Opinion for the court filed by Circuit Judge BUCKLEY.
BUCKLEY, Circuit Judge:The Trans Alaska Pipeline System (“TAPS”) provides the sole means of shipping petroleum produced from the North Slope of Alaska south to the Port of Valdez, Alaska. Because there are multiple shippers and only a single pipeline, TAPS commingles the various shippers’ petroleum. Necessity dictates that TAPS return to shippers a portion of that “common stream” at Valdez, regardless of whether their contributions were more or less valuable than the resulting mixture. The TAPS “Quality Bank” is an accounting arrangement approved by the Federal Energy Regulatory Commission (“FERC” or “Commission”) that makes monetary adjustments between shippers in an attempt to place each in the same economic position it would enjoy if it received the same petroleum at Valdez that it delivered to TAPS on the North Slope. To accomplish this, the Quality Bank charges shippers of relatively low-quality petroleum who benefit from commingling and distributes the proceeds to shippers of higher quality petroleum whose product is degraded by commingling.
While the concept is simple enough, the devil is in the details: it is difficult to determine which contributions improve or degrade the value of the common stream, and to what extent. The operators of the pipeline must employ a method of estimating the value of various contributions to the common stream and for determining the relative values of the petroleum products delivered at Valdez. This methodology, which the Commission must approve pursuant to its authority under the Interstate Commerce Act (“ICA”), 49 U.S.C.App. §§ 1 et seq. (1988); see also 42 U.S.C. § 7172(b) (1988) (transferring authority to regulate oil pipeline rates under the ICA from the Interstate Commerce Commission to FERC); Exxon Pipeline Co. v. United States, 725 F.2d 1467, 1468 n. 1 (D.C.Cir.1984) (explaining transfer of authority), is embodied in tariffs filed by the owners of TAPS (“TAPS Carriers”).
In 1993, FERC determined that due to changed circumstances the existing Quality Bank valuation methodology was no longer just and reasonable; and it consequently ordered a new one to be implemented. Trans Alaska Pipeline System, 65 F.E.R.C. ¶ 61,-277 (1993) (“1993 Order”). Various shippers filed petitions for review, claiming that aspects of the new methodology violated substantive provisions of law or were arbitrary and capricious and thus violated the Administrative Procedure Act (“APA”), 5 U.S.C. § 706(2)(A) (1994). We consolidated these *685petitions and now grant them in part and deny them in part. We find that the Commission was justified in ordering a change in the Quality Bank valuation methodology and in declining to order certain refunds. We also find, however, that two aspects of the new methodology and the Commission’s claim that it lacked jurisdiction to consider one shipper’s complaint do not comport with the APA’s requirement of reasoned decision-making.
I. Background
A. The TAPS Quality Bank
TAPS is a 48-inch diameter pipeline that extends nearly 800 miles from its origin on Aaska’s North Slope near Prudhoe Bay to its terminus at Valdez on Aaska’s south central coast. The pipeline is jointly owned by seven TAPS Carriers. Affiliates of some of the TAPS Carriers constitute a subset of the group of companies that ship petroleum through the line. TAPS carries a mixture of crude oils and natural gas liquids (“NGLs”) from a series of North Slope oil fields. The Quality Bank makes monetary adjustments among the shippers to compensate for the commingling of differing qualities of crude oil.
