S N J M M J is N
1984 1985 1986 1987 1988 1989 1990
I I > 0.2 mg/l Dissolved Oxygen H 0.1-0.2 mg/l
0 mg/l
Rgure 8. Anoxic conditions at Station CB4.1C.
16
-------
The extent and duration of low dissolved oxygen are more easily characterized
and visualized in terms of the number of "cubic meter days" at particular
concentrations (Figure 9). The lowest oxygen concentrations generally
occur in the deepest parts of the Bay. These graphs show that appreciable
water masses, with concentrations under 1 and 0.2 mg/1, occurred in
segments CBS through CBS. The volume of water in each of these two
concentration categories has generally increased overtime in segment CB4
and varied without distinct trend in segments CB3 and CBS. Water volumes
with less than 3 mg/1 occurred in CB2 through CB7; these volumes generally
increased in segments CBS through CB7. Water volume with concentrations
under 5 mg/1 tended to decrease in the upper Bay segments down through
segment CBS and increase in the lower Bay segments. The Bay as a whole
showed no distinct trend in any of the concentration categories (Figure 9).
The plots of dissolved oxygen deficit for the total Bay showed no definitive
trend through time (Figure 10). In the upper Bay, oxygen deficit decreased
in segments CB1 and CB2. In segment CBS, oxygen deficit increased, while
in segments CB3 through CB7, oxygen deficit was variable without distinct
trends (Figure 10).
Dissolved Oxygen Trends and Other Water
Quality Parameters
The causes of the apparent trends are not clear. Because dissolved oxygen
levels depend on so many factors, relating dissolved oxygen to nutrient
loadings, freshwater flow, or any individual factor is likely to be unsuccessful.
Nevertheless, Jt is helpful to consider concurrent trends in some of the
possible contributing factors.
Average water temperature calculated with the interpolator, showed a
general increase since the beginning of the Monitoring Program (Figure 11).
The increase was primarily attributable to warmer winter temperatures,
although summer temperatures have increased slightly as well. Average
salinity showed freshening trends of varying degree in most segments (e.g.,
segments CB4 and the total Bay, Figures 12 and 13, respectively). This
finding is surprising, considering the general decrease in freshwater
Susquehanna River flow over the period (Figures 5 and 14). Decreases
occurred in all but the winter seasons and do not seem to be correlated with
changes in seasonal freshwater inflow. Spring rains and melting snow
typically raise Susquehanna flow in March and April. The spring pulses
have decreased in volume over the years since 1984 (Figure 14) and they
have occurred somewhat later in the season in recent years. The flow regime
in the Susquehanna, however, may not reflect flows of the major lower Bay
tributaries.
The depth of the pycnocline is affected by flow and salinity. In segment
CB4, for example, within the summer season of an individual year, the depth
of the pycnocline varied inversely with salinity; that is, the pycnocline was
found deeper in the watercolumn as salinity increased (Figure 15). However,
between years, the average depth of the pycnocline tended to vary directly
17
-------
Cubic Meter Days of
CB8
Other
CB7
CB6
600000-
600000-
400000-
CB4
CBS
Relative Volume of
Water in Bay Segments
CB7
1985 1986 1987 1988 1989 1990
CB1
15
125000-
100000-
75000.
50000-
25000-
0
1985 1986 1987 1988 1989 1990
CB4
1000000
800000-
600000-
400000-
200000-
1985 1986 1987 1988 1989 1990
CBS
150000.
125000-
100000-
75000-
50000-
25000-
0
at a
1985 1986 1987 1988 1989 1990
18
-------
Low Dissolved Oxygen
CB2
CB3
isoooa
125000-
100000-
75000-
50000
25000-
0
1985 1986 1987 1988 1989 1990
CBS
1000000
800000-
600000-
400000-
200000-
1985 1986 1987 1988 1989 1990
3000000
2500000-
2000000-
1500000-
1000000-
500000-
0
Total Bay
1985 1986 1987 1988 1989 1990
150000
1985 1986 1987 1988 1989 1990
CB6
1000000
800000-
600000-
400000-
200000-
1985 1986 1987 1988 1989 1990
Figure 9. Cubic meter days of low dissolved oxygen.
