This case is concerned with the extent of environmental cleanup or remediation that defendant must perform as a result of its violation of the Water Resources Commission Act (wrca). MCL . 323.1 et seq.; MSA 3.521 et seq. Plaintiffs appeal from an Alger Circuit Court order which recited the parameters of defendant’s remedial action plan.
Plaintiffs argue on appeal that the trial court erred in concluding that the hazardous substance *151within the aquifer1 was being removed by naturally occurring biological or chemical processes. They assert, also, that the trial court erred in concluding that the substance in the aquifer and soil was not a carcinogen. We reverse.
i
Defendant owns a wood treatment plant. It pressure-treats wood using, over the years, various preservative compounds including pentachlorophenol (pcp). Defendant stopped using pcp in 1980. In 1981, plaintiffs Michigan Water Resources Commission and Michigan Department of Natural Resources (dnr) entered into a consent order with defendant.
The order prohibited defendant from using pcp as a wood treating agent. It mandated defendant to develop and complete a hydrogeologic study of the area where defendant had allowed chemicals to enter the ground. It required a remedial action plan. The plan had to include alternative remedial actions for removal, containment and treatment of all contaminated soils and groundwater identified by the hydrogeological study.
In 1986, plaintiffs filed a complaint against defendant for injunctive relief alleging that defendant failed to comply with the consent order. Plaintiffs complained that defendant’s discharge of hazardous substances into the groundwater violated the wrca, the Hazardous Waste Management Act and the Michigan Environmental Protection Act. MCL 323.1 et seq.; MSA 3.521 et seq.; MCL 299.501 et seq.; MSA 13.30(1) et seq.; MCL 691.1201 et seq.; MSA 14.528(201) et seq._
*152Defendant agreed that it had violated § 6(1) of the wrca and that remediation was necessary-under § 6(4). MCL 323.6; MSA 3.526. In exchange for defendant’s admissions, plaintiffs dismissed their other charges. The only issue left for the trial court’s determination was the extent of remediation needed from defendant.
ii
A
The trial court found that defendant had satisfied the remedial requirements of the Environmental Contamination Response Activity rules. 1990 AACS, R 299.5101 et seq. On appeal, plaintiffs argue that the finding was erroneous.
The rules, which were promulgated under the Emergency Response Act,2 are relevant to violations of the wrca; they apply to all known sites of environmental contamination. See MCL 323.6(4); MSA 3.526(4). Rule 7053 establishes the general characteristics of remedial action needed at a contaminated site.
Plaintiffs assert that defendant failed to document that pcp was decreasing as a result of naturally occurring biological or chemical processes. Plaintiffs’ witness Nadine Romero, a dnr hydro-geologist, testified that she had been involved in the comprehensive monitoring of defendant’s facility. Seventeen monitoring wells were drilled, and data from the wells were used to map the pcp concentrations in the groundwater. A contami*153nated pcp plume4 was identified. The point of most serious pcp contamination contained concentrations of 6000 parts per billion. The further south the well, the less the pcp concentration. Groundwater in the area flowed southerly to southwesterly.
Romero testified that pcp could be biologically degraded into various components including trichlorophenols, dichlorophenols and phenols. None of these components was detected during sampling. Romero did not believe that pcp could biodegrade naturally. Instead, some type of bacteria needs to be introduced into the groundwater to begin the breakdown process.
Defendant’s expert Dr. Neil Hutzler testified that pcp is biodegradable, meaning that micro-organisms could break its chemical structure. He stated that the concentration of pcp in the groundwater decreased as it moved southward. This decrease "probably” had two causes. One was dilution through infiltration of water coming from the surface. The other was dispersion into existing groundwater.
Hutzler detected no pcp contamination near the southern end of defendant’s property. One explanation for this was that there were no wells drilled along the center line of the contaminated pcp plume. Another was that the pcp concentration had been diluted to a point where it was no longer detectable. Although he tested for biodegradation of pcp in the soil, he did not test for it in the groundwater.
Defense witness David Williams testified that the concentration of pcp in the groundwater was declining due to dilution and possibly biodegrada*154tion. Williams had done no studies to determine whether pcp was being biodegraded in the aquifer. He assumed that biodegradation was taking place in the groundwater. This assumption was based on the fact that the concentration of pcp in the soil was being biodegraded, and the source of the pcp in the groundwater was the leaching of pcp from the soil. Williams never found any breakdown products of pcp in the groundwater.
