John Fleck’s Jan. 14 front-page article on the recent EPA report about the Kirtland spill undoubtedly caused many Albuquerque homeowners to breathe a sigh of relief. Thirty years until the potent carcinogen EDB, ethylene dibromide, contaminates our drinking water; plenty of time to clean it up. “It takes it out of crisis mode” said Tom Blaine of the New Mexico Environment Department.
However, analysis of the draft EPA report reveals significant limitations that imply that this relief may be premature. The NMED is on record that the toxic plume of contaminants is likely to reach the nearest Ridgecrest well in five to seven years. How can there be such a large discrepancy between the NMED estimate and the new EPA report’s conclusions?
The EPA report is based on a three-dimensional numerical model of groundwater flow and EDB transport in the area surrounding Kirtland. As the report says, such models are “powerful tools when enough site-specific high-quality data are available.” Unfortunately, in this case, the data used for the model was of questionable validity for several reasons.
First, the model used in the report is a steady-state model that assumes that groundwater flow conditions are constant. In actuality, the groundwater flow conditions between the spill site and the nearest Ridgecrest drinking wells are complex, variable and not clearly specified.
Variation in groundwater velocity occurs based on the amount of water pumped from the drinking wells, with more pumping increasing the water velocity. Since more water is pumped during the summer, the velocity of the groundwater is faster during the summer months. Groundwater speed increases as it approaches the drinking wells. Different geologic layers – sand, clay, gravel – create different levels of hydraulic conductivity, or ease of water flow through the soil.
The central limitation of the EPA report is the fact that its primary sources of data are the 115 Kirtland monitoring wells, which are all clustered around the spill site, and the closest drinking wells, which are between 2,000 feet and a mile away. The model therefore incorporates no real data about the crucial area between the spill site and the nearest municipal drinking wells – the area through which the contamination must travel in order to reach the drinking wells. Without any data from this area, any conclusion about how fast the contamination may reach the drinking wells is suspect.
In the absence of this important data, the model makes assumptions, many of which are highly questionable. For instance, the crucial variable of hydraulic conductivity was calculated using inappropriate data from the drinking wells and water flow through their screens. In fact, we need to know the hydraulic conductivity of the area between the spill site and the drinking wells.
In addition, the drinking well pumping rates that were used for the model were winter rates, thus significantly underestimating both the average yearly pumping rate – because approximately 50 percent more water is pumped during the summer – and the speed of the groundwater.
The model also assumes that the amount of EDB in the aquifer will decrease by 10 percent per year, which is a very questionable assumption, given that it is not based on any data, that EDB does not degrade naturally and that no contaminated water has been removed from the aquifer or is projected to be removed in the near future.
The Kirtland spill is the largest toxic contamination of an aquifer in U.S. history, much larger than the other 42 Air Force Superfund sites. Albuquerque’s drinking water is in jeopardy. How soon will the contamination reach the nearest drinking well? No one really knows.
At this point, we need more monitoring wells in the area between the spill site and the drinking wells, not only to generate more data for better modeling, but also in order to devise an appropriate remedy to this looming disaster. As hydrogeologist Robert Gilkeson says, given the significant limitations of the draft EPA report, “the crisis mode remains.”