| Below .2
il Below 1
E3 Below 3
H Below 5
19
-------
Dissolved Oxygen Deficit
CB1
CB2
•86 '87 '88 •» '90
WatwYw
•85 -86 -87 -88 '89 '90
WotwYMr
'85 '88 '87 '88 '89 '90
Water YMr
'85 '86 '87 '88 '89 '90
W«Mr YMT
j 400000
300000
-400000
300000-
200000-
p
O 100000-
•87 '88 '89 '90
Water Y«ar
'85 '88 -87 '88 '89 '90
Water YMr
u 20000-
O
0 10000-
•85 '86 '87 '88 '89 '90
WltwYMr
•85 '86 '87 -88 '89 '90
W«t»r Yaar
0
O 100000-
•85 '86 '87 '88 '89 '90
Rgure 10. Total oxygen deficit mass su mmed over each water year. The numbers above the bars indicate the deficit as a percentage
of potential dissolved oxygen mass at saturation.
20
-------
Figure 11. Average temperature in
Chesapeake Bayfrom June 1984 through
September 1990.
Summer Trend
Winter Trend
O Summer
D Winter
• Spring or Fall
Figure 12. Summer salinity in Segment CB4.
Figure 13. Average monthly salinity in
Chesapeake Bay (Trend line computed
.using data from October through
September 1990).
Water Temperature
Oct84 OctSS Oct86 Oct87 Oct88 Oct89 Oct90
Salinity at CB4
25
20 J
I KM
Oct84 Oct85 Oct86 Oct87 Oct 88 Oct89 Oct90
Salinity in Chesapeake Bay
25
20 J
I'
Oct 84 Oct 85 Oct 86 Oct 87 Oct 88 Oct 89 Oct 90
21
-------
Rgure14. Average monthly
Susquehanna River flow from October
1983 through September 1990.
Susquehanna River Monthly Flow
120000-
"i 100000.
§
§ 80000-
Z 60000.
1
O 40000-
20000-
•
A
:=:'
ill!
j:!:l
pi
Jl
i! 11
Sill
I1IM i II f.
II III1
Us !• i!
.:' :: ::
p ¥
M
1
1
A
* ' ...i ::
A iiiii ii
i: ill :l
i ii ii
:: :• : ::.
1: il "
\j
A
:i: iiii
::: M
1! 1 !
i iiii ii .. :
L Ii '!! 1 II
if ii
'
1
III :
:« £
11 I
/Mi H Ii il A"
iAj f i.==!Ml
|( i; jy? ijfgl "I;
'%M **
1984 1985 1986 1987 1988 1989 1990
Water Year
Pycnocline Depth
figure 15. An example of the change in
depth of the pycnocline with changes in
salinty at Segment CB4.
A Pycnocline Depth
• Summer Salinity
H Dissolved Oxygen Deficit
JC 400000-
o
c 300,000-
§
2 200,000-
1
5 100,000-
0-
25-
l20^
X15-
c ;
"3
"10.
5-
„:
A
A
;
•j
A A
A
::i
*
A •
••
A4i
'•'•
jj!i
A
•*•
:::*
..A-
ffi
A
A
•A--
'•'•m-
A .
A
:4:
I*
w
•::
:*:
A; •
*•
-5
-10
-15
-20
-25
-30
1984 1985 1986 1987 1988 1989 1990
with salinity. Lowering the depth of the pycnocline reduces the volume of
subpycnocline water, theoretically lessening the volume of water subject to
oxygen depletion. Since oxygen deficit is the measure which accounts for
the effect of temperature and salinity on oxygen saturation, it possibly could
show the effect of differences in pycnocline depth alone. Oxygen deficit in
segment CB4 does appear to parallel yearly changes in pycnocline depth; the
oxygen deficit is higher when the pycnocline is shallower and vice versa
(Figure 15).