The trial court made the following findings:
Clearly, some process is taking place which significantly reduces the groundwater Penta [pcp] over a fairly short distance. At dnr 13 the 1990 concentration was 6,000 parts per billion; some 250' southwest along the downward gradient of the groundwater, the concentration of Penta at dnr 9 and 12 was 1,000 parts per billion; approximately 200' further southwest at dnr l, 2, 3, the level was 530 parts per billion; approximately 125' further southwest at ow ll, Dr. Hutzler found 260 parts per billion. The next test well is some 500' further southwest along the gradient (dnr 4) and of course no Penta was found there or at any of the wells south of ow 11; ow 11 is more than 1,000' northeast of the south boundary of the site. The 6,000 parts per billion to 260 parts per billion over 600', and the absence of any detectable Penta in the 1,000' south of ow 11, could be explained at least in part by dilution; rainwater and melted snow can wash down through soil uncontaminated with Penta, thus diluting the contaminated plume coming from the northeast; further, the parties hypothesize that some Penta has leached out of the soil since 1980, such that the plume is probably not being renewed to the same level as might have been true 10 years ago. Although Ms. Romero tends to discount an additional possibility, both Dr. Hutzler and Mr. Williams are of the opinion that natural biodegradation, at least in the soils, might also be contributing to the sharp decrease in concentration of Penta in the groundwater. The re*155quirements of R 299.5705(6) are being satisfied, albeit marginally.
II
B
We must determine whether it was erroneous to conclude that there was documented evidence establishing that the concentration of pcp was decreasing due to naturally occurring biological or chemical processes. Rule 705(6). The trial court’s findings of fact in an equity action can be set aside only if they are clearly erroneous. Beason v Beason, 435 Mich 791, 803; 460 NW2d 207 (1990). However, where a finding is derived from an erroneous application of law to fact, we are not limited to review for clear error. Id., 805.
In construing administrative rules, courts apply principles of statutory construction. Coalition for Human Rights of the Handicapped v Dep’t of Social Services, 431 Mich 172, 185; 428 NW2d 335 (1988), citing General Motors Corp v Bureau of Safety & Regulation, 133 Mich App 284; 349 NW2d 157 (1984). However, if the agency’s rule is clear and unambiguous, judicial interpretation is neither required nor permitted, and the court should not look beyond the ordinary meaning of the unambiguous language. Jordan v Dep’t of Corrections, 165 Mich App 20, 28; 418 NW2d 914 (1987); Michigan Mutual Ins Co v Allstate Ins Co, 146 Mich App 475; 382 NW2d 169 (1985), aff'd 426 Mich 346 (1986). If construction is warranted, we must construe the rule according to the common and approved usage of the language, resorting to dictionary definitions where appropriate. See In re Forfeiture of One Front-End Loader, 192 Mich App 617; 481 NW2d 791 (1992).
*156The crucial question in this case is whether defendant properly documented that pcp was being removed through naturally occurring biological or chemical processes. We find that the term "documented” is clear and unambiguous. It requires more than conjecture and speculation. A conjecture is simply an explanation, consistent with known facts or conditions, but not deducible from them as a reasonable inference. Kaminski v Grand Truck W R Co, 347 Mich 417, 422; 79 NW2d 899 (1956).
In this case, both Hutzler and Williams opined that pcp was diluting in the aquifer. Williams also believed the level of pcp was decreasing due to biodegradation. Hutzler insinuated that pcp could not naturally biodegrade in the aquifer. Both failed to provide the trial court with any documented evidence supporting their opinions. Neither witness tested the groundwater to determine if the pcp was being biodegraded or diluted.
A decrease in the concentration of pcp is consistent with a finding of dilution5 or biodegradation. However, a decrease could also be consistent with other findings, for example, as Hutzler suggested, that the wells were not in the proper location to detect contamination. The only test for biodegradation was performed by Romero. She found nothing to suggest that pcp was naturally biodegrading in the aquifer. Therefore, we conclude that defendant *157failed to document that the pcp was being removed by naturally occurring biological or chemical processes.
We disagree with defendant’s argument that the trial court did mandate active remediation and that it was sufficient to meet the requirement of Rule 705(6). The rule specifically provides that remedial actions "shall provide for the removal” of the hazardous substance from the aquifer. The capping of the soil and the installation of additional groundwater monitoring wells does not provide for the removal of pcp. Capping of the soil may prevent the further leaching of pcp into the groundwater, but does nothing to "remove” the pcp which currently exists.
