Appearing before Bristol magistrates on 15 November, Sevalco pleaded guilty to six specimen charges of exceeding authorised emission limits, falsifying records and failing to notify the Agency (see box ).
Because of the severity of the charges, the magistrates referred the case to the Crown Court. The magistrates were "absolutely appalled by the findings of this case and consider that individual employees, Sevalco and indeed the parent company auditors as a whole were grossly negligent in their duty in effectively monitoring the release of cyanide into the environment."
Cyanide is acutely poisonous even in low concentrations, although it is not persistent or bioaccumulative. An Agency study found that the impact on the estuary was limited to local invertebrate life.
Questions on self-monitoring
On 8 December, Bristol Crown Court fined Sevalco a total of £240,000 and ordered it to pay costs of £70,000. The fine appears to be the second largest ever handed out for offences under the integrated pollution control (IPC) regime. The judge described the concealment of discharge data as "astonishing".
The case raises serious concerns that the principle of self-regulation and self-monitoring is potentially vulnerable to abuse - even by large, established companies. Sevalco is a member of the Chemical Industries Association and a signatory to its "Responsible Care" initiative on environmental best practice.
Nor is the case unique. In 2000, a similar case involving management failures and the falsification of data forced leading waste management group Leigh Environmental to pay out £108,000 in penalties (ENDS Report 300, pp 22-24 ).
In enforcing both IPC and the replacement IPPC regime, the Agency relies heavily on monitoring data collected by operators, and on permit conditions which require operators to report any breaches. The Agency does conduct check monitoring tests and site inspections - but these activities are being squeezed by an increased workload and tightening budgets.
The Sevalco case also highlights differences in the penalties attached to environmental crime in the UK compared with the USA. In June, Phelps Dodge - Sevalco's ultimate parent company - agreed to pay $1.4 million in a settlement over alleged violations of the federal Clean Air Act. The Environmental Protection Agency estimates that more than 1,000 tonnes of sulphur dioxide was emitted, illegally by-passing pollution control and monitoring equipment on hundreds of occasions, from a copper mine and ore processing plant in Arizona now owned by Phelps Dodge.
Sevalco's Avonmouth site employs some 140 people and each year produces around 100,000 tonnes of carbon black, a key constituent in the manufacture of tyres, rubber, plastics and inks. The company is a subsidiary of US firm Columbian Chemicals, itself part of the Phelps Dodge Corporation.
Sevalco operates several carbon black processes at Avonmouth and has been regulated under IPC since 1993. Tim Loveday, the Agency's inspector for the site, says the facility has been "challenging". For years, unstable operating conditions have caused minor, unauthorised releases to air of carbon monoxide and carbon black.
Emissions of particulate matter have been a long-standing concern - and Mr Loveday says that trees near the site are black as a result of historic emissions.
The prosecution related to the site's former "medium thermal process" in which oil was pyrolysed to form carbon black, as well as various gases including hydrogen cyanide. Prior to 1997, emissions from the process were abated using a wet scrubber. However, emissions of particulates were still high and the abatement plant consumed significant volumes of water.
Sevalco was authorised to discharge wastewater containing low levels of cyanide into the Severn estuary at high tide. For the site's main discharge, which includes wastewater from the abatement plant, the Agency set a concentration limit for cyanide of 1mg/l and an annual mass release limit of 0.07 tonnes. A second discharge sends water from the surface water collection system, including some process water, into the Severn via a ditch. The limit was set at 0.1mg/l, with an annual mass release limit of 0.05 tonnes.
The authorisation required Sevalco to conduct weekly in-house sampling for cyanide, ammoniacal nitrogen and pH, and to report the results to the Agency each quarter. The company was required to notify the Agency immediately if any discharge limits were exceeded.
Sevalco was also required to conduct monthly sampling for other substances including metals and suspended solids. It contracted this work out to a laboratory run by Wessex Water and accredited by the UK Accreditation Service. Wessex's monitoring also included sampling for cyanide.
In 1997, as part of a four-year, £17 million programme to develop the site, Sevalco built a new production unit to increase capacity. The work included installing taller stacks and a more efficient bag filter to replace the wet scrubber. The company also installed a surface water collection system which enabled water to be reused within the process.