22
CECSH.M1
-------
Production of new organic material (primary production) in the Bay, as
indicated by chlorophyll a, was highest in the spring and early summer most
years (Figure 16). Spring "blooms" were observed in at least a few segments
every year except in 1989. (Either there was no spring bloom that year or
it was missed by the Monitoring Program). Interestingly, 1989 was among
the worst years for dissolved oxygen since the Program's beginning, as
gauged by most measures used in this analysis. A relationship between
primary production and trends in dissolved oxygen is otherwise difficult to
establish in these data. Additionally, because chlorophyll a is highly
variable in both time and space, the error associated with interpolator
estimates is likely to be higher than for other parameters.
What The Results Mean
This volumetric analysis shows that the amount of hypoxic and anoxic water
varies with hydrologic condition. Under the conditions existing since the
Monitoring Program began, anoxic waters have appeared each summer.
This finding was in contrast to the higher levels found in the 1950s under
similar hydrologic conditions (CBP, 1982). Schubel( 1972) examined the
CBI data available at that time and noted that mid-summer levels of
dissolved oxygen below 12 meters were often less than 0.2-0.1 mg/1, yet
complete anoxia (0 mg/1) had not been observed in the main Bay.
The volume of hypoxic and anoxic waters in the Bay is vitally important, as
is the concentration of dissolved oxygen. However, oxygen deficit is a better
measure of trends in dissolved oxygen, since it compensates for the direct
effects of upbay-downbay gradients in temperature and salinity as well as
seasonal and annual differences. The apparent opposite trends in oxygen
deficit in upper and lower Bay segments are particularly interesting.
However, the affected segments (CBI and CB2 in the upper Bay and CBS
in the lower Bay) are small relative to the others (Figure 1) and the
magnitude of change in oxygen deficit is also relatively small.
One question central to the trend analysis is whether the sampling frequency
of the Monitoring Program is sufficient to capture the spatial extent and
duration of hypoxia and anoxia in Chesapeake Bay. This topic is discussed
in more detail in Appendix A and only briefly here.
Experimental data from remote sensing buoys positioned near the Bay
bottom suggest that individual subpycnocline locations can experience
large variations in dissolved oxygen over short periods. This is due largely
to horizontal movement of water masses of different oxygen content,
vertical movement of the pycnocline, or irregular re-aeration events, such as
storms. Nevertheless, comparisonsof such continuous data withMonitoring
Program data from the same region yield similar measures of average
conditions (Figure 2, Appendix A). The extent of the Monitoring Program's
geographic coverage can apparently compensate, in large part, for a lesser
sampling frequency by adequately assessing dissolved oxygen in the spatially
variable but temporally stable water masses of a region. Short-lived changes
which might be caused by storms destratifying a region of the B ay, however,
23
-------
Chlorophyll a
CB8
CB7
CB6
CB4
CBS
Relative Volume of
Water in Bay Segments
70.
60.
i SO-
40.
30-
20-
10.
0.
CB1
Oeta4 Octas octae octa? Octaa Octas octno
70.
CB4
60-
50-
C*
?
M
-*
?
Oc»84 OCI8S Ocl86 Oc»87 OCI88 Od89 OctW
CB7
CBS
60-
40.
M.
20.
10.
0.
oet84 octas ootae oota? octea ootas oetao
60-
f »•
• 40.
| 30.
§ 20-
10.
0.
0 « •*. o
••••.» c°%. »« 4pV °
oct 84 octas octsa oct87 octaa oetag octoo
24
CSCMi.Ml
-------
CB2
CB3
70-
60-
% 50-
I*"'
f
| 30.
S 20-
10.
0.
o
o
o o
0 •
000 0
o> 0° o o f
*° °%<*Jo \> "X**"^* V^o**
n
60.
§ 50-
f 40.
JE
g- 30.
0 20-
10.
0.
0
0
0 0
• .-. . •*. -. .' .
**"*> ° "C- "^IA %«°^ o/°^^o^'^<>ff'S0
o oooo° V o° * o
OctB4 OctSS Oct86 OctB7 Oct88 Oc189 Oct90
Oct84 OctSS Oct86 Oct87 OctSS Oct89 Oct90
CBS
60-
? 5°"
* 40.
