On July 13, EPA published in the Federal Register its final rule on the Total Maximum Daily Load (TMDL) program, aimed at establishing “a process for making decisions in a common sense, cost effective way on how best to restore polluted waterbodies. . . . [by] identifying and implementing necessary reductions in both point and nonpoint sources of pollutants as expeditiously as practicable.”
In its final TMDL rule, EPA noted it is not taking action at this time on proposed regulations of animal feeding operations (AFOs) and corporate animal feeding operations (CAFOs). This is good news. A data analysis conducted for the American Farm Bureau Federation has exposed a foundation of very weak data upon which the entire TMDL program—and particularly the agency’s AFO and CAFO effort—is built.
The appendix to EPA’s National Water Quality Inventory (NWQI) 1996 Report to Congress reveals a gaping black hole of water quality misinformation. While the agency’s portrayal of those numbers has led the public and policymakers to believe that agriculture is polluting 70 percent of the nation’s streams and rivers, agriculture’s contribution is more likely in the range of 5 to 25 percent.
The 1998 NWQI was released in July without fanfare. It does not support EPA’s case for increased regulation of agriculture, as the sector’s contribution to water impairment is reported to have fallen from 70 percent to 58 percent. And more realistically, the new data suggest agriculture’s share is more likely between 4.7 and 20.3 percent.
Data poor or simply lacking
Table 1 shows the seven agricultural subcategories EPA has identified as sources of water pollution. In the 1996 EPA report’s appendix, each subcategory is characterized as a Major, Moderate, Minor, or Unknown problem. In its guidance document advising how the NWQI is to be interpreted, the agency acknowledges “. . . these designations are difficult to quantify and will continue to reflect the best professional judgement (BPJ) of the data analyst.”
The data in Table 1 were taken directly from the Appendix of EPA’s 1996 report. The EPA spreadsheet has been re-sorted to show where information is totally lacking in the seven agricultural subcategories; where data quality is poor (“evaluated” based on someone’s best professional judgement); and where states are double-, triple-, and even quadruple-counting impaired stream miles.
In 1996, 27 states, plus Puerto Rico and several Indian tribes, territories, and river basins, failed to report to EPA in a quantifiable format the data for the seven agricultural subcategories, or simply indicated the data were unknown. Yet 22 of those 27 states reported a total number for river miles impaired by agriculture. The numbers improved only slightly in 1998: 26 states reported no subcategory data, but 21 of them nevertheless reported a total for stream miles impaired by agriculture.
How could state officials report a total number for miles of rivers impaired by agriculture if no data are reported in the agricultural subcategories? We have labeled this area of Table 1 “EPA’s Black Hole of Data.”
Column 2 of the table shows the reporting states surveyed 693,905 miles of rivers for their 1996 reports–just 19 percent of the nation’s 3.6 million miles of streams and rivers. More often than not, those surveyed waterways were not selected by random sample, but rather because there was reason to suspect water quality problems existed. Moreover, water quality monitoring, if it was done at all, was often performed downstream of a sewage treatment plant.
In 1998, states assessed 842,426 miles, but EPA’s report acknowledged many assessment problems remained.
The problem with “evaluated” data
The next item to consider in Table 1 is in Column 3, under the heading “Percent of Impaired Miles Evaluated.” The data become further suspect because so much of it is “evaluated”: based not on actual water quality monitoring, but on someone’s best professional judgement (BPJ).
The 14 states that made extensive use of “evaluated” information to assess water quality were 3.5 times more likely to declare agriculture as the source of pollution than were states that relied exclusively on water quality monitoring data for their assessments. In 1998, 10 additional states had eliminated their use of “evaluated” information, but three states (Massachusetts, New Hampshire, and Oklahoma) increased their use of evaluated information.
Even the data reported from actual water quality monitoring are suspect. Much of it is more than five years old. In at least three states whose data we have examined closely, some monitoring data are between 15 and 20 years old.