The Quality Bank operates at three locations. At Pump Station No. 1, located at the Prudhoe Bay origin of the pipeline, the Bank values the petroleum streams delivered to TAPS by the various shippers. It charges some shippers and makes payments to others based on the difference in value between their individual contributions and the weighted average of all incoming streams. More than 400 miles south of Prudhoe Bay, at the junction of TAPS and the Golden Valley Electric Association pipeline (“GVEA”) near Fairbanks, refineries operated by petitioners MAPCO Aaska Petroleum, Inc. (“MAPCO”) and Petro Star, Inc. (“Petro Star”) divert a portion of the common stream and remove certain petroleum products from it. That portion of the common stream less the products removed, known as the refinery “return stream,” is then returned to TAPS. At GVEA, the Quality Bank compares the value of the diverted portion of the common stream to that of the return stream, charging the refiners and compensating other shippers for the reduction in the common stream’s value caused by the removal of the refinery products. Finally, at the Port of Valdez, TAPS returns the common stream to the shippers in amounts proportionate to the quantity of petroleum they originally delivered to the pipeline. Because there are minor daily fluctuations in the value of the petroleum delivered at Valdez, the Quality Bank makes price adjustments based on the difference in value between the petroleum received by a shipper on a given day and the average value of the common stream at Valdez over the course of the month. Thus shippers who receive a tanker-full of oil of a higher-than-average quality will make a payment to the Quality Bank so that it may in turn compensate those who receive oil of a lower-than-average value.
In 1984, following years of litigation between the TAPS Carriers and MAPCO over the valuation methodology used by the Quality Bank, FERC approved a settlement between the parties that embodied a notably simple approach. Trans Alaska Pipeline System, 29 F.E.R.C. ¶ 61,123 (1984) (“1984-Order”). Because lighter, high gravity crude oil (as gravity is measured on the American Petroleum Institute (“API”) scale) is generally more valuable than a heavier, low gravity crude, the settlement proposed to equate the gravity of the petroleum with its value: contributors of petroleum having a gravity higher than that of the TAPS common stream would receive payments from the Quality Bank while contributors of petroleum having a gravity lower than that of the stream would make payments to the Bank. Under this system, known as the “intra-field gravity differential” methodology, the amounts of these payments were calculated using the adjustments to the posted prices for variations in gravity appearing in the postings for a number of Texas and California crude oils having a range of gravity that includes the average API gravity of the TAPS commingled stream. Id. at 61,289.
Tesoro Aaska Petroleum Co. (“Tesoro”), a TAPS shipper, contested the settlement on the ground that the gravity of petroleum is an inaccurate measure of its value. Tesoro favored a “distillation” methodology that *686would value the petroleum based on the boiling point of various hydrocarbons in the streams. Id. In approving the settlement over Tesoro’s objection, FERC conceded that there is no perfect valuation methodology and that other approaches might produce more accurate measurements than the one proposed by the settlement. Nevertheless, the Commission found that the proposed gravity method passed the threshold test of being “just and reasonable.” Id. The Commission noted that Tesoro or any other interested party had the ability to propose another methodology in the future. Id. at 61,240.
B. The Challenge to the Gravity Methodology
In 1989, the issue of the Quality Bank valuation methodology reemerged. Following routine practice, the TAPS Carriers filed new tariffs, to be effective July 1,1989, which proposed slight adjustments in the value of TAPS streams consistent with changes in market prices of crude oils of various gravities. See Amerada Hess Pipeline Corp., 47 F.E.R.C. ¶ 62,336 at 63,654-56 (1989) (“Pipeline Board Order”). Petitioner OXY USA, Inc. and Conoco, Inc. challenged the filings as unjust and unreasonable, alleging both that the Carriers improperly calculated the new gravity values and that they had violated the terms of their previously filed tariffs by permitting shippers to include NGLs in the petroleum they shipped through TAPS. Id. at 63,655.
Underlying the OXY/Conoco challenge, and critical to the petitions for review now before us, is the concern of some shippers over the increase in the amount of NGLs shipped through the pipeline between 1984 and 1989. NGLs have a very high gravity relative to other petroleum products but, according to the critics of the gravity methodology, they do not have as high a value as that attributed to them by that methodology. The critics believe that shippers with a high percentage of NGLs in their petroleum streams were actually reducing the value of the common stream but were being compensated by the Quality Bank as if they were increasing its value.
FERC’s Oil Pipeline Board suspended the new rates for one day and then permitted them to become effective subject to refund following a hearing concerning their lawfulness, id. at 63,656, pursuant to the Commission’s authority under section 15(7) of the ICA. That section provides that
in case of a proposed increased rate or charge ... the Commission may by order require the interested carrier or carriers to keep accurate account ... and upon completion of ... hearing and decision may by further order require the interested carrier or carriers to refund, with interest ... such portion of such increased rates or charges as by its decision shall be found not justified.