J"
| 30.
j2
§ 20-
10.
o
• 0
o
o
0 °°
a
0 » °°
0 /. o *~ °* °o°°oV oV«0° °
•70.
CB6
OctS4 OctSS OCI86 Oct87 OctSS OctSS OCI90
1 5°"'
' 40.
|- 30.
O 20.
10-
0
OctS4 OetBS OctBS Oct87 OctSS Oct89 Oct90
70
60
so.
40
30
20.
10
0
Total Bay
OctB4 OctSS OctS6 Oct67 OctSS OctS9 Oct90
Figure 16. Chlorophyll a in Chesapeake Bay.
25
-------
Conclusions
are likely to be missed by Monitoring Program sampling. Such events have
important, immediate ecological effects, but their effect on trends is likely
negligible. The findings encourage confidence in the use of Monitoring
Program data to estimate long-term trends.
The trends in dissolved oxygen described by the different measures are
consistent with one another, suggesting that the volume of water with low
dissolved oxygen has increased from 1950to 1990. Since 1984, the average
dissolved oxygen concentration decreased slightly over most of the Bay,
with some improvement in upper Bay segments. Oxygen deficit declined
in upper segments, increased in the lower Bay, while the major portion of the
segments exhibited considerable interannual variability with no discernible
trend. The extent and duration of hypoxic and anoxic conditions in the deep
regions of the middle Bay showed no trend. Although significant reductions
in phosphorus levels have been achieved, corresponding improvements in
dissolved oxygen have not yet been seen.
References
Chesapeake Bay Program. 1982. Chesapeake Bay Program Technical
Studies: A Synthesis. EPA Region III, Philadelphia, PA.
Chesapeake Bay Program. 1983. Chesapeake Bay: A Profile of
Environmental Change. EPA Region m, Philadelphia, PA.
Chesapeake Bay Program. 1991a. Trends in Phosphorus in Chesapeake
Bay October 1984-September 1990. Computer Sciences Corporation,
Chesapeake Bay Program, Annapolis, MD. In preparation.
Chesapeake Bay Program. 1991b. Trends in Nitrogen in Chesapeake Bay
October 1984-September 1990. Computer Sciences Corporation,
Chesapeake Bay Program, Annapolis, MD. In preparation.
Chesapeake Bay Program. 199Ic. Chesapeake Bay Dissolved Oxygen
Restoration Goals. CBP/TRS/xx/91 Annapolis, MD. Draft
Clark, I. 1979. Practical Geostatistics.
London, England.
Applied Science Publishers,
Monmonier.M.S. 1982. Computer Assisted Cartography. Prentice-Hall,
Inc., Englewood Cliffs, N.J.
Reynolds, R.C. and L.H. Banner. 1989. A three-dimensional interpolator
for estimating water quality conditions in the Chesapeake Bay:
Description with preliminary application to dissolved oxygen. Computer
Sciences Corporation. Annapolis, MD.
SChubel.J.R. 1972. The physical and chemical conditions of Chesapeake
Bay: An evaluation. Johns Hopkins University, Chesapeake Bay
Institute. Special Report #21, Ref. 72-1.
26
-------
Appendix A
Semi-Continuous Oxygen Records
and Dissolved Oxygen
Measured by the Chesapeake Bay
Water Quality Monitoring Program
27
-------
Introduction
Low levels of dissolved oxygen are a major concern in Chesapeake Bay.
The extent and duration of low dissolved oxygen directly affect the Bay's
living resources, requiring that they be accurately assessed to monitor trends
and progress in the restoration efforts.
Modern remote-sensing devices allow the collection of relatively long-term
semi-continuous or close-interval oxygen data. These records have revealed
unexpected scales of variation—large fluctuations in dissolved oxygen over
days and even hours. Dissolved oxygen is measured routinely in the
Chesapeake Bay Water Quality Monitoring Program; however, these
measurements represent momentary conditions at a single point During the
summer, when sampling is most frequent, measurements are made twice
monthly.
