Comparison of Temporal Trends in Ambient and Compliance Trace Element and PCB Data in Pool 2 of the Mississippi River, 1985-95

Jesse Anderson, Biologist

U.S. Geological Survey, 2280 Woodale Drive, Mounds View, MN 55112

Jim Perry, Professor, Department of Forest Resources

University of Minnesota, 1530 North Cleveland, St. Paul, MN 55108

 

Abstract

One goal of the Intergovernmental Task Force on Monitoring is to integrate the efforts of agencies that collect ambient and compliance water-quality data. The similarity in temporal trends between retrospective ambient (in-stream) and compliance (wastewater) water-quality data collected from Pool 2 of the Mississippi River was determined for 1985-95. Constituents studied included the following trace elements: arsenic (As), cadmium (Cd), chromium (Cr), hexavalent chromium (Cr+6), copper (Cu), lead (Pb), mercury (Hg), nickel (Ni), selenium (Se), zinc (Zn), and polychlorinated biphenyls (PCBs). Water-column, bed-sediment, and fish-tissue (fillets) data collected by five government agencies comprised the ambient data set; effluent data from five registered facilities comprised the compliance data set. The non-parametric Mann-Kendall trend test indicated that 33% of temporal trends in all data were statistically significant (p<0.05). Possible reasons for this were low sample sizes, and a high percentage of samples below the analytical detection limit. Seven trace elements (Cr, Cd, Cu, Pb, Hg, Ni, and Zn) had statistically significant decreases in wastewater and portions of either or both ambient water and bed sediment. No trends were found in fish tissue. Inconsistency in trends between ambient and compliance data were often found for individual constituents, making overall similarity between the data sets difficult to determine. Logistical differences in monitoring programs, such as varying field and laboratory methods among agencies, made it difficult to assess ambient temporal trends. Trends in compliance data were more distinct; most trace elements decreased significantly, probably due to improvements in wastewater treatment.

Introduction

Ambient (instream) and compliance (wastewater) water-quality monitoring have been conducted in separate, often unrelated programs for different purposes (U.S. Geological Survey, 1995). Comparison of these data may be difficult because of varying field, analytical, and laboratory methodologies and precision among the collecting agencies (Powell, 1995; Knopman and Smith, 1993; Ward et al., 1990). In addition, interpretation of data from multiple agencies can be difficult because of variations in sampling frequencies, sampling locations, uncertainty in measurement, multiple or changing censoring levels, and outliers (Ward et al., 1990). These limitations can be especially pronounced when attempting to describe trends in water-quality (Powell, 1995; Knopman and Smith, 1993).

The Intergovernmental Task Force on Monitoring (ITFM) was formed in 1992 to review and evaluate water-quality monitoring activities nationwide and to recommend improvements (U.S. Geological Survey, 1995). One recommended improvement is the linkage of ambient and compliance monitoring. The ITFM suggests forming partnerships among ambient and compliance monitors; this cooperation can lead to more cost-effective ways of protecting the environment. Often it is necessary to understand contaminant loading effects on ambient conditions as well as the effects of ambient characteristics on regulatory decisions and water uses (U.S. Geological Survey, 1995). The ITFM suggests studies to determine whether ambient or compliance sampling can be reduced locally or nationally. The amount of reduction would depend on the degree of similarity between the ambient and compliance data sets. If patterns in ambient and compliance water-quality data are similar, such as trends over time or direct relationships, it may be concluded that the sampling frequency of ambient and/or compliance monitoring may be reduced, saving the public and private sector considerable resources; or it may allow the water-quality monitoring community to make better decisions with the same resources. Trend comparisons are important water-quality evaluations. Numerous data have been collected in the past but more evaluations of data would help the water management community avoid the "data rich but information poor" dilemma (Ward et al., 1990).