Divide and conquer
The last column of Table 1 totals the number of waterway miles said to be impaired by each of the seven agricultural subcategories. It should add up to the same number as is reported in Column 4 (River Miles Impaired by Agriculture). Yet 11 of 23 states report a larger number in the last column than in Column 4, suggesting those states are double-, triple-, and even quadruple-counting some impaired miles in the seven subcategories. (Only 8 of 24 are apparently double counting in 1998.)
EPA itself also double-counts impaired miles in its graph on page 33 of the 1996 NWQI Report. This practice of showing total miles or percentages of pollutants by each source is not a useful method of depicting the scope or severity of the true problem. It is an inappropriate accounting scare tactic that leads to a divide-and-conquer mentality.
Illinois is one of few states to organize water quality data in such a manner as to allow an accurate depiction of the scope of the water quality problem in the state’s streams. Data from Illinois EPA, which counts each stream mile only once, gives the reader a clear picture that cleaning up a mile of stream polluted by 11 different sources will be a much different project than cleaning up a stream polluted by just one source. The data also clearly show that, more often than not, agriculture is not the sole source of pollution within a given stream mile, as U.S. EPA’s graphic would lead readers to believe.
Water quality: A risk to public health?
Section 2.2 of the Unified National Strategy for Animal Feeding Operations, released on March 9, 1999, by EPA and the U.S. Department of Agriculture, is titled “Water Quality and Public Health Risks.” It uses such phrases as “can pose,” “have the potential to,” “can result,” “may be harmful,” “can contaminate,” “have been associated,” and “may produce odors.”
Not once does the discussion report scientifically documented water quality problems or water quality-related public health risks due to livestock operations. The report merely implies that livestock agriculture is big, and that big is bad . . . and therefore must be regulated.
EPA and USDA also note in the Unified AFO Strategy that “pathogens, such as Cryptosporidium, have been linked to impairments in drinking water supplies and threats to human health.” The Strategy does not itself conclude that livestock operations are responsible, but it certainly gives that impression.
In a further misrepresentation by omission, the Strategy fails to acknowledge that a Centers for Disease Control report concluded agriculture was not to blame for the widely reported cryptosporidium outbreak that hit the city of Milwaukee several years ago. Neither does it acknowledge that the 10 ppm federal standard for nitrate-nitrogen in drinking water has now been found to be without scientific basis. New information shows bacterial infection in infants is the prime cause of methemoglobinemia (blue baby disease), once attributed to cryptosporidium.
Unfortunately, perception and reality often lead to much different public policy conclusions. Despite the perceptions, all indications are that surface water quality in our streams and rivers is improving and the trend will more than likely continue in that direction for some time, even without the costly AFO strategy regulations. But, that good news will never see the light of day until EPA’s black hole of data is corrected.
Correcting the black hole
For public policy on water quality problems to proceed in the right direction, it is incumbent on the agricultural community itself to identify water quality problems traceable to agriculture activities, and to eliminate them. At the same time, it is important that policy analysts and concerned citizens join the agricultural community to help state environment agencies clean up the data mess:
- States must be encouraged to use reliable water quality monitoring data less than five years old.
- States should follow Texas’s example and report waterbodies as impaired only when water quality monitoring data clearly show the water does not meet state standards.
- Data like those from Illinois should be developed to accurately depict the extent and scope of river and stream impairment.
- The responsible state agency should produce an accurate Map of Change every two years as part of the water quality assessment it must submit to EPA. There should be two maps: one that shows all waterbodies that have improved and are meeting standards appropriate for their designated uses, and one that shows all waterbodies that have degraded and are considered impaired.
Our examination of state 305(b) Water Quality reports indicates no state is producing such maps . . . yet without them, there is no clear way of showing whether all the money spent on water quality is doing any good.
The states’ 305b reports are the “on-ramps” to regulation under the TMDL program (303d lists). A TMDL in a watershed can preclude further economic growth and development. States that report “evaluated ” information—or worse yet, no data at all—should not be allowed to put a watershed on the 303d lists.
Jim Porterfield is technical specialist for land, water, and forestry Resources at the American Farm Bureau Federation.
For more information
EPA’s final TMDL rule, published July 13 in the Federal Register, is available on the Internet at http://www.epa.gov/owow/tmdl/finalrule/.