Dissolved oxygen data from semi-continuous datasets were compared to
that collected under the Monitoring Program to evaluate the program's
estimates.
Semi-Continuous Dissolved Oxygen Data
In recent years, several investigators have collected close-interval dissolved
oxygen data in Chesapeake Bay (Mountford et al. 1989; Breitburg 1990;
Sanford et al. 1990; Diaz et al. in press; R. Summers (Maryland Dept. of
Environment); S. Weisberg (Versar)). The Sanford data were best suited for
this comparison because they provided synoptic records at four locations in
the vicinity of Monitoring Program stations (Rgure 1). In addition, these
locations were in the middle Bay where low dissolved oxygen is a chronic
problem.
Buoys bearing the monitoring devices were deployed for 28 days—from
August 12 to September 9,1987. At each location (each differing in total
depth), a sensor was fixed approximately 1 meter off the bottom. At the
westernmost site, a second meter was fixed at approximately mid-depth.
Sanford used Endeco (R) pulsed sensors to record dissolved oxygen,
temperature, and salinity every 5 or 15 minutes, depending on the site. To
standardize the datasets for this comparison, only data at 15-minute intervals
were used. The depths thus monitored were two at 6 meters and one each
at 9,13, and 19 meters. Only data from the mainstem Bay were used; the
6. meter data from the site inside the Choptank River were omitted. Figure
1 shows the time-series records from each site.
28
-------
Remote-Sensing Locations and
Monitoring Program Stations
Rgure 1.
Grey Dots = Remote-Sensing Sites
Black Dots = Monitoring Stations Used inthe Comparison
29
-------
Dissolved Oxygen Data from the CBP
Monitoring Program
Monitoring Program data used in the comparison consisted of measurements
collected at stations in segment CB4 (Figure 1) on routine monitoring
cruises, August 16-17 and September 2-3,1987. These cruises took place
during the same time the Sanford buoys were deployed. In the Monitoring
Program protocol, dissolved oxygen is typically measured at one- to two-
meter intervals between the surface and bottom at each station.
Comparison of the Datasets
The Monitoring Program dissolved oxygen data from both cruises were
plotted to display the distribution of concentrations by depth (Figure 2). The
dissolved oxygen measurements at depths comparable to the depths of the
four buoy records were grouped separately and the mean, standard de-
viation, minimum, and maximum concentrations at each depth were
determined (Table 1). Comparable depths in the Monitoring Program data
were defined as the region one meter above to one meter below the fixed
depths of the four buoys. The mean, standard deviation, minimum, and
maximum concentrations of the buoy records were overlayed on the plot
(Figure 2) and tabulated (Table 1) for comparison.
Figure 2. Comparison of monitoring data
and buoy data.
-5-
-10-
-25-
* * *****
* * * ** *
_jk n-tft *
* ***** * ** **
***** *** **
#t* * **** **
*** ***** **
** *******
3pfc HC'Mt'^Ciytfs
#* **** *
** *****
** * ****
* ***
* *********** ***** * *
* * ********)***** *
* ** ********* *********
* * * ****************** *
* * * *t****>***^****** *
******* ** ***
** ** **$t* ** * *
A * ***** *
-9 * *:
** *
* *
*
0.0
2.0
4.0 6.0 8.0
Dissolved Oxygen (mg/l)
10.0
****: Monitoring Data
-1 SD
+1 SD
Buoy Data
30
CSCSA1.M1
-------
The statistics for the semi-continuous data at 6, 9 and 19 meters closely
matched those for comparable depths at the Monitoring Program stations.
Minima and maxima of the semi-continuous data in most cases exceeded
those of the Monitoring Program data. The standard deviations were also of
approximately the same magnitude (Table 1).
Table 1. Comparison of continuous and Monitoring Program dissolved oxygen datasets.
Depth
6m
9m
13m
19m
Number of
Observations
CON MON
2672 47
2672 39
2659 30
2671 17
Average
D.O.
CON MON
6.7 6.3
3.8 4.2
4.7 1.7
0.6 0.8
Standard
Deviation
CON MON
1.0 1.5
2.0 1.5
1.7 1.4
0.6 . 0.5
Minimum
D.O.