To test the strategy of integrating ambient and compliance water-quality data, ITFM has encouraged pilot studies in selected watersheds in the United States. This paper describes one such effort as part of the U.S. Geological Survey’s (USGS) National Water-Quality Assessment (NAWQA) Program. As part of the Upper Mississippi River Basin NAWQA study, an ITFM pilot study was conducted that assessed the temporal trends in ambient and compliance trace element and polychlorinated biphenyl (PCB) data in Pool 2 of the Mississippi River (figure 1). This study assessed the similarity in temporal trends between ambient and compliance trace element and PCB data in water, bed sediment, and fish tissue from 1985-95. This time period was chosen because: 1) compliance data from the Permit Compliance System (PCS) data base are available starting in 1985 (Mary Kimlinger, Minnesota Pollution Control Agency, oral comm., 1996), 2) many years of data are needed to effectively detect statistically significant trends in water-quality data sets (Gary Oehlert, Department of Applied Statistics, University of Minnesota, oral comm., 1996), and 3) some ambient trace element and PCB water-quality samples prior to 1985 may have been contaminated, resulting in data unsuitable for trend analysis.

The constituents addressed by this study were arsenic (As), cadmium (Cd), chromium (Cr), hexavalent chromium (Cr+6 ), copper (Cu), lead (Pb), mercury (Hg), nickel (Ni), selenium (Se), zinc (Zn), and PCBs. All have been classified as "category 1" priority pollutants because they are persistent in the aquatic environment and may bioaccumulate or enter food chains (Chapman et al., 1982). Trace elements and PCBs are ubiquitous in the modern industrial environment (Taylor and Shiller, 1995). Discharges from wastewater and industrial activities cause increases in heavy metal concentrations in receiving waters near major urban areas (Meade, 1995). Point-source discharges from industrial facilities and municipal wastewater treatment plants are largely responsible for many of the toxic elements in the Upper Mississippi River (UMR) system (Upper Mississippi River Water Quality Initiative, 1993). Despite being banned in 1979, PCBs are still relatively widespread in the aquatic environment (Sullivan, 1988; Eisler, 1986).

Description of Study Area

The Minneapolis-St. Paul metropolitan area (TCMA) is bisected by the Mississippi River. It is one of the largest population centers in the UMR Basin, with an estimated population of 2.3 million people (Stark et al., 1996). The Mississippi River is an integral resource to the TCMA. It provides a public water supply, electrical power, commercial transportation, wastewater dilution, a diverse fishery, and recreational and aesthetic value. The UMR, the length of river between Minneapolis, Minnesota and St. Louis, Missouri, contains a series of 29 locks and dams operated by the U.S. Army Corps of Engineers. These locks and dams create ‘pools’, defined as the stretch of river between two adjacent locks and dams (Wiener et al., 1984). Pool 2 of the Mississippi River extends from St. Paul, Minnesota to just upstream of Hastings, Minnesota (Figure 1). This pool is located in the TCMA, and receives a majority of the major industrial and municipal discharges to the UMR in the TCMA.

Ambient Monitoring in the Study Area

Ambient monitoring describes instream water-quality conditions. Ambient monitoring can include sampling for physical and biological parameters, in addition to chemical concentrations in water or bed sediment samples. In the study area, ambient monitoring is conducted less frequently than compliance monitoring (typically, quarterly or annually versus monthly for compliance monitoring). Ambient water-quality samples typically are collected by Federal, state, and local government agencies. In the TCMA, these agencies include the Metropolitan Council Environmental Services (MCES), Minnesota Pollution Control Agency (MPCA), Minnesota Department of Natural Resources (DNR), U.S. Army Corps of Engineers (COE), and the USGS. Data are stored in data bases such as the USGS National Water Information System (NWIS), and the U.S. Environmental Protection Agency (USEPA) Storage and Retrieval (STORET). Constituents measured and frequency of monitoring vary among agencies depending on the scope of the monitoring program. As a result, monitoring programs are often not coordinated among agencies.

Compliance Monitoring in the Study Area

The Federal Water Pollution Control Act of 1972 (in later laws renamed The Clean Water Act) established the National Pollutant Discharge Elimination System (NPDES) to regulate and monitor sources of wastewater to the Nation’s waterways. Permits from the USEPA for discharge of contaminants into National waters are required as part of the NPDES program; these permits require monitoring of influent to and effluent from these facilities because municipalities and industries must have their wastewater effluent comply with Federal or state standards (Tyler, 1992). Permitted facilities send monthly Discharge Monitoring Reports (DMR) to a supervisory government agency; the DMR describes the quality of the wastewater effluent. These data are then stored in the USEPA PCS data base. DMR data serve the following management functions: 1) track permit issuance and reissuance, 2) identify effluent limits and report violations, 3) determine compliance statistics at a state or national level, 4) track enforcement actions and the resulting compliance schedules and interim limits, and 5) respond to requests for information from Congress, state legislatures, and the general public (Mary Kimlinger, Minnesota Pollution Control Agency, written comm., 1997).