However, the change had unintended consequences. Waste gases now needed cooling before being passed to the bag filter - and the cooling water was recirculated, greatly increasing the amount of cyanide it absorbed from the gases. As a result, the Agency estimates, cyanide concentrations in the storage tank used to hold the wastewater prior to discharge increased from 1ppm to 100ppm.
The prosecution focused on efforts by Sevalco staff to conceal the increased discharge of cyanide over four years - from April 1998 to September 2002.
Details reported in this article relating to the actions of individual employees reflect the evidence presented in court by the Agency. However, the Agency did not press charges against any individual.
The man at the heart of the case
According to the Agency, the key figure in the deception was Stephen Buckner. Mr Buckner worked for Sevalco for 25 years, starting in the personnel department and rising to become the plant's general manager, environment manager and a director of the company despite having no formal qualifications. Speaking after the case, Tim Loveday described Mr Buckner as "a very confident man who always gave the right impression in relations with the Agency."
In court, the Agency set out Sevalco's procedures for monitoring and reporting, based on statements from a number of key employees.
Sevalco's works chemist, Roger Stevens, was responsible for taking the samples and recording them in a special log called the "blue book". He would then pass the book to Mr Buckner, who would compile the quarterly report to the Agency.
Mr Stevens also received a monthly report from Wessex Water which he usually passed directly to Mr Buckner. Mr Stevens told the Agency that he could not remember routinely checking these reports.
The Agency said that, at interview, Mr Buckner was not entirely clear about his involvement with the monitoring data, but denied any wrongdoing. However, he confirmed that he was responsible for compiling the quarterly returns. The court heard that to do this he must have had access to the data from Wessex Water.
In court, the Agency alleged that Mr Buckner had told Mr Stevens that, where data indicated a breach, he should not enter the true figure in the blue book. Mr Stevens' account at interview was contradictory. At one point, he suggested the falsification happened only "a couple of times" - but on another occasion he said that the practices "continued throughout 1998-2002."
As well as the apparent tinkering with data records, Sevalco's monitoring procedures were also "severely flawed", the Agency said. The company used a "Hanna" kit to analyse for cyanide - but because this measured free cyanide, rather than total cyanide as required by the authorisation, the company's own tests would have underestimated the level of cyanide by about a third. Moreover, the Agency found that the equipment was not properly calibrated.
Conspiracy to conceal
By 14 January 1998, the court heard, Mr Stevens appeared to be aware of increased levels of cyanide in the cooling water. A note sent by Mr Stevens to Mr Buckner and others set out concerns that the process change had led to a "marked increase" in cyanide levels - and jeopardised compliance with the site's discharge limits.
According to the Agency, Mr Buckner's attempt to bury the problem appears to have started soon afterwards. In April 1998, he submitted a quarterly return which accurately showed a breach on 19 January, with cyanide levels in the second discharge of 0.66mg/l as against a limit of 0.1mg/l.
However, Sevalco failed to notify the Agency of the breach immediately, as required by the authorisation. It later emerged that a letter had been drafted which drew attention to the breach, and explained that the new process was having an adverse impact on cyanide discharges. This letter appears to have been distributed to a number of middle managers - but was never sent to the Agency. The Agency told the court that this was evidence that the company was "on notice" that there was a problem.
In October 1998, Mr Buckner was promoted to general manager and director of the company. However, he continued to compile and submit the quarterly returns to the Agency.
In its investigation, the Agency compared the weekly results in the company's blue book with monthly data from Wessex Water and figures declared in quarterly returns. The exercise offered only a snapshot of the scale of the deception, because comparable data were available only for one in sixty discharges.
Nevertheless, on four occasions the quarterly returns bore "no resemblance" either to Wessex's data or to those in the blue book.
For instance, on 25 January 2000 the actual level was 17mg/l - seventeen times the authorised limit - while the quarterly returns claimed a figure of 0.91mg/l. Furthermore, the annual discharge in 2002 was found to be almost 0.167 tonnes, exceeding the authorised limit by 138%. The company failed to notify the Agency of the breaches.
The plot thickens
The Agency told the court how its investigation had found that Mr Stevens, the works chemist, was manipulating the operation of the plant in an attempt to distort Wessex Water's data.
Between March 2001 and June 2002, when he retired, Mr Stevens gave instructions to hold back the flow of wastewater from the medium thermal plant for 24 hours before any visit by Wessex Water - with the result that samples would be lower in cyanide. The Agency said that Mr Stevens' replacement, Jason Mitchell, continued with the practice.
