In a bench ruling, the United States Court of Federal Claims awarded P.W. Construction, Inc. (PWCI) money damages in connection with a construction contract. Because the record does not support the grant of damages for lost productivity, additional landscaping, and additional pipe, this court vacates those awards. Because the trial court erred in holding the Government liable for PWCI’s above-bid costs incurred to secure valves under the contract, this court reverses the grant of damages on that issue. Accordingly, this court vacates-in-part, reverses-in-part, and remands.

The Government awarded PWCI a fixed-price contract for the installation of natural gas pipe, valves, meters, and pressure stations at Vandenberg Air Force Base in California. During performance of the contract work, PWCI encountered unexpected underground obstructions. PWCI alleged that these obstructions delayed project completion and increased completion costs. PWCI sued the Government for $1,700,000 in damages. Following closing arguments, the Court of Federal Claims ruled from the bench, finding the Government hable for lost productivity, additional landscaping, and additional pipe used to route around the obstructions. The court also found the Government liable for additional costs incurred because the Government refused to allow PWCI to substitute KOSO earthquake valves for the required Quake Master valves.

DISCUSSION

This court reviews the Court of Federal Claims’ factual findings for clear error, Hankins Const. Co. v. United States, 838 F.2d 1194, 1195 (Fed.Cir. 1988), and its legal determinations without deference, Dairyland Power Coop. v. United States, 16 F.3d 1197, 1201 (Fed.Cir. 1994). The trial court’s selection of a method for determining damages is a legal question, which this court reviews “non-deferentially on the basis of reasonableness.” Dawco Const., Inc. v. United States, 930 F.2d 872, 880 (Fed.Cir. 1991), rev’d in non-relevant part by Reflectone, Inc. v. Dalton, 60 F.3d 1572 (Fed.Cir. 1995). This court reviews damage calculations by the Court of Federal Claims for an abuse of discretion. Hughes Communications Galaxy, Inc. v. United States, 271 F.3d 1060, 1065 (Fed. Cir. 2001). A trial court abuses its discretion when: “(1) the court’s decision is clearly unreasonable, arbitrary or fanciful; (2) the decision is based on an erroneous construction of the law; (3) the trial court’s factual findings are clearly erroneous; or (4) the record contains no evidence *557upon which the trial court rationally could have based its decision.” Id.

Court of Federal Claims Rule 52 states that “[i]n all actions tried upon the facts, the court shall find the facts specially and state separately its conclusions of law thereon.” See also Fed.R.Civ.P. 52(a) (2002). The purpose of this rule is, inter alia, to provide this court with an adequate basis for meaningful review. Cf. Pretty Punch Shoppettes, Inc. v. Hauk, 844 F.2d 782, 784, 6 USPQ2d 1563, 1565 (Fed.Cir. 1988). Thus, the Court of Federal Claims may not merely state its findings in conclusory terms, but must provide sufficient explanation as to how it reached its conclusions on pertinent factual and legal issues. This court may not speculate about the trial court’s ultimate findings of fact or conclusions of law. If the Court of Federal Claims does not state a decision in terms sufficient for meaningful review, this court generally will vacate. Cf. Gechter v. Davidson, 116 F.3d 1454, 1458 (Fed.Cir. 1997).

Using the measured mile method, PWCI’s expert, David Ough, presented an estimate of PWCI’s lost productivity attributed to the underground obstructions. Based on his calculations, he testified that PWCI was entitled to $505,120.37 in lost productivity. Using Mr. Ough’s productivity rates, the Court of Federal Claims determined that PWCI was entitled to $260,830.10 in lost productivity. Although the trial court’s calculations corrected a mathematical error in Mr. Ough’s estimate, the trial court still relied on Mr. Ough’s productivity rates for its calculations. Based on Mr. Ough’s rates, the trial court found the Government liable for 48% of PWCI’s labor hours. Those rates, however, have several problems.

The primary problem with the productivity rates is that they do not compare equivalent types of work. See Thomas E. Shea, Proving Productivity Losses in Government Contracts, 18 Pub. Cont. L.J. 414, 426 (1989) (“The type of work must be equivalent for the pre- and post-disruption periods in order to allow a fair comparison. It would not be valid to compare the rate for the production of large widgets during the pre-disruption (normal) period with the production of smaller widgets during the disruption period.”). The pre-disruption period rate in this case did not include intensive welding and trenching work performed by subcontractors, but the impaired rate included such work. Additionally, the rates did not consider whether productivity would be affected if PWCI used steel versus polyethylene pipe. The trial court recognized this problem, but did not correct for the error.

The record shows that welding in the impaired period was butt-welding on polyethylene pipes, which takes only 15 seconds to 2 minutes per weld, whereas the welding done in the pre-disruption period was steel welding, which may take up to 2.69 hours per weld. Without more, this evidence suggests that a comparison of the pre-and post-disruption periods must take into account the difference in welding costs. To account for this difference, Mr. Ough deleted both the welding work and the trenching work from the pre-disruption rate and left both kinds of work in the post-disruption period. Mr. Ough’s deletion thus yields a rate that reflects a higher productivity than actually enjoyed by PWCI. PWCI argues that this deletion from the ideal rate has a negligible impact on the overall productivity rates.