49 U.S.C.App. § 15(7) (1988). On appeal, the Commission affirmed the Oil Pipeline Board’s decision and explained that it would order refunds, retroactive to the date the rate adjustments were filed, should a hearing before an administrative law judge (“ALJ”) reveal that the adjustments were incorrectly calculated or that the TAPS Carriers were in violation of the terms of their tariffs by transporting NGLs, but stated that “because the TAPS owners ha[d] not proposed to change the existing methodology, any change in [the Quality Bank valuation] methodology [would] be effected prospectively.” Trans Alaska Pipeline System, 49 F.E.R.C. ¶ 61,-349 at 62,264-65 (1989) (“1989 Order”)-, see also Trans Alaska Pipeline System, 51 F.E.R.C. ¶ 61,062, at 61,137 (1990) (“1990 Order”). That said, the Commission ordered an investigation into the lawfulness of the Quality Bank gravity methodology pursuant to section 13(2) of the ICA, 1989 Order, 49 F.E.R.C. ¶ 61,349 at 62,265, which grants the Commission
full authority and power at any time to institute an inquiry ... concerning ... any of the provisions of this chapter ... on its [own] motion ... including the power to make and enforce any order or orders ... excepting orders for the payment of money-
49 U.S.C.App. § 13(2) (1988).
In November 1991, an ALJ assigned to investigate both the OXY/Conoco objections to the new tariff filings and the underlying *687reasonableness of the Quality Bank valuation methodology issued his opinion on all of the issues raised. First, he determined that the Quality Bank’s past charges were proper and that there had been no violations of the TAPS Carriers’ tariffs because the shipment of NGLs was consistent with the terms of the FERC-approved 1984 settlement. Trans Alaska Pipeline System, 57 F.E.R.C. ¶ 63,-010 at 65,041-42 (1991) (“ALJ Decision”). After reexamining the Quality Bank’s valuation methodology, however, he determined that “the evidence indicates that the current straightline gravity basis for valuing crude oil does not assign an accurate value for NGLs.” Id. at 65,050. Coupled with his finding that shipments of NGLs had increased precipitously since 1984, this led the ALJ to conclude that circumstances surrounding the TAPS Quality Bank had “changed significantly,” that the evidence “strongly establishes the distortion of value caused” by the NGLs delivered at Pump Station No. 1 and present in the GVEA return stream, and therefore that the gravity methodology as applied at Pump Station No. 1 and GVEA “no longer yield[ed] a just and reasonable result.” Id. at 65,049-50, 65,052-53. To remedy this problem, he recommended a prospective alteration of the gravity methodology at those two locations, the details of which are not relevant to the claims before us. Id. at 65,069. The ALJ also found, however, that “[tjhere is no evidence of any change in conditions at Valdez,” and thus determined that the current gravity valuation method continued to be appropriate at that location. Id. at 65,053, 65,066.
C. The Commission’s Orders
Rather than adopting the ALJ’s proposed changes in the gravity methodology, the Commission referred the case to a settlement judge to see if the affected parties could negotiate an agreement amongst themselves. In 1993, the settlement judge submitted to FERC, for its consideration, a settlement agreement executed by many, but not all, of the TAPS Carriers and affected shippers, which largely abandoned the gravity methodology. 1993 Order, 65 F.E.R.C. ¶ 61,277 at 62,285-86.