The major facilities (those which discharge more than 1 million gallons per day) included in this study, which discharge treated wastewater into Pool 2, are the Metropolitan Wastewater Treatment Plant, Cottage Grove Wastewater Treatment Plant, Koch Refinery, Ashland Oil Company, and 3M Chemolite.

Methods of Evaluation

Retrospective ambient and compliance water-quality data, collected from 1985-95 in Pool 2 of the Mississippi River, were obtained from various government agencies and industrial and municipal facilities for the constituents of concern. To enhance the evaluation of ambient water-quality in the study area, data were complied for three media: the water column (both filtered and unfiltered samples), bed sediments, and fish tissue (fillets). Water-quality data were complied from the USGS and the MCES. Surficial bed-sediment data were complied from the MCES and COE. Fish fillet contaminant data were complied from the DNR and MCES for two species, common carp (Cyprinus carpio) and smallmouth bass (Micropterus dolomieui). These two species were chosen because they are commonly found in the study area, they have different diet and foraging patterns, and have dissimilar lipid contents in their tissues, which is a factor in the bioaccumulation of contaminants.

The Mann-Kendall trend test (Mann, 1945; Kendall, 1975) was used to determine temporal trends in the ambient and compliance data sets. The Mann-Kendall trend test can be stated most generally as a test of whether Y (constituent concentration) values increase or decrease with time (Helsel and Hirsch, 1992). This nonparametric test has several advantages: 1) the data do not need to conform to a prespecified distribution; 2) missing data are allowed in the data set if they are missing at random; and 3) data values reported as ‘less than the detection limit’ can be used by assigning them a common value smaller than the smallest measured value in the data set (Gilbert, 1987).

All ‘below detection limit’ observations were assigned a common value smaller than the smallest detected value in the data set. In the event that multiple detection limits were present in a data set, all nondetect values were assigned the highest detection limit.

Compliance water-quality data were reported in daily loads, often in kilograms of constituent discharged per day. These data were converted to a concentration (mass of constituent per volume of water), because ambient water data were expressed as concentrations, either milligrams per liter (mg/L), or micrograms per liter (µg/L). Therefore, the Mann-Kendall trend test was computed on constituent concentrations so direct comparisons between ambient and compliance data could be made. ‘Below detection limit’ measurements in a compliance data set were often reported as a load or concentration of zero. These data were considered as ‘below detection limit’ measurements in the evaluation.

A significance level of 0.05 was chosen for this evaluation. The graphing and data analysis program Plotit®* (Scientific Programming Enterprises, 1994) was used to calculate Kendall’s tau and its associated p-value for all compiled data. The null hypothesis was that constituent concentrations did not change significantly through time (i.e., no trend was evident).

Results

Selected ambient and compliance trace element and PCB data collected from Pool 2 of the Mississippi River from 1985-95 were statistically analyzed. Analyses using the Mann-Kendall trend test showed that there was a statistically significant trend (p<0.05, ) in 33% of all tests (considering all constituents, all media, and both ambient and compliance data).

Only 25% of all possible trends in ambient data were statistically significant. Statistically significant trends were detected in ambient bed sediment for all constituents analyzed except PCBs, Cd, and Se. Statistically significant decreasing trends were detected for the following constituents: Arsenic, Cr, Cu, Pb, Hg, Ni, and Zn. One statistically significant increasing trend was detected: Cu in MCES data for ambient bed sediment. Often in the ambient bed-sediment data set, for a given constituent, a significant decrease was found in COE data, whereas no trend was detected in MCES data. In unfiltered ambient water samples, statistically significant increasing trends were found for Arsenic, Hg, and Se; statistically significant decreasing trends were detected in Cd, Cr, Cu, Pb, and Ni. Zn, Cr+6, and PCBs were the only constituents for which statistically significant trends were not detected in unfiltered samples. One statistically significant trend was found in filtered water, a decrease in Ni concentration. Fish tissue contaminant data showed no statistically significant trends in smallmouth bass or carp fillets for PCBs or any trace element.

