On appeal from the Department of Environmental Protection, issuance of certain regulations controlling atmospheric emissions of mercury, 36 N.J.R. 5406(a) (December 6, 2004).
The opinion of the court was delivered by: Parrillo, J.A.D.
NOT FOR PUBLICATION WITHOUT THE APPROVAL OF THE APPELLATE DIVISION
Before Judges Lefelt, Parrillo and Sapp-Peterson.
On November 4, 2004, the New Jersey Department of Environmental Protection (DEP) issued the first regulations ever to control mercury emissions from iron and steel melters. The only prior controls on mercury emissions were in the melters' operating permits, which are specific to each facility and widely different. Moreover, none of the iron and steel facilities in New Jersey is currently covered by federal regulation. The new rules require a 75% reduction in mercury emissions starting in 2010, achieved through source separation measures and, if necessary, installation of additional exhaust controls. These rules, which exceed federal requirements, affect electric arc furnace steel manufacturers who are members of appellant Steel Manufacturers Association (SMA), including appellant Gerdau AmeriSteel, who owns and operates two mini-mills in New Jersey that utilize electric arc furnaces to recycle scrap metal to produce new steel products. Appellants support source separation, but object to the mandatory installation of additional exhaust controls should source separation measures fail to achieve the DEP's reduction goals. Consequently, in these consolidated appeals, they challenge the newly promulgated regulations on a variety of grounds, arguing, among other things, that DEP exceeded its statutory authority; acted arbitrarily, unreasonably and without a legitimate technical basis by requiring use of emission control technologies that are neither commercially available nor of proven effectiveness for the forms of mercury generated by the mini-mills; and violated the rulemaking requirements of the Administrative Procedure Act (APA), N.J.S.A. 52:14B-23.
By way of background, the record discloses that there are six "melters" or "mini-mills" in New Jersey, which are facilities that turn mercury-containing scrap metal into iron or steel. Three of them, including the two operated by Gerdau, utilize electric arc furnaces (EAF) to recycle or melt the scrap metal, and the other three use a vertical furnace called a cupolo. Collectively, their permitted production capacity is approximately three million tons per year. Together, they constitute the State's largest single "source category" for atmospheric emissions of mercury, and DEP estimates that they discharge about 1000 pounds of mercury per year as a byproduct of the melting process, even after treatment by some form of emission-control technology.
Mercury is a heavy metal which is regulated because it is a "hazardous air pollutant" (HAP) under both federal and state law. 42 U.S.C.A. § 7412(b); N.J.S.A. 26:2C-2. Mercury has high toxicity and a tendency to persist in the environment and become concentrated in food sources, particularly fish. In fact, mercury has far greater toxicity compared to the air pollution "criteria pollutants." 36 N.J.R. 5412, response to comment 31. Inorganic mercury is commonly released to the environment by burning fuels and wastes containing mercury. The likely source of mercury in New Jersey is from "feedstock ferrous scrap", which includes recycled motor vehicles, home appliances, and waste metal from demolished buildings. More specifically, the mercury in switches, sensors and thermostats that those items contain is the likely source of the mercury in a melter's exhaust.
To date, only a modest proportion of the reduction in mercury emissions by melters has been achieved through source separation, namely, having scrap dealers and recyclers, which supply mini-mills with the end-of-life vehicles that are the largest source of their feedstock, remove electrical switches that contain mercury from the vehicles before processing them.
To be truly effective, source separation has to occur before scrap is processed for delivery to the mini-mill, because the receipt of scrap in bulk quantities prohibits a mini-mill from identifying and removing the "relatively minute quantities of mercury-containing components". Yet, the regional scrap metal industry, which has been largely unregulated in New Jersey, has very little incentive to comply. Devices such as switches, sensors and thermostats simply do not have enough intrinsic value for dismantlers or recyclers to remove them before processing, so even if the use of mercury in new devices were phased out, the old ones would continue to appear in the feedstock for years to come. In addition, scrap processors sometimes lack necessary information from vehicle manufacturers about the exact location of those devices. As a representative for one of the melters, United States Pipe and Foundry*fn1 , testified at public hearings on DEP's proposed regulations at issue here, his company spent more than a year demanding mercury-free scrap from its eleven suppliers and offering "an incentive," but only one of them even tried to remove switches containing mercury. According to the representative, in the currently tight market for scrap metal, "if you can get it anywhere, even from overseas, you need to try to get it or you don't operate". Thus, while source removal may be a reasonable and cost-effective emissions control measure -- DEP having estimated the cost of switch removal at $2 per switch, or $1140 per pound of mercury removed -- DEP also concluded from the results of a pilot project demonstrating only a 50% reduction, that such an approach by itself "will not necessarily" achieve a sufficient reduction in mercury emission. 36 N.J.R. 5413-14, response to comment 34; 36 N.J.R. 5411, response to comment 25.
In addition to source removal, four of the six melters in New Jersey (three EAFs and the one cupola), including Gerdau's facilities in Raritan and Sayreville, use an emission control technology called a "baghouse," which is essentially a fabric filter. However, in one melter's operations, the baghouse reduced mercury emissions by only 34%. In fact, recent stack test data show that the mean mercury emission rate from iron and steel melters, weighted based on production capacity, was 137 milligrams of mercury emissions per ton (mg/ton) of iron or steel production, an emissions rate deemed by DEP to be causing an unacceptable degree of human exposure to mercury.