In the first place, Mr. Ough’s calculation deleted trenching work from the ideal period, but included the same work in the impaired period. Furthermore, Mr. Ough’s method deleted all welding time from the ideal period, but included the butt-welding time in the impaired period. *558The two periods thus have not made a correct comparison, as acknowledged by the trial court, which nonetheless used these rates to calculate damages. Indeed, the Government submitted unrebutted evidence that adding the welding costs back into the ideal period yields an ideal rate of 0.171 hours per linear foot, a rate very close to the impaired rate. This testimony suggests that the inclusion of welding and trenching costs for the different pipes is not negligible.

Because the impaired rate accurately reflects productivity during the impaired period, but the ideal rate does not accurately reflect productivity during the ideal period, this court vacates the damage award on lost productivity. The rates must account for the differences in welding and trenching costs for the different pipes. On remand, the trial court will have the opportunity to adjust Mr. Ough’s rates to more accurately reflect productivity during both the ideal and impaired periods.

The Court of Federal Claims appears to have applied a modified total cost method to calculate the additional landscaping damage award. The trial court made no express determination that PWCI was entitled to prove damages using the modified total cost method. Although the record shows that PWCI spent substantially less for landscaping than it expected when it bid, PWCI made a claim for landscaping damages based on the total cost method. Without explanation, the trial court awarded $20,000 in landscaping damages.

The Government asserts that PWCI did not show the reasonableness of its bid and thus cannot claim damages under a modified total cost method. PWCI responds that Mr. Marshall, its president, explained the reasonableness of its bid in his testimony. Despite PWCI’s assertions, Mr. Marshall’s testimony does not explain the reasonableness of PWCI’s bid. Because the record on appeal does not show that PWCI’s bid was reasonable, this court vacates the landscaping damage award. On remand, the trial court will have the opportunity to test the evidence relative to the reasonableness of PWCI’s bid.

The bench ruling does not indicate whether the trial court applied a modified total cost or jury verdict method to award damages for additional pipe. PWCI based its bid on an estimate of 165,000 feet of pipe, but it actually purchased about 185,-000 feet of pipe. Thus, Mr. Ough’s report listed an overrun of about 20,000 feet. Finding that the Government had already compensated PWCI for some of the additional pipe, the trial court granted PWCI an award based on an additional 12,000 feet of pipe at a blended rate of $16.17 per foot, for a total award of $194,040. The trial court, however, did not state how it arrived at the 12,000 feet overrun or the $16.17 blended rate.

Based on the record, this court is unable to determine how the trial court reached the price of $16.17 per foot. According to Mr. Ough’s report, the pipes cost between $0.14 and $18.20 per foot. The total average cost of the pipe used is slightly over $2.00 per foot. The total average cost of the pipe purchased in 1996 (the impaired period) is about $3.63 per foot. The record shows that only about 950 of the 185,-000 feet of pipe cost over $7.00 per foot. Further, only 121 feet cost over $16.00 per foot. Thus, the record does not support a blended rate of $16.17. This blended rate is skewed toward the much higher cost of steel pipe (66 cents to $18.20 per foot) when PWCI primarily used polyethylene pipe (14 cents to $10.35 per foot) during the impaired period. Given the lack of evidentiary support and the lack of any explanation by the trial court, this court *559holds that the trial court abused its discretion in setting the cost of the pipe at $16.17 per foot.

Additionally, the record evidence does not support a finding of 12,000 feet of additional pipe. Ronald Marchelletta, a consultant for PWCI, estimated from the as-built plans that an additional 7,000 feet of pipe was installed beyond that depicted in the contract plans. Mr. Marshall testified that downward and upward deflections of pipe to avoid the 700 to 800 unanticipated obstructions added another four feet per obstruction. Based on Mr. Marshall’s testimony, a total of 3,200 feet of pipe (800 x 4 = 3,200) was used to route around obstructions. Even construing the record in favor of PWCI, the record shows that PWCI accounted for at most 10,200 feet of additional pipe.

Given the arbitrary nature of the additional pipe award, this court vacates and remands for the trial court to recalculate that award. On remand, the trial court will have the opportunity to ascertain the type of additional pipe used and the cost of that pipe. The trial court also may establish more accurately the amount of additional pipe.