More statistically significant trends were evident in the compliance data set (56% of tests). Statistically significant decreasing trends were found for Cr, Hg, Se, and Zn in wastewater from more than one facility. All statistically significant trends in this data set were decreasing, no statistically significant increasing trends were detected. The effluent from 3M Chemolite had statistically significant decreases in Hg, Se, and Zn. Statistically significant decreases in Cr, Cr+6, and Se were found in the wastewater from Ashland Oil. The effluent from Koch Refinery had a statistically significant decrease in Cr over the period of record. No statistically significant trends were detected in effluent data collected at the Cottage Grove Wastewater Treatment Plant. The wastewater effluent data from the Metropolitan Wastewater Treatment Plant had statistically significant decreases in all trace elements monitored except As and Cr+6. The effluent also had a statistically significant decrease in total PCBs, the only detectable trend in all analyzed PCB data.

The principal goals of this study were to determine if there were statistically significant temporal trends in ambient and compliance trace element and PCB data, and if any temporal trends in these data sets were statistically similar. For no single constituent were trends statistically and directionally similar among all the ambient and compliance data sets. There were statistically significant decreases in the concentrations of nine constituents in the compliance data set. Seven of these constituents (Cr, Cd, Cu, Pb, Hg, Ni, and Zn) had similar statistically significant decreases in either or both ambient water and bed sediment.

Discussion

Ambient Water Quality in Pool 2 of the Mississippi River

Only 25% of all possible trends in ambient water quality data were statistically significant. Three possible reasons for this were: low sample size, the high percentage of ‘below detection limit’ measurements in the data sets, or there were no temporal trends in the majority of the data.

One factor that affects statistical tests is the number of data points analyzed (sample size). As sample size increases, a more accurate estimation of the true environmental condition emerges. When sample size is low, differences may be real but masked by variance. In data analyzed for this study, sample size was low (<30) for several constituents. This was most common in the ambient data set, especially in the fish-tissue data, for which no trends were evident.

A high (>50) percentage of samples analyzed were below the detection limit. This also may have been a factor in the instance of few detectable water-quality trends in the ambient data set. The occurrence of values below the detection limit in environmental data sets is a major statistical complication (Newman et al., 1989; Helsel and Cohn, 1988), and presents serious interpretation problems for data analysts (Helsel, 1990). The constituents studied for this analysis were trace substances in the aquatic environment. More samples ‘below detection limit’ were found in filtered water samples. In these samples, water was filtered before analysis, which removed particulate matter and its associated trace elements and PCBs. This process may help explain why only one statistically significant trend was found in filtered ambient water. Additionally, the majority of PCB samples compiled for this study were ‘below detection limit’. During 1985-95, PCBs were detected in 1% of ambient water samples and 6% of ambient bed-sediment samples. The only detectable temporal trend in PCBs was in the effluent from the Metropolitan Wastewater Treatment Plant. This data set had 52% of samples determined to be ‘below detection limit’; however, there were no detected concentrations from August 1992 to December 1995. This pattern of nondetection values in the latter portion of the data set is probably what led to the statistically significant decrease.

An Example of Variability in Ambient Water Quality in Pool 2

The disagreement in trends among the ambient media was possibly due to a combination of logistical differences in monitoring programs and natural variation in the river. Cu concentrations in ambient bed sediment provide an example. From 1985 to 1995, COE data indicated a statistically significant decrease, while MCES data indicated an increase. Data from these agencies are collected for different reasons. The MCES has established sampling locations in the study area, while COE sampling locations vary with navigation channel dredging activities. Additionally, during the period of record studied, the COE started sampling bed sediment in 1988, this same year the MCES stopped bed-sediment sampling. COE’s bed-sediment monitoring program was not designed to detect spatial and temporal trends on a pool-wide basis (Dan Wilcox, U.S. Army Corps of Engineers, oral comm., 1997). Trends in these data should not be expanded to make broad generalizations about the bed sediment quality of Pool 2. MCES bed sediment data collection ended in 1988. Trends in these data should not be expanded to describe current trends in the study area, or compared to COE data. Variability in trends in these data illustrate the difficulty in compiling data from different sources in an attempt to determine holistic trends in Pool 2 water quality.