In September 2001, Mr Buckner fell ill and went on sick leave. However, his replacement as general manager, Bob Sale, did not understand the system for reporting data, so he visited Mr Buckner at home. The Agency said that, in effect, Mr Bucker was still reporting.
Mr Mitchell told the Agency that he was instructed by Mr Stevens to record false data if necessary. Mr Mitchell complied, but said he was deeply troubled and recognised it was illegal to do so. Mr Mitchell told the Agency that he expressed concern on several occasions to his manager, Michael Shaddick, the quality assurance manager. According to the Agency, Mr Shaddick apparently said that nothing much could be done about it, that the discharge would not cause any harm and that the limit was too tight anyway.
Mr Mitchell later raised concerns with Roger Vowels, who became environmental manager in August 2002. The Agency alleged that Mr Vowels told Mr Mitchell to carry on falsely recording data.
In September 2002, Jeremy Bastin took over as quality manager. He was also concerned by the deception and told Mr Mitchell not to change the results. However, in October, Mr Vowels allegedly persuaded Mr Bastin to alter the records. After a pang of guilt, Mr Bastin restored the correct figure. Mr Vowels denied this version of events.
Eventually the game was up. In April 2002, Mr Sale was replaced as general manager by Nigel Cheeseright, a newcomer to Sevalco. The Agency accepts that Mr Cheeseright had no hand in the deception - and commended him for helping to raise the alarm. Indeed, when he discovered the situation he took prompt action to prevent further breaches of the cyanide limit.
On 15 October, Mr Cheeseright reviewed the cyanide data, realised something was wrong and started asking questions. On 22 October, he informed the Agency that there were discrepancies in the reported data. Eight days later, he revealed the results of an internal investigation at a meeting with the Agency.
Messrs Buckner, Vowels and Shaddick are no longer employed by Sevalco. Mr Mitchell and Mr Bastin were both disciplined but still work for the firm.
"Failure to act with due diligence"
The saga contrasts with the corporate values espoused by Columbian Chemicals, Sevalco's parent company. The company says it "strives to be a good neighbour and an environmental steward. Environmentally, this means that all our plants comply with or exceed all local and governmental environmental regulations." Its corporate "vision" says: "Doing what's right is being honest, ethical, and having personal and professional integrity."
In mitigation, Sevalco's barrister argued that "this is not a case where the company's systems were so inadequate that breaches of the authorisation were bound to occur and to remain undetected." The falsification of data occurred, he suggested, because senior employees breached the trust the company had placed in them.
However, in passing sentence at Bristol Crown Court, Judge Darwall-Smith disagreed. The company "had failed and failed badly to have the proper systems in place" to ensure that "such blatant behaviour was picked up", he said. Sevalco had "failed to act with due diligence".
In a statement released after the sentence, Sevalco said it "deeply regrets the events that took place between 1998 and 2002 under the previous management team."
Mr Cheeseright says that the company has now ensured clearer responsibilities for each managerial position to ensure the separation of operational and technical roles. Stephen Buckner had held several roles - presenting a conflict between the operational and technical compliance aspects of his job. Moreover, he had no formal qualifications in environmental management.
In court, Sevalco's own defence team speculated that Mr Buckner may have been motivated by his responsibility, as plant manager, for suggesting and promoting the process changes that led to the excessive levels of cyanide. As compliance manager, he was perfectly placed to conceal the problem.
However, throughout the case Mr Buckner has consistently denied any wrongdoing.
The Agency says that Sevalco had no written procedures relating to cyanide monitoring. Nor was there a written procedure for auditing the blue book or preparing quarterly reports.
Speaking after the case, Mr Cheeseright insisted that Sevalco's parent company, Phelps Dodge, takes its environmental responsibilities seriously and has an "extensive audit procedure". Every two to four years, a team of up to six auditors visits each site for five days to check processes, procedures and performance.
However, at least one such audit was carried out at Sevalco's Avonmouth site during the period covered by the prosecution, which apparently failed to uncover any of the management failures which subsequently came to light.
Mr Cheeseright said that the company has put in place "robust procedures" as a result of the case. These include a new environmental management system which, he hopes, will achieve ISO14001 certification early in 2005.