The proposed “assay methodology,” which was similar to the methodology Tesoro favored in 1984, would divide each petroleum stream entering TAPS into eight components or “cuts” based on the temperature at which particular petroleum products boil out of the stream. Each of the eight cuts would be individually valued, and then combined to determine the stream’s value. Under the proposed settlement, the five lightest cuts (those with the lowest boiling points) — propane, isobutane, normal butane, natural gasoline, and naphtha — would be valued at published market prices for those products. Because there were no readily available market prices for the three heaviest cuts, namely distillate, gas oil, and residual fuel oil (“re-sid”), the settlement suggested the use of market prices of similar products adjusted to take account of product differences. The Quality Bank would continue to calculate debits and credits by comparing the value of the common stream with that of each shipper’s contribution to it. Id. Consistent with the ALJ’s proposal, the parties would implement the new methodology at Pump Station No. 1 and GVEA while leaving the gravity methodology in place at Valdez. Id. The settlement agreement also provided a mechanism for resolving a dispute between the co-owners of the Prudhoe Bay Unit petroleum stream over the allocation of payments made to them by the Quality Bank. Id.
In November 1993, FERC adopted the settlement with modifications. 1993 Order, 65 F.E.R.C. ¶ 61,277. Various combinations of petitioners now challenge five specific aspects of the Commission’s Order. We briefly review the position the Commission assumed concerning these five contested issues and then evaluate the merits of each petitioner group’s contentions.
1. The assay methodology
FERC accepted the settlement’s proposal to replace the gravity valuation methodology with an assay methodology. It observed that the purpose of the Quality Bank is to
establish the relative value of the different quality oils that are tendered to TAPS. As such, it must incorporate a valuation methodology that is a reasonable proxy for *688the differences in the market value of the TAPS streams.
Id. at 62,286. The Commission agreed that due to the changed circumstances noted by the ALJ, the gravity methodology was no longer acceptable and that a new methodology was required. Id. at 62,287. It then found that the proposed assay methodology, with modifications to the methods of valuing two of the eight petroleum cuts, was “just and reasonable.” Id. at 62,290.
2. The distillate cut
The settlement agreement proposed to label petroleum that boils out of a stream between 350 and 650 degrees Fahrenheit as “distillate” and to value it at the market price of No. 2 fuel oil plus an adjustment of .001 cents per barrel. FERC decided to deviate from this proposal in two respects. First, it modified the settlement to split the distillate cut into two parts, light distillate (350-450 degrees) and heavy distillate (450-650 degrees), on the ground that light and heavy distillates “are distinctly different and marketed separately.” Id. at 62,288. Second, the Commission
believe[d] that market prices, uncomplicated by subjective adjustments, must be used for the Quality Bank adjustments to be non-discriminatory, in appearance as well as in fact. Market prices have the advantage of being objective, non-discriminatory, easily ascertainable, and generally not susceptible to manipulation.
Id. at 62,289. Accordingly, it rejected the settlement agreement’s addition of an adjustment factor to a market price to arrive at a value. Instead, it ordered the Quality Bank to value light distillate at the market price of jet fuel and heavy distillate at the market price of No. 2 fuel oil, the finished products into which those cuts are often refined. Id. at 62,290.
3. The resid cut
The settlement proposed to classify oil with a boiling point above 1050 degrees Fahrenheit — the heaviest, most viscous portion of the stream — as resid. Because there is no market price for resid, which is what is left of the petroleum stream after the more valuable products are removed, the settlement proposed to value it at the average market price of two lighter products, No. 6 fuel oil and fuel oil 380 (“FO-380”), adjusted to account for the cost of blending agents that would be needed to actually convert the resid into these proxy products. Id. at 62,-289. FERC discarded the proposed blending approach and ordered that the Quality Bank value resid at the average market price of No. 6 fuel oil and FO-380. Id. at 62,290. After several shippers argued that to do so would significantly overvalue heavy, resid-laden petroleum streams, FERC expanded the resid category to encompass oil with a boiling point higher than 1000 degrees. This had the effect of reducing the size of the more valuable gas oil cut (previously consisting of oil with a boiling point between 650 and 1050 degrees), and thus potentially reducing the relative value of heavier streams depending, of course, on the precise composition of those streams. FERC also decided to value resid with a boiling point higher than 1050 degrees at the price of FO-380 while valuing the lighter resid with a boiling point between 1000 and 1050 degrees at the price of the more expensive No. 6 fuel oil. Trans Alaska Pipeline System, 66 F.E.R.C. ¶ 61,-188 at 61,419-20 (1994) (“Rehearing Order”).
A Prospective application
Consistent with FERC’s prior announcements that any change in the Quality Bank methodology would be prospective, the settlement agreement contained no provision for refunds to shippers who had paid more into the Quality Bank under the gravity methodology than they would be assessed under the assay methodology, even though the former had been found to be no longer just and reasonable. FERC affirmed the prospective nature of the shift in valuation methodology, noting that because the gravity methodology had been approved by FERC, the TAPS Carriers were justified in relying on it until it was changed. 1993 Order, 65 F.E.R.C. ¶ 61,-277 at 62,291-92. According to the Commission, the retroactive application of the assay methodology would violate the filed rate doctrine. Id. at 62,292.
*689 5. Prudoe Bay Unit dispute
FERC rejected the settlement agreement’s proposed resolution of the dispute among the co-owners of the Prudhoe Bay Unit stream concerning the Quality Bank’s payments. It reasoned that this was a matter of private contract beyond the Commission’s jurisdiction because it concerned events that occur before the commingled Prudhoe Bay Unit crude is delivered to the pipeline. Id. at 62,291. Accordingly, FERC ordered that the proposed dispute resolution procedures not be included in the TAPS Carriers’ tariff filings. Id.
II. Discussion
A. Change in Methodology
Petitioners Exxon Company, U.S.A. (“Exxon”), MAPCO, and Petro Star challenge the most basic aspect of the Commission’s Order: the decision to abandon the Quality Bank’s gravity valuation methodology and replace it with an assay methodology. Intervenors State of Alaska (“Alaska”), ARCO Alaska, Inc. (“ARCO”), BP Exploration (Alaska), Inc. (“BP”), OXY USA, Inc. (“OXY”), and Tesoro urge, as an initial matter, that these petitioners are estopped from raising their challenge because all three were signatories to the settlement agreement that proposed the methodological change. We find that petitioners’ challenge is properly before us, but that it lacks merit.
1. Estoppel/standing
Although the intervenors label their challenge to petitioner’s right to raise their claim as an “estoppel” argument, their argument is more properly characterized as a claim that petitioners lack standing to raise their challenge. Only parties “aggrieved” by a final Commission order issued under the ICA may bring a petition for review. 28 U.S.C. § 2344 (1988); Shell Oil Co. v. FERC, 47 F.3d 1186, 1200 (D.C.Cir.1995). The intervenors contend that because petitioners supported the change in methodology as it was embodied in the proposed settlement agreement, they cannot now claim to be aggrieved by the change.
The intervenors’ argument mischaraeter-izes the factual background of the dispute. “It is the general rule that a party may not appeal from a disposition in its favor,” Showtime Networks, Inc. v. FCC, 932 F.2d 1, 4 (D.C.Cir.1991), but the Commission’s order cannot be fairly characterized as being in petitioners’ favor. Petitioners supported a change in methodology, but not the change ultimately ordered. The proposed settlement would have valued the distillate and resid cuts quite differently than does the Commission’s methodology, thus altering the relative value assigned to various petroleum streams to the intervenors’ detriment. By advocating a specific settlement, petitioners did not forfeit their standing to object to elements of the settlement to which they had agreed if changes made in others by the Commission work to their overall disadvantage. This recognizes the reality that businessmen will yield on particular points if they are satisfied that the net results of an agreement will accrue to their benefit.
The sole ease cited by the intervenors to support their position, Southern Natural Gas Co. v. FERC, 877 F.2d 1066 (D.C.Cir.1989), is not to the contrary. In that case, we found that the petitioner could not be “aggrieved” by FERC’s denial of an alternative proposal when the Commission granted the one that was actually its first choice. Id. at 1070-71. Here, petitioners’ first choice was clearly the settlement as proposed-, and Exxon made it clear to the Commission that it preferred the retention of the gravity methodology to the adoption of the assay method as modified by FERC. See Rehearing Order, 66 F.E.R.C. ¶ 61,188 at 61,417. Because the Commission rejected the settlement to which they had agreed