The trial court also granted PWCI damages flowing from the Government’s refusal to allow PWCI to substitute KOSO earthquake valves for the required Quake Master valves. The project specification called for seismic shutoff valves, also known as earthquake valves, which stop the flow of gas during seismic activity. The specification required that “[t]he maximum allowable working pressure [for the valves] shall be 410 kPa (60 psi).” The specification included a Material and Workmanship Clause. This clause provides that references to brand names in the specification establish only a standard of quality. FAR § 52.236-5(a). The clause further states: “The Contractor may, at its option, use any equipment, material, article, or process that, in the judgment of the Contracting Officer, is equal to that named in the specifications, unless otherwise specifically provided in this contract.” Id. (emphasis added). Thus, substitute goods may satisfy contract requirements if they are equal to the specified items in the judgment of the contracting officer. Where only one manufacturer’s product or material could literally meet the specifications, the contractor can rely on this clause to support the use of a substitute. Manning Electric & Repair Co. v. United States, 22 Cl.Ct. 240, 245 (1991).

In preparing its bid, PWCI discovered that the specification could be satisfied using only Quake Master valves. PWCI sought permission from the Government to substitute KOSO values for the required Quake Master valves. The Government’s architectural and engineering (A/E) firm, J.C. Chang & Associates, informed the Government that only the Quake Master valves meet the specification as written, but that the KOSO valves “[were] acceptable substitutions.” An internal Government memorandum stated that the KOSO valves were “of a lesser quality” because Quake Master is a 60 PSI rated valve while KOSO is a 20 PSI rated valve. The memo noted, however, that the A/E firm had said the KOSO 20 PSI rating was “satisfactory.” The contracting officer refused to approve the KOSO substitute, and PWCI spent an additional $71,158.63 over its bid to secure Quake Master valves. The Court of Federal Claims awarded the full $71,158.63 to PWCI.

When a contractor proposes a substitute for an specified product, it must show: 1) that the specification is proprietary; 2) that it submitted information showing an equal substitute; and 3) that the substitute is the functional equivalent or of the same *560standard of quality as the specified product. Manning, 22 Cl.Ct. at 245; American Renovation and Const. Co. v. United States, 45 Fed. Cl. 44, 49-50 (1999). The Manning opinion also states that a CO’s decision to reject materials as not compliant with specifications will be upheld unless it was arbitrary and capricious. Id. An earlier Claims Court opinion states: “[A] contracting officer does have to exercise judgment to determine whether the item proposed to be substituted is equal in quality and performance to the designated proprietary product, but his power does not extend to a refusal to allow a replacement which is equal in these respects.” Jack Stone Co. v. United States, 170 Ct.Cl. 281, 344 F.2d 370, 374 (Ct.Cl. 1965). Given this earlier precedent, the Manning reference to arbitrary and capricious applies only to the third “functional equivalent” prong. In other words, the contracting officer (CO) must exercise discretion in deciding whether the proposed equivalent is in fact equal to the specified item, but the CO cannot arbitrarily reject a substitute shown to be equal to the specified item. Although “equal” does not mean identical, it does mean equal in permanence, quality, and other essential features. In deciding whether the proposed substitute is equal to the specified item, the CO must look to the requirements specified in the contract or project specifications and may not rely on undisclosed requirements. Manning, 22 Cl.Ct. at 246-47.

“It is settled that the Government is entitled to obtain precisely what it contracts for as long as it does not mislead the contractor.” J.L. Malone & Assoc., Inc. v. United States, 879 F.2d 841, 845 (Fed.Cir. 1989) (quoting American Elec. Contracting Corp. v. United States, 217 Ct.Cl. 338, 579 F.2d 602, 608 (Ct.Cl. 1978)). In this case, the parties dispute whether the KOSO value is equal to the required Quake Master valve (i.e., the “functional equivalent” prong). The record shows that the KOSO valve was not equal or functionally equivalent to the Quake Master valve because of the KOSO valve had a lower pressure rating. The specification expressly stated that the project valve must have a maximum pressure rating of 60 PSI. PWCI was not misled here because the specification expressly recited a 60 PSI rated valve. A 20 PSI valve is not functionally equivalent to a 60 PSI valve. The Government is entitled to refuse substitution with a valve that is not equal to the specified 60 PSI valve.

Although the A/E firm considered the KOSO valve to be an “acceptable substitution!],” the Government found the KOSO valves to be “of a lesser quality.” “Acceptable” clearly is not the same as “equal” or “functionally equivalent.” If the Government contracts for a Cadillac, the Government is entitled to a Cadillac or another car that is equal to a Cadillac. Although any vehicle that runs may be “acceptable,” that does not make every operating vehicle equal in performance or quality to a Cadillac. As recognized by both Manning and Jack Stone, a CO “does have to exercise judgment to determine whether the item proposed to be substituted is equal in quality and performance to the designated proprietary product.” Jack Stone, 344 F.2d at 374. It is that exercise of judgment that should be upheld unless arbitrary and capricious. Because the trial court clearly erred in finding that the Government’s determination of inequality was arbitrary and capricious, this court reverses the damages award on this issue.

With the instructions noted above, this court vacates in part, reverses in part, and remands to the trial court for further proceedings.

*561COSTS

Each party shall bear its own costs.