Summary And Conclusions

The purpose of this chapter is to address the second leg of the three-legged stool approach for answering the question posed in Chapter 1: How has the nation's water quality changed since implementation of the 1972 CWA's mandate for secondary

United States: 01-18

Percentile

Figure 3-51 Before- and after-CWA frequency distributions of worst-case DO aggregated by all major river basins or n = 311 RF1 reaches with paired before and after data sets: 48 States. Source: USEPA STORET.

Percentile

Figure 3-51 Before- and after-CWA frequency distributions of worst-case DO aggregated by all major river basins or n = 311 RF1 reaches with paired before and after data sets: 48 States. Source: USEPA STORET.

treatment as the minimum acceptable technology for POTWs? Recall that the basic goal of the second leg was to determine the extent to which water quality improvements could be linked to the CWA's push for secondary and greater levels of treatment in the nation's POTWs. If evidence showed that worst-case DO concentrations improved at broad, as well as localized, spatial scales, the second leg of the investigation could add cumulative support for the conclusion that the CWA's mandates were successful. The following objectives were established to guide this part of the study:

• Develop before- and after-CWA data sets composed of DO summary statistics derived from monitoring stations screened for worst-case conditions.

• Develop a worst-case DO summary statistic for each station for each before- and after-CWA time period and then aggregate these data by sequentially larger spatial scales (reaches, catalog units, and major river basins).

• Conduct an analysis of the spatial units having both a before- and after-CWA summary statistic and assess the magnitude of worst-case DO change between the two time periods.

• Assess the before- and after-CWA change in the point source discharge/downstream DO signal over the progressively larger spatial scales.

Key Points of the Background Section

Section A provided background concerning the source of DO data used in this study, why worst-case conditions are an appropriate screening tool for developing the be fore- and after-CWA data sets, and the role spatial scale played in the second leg of this study. Key points include the following:

• The sharpest signal linking point source loading and downstream DO inherently resides in data collected in worst-case (high temperature and low flow) conditions. These worst-case conditions typically occur in the summer months (July through September) during consecutive runs of dry years (persistent drought).

• Widespread persistent drought was most pronounced in the summers in 19611965 (before the CWA) and 1986-1990 (after the CWA). These time-blocks were used to define the before- and after-CWA time periods for the comparison analysis.

• From a spatial perspective, worst-case critical, or minimum, DO below a point source occurs in the "degradation" or "active decomposition" zone of the oxygen sag curve. However, screening rules were not developed to select monitoring stations located within these zones because the goal of this second leg is to examine changes in the point source discharge/downstream DO at broad scales as well as localized scales. Consequently, the only screening rule regarding location of stations eligible for the before- and after-CWA analysis is that the station must be somewhere downstream from, and therefore potentially influenced by, a municipal and/or industrial point source discharge.

Key Points of the Data Mining Section

Section B presented the six-step data mining process used to create the before- and after-CWA data sets to be used in the comparison analysis. The screening rules associated with each step are listed below:

Step 1—Data Selection Rules

• DO, expressed as a concentration (mg/L), will function as the signal relating municipal and industrial point source discharges of oxidizeable organic matter to downstream water quality responses.

• DO data are extracted only from the July-September (summer season) time period.

• Only surface DO data (DO data collected within 2 meters of the water surface) are used.

Step 2—Data Aggregation Rules from a Temporal Perspective

• 1961-1965 serves as the time-block to evaluate persistent drought before the CWA and 1986-1990 serves as the time-block to evaluate persistent drought after the CWA.

• To remain eligible for the before- and after-CWA comparison, DO data must come from a station residing in a catalog unit that had at least one year classified as dry (streamflow ratio 75 percent of long-term summer mean) out of the 5 years in each before- and after-CWA time-block.

Step 3—Calculation of the Worst-Case DO Summary Statistic Rules

• For each water quality station, the tenth percentile of the DO data distribution from the before-CWA time period (July-September, 1961-1965) and the tenth percentile of the DO data distribution from the after-CWA time period (July-September, 1985-1990) are used as the station's DO worst-case statistics for the comparison analysis.

• To remain eligible for the before- and after-CWA statistical comparison, a station must have a minimum of eight DO measurements within each of the 5-year time-blocks.

Step 4—Spatial Assessment Rules

• Only water quality stations located on streams and rivers affected by point sources are included in the before- and after-CWA comparison analysis.

