Upper Mississippi River Water Quality Assessment,  March 2002
Upper Mississippi River Conservation Committee, Water Quality Technical Section
(Click here to view the entire report)

                                                       Summary    
Upper Mississippi River (UMR) water quality data were compiled from federal, state and
local agencies that conducted monitoring on the river over the past two decades (1980-
1999). The major objectives of this effort are to increase coordination and cooperation
among monitoring agencies, develop a unified database of relevant water quality
information, and to use these data to produce a systemic assessment of the water quality of
the UMR. This effort is particularly important for the Mississippi River, which forms the
boundary with five states and is monitored and managed by many federal, state, and local
resource agencies. The river reach for this evaluation extends from Anoka, Minnesota
(just upstream of the Twin Cities) to the Ohio River, a distance of 872 river miles(RM).

Two databases were compiled. The first database includes field and laboratory inorganic
chemistry data from samples collected near or in the main channel of the river. The
primary focus of the assessment is on summer (June 1 to September 15) water quality data
collected over a twenty-year period (1980 to 1999), which resulted in the creation of a
summer data subset of the entire compiled database (universe). The second database
includes fish contaminant data on polychlorinated biphenyls (PCBs), chlordane, and
mercury collected throughout the UMR. These later data were primarily obtained by
agencies responsible for providing fish consumption advice for sport anglers on the
Mississippi River.

Summer water quality data were provided by six agencies and included more than 5,800
records for the twenty-year summary period. Monitoring was not equally distributed
throughout the study reach and tended to be concentrated in certain reaches, especially in
the upper river reach from Anoka, Minnesota (RM 871.6) to Minneiska, Minnesota (RM
738), where three agencies have been sampling the river for many years. Since the late
1980s, the U.S. Geological Survey's Long Term Resource Monitoring Program (LTRMP)
has been a major provider of water quality data, and the LTRMP information represents a
substantial portion (46%) of the summer data.

Temporal and spatial (longitudinal) evaluation of 11 water quality parameters was
conducted by plotting the entire summer data set by river mile over four, 5-year intervals.
In addition, data were segmented into 15 river reaches by utilizing hydrologic unit codes
defined by USGS. Boxplots were prepared of the segmented data to broadly compare
water quality differences over different reaches of the river.

No attempts were made to assess potential field or laboratory method differences between
the monitoring agencies and their laboratories. As a result, the interpretation of the
compiled water quality or fish contaminant data must be treated with some caution.
However, several parameters exhibit distinct longitudinal profiles and changes over time
that likely overshadow any potential field or laboratory bias.

The average summer river flow increases substantially along the north-south longitudinal
gradient, from 30,000 cubic feet per second (cfs) at Winona, Minnesota to more thanx
200,000 cfs at Thebes, Illinois. This flow increase parallels the large increase in drainage
area over this reach.

River temperature is influenced by climatic conditions, and the river temperature
generally increased about 5 °C from north to south during the summer period.

Nonpoint source inputs from tributary streams, major point source discharges, and river
flows are the dominant factors influencing the observed longitudinal water quality
patterns. This was especially apparent in Pool 2, where the river flow is relatively low
and nonpoint source pollution from the Minnesota River and wastewater discharges from
the Twin Cities Metropolitan Area have a strong influence on the river's quality. Large
changes in the river's quality are also observed in the lower portion of the UMR, where
nonpoint source pollution from large agricultural watersheds, including the Missouri
River, contributes to high nutrient and suspended solids concentrations.

Point source pollutant abatement activities implemented in the 1980s have resulted in
noticeable reductions in total and un-ionized ammonia nitrogen concentrations and
increases in dissolved oxygen (DO) concentrations below the Twin Cities Metropolitan
Area. Widespread infestations of zebra mussels in the river reach extending from Pool 9
(RM 648) to Pool 14 (RM 494) in the late 1990s are believed to have had some influence
on water quality during some summers, and may partly explain the lower DO
concentrations reported during this period. Nitrite+nitrate nitrogen concentrations
throughout the river increased to higher levels in the 1990s, compared to concentrations
observed during 1985-89. For the upper river, this response may have been partly
associated with changes in municipal wastewater treatment technology (nitrification).
However, changes in precipitation and river flow are additional factors associated with
river-wide increases in nitrite+nitrate nitrogen concentrations. The drought conditions of
the late 1980s reduced nonpoint source runoff and increased utilization of inorganic
nitrogen within the riverine pools. Increased nonpoint source runoff in the 1990s likely
favored mobilization of nitrite+nitrate nitrogen from agricultural watersheds, resulting in
high nitrogen concentrations in the river during this period.

Fish contaminant data were compiled from six agencies and one industrial source and
included the river reach from Anoka, Minnesota to Memphis, Tennessee. For the reach
from Anoka to the Ohio River, a total of 3,647 records was obtained. The fish tissue data
summarized in this assessment were obtained from skin-on or skin-off fillets, which
comprised more than 80 percent of the data. Most of the data were confined to three
separate reaches or areas. These include Pools 2 through 10, Pool 15, and the open river
reach. The Minnesota Department of Health and the Wisconsin Department of Natural
Resources data account for two-thirds of the total fish tissue data, resulting in a large
portion of samples from these states' border waters. Carp were the most frequently
sampled fish species. Due to their high lipid content, carp are a suitable fish for assessing
contamination of PCBs and chlordane, which accumulate in fatty tissue. An evaluation of
mercury contamination was limited to fillet samples from channel catfish, walleye, andxi
white bass, since these were the most frequently analyzed fish, as well as the species
which typically exhibited higher mercury concentrations.

Median PCB concentrations in fish fillets were greatest in the upper reach of the UMR,
with the highest levels reported in the Pool 2 to 4 reach in the early 1980s. PCB sources
in this river reach are generally attributed to diffuse inputs from the Twin Cities
Metropolitan Area. Intensive sampling at a local PCB source (Alcoa, Inc.) in lower Pool
15 in the late 1980s and early 1990s resulted in elevated PCB tissue concentrations in Pool
15 that are likely not representative of this entire reach. Fish tissue PCB concentrations
have decreased noticeably river-wide from the early 1980s to the 1990s. These reduced
PCB tissue concentrations likely reflect use restrictions, reduced point source
contributions, and reduced nonpoint source inputs associated with soil or sediment
cleanup activities.

In contrast to PCBs, median chlordane concentrations were highest in carp fillets obtained
from the lower reaches of the UMR. Concentrations reported in the upper reaches were
often near or below the reported detection level. This spatial trend is likely the result of
greater chlordane use in states bordering the lower river. Chlordane concentrations in
carp fillets appear to be decreasing with time, likely as a result of use restrictions and
decreased inputs.

A systemic assessment of mercury contamination in UMR fish was more difficult due to
fewer samples. Most of the mercury data were available for the river reaches bordering
Minnesota and Wisconsin. Median mercury concentrations in channel catfish and white
bass fillets were greater in the upper portion of the river (Pools 2 to 6) as compared to
samples collected below Pool 14. Compared to fish samples collected nationally (Bahnick
et.al., 1994), mercury concentrations in channel catfish fillets from the UMR were slightly
higher than the national average. Walleye were a frequently sampled species in the upper
river, and yielded average mercury fillet concentrations about three-fold lower than the
national average. Based on walleye fillet data collected between the late 1980s and 1998
in Pools 2 to 9, median mercury concentrations appear to be exhibiting a declining trend,
consistent with reduced mercury inputs as documented by recent sediment coring studies
of Lake Pepin (Balogh et.al., 1999).