It is clear to us that the trial court erred in applying the law to the facts of this case when it determined that the requirements of Rule 705(6) had been met. See Beason, 805. There was no documented evidence that the pcp concentration was decreasing due to naturally occurring biological or chemical processes. Moreover, there was no evidence establishing that the contemplated remediation could remove the pcp from the acquifer. Therefore, we reverse the trial court’s order on this issue.
m
A
Plaintiffs argue also that the trial court erred in determining that pcp is not a carcinogen. The level of cleanup at a contaminated site is determined in part by the carcinogenicity of the contaminant. The level of remediation necessary for carcinogenic substances in an aquifer is much higher than for noncarcinogenic substances. Cf. 1990 AACS, R 299.5709-299.5717.
*158Plaintiffs’ expert, Bradley Venman, testified that pcp is a carcinogen. His opinion was based upon two separate lifetime bioassay6 studies for cancer research conducted by the National Toxicology Program (ntp). A lifetime bioassay study attempts to expose animals to a contaminant for a significant portion of their lifetime. Venman recognized that five previous studies showed that pcp is not a carcinogen. However, he opined that only two of them were lifetime bioassay cancer studies and are of limited value due to their design and implementation.
The trial court provided the following reasoning in concluding that Venman failed to establish that pcp is a carcinogen:
First, his conclusion is supported only by a single report done by a toxicology testing laboratory for the National Toxicology Program. I doubt that such a single report qualifies for the "weight of evidence” conclusion reached by Mr. Venman. . . . There are several other points which would lead to the conclusion that Penta is not a carcinogen. First, and this is perhaps an ad hominem argument, Mr. Venman prepared draft rules in February of 1990, which have evolved into R 299.5101 et seq., in which Penta was listed in a category of non-carcinogens. All of the longer term studies done prior to the National Toxicology Program were not positive for carcinogenic affects [sic]. In 1980, the epa ambient water quality criteria permitted Penta concentrations of 1,000 parts per billion, and the epa has not changed that criteria; it would have done so if Penta is a carcinogen. As noted above, the Office of Drinking Water of the epa apparently does not treat Penta as a carcinogen, since it allows toxicity levels of not more than 200 parts per billion, and aesthetic *159levels of not more than 30 parts per billion. Further, in 1989 the Office of Drinking Water re-proposed a Maximum Contaminant Level goal of 200 parts per billion; the office only requested comments on a proposed mclg of zero, appropriate if Penta is a carcinogen.
Given this state of scientific knowledge and opinion, I cannot conclude that Penta is a carcinogen. [Citations omitted.]
Ill
B
We will determine that a finding is clearly erroneous if, on all the evidence, we are left with a definite and firm conviction that a mistake has been made. Beason, 805. In this case, after reviewing the entire record, we conclude that a mistake was made.
Venman was the only individual testifying to the carcinogenicity of pcp. Contrary to the trial court’s finding, Venman relied on two separate lifetime bioassay studies, not one. He distinguished the earlier studies which had concluded that pcp is not a carcinogen. He testified about action by the United States Environmental Protection Agency (epa) Office of Drinking Water. He stated that the office had requested comments on changing the maximum contaminant level goal (mclg) of pcb to zero based on the proposed reclassification of pcp as a carcinogen.
After the entry of judgment, plaintiffs filed a motion for new trial alleging newly discovered evidence which proved that pcp is a carcinogen. They attached a portion of the Federal Register for December 6, 1990, to their motion for new trial. 55 Fed Reg 50450, 50458 (1990).
The epa, relying on the same bioassay study *160used by Venman, had promulgated a final rule and published it in the Federal Register. It designated pcp as a carcinogen. Id. Plaintiffs also attached a portion of the Federal Register for January 30, 1991. 56 Fed Reg 3526, 3607 (1991). In that Federal Register, the Office of Drinking Water proposed final rules establishing the mclg for pcp at zero. Id. The trial court relied on the previous mclg of 200 parts per billion when making its findings in this case.
The trial court denied the motion for new trial, but without prejudice to plaintiffs’7 right to seek further review and potential relief pursuant to the court’s continuing jurisdiction. The trial court noted that, if new evidence became available such that the remedial action plan no longer accorded with current scientific knowledge, modifications to the plan may become necessary. The court stated it did not view further proceedings as being a new trial of issues already litigated, but rather an exercise of continuing jurisdiction.
Based on our review of the entire record, we are left with a definite and firm conviction that a mistake has been made. It is clear to us that pcp is a carcinogen. The fact that the trial court had continuing jurisdiction of this case does not resolve the matter. The trial court suggested in its order denying a new trial that the evidence presented in support of it was insufficient to mandate a revision to the remedial action plan. We disagree. The trial court should have required defendant to remediate the site in accordance with the rules for carcinogens. Rules 709-711.
Reversed and remanded. We do not retain jurisdiction._