Step 5—Data Aggregation Rules from a Spatial Perspective

• The before- and after-CWA data sets are collections of DO summary statistics that characterize worst-case DO at individual water quality monitoring stations across the United States for the 1961-1965 time-block and the 1986-1990 time-block, respectively (one DO summary statistic per station per time-block).

• For each data set and time-block, the tenth percentile value from each eligible station is aggregated within the spatial hydrologic unit. (Since the scales are hierarchical, a station's summary statistic is effectively assigned to both a reach and a catalog unit.) A summary statistic is then calculated and assigned to the spatial unit for the purpose of characterizing its worst-case DO. If a spatial unit has only one monitoring station within its borders that meets the screening criteria, the tenth percentile DO value from that station simply serves as the unit's worst-case summary statistic. If, however, there are two or more stations within a spatial unit's borders, the tenth percentile values for all the eligible stations are averaged, and this value is used to characterize worst-case DO for the unit.

• The mean tenth percentile value is computed from the eligible station's tenth percentile values for the before- and after-CWA periods.

Step 6—Development of the Paired Data Sets (at Each Spatial Scale)

• To be eligible for the paired comparison analysis, a hydrologic unit must have both a before-CWA and an after-CWA summary statistic assigned to it.

Key Points of the Comparison Analysis Section

Section C presented the results of the comparative before- and after-CWA analysis of worst-case DO data derived using the screening criteria described in Section B and aggregated by spatial units defined by three scales (RF1 reach, catalog unit, and major river basin). Listed below are key observations for each spatial scale:

RF1 Reach Scale

• Sixty-nine percent of the reaches evaluated showed improvements in worst-case DO after the CWA. [Three hundred eleven reaches (out of a possible 12,476 downstream of point sources) survived the data screening process with comparable before- and after-CWA DO summary statistics. The number of reaches available for the paired analysis was limited by the historical data for the 1961-1965 period].

• These 311 evaluated reaches represent a disproportionately high amount of urban/industrial population centers, with approximately 13.7 million people represented (7.2 percent of the total population served by POTWs in 1996). The top 25 improving reaches saw their worst-case DO increase by anywhere from 4.1 to 7.2 mg/L!

• The number of evaluated reaches characterized by worst-case DO below 5 mg/L was reduced from 167 to 97 (from 54 to 31 percent).

• The number of evaluated reaches characterized by worst-case DO above 5 mg/L increased from 144 to 214 (from 46 to 69 percent).

• The long-term trends of worst-case DO presented for selected RF1 reaches identified with the greatest before and after improvement document progressive improvements in DO as sewage treatment plants were upgraded to comply with the CWA requirements for a minimum level of secondary treatment.

Catalog Unit Scale

• Sixty-eight percent of the catalog units evaluated showed improvements in worst-case DO after the CWA. [Two hundred forty-six catalog units (out of a possible 1,666 downstream of point sources) survived the data screening process with comparable before- and after-CWA DO summary statistics].

• The number of evaluated catalog units characterized by worst-case DO below 5 mg/L was reduced from 115 to 65 (from 47 to 26 percent). The number of evaluated catalog units characterized by worst-case DO above 5 mg/L increased from 131 to 181 (from 53 to 74 percent).

• Fifty-three of the 167 improving catalog units (32 percent) improved by 2 mg/L or more, while only 10 of 79 degrading catalog units (13 percent) degraded by 2 mg/L or more.

• These 246 evaluated catalog units represent a disproportionately high amount of urban/industrial population centers, with approximately 61.6 million people represented (32.5 percent of the total population served by POTWs in 1996).

• The long-term trends of worst-case DO and BOD5 presented for selected catalog units identified with the greatest before and after improvement document progressive increases in DO with corresponding decreases in BOD5 as sewage treatment plants were upgraded to comply with the CWA requirements for a minimum level of secondary treatment.

Major River Basin Scale

• A total of 11 out of 18 major river basins had sufficient reach-aggregated worst-case DO data for a before- and after-CWA comparison analysis.

• Based on two statistical tests, 8 of the 11 major river basins can be characterized as having statistically significant improvement in worst-case DO levels after the CWA. The three basins that did not statistically improve under either test also did not have statistically significant degradation.

• When all the 311 paired (i.e., before vs. after) reaches were aggregated and the statistical tests run on all 18 of the major river basins of the contiguous states as a whole, worst-case DO also showed significant improvement.