Compliance Water Quality in Pool 2 of the Mississippi River

Results from the compliance data were more statistically distinct than those from ambient data. A total of 56% of all possible trends in the compliance data set were statistically significant, and all were decreasing. Trends may have decreased because of improvements in wastewater treatment. Four of the five major compliance facilities in the study area (3M Chemolite, Ashland Oil, the Metropolitan Wastewater Treatment Plant, and Koch Refinery) made improvements in wastewater treatment during 1985-95. The 3M Chemolite plant upgraded its pH control system for better precipitation of zinc (Tom Baltutis, 3M Chemolite, oral communication, 1997). Ashland Oil upgraded its wastewater treatment facility in 1994 (Alan Mayo, Ashland Oil, oral communication, 1997). Koch Refinery has removed all chromium from its cooling towers, and has decreased its usage of mercury instrumentation (Heather Faragher, Koch Refinery, oral communication, 1997).

The MCES Industrial Pretreatment Program is another example of an improvement in wastewater treatment. The load of trace elements to the UMR from the Metropolitan Wastewater Treatment Plant has been substantially reduced in recent years because of the MCES Industrial Pretreatment Program. This program controls and regulates facilities that discharge their effluent to the sewer system to ensure compliance with local and Federal regulations (Leo Hermes, Metropolitan Council Environmental Services, written comm., 1997). The load of trace elements to the river has been reduced by about 91,000 kg annually (Leo Hermes, Metropolitan Council Environmental Services, oral comm., 1997).

Comparisons Between Trends in Ambient and Compliance Water-Quality Data in Pool 2

In the majority of constituents studied here (7 of 11; Cd, Cr, Cu, Pb, Hg, Ni, and Zn), there were statistically significant decreasing trends in parts of both the ambient and compliance data sets. For these 7 constituents, there were statistically significant decreases in either ambient water, bed sediment, or both, and a parallel decrease in a portion of the compliance data set. The majority of trends in the compliance data set were significantly decreasing. In the ambient data sets with larger sample sizes, results were similar to those in wastewater (i.e., statistically significant decreases). For instance, COE bed-sediment data had statistically significant decreases in 7 of 9 trace elements. Similarly, unfiltered-water data had statistically significant trends in the majority of constituents, and most statistically significant trends were decreasing.

This retrospective analysis indicates the difficulty in both the interpretation and detection of definitive water-quality trends. No conclusive trend was the most common result in this analysis, and disagreements in the statistical significance of trends in data collected by the various agencies or facilities were often found (Table 1). Most trends in the ambient data set were not statistically significant, whereas a majority of trends in the compliance data set were statistically decreasing (Table 1). On a broad scale, this fact shows that the trends in the data sets were dissimilar. The fact that there was no statistical significance among all ambient and compliance data for any single constituent reinforces this point (Table 1).

The numerous statistically significant decreases of constituents in wastewater (due to improvements in wastewater treatment) apparently have not yet resulted in significant changes in ambient water, bed sediment, and fish fillets. The dilution effects of the Mississippi River may be responsible for the lack of trends in ambient water, and therefore the lack of similarity between ambient and compliance water-quality trends. It is possible that, in time, the reductions in wastewater loads will also be seen in bed sediment and fish fillets.

This study indicated that trends between the ambient and compliance data sets were generally different. Part of this may be due to the attributes of the ambient monitoring programs that collected trace element and PCB data in Pool 2 of the Mississippi River from 1985-95. Many of these monitoring programs contain data that may not be suitable for trend analysis in the study area. For example, COE data was collected in the context of dredging operations. The fish fillet data were few and were not well distributed over the 11-year period. The filtered-water data sets also were small. The largest ambient data set (MCES unfiltered-water data) had data which was well distributed over the period of record, but the samples were taken quarterly, often at several locations on the same day.