In November, Sevalco closed the medium thermal plant at the centre of the saga. Mr Cheeseright said that under current market conditions it was not financially viable to carry out the improvements needed to address the problem with the process.
Spotlight on the Agency
The tale also raises important questions about how the company pulled the wool over the Agency's eyes with such apparent ease for more than four years.
The Agency inspected the site on three or four occasions per year from 2000 to 2002. As a result of the case, it has stepped up its oversight of the works - with eight inspections in 2004.
Embarrassingly, in 2002 the Agency had rated Sevalco as a "well managed" site under its operator performance risk appraisal (OPRA) scheme. OPRA is the centrepiece of the Agency's efforts to "modernise" its oversight of industrial processes through introducing risk-based regulation.
Surprisingly, the Agency's OPRA score also gave the company maximum points for its procedures for "recording and use of information". In 2003, after the truth came out, the Agency reduced the site's overall management score from a B rating to a poor C rating - and revised the score for recording information to the minimum available.
It is surprising, however, that the Agency failed to ensure that the company had effective environmental management systems and procedures in place.
The Agency also failed to recognise the technical problem arising from increased levels of cyanide in the cooling water system for the dry filtration equipment. This change to the process was a substantial variation to Sevalco's authorisation, and the company was required to submit a detailed application to the Agency for approval.
Indeed, the Agency praised Sevalco for the changes to the process in a report on industry in Avonmouth, published in June 2002. The report quotes Stephen Buckner as saying: "The major improvements we made to air quality have been followed by major improvements to water quality. We have taken a big step forward in safeguarding and improving the environment."
The Agency also did not pick up on the inadequacies in Sevalco's testing procedures - notably its measurement of free cyanide rather than total cyanide. Indeed, the Agency carried out no independent "check" monitoring of the site's discharges to water in the five years before the falsification came to light.
With hindsight, the Agency also missed an early warning of the problems to come. In May 1998, Ron Owen, the inspector for the site at the time, wrote to Mr Buckner about the breach on 19 January. In the letter, Mr Owen accepted Mr Buckner's "assurances that the steps you have taken should prevent recurrence." He expressed concern about the failure to notify the Agency about the breach and warned that a repetition could result in enforcement action - but accepted "that there was no intention to hide the problems."
After the deception came to light, the Agency mounted an investigation into its regulation of Sevalco's site. It refused to let ENDS see the report on the grounds it was an "internal management document" containing "personal information" - but it disclosed limited information on the broad findings.
The main conclusion was that the Agency should have included sampling for cyanide and ammoniacal nitrogen in its water quality monitoring of the Severn estuary. This could have included a sample from Sevalco's discharge.
The investigation also recommended that the operator monitoring assessment (OMA) scheme should be extended to cover monitoring of liquid discharges and used in all Agency regions. The scheme, introduced three years ago, is a risk assessment tool designed to allow inspectors to check the effectiveness of companies' air emission monitoring programmes under IPC. Operators with poor scores are more likely to be subject to independent check monitoring (ENDS Report 313, pp 39-40 ).
Likewise, the investigation found that "compliance assessment plans" would provide a useful tool for helping to ensure compliance. Such plans will be linked to the OPRA scheme and used to specify the level of inspections and other regulatory oversight. The plans are due to be in place by April 2005 for sites regulated under IPPC.
However, it is not clear that the Agency's investigation addressed any of the wider shortcomings in its regulation of Sevalco. A spokesman refused to discuss the scope of the investigation, but claimed that it had been "very thorough".
The Sevalco case raises awkward questions about the Agency's effectiveness as a regulator on the ground. Similar concerns were flagged up two years ago after the exposure of serious shortcomings at Great Lakes Chemical's works in Newton Aycliffe (ENDS Report 334, pp 19-22 ).
The Agency is under enormous pressure to demonstrate that risk-based regulation can deliver more effective results at reduced cost. The idea is that well-managed, responsible companies should benefit from a "lighter regulatory touch", leaving inspectors free to focus on high risk sites. This was the theme of the Agency's recent "spotlight" report on business environmental performance (ENDS Report 355, pp 25-28 ).
But in implementing this agenda, the Agency will need to be confident that new tools such as OPRA and OMA are as robust as possible - and backed up by sufficient resources on the ground to ensure that the regulator is not so easily hoodwinked again.