Because of the threat to the public health and welfare posed by the inadequacy of then current mercury pollution policies, DEP established the New Jersey Mercury Task Force (Task Force) on March 9, 1998. Its charge was to "review innovative and low cost emission reduction strategies available in various industrial sectors", and to "[r]ecommend mercury emission controls and standards for in-state sources."
In carrying out its duties, the Task Force looked to mercury control technologies in related industries, particularly coal power plants and municipal solid waste (MSW) incinerators, as well as European mini-mills. To be sure, there were differences: the levels of mercury in the melter's scrap feed were highly variable and unpredictable compared to the far less variable levels in coal plants due to the homogenous nature of their fuel supplies. There were also differences in the balance or "speciation" between ionic and elemental forms of mercury in the source material in the exhaust stream as well as the temperatures of the exhaust or "flue gases." Indeed, DEP agreed that elemental mercury might represent "a relatively large portion" of a mini-mill's mercury emissions and that ionic forms of mercury are easier to capture than the elemental mercury that predominates in melting operations. Despite these differences in characteristics, however, studies assessing the prospects for adapting the mercury control techniques used in other industries concluded that the technology is readily transferable.
On November 30, 2000, the consulting firm of BE&K Terranext issued a report analyzing the "best available control technology" and the "state of the art" (SOTA) for reducing mercury emissions from a mini-mill. Terranext noted the "potentially applicable" technology of injecting activated carbon into the melters' exhaust gas streams, which MSW incinerators were doing and which coal power-plants were testing.*fn2 In fact, one company, ADA Environmental Technology, advised Terranext that its experience relating to coal power-plants "is expected to be similar to what may be encountered in the steel industry". ADA accordingly assumed that a baghouse would collect "essentially 100% of the mercury associated with particulate matter," and that "[t]he relatively low temperature of 200°F should lend itself to effective capture of vapor phase mercury by carbon based sorbents". The only variable, depending on the mercury vapor's speciation, was the amount of carbon-based sorbent that would have to be injected to capture 90% of total mercury vapor.
The Task Force agreed that activated carbon injection (ACI) technology is commercially available, having been used for well over a decade by operators of MSW incinerators to significantly reduce mercury emissions, and that this technology was readily transferable to New Jersey's iron and steel melters, particularly those that already have baghouses for control of particulate emissions melters. Thus, recognizing that source reduction measures alone may not fully achieve the Task Force's mercury emission reduction goals of 75%, the Task Force issued Volume I of its Report in December 2001, recommending a cooperative three-year effort "to reduce mercury contamination of scrap through elimination and separation measures," followed by the mandatory installation of exhaust controls if the source-separation measures by themselves did not achieve the Task Force's emission-reduction goal of 75%. In Volume III, issued in January 2002, the Task Force, obviously impressed by the 95% reduction of exhaust mercury levels for MSW incinerators, and finding that both the volume of a melter's flue gas flow and the concentrations of mercury in it were not outside the range for MSW incinerators or coal power-plants, concluded that, for the three EAFs and the one cupola already using a baghouse, the amount of carbon injection needed for additional reductions to achieve compliance would be "a relatively low capital cost option". The two cupolas that used scrubbers "would need to rely on scrap management or evaluate measures to remove mercury switches," or develop another way to oxidize the mercury vapor into an ionic form before it enters the scrubber.
Subsequent studies confirmed this view. A February 27, 2004 EPA report about mercury-control technology for coal power-plants described how the mercury, which "in the high temperature regions of" the boilers becomes elemental vapor, "begins to convert" to ionic or other non-elemental forms, and then turns solid or gets "absorbed onto the surface of other particles". The report related a "very limited set of short-term full scale trials" of activated carbon injection for power-plants using subbituminous or lignite coal, and noted that such injection, in conjunction with an electrostatic precipitator, "has the potential to achieve" a 70% reduction, or even 90% if a fabric filter is also used.
Along similar lines, a March 11, 2002 DEP memo explained that the mercury vapor at the U.S. Pipe and Foundry facility was three-quarters elemental at the baghouse inlets, where the temperature averaged 718°F. However, at the baghouse outlets, where the average temperature was 170°F, only three-eighths of the mercury vapor was elemental, meaning that half of the elemental mercury had changed to ionic, which DEP presumed was "likely related in part" to the temperature change.
A subsequent DEP memo, dated June 10, 2003, acknowledged that source separation could reduce emissions "to a level in the range of 35 mg" per ton, but only if it resulted in feedstock that was entirely fee of mercury. Accordingly, the memo recommended that New Jersey's four facilities with baghouses (which included Gerdau's in Sayreville and Raritan) might have to achieve further reductions by injecting activated carbon into their exhaust gas stream. The memo also referred to other injectable sorbents, including "oxidized hydrated lime and inexpensive silicate substrates impregnated with chemicals that possess a strong affinity for mercury," which, in addition to activated carbon, were being used in the MSW incinerators and tested at coal power-plants. The memo stated that "amended silicates" were "projected to be commercially available within 1 year with lower price[s] than" activated carbon. It also alluded to some success in Europe with the injection of sodium tetrasulfide, either alone or in ...