Continued ambient trace element and PCB monitoring, free of the shortcomings discussed previously, will be useful for determining trends in ambient concentrations in Pool 2. A properly designed, holistic monitoring program can help the water-management community better determine if, or how, ambient and compliance monitoring data can be integrated.

National Implications of Ambient and Compliance Data Integration

Results from Pool 2 point to the importance of effective monitoring integration, especially if certain analysis of the data (trends) are desired. The results from this study have implications beyond Pool 2 of the Mississippi River. For integration among the data sets to be successful, the amount and usefulness of ambient and compliance data should improve.

Collaboration is important because few single organizations can collect all the information needed for informed decision making (U.S. Geological Survey, 1995). Data sharing can only be accomplished if ambient and compliance data are readily available. Many Federal and state agencies store their water-quality data in agency-specific data systems that other agencies cannot easily access or in files that are not yet automated (Powell, 1995). Much of the compliance and ambient data generated by the regulated community are unavailable for other uses because of differing designs and goals in collecting the data and also because no one has asked for them in a systematic way beyond their narrow compliance context (U.S. Geological Survey, 1995). Before ambient or compliance data can be shared and integrated, they should be readily available.

Data produced by existing monitoring programs do not always meet existing needs, such as determining ambient water-quality status and trends (Powell, 1995). If the ambient monitoring community (Federal, state, and local government agencies) cannot meet current needs, the water-quality data collected by compliance monitors may assist ambient monitoring efforts. Public access to these PCS data is provided through the Freedom of Information Act (U.S. Environmental Protection Agency, 1993). These data have implications beyond the immediate purpose of the compliance data. The quality of a wastewater effluent can have profound consequences for the ambient conditions in a stream because wastewater generally has much higher concentrations of dissolved constituents. These constituents can be dispersed into the water column, assimilated into the bed sediment and biota, and possibly degrade the ambient water quality. It is important to learn the amount and quality of wastewater being discharged to an ambient receiving water. After reviewing these data it can be determined if, or how, they can contribute to ambient monitoring efforts for determining status and trends. Accordingly, the use of compliance water-quality data can be very beneficial in ambient monitoring efforts.

Ambient water-quality data from various local, state and Federal agencies can help the monitoring community holistically determine the quality of our waters. Compiling ambient data from all available sources can identify impediments to both the integration of ambient data and the possible integration with compliance data. These data could reinforce or contradict established ambient trends, and put new results in context with historical data, therefore facilitating better water-quality management.

Improved data-sharing techniques can lead to another benefit: better communication among those who monitor water quality on all levels. While gathering retrospective data from a particular organization, ancillary information describing the scope of the monitoring program can be collected. Examples of ancillary information include: specific reasons why the data were collected, sampling locations, laboratory and field methodologies, quality-control and quality-assurance data, and planned analyses of the data. This information can put the water-quality data in context with the data collected from other sources. Additionally, these data may ultimately lead to reduced sampling duplication by those agencies who monitor ambient water quality.

Acknowledgments

This effort was part of the Upper Mississippi River Basin study. Thanks to Jim Stark, Ginger Amos, Dave Lorenz, George Garklavs, Keith Robinson, Paul Capel, Kathy Lee, James Fallon, and Paul Hanson of the U.S. Geological Survey and Gary Oehlert of the University of Minnesota for reviewing this paper and providing technical assistance. Also, thanks to the following individuals for providing data from their respective agencies/facilities: Mary Kimlinger, Mary Dzerik, Linda Nelson, and Eudale Mathiason at the Minnesota Pollution Control Agency; Mark Briggs at the Minnesota Department of Natural Resources; Dennis Anderson at the U.S. Army Corps of Engineers; Jahna Lindquist at the Metropolitan Council Environmental Services; Tom Baltutis at 3M Chemolite; Alan Mayo at Ashland Oil; and Heather Faragher at Koch Refinery.

Literature Cited

Cavanaugh, T.M., and W.J. Mitsch. 1989. Water quality trends of the Upper Ohio River from 1977-1987. Ohio Journal of Science 5:153-163

Chapman, P.M., G.P. Romberg, and G.A. Vigers. 1982. Design of monitoring studies for priority pollutants. Journal of the Water Pollution Control Federation 54:292-297.

Eisler, R. 1986. Polychlorinated biphenyl hazards to fish, wildlife, and invertebrates: A synoptic review. US Fish and Wildlife Service Biol. Rep. 85(1.7), 72 pp.

Frenzel, S.A. 1996. Occurrence of selected contaminants in water, fish tissue, and stream-bed sediments in central Nebraska, 1992-1995. U.S. Geological Survey Open File Report 96-223, 6 pp.

Gilbert, R.O. 1987. Statistical methods for environmental pollution monitoring. Van Nostrand Reinhold, New York, 320 pp.

Helsel, D.R. 1990. Less than obvious statistical treatment of data below the detection limit. Environmental Science and Technology 24:1766-1774.

Helsel, D.R., and T.A. Cohn. 1988. Estimation of descriptive statistics for multiply censored water quality data. Water Resources Research 24:1997-2004.

Helsel, D.R., and R.M. Hirsch. 1992. Statistical Methods in Water Resources. Elsevier Science Publishers, New York, 522 pp.

Kendall, M.G. 1975. Rank Correlation Methods, 4th ed. Charles Griffin, London, 202 pp.

Knopman, D.S. and R. A. Smith. 1993. Twenty years of the clean water act: Has U.S. water quality improved? Environment 35:17-41.

Mann, H.B. 1945. Non-parametric tests against trend. Econometrica 13:245-259.

Meade, R.H. ed. 1995. Contaminants in the Mississippi River, 1987-92. U.S. Geological Survey Circular 1133, 140 pp.

Newman, M.C., P.M. Dixon, B.B. Looney, and J.E. Pinder III. 1989. Estimating mean and variance for environmental samples with below detection limit observations. Water Resources Bulletin 25:905-915.

Powell, M. 1995. Building a national water quality monitoring program. Environmental Science and Technology 29: 458A-463A.

Scientific Programming Enterprises. 1994. PlotIT for windows reference manual, Version 3.1.

Stark, J.R., W.J. Andrews, J.D. Fallon, A.L. Fong, R.M. Goldstein, P.E. Hanson, S.E. Kroening, and K.E. Lee. 1996. Water-quality assessment of part of the Upper Mississippi River Basin, Minnesota and Wisconsin- environmental setting and study design. U.S. Geological Survey Water-Resources Investigations Report 96-4098, 62 pp.

Sullivan, J.F. 1988. A review of the PCB contaminant problem of the Upper Mississippi River system. Wisconsin Department of Natural Resources, 50 pp.

Taylor, H.E., and A.M. Shiller. 1995. Mississippi River comparison study: Implications for water quality monitoring of dissolved trace elements. Environmental Science and Technology 29: 1313-1317.

Tyler, E.L. 1992. Reauthorizing the federal water pollution control act. Water Resources Update 88: 7-16.

U.S. Army Corps of Engineers. 1990. Mississippi River headwaters lakes in Minnesota low flow review, variously paged.

U.S. Environmental Protection Agency. 1993. Permit compliance system, public access to PCS data products. EPA/831/F-93-001, 12 pp.

U.S. Geological Survey. 1995. The strategy for improving water-quality monitoring in the United States- final report of the intergovernmental task force on monitoring water quality. U.S. Geological Survey Open File Report 95-742, 25 pp.

Upper Mississippi River Water Quality Initiative. 1993. Report of the toxic pollution workshop, Feb 17-18, 1993, Minneapolis, Minnesota, 14 pp.

Ward, R.C., J.C. Loftis, and G.B. McBride. 1990. Design of water quality monitoring systems. Van Nostrand Reinhold, New York, 230 pp.

Wiener, J.G., Jackson, G.A., May, T.W., and Cole, B.P. 1984. Longitudinal distribution of trace elements (As, Cd, Cr, Hg, Pb, and Se) in fishes and sediment in the Upper Mississippi River, pages 139-170 in J.G. Wiener, R.V. Anderson, D.R. McConville (eds). Contaminants in the Upper Mississippi River. Butterworth Publishers, Boston, 368 pp.

 

 

 

Table 1. Comparison of Ambient and Compliance Trends


  •