Conclusions

The statistical analyses developed for this study are not ideal. One major concern is the potential bias introduced in the ambient monitoring programs used to collect the data archived in STORET. It is believed that the analysis of data sets with data in the before and after time periods alleviates some of these concerns and that results are generally comparable for the two different statistical tests. Based on the systematic, peer-reviewed approach designed to identify and evaluate the national-scale distribution of water quality changes that have occurred since the 1960s, this study has compiled strong evidence that the technology-based and water quality-based policies of the CWA for point source effluent controls have been effective in significantly reducing loads and improving DO. In this retrospective analysis, DO was used as the key indicator because the reduction of organic carbon and nitrogen (BODu) loading from municipal and industrial point sources was one of the major goals of the CWA's technology-based policy, which required industrial effluent limits and a minimum level of secondary treatment for municipal facilities. Based on ambient DO records, significant before and after improvements in many rivers and streams have been identified over national, major river basin, catalog unit, and RF1 reach-level spatial scales.

The "signal" of downstream water quality responses to upstream wastewater loading and the changes in this signal since the 1960s have been successfully decoded from the "noise" of millions of archived water quality records. Given the very large spatial scale of the major river basins, it is remarkable to observe statistically significant before and after DO improvements as detected using the systematic methodology described in this book. Previous evaluations of the effectiveness of the CWA (e.g., Smith et al., 1987a, 1987b; Knopman and Smith, 1993) were not able to report conclusively significant improvements in DO. In these earlier studies, however, the methodologies used were not specifically designed to separate the signal of downstream water quality response from the noise within large national databases. Using appropriate data screening rules and spatial aggregations, it has been demonstrated that improvements in water quality, as measured by improvements in worst-case DO, have been achieved since the 1960s.

The findings of this national-scale water quality assessment demonstrate three important points:

• As new monitoring data are collected, it is crucial for the success of future performance measure evaluations of pollution control policies that the data be submitted, with appropriate QA/QC safeguards, to accessible databases (e.g., Alexander et al., 1998). If the millions of records archived in STORET had not been readily accessible, it would have been impossible to conduct this analysis to identify the signals of water quality improvements that have been achieved since the early 1960s.

• Significant after-CWA improvements in worst-case summer DO conditions have been quantitatively documented with credible statistical techniques in this study over different levels of spatial data aggregation from the small sub-watersheds of RF1 river reaches (mean drainage area of approximately 115 square miles) to the very large watersheds of major river basins (mean drainage area of 434,759 square miles).

• The data mining and statistical methodologies designed for this study can potentially be used to detect long-term trends in signals for water quality parameters other than DO (e.g., suspended solids, nutrients, toxic chemicals, pathogens) to develop new performance measures to track the effectiveness of watershed-based point source and nonpoint source controls. The key element needed to apply the data mining methodology to other water quality parameters is the careful specification of rules for data extraction that reflect a thorough understanding of the hydrologic processes that influence the spatial and temporal distributions of a water quality constituent, as well as the relevant sources of associated pollutants (see Hines et al., 1976).

Population Affected by Reaches with Improved DO To quantify, in financial terms, the environmental benefits derived from various environmental policy decisions, USEPA developed the National Water Pollution Control Assessment Model (NWPCAM) (Bondelid et al., 2000), which includes a link between 1990 population and RF1 river reaches. As discussed in Section E, this model does not include all estu-arine and coastal waters, and as a result, does not account for the entire U.S. population. It is estimated that about one-third of the U.S. population is not accounted for in the model. At the same time, if a person is located near two rivers, that person is counted twice, since he or she can derive a benefit from environmental improvements in either river.

Recognizing these limitations of this accounting procedure, the model accounts for 197.7 million people in 23,821 reaches. In the 311 reaches analyzed here (1.3 percent of reaches in the model), the model accounts for 13.7 million people (6.9 percent of the population in the model). The ratio of the percent population to percent reaches in the model demonstrates that the screening process developed for this analysis is reasonably successful in finding reaches with data near urban centers, although 57 of the 311 reaches did not have population associated with them. Of the 13.7 million people represented by the 311 reaches, 11.8 million of them (86 percent) are associated with reaches that have an increased worst-case DO from before to after the CWA. Almost one-half (45 percent) of the selected population are associated with

Was this article helpful?

0 0

Post a comment