leaded gasoline. 



For copper, cadmium, silver, and zinc 

 the directions of change were not over- 

 whelmingly in a single direction. How- 

 ever, for cadmium, there were 12 sites 

 where the 1970s data were statistically 

 different from the NS&T data, and in 1 1 

 of those cases the 1970s concentra- 

 tions were higher. Conversely, for 1 8 of 

 the 22 sites where copper concentra- 

 tions were statistically different they were 

 lower in the 1970s. As pointed out by 

 Lauenstein et al. (1990), the copper 

 increase may reflect the fact that of all 

 the metals measured in both programs, 

 copper is the only one whose annual 

 use in the United States has shown an 

 increase since the mid-1970s. 



Changes in concentrations of chemi- 

 cals in mollusks at a single site may 

 indicate trends, but they could be ran- 

 dom variations unrelated to chemical 

 inputs. Differences over a 1 0-year time 

 span were considered trends rather 

 than random variations because they 

 occurred at many sites. The data from 

 sediment cores and from mussels at 

 single sites were considered trends 

 because there were data for many years 

 and there were consistent relationships 

 between concentration and time. 



Changes over short time scales may be 

 found in data on mussels and oysters 

 that have been collected by NOAA at 

 1 2-month intervals (in the late winter to 

 early spring) at NS&T sites since 1986. 

 This is the beginning of what will be- 

 come an increasingly valuable set of 

 data. At the early stages, however, it is 

 difficult to separate changes that signify 

 trends from those that do not. 



The first three years of concentration 



data for 132 sites have been examined 

 for cases where there were statistically 

 significant differences and where the 

 concentration in the middle year was not 

 the highest or lowest (NOAA, 1989). 

 This restriction was imposed so that 

 possible trends cou Id only be those cases 

 where concentrations had continuously 

 increased or decreased over the three 

 years. Even with that restriction, trends 

 cannot be confidently identified at single 

 sites with only three years of data. Some 

 confidence may be associated with 

 cases where groups of sites show simi- 

 lar changes. 



On the whole, concentrations of each 

 trace metal or group of organic com- 

 pounds changed in a statistically signifi- 

 cant and uniform fashion at about 10% 

 of the sites. There were a few groups of 

 sites showing a common change. For 

 example, cadmium and chlordane con- 

 centrations were decreasing in mussels 

 at sites in Long Island Sound; copper 

 and mercury were increasing at sites in 

 the Hudson/Raritan Estuary. Neverthe- 

 less, without data for more years, re- 

 sults from NS&T analyses, by them- 

 selves, cannot confidently identify trends. 



Conclusions 



The spatial distribution of contamination 

 throughout the coastal United States, 

 revealed through analyses of surface 

 sediments, shows the higher levels to 

 be characteristic of urbanized estuar- 

 ies. Those high levels, however, are 

 generally lower than those expected to 

 cause sediment toxicity, and among the 

 NS&Tsites, biological responses to con- 

 tamination, such as liver tumors in fish 

 or sediment toxicity, have not been 

 commonly found. Temporal trends in 

 contaminant levels at NS&T sites are 



16 