CON MON
1.9 2.0
0.0 1.4
0.4 0.1
0.0 0.1
Maximum
D.O.
CON MON
9.7 8.6
8.1 8.3
8.1 6.3
5.5 1.6
CON = continuous monitoring data
MON = Monitoring Program dissolved oxygen data
The data from the sensor moored on the eastern side of the Bay at 13 meters
didnothave the same characteristics as the 13-meter data from the Monitoring
Program. The bottom at the eastern location apparently experienced more
aeration (i.e., the mean dissolved oxygen concentration of the continuous
data was considerably higher than the mean from the Monitoring Program
at that depth). The 13-meter depth is approximately the depth of the
pycnocline in that regionof the Bay and may alternately show characteristics
of above or below pycnocline environment. The 13-meter data from the
Monitoring Program are more like those from below the pycnocline; the data
from the continuous monitor may be more like areas above the pycnocline.
Discussion
In the middle region of the Bay, where these data were collected, the average
depth of the pycnocline is usually between 6 and 12 meters during the
summer. Areas in the shallow littoral zone where the bottom deepens to 6
meters or greater are subject to the intrusion of highersalinity, subpycnocline
waters from the deep channel. Waters below the pycnocline typically have
little or no dissolved oxygen for much of the summer. Movement of these
oxygen-depleted bottom waters into shallower areas occurs when wind-
forcing of surface water causes the pycnocline to tilt Such episodes may last
from a few hours to several days. The sharp spikes and troughs in the semi-
continuous records shown in Figure 1 are signsof these ephemeral intrusions.
31
-------
Breitburg (1990) addressed the problem of dissolved oxygen variablity as
it relates to the monitoring and assessment of nearshore bentnic habitats.
She found that measurements taken at the frequency of the Monitoring
Program were unlikely to adequately assess low dissolved oxygen exposures
at particular sites. However, the comparisons above suggest that the
Monitoring Program data do provide an adequate assessment of bottom
dissolved oxygen conditions on a regional basis. The spatial coverage of the
MonitoringProgramstationnetwork in this portion of the Bay can apparently
compensate for a lesser sampling frequency. The various water masses of
differing dissolved oxygen content—the source of deepwater variability at
individual sites—-are represented within the sampling station desiga The
extreme conditions within a region to which a particular site may be exposed
are identified for the most part However, the minima and maxima of the
continuous data—the result of short-lived events—exceeded the minima
and maxima of the Monitoring Program data.
Semi-continuous records from different locations in the Bay and different
time periods have been acquired recently and further analysis is underway.
In the meantime, the comparison with the semi-continuous data has
encouraged confidence in using Monitoring Program data to look at status
and trends in dissolved oxygen regionally and Baywide.
References
Breitburg, D. 1990. Monitoring dissolved oxygen in a variable
environment: Detecting biologically meaningful events. Presented at
the Chesapeake Research Conference, "New Perspectives in the Chesapeake
System." December 4-6, 1990. Baltimore, MD.
Diaz, R., J. Neubauer, L. Schaffher, A. Piehl, S. Baden. March, 1990.
Continuous monitoring of dissolved oxygen in an estuary experiencing
periodic hypoxia and the effects of hypoxia on macrobenthos and fish.
Proceedings of the Symposium on Coastal Eutrophication, Bologna,
Italy. In press.
Mountford,K.,R. Reynolds,N. Fisher. 1989. A telemetering environmental
monitoring data buoy (in) Chesapeake Bay. Proceedings: Conference
on Marine Data Systems. Marine Technology Society, New Orleans, LA.
Sanford, L.t K. Sellner, D. Breitburg. 1990. Covariability of dissolved
oxygen with physical processes in the summertime Chesapeake Bay. J.
Mar. Res. 48(3) 567-590.
Weisberg, S. B.t J. K. Summers, A. F. Holland, J. Kou, V. D. Engle, D.
L. Breitburg, R. J. Diaz. Characterizing dissolved oxygen conditions in
estuarine environments. In preparation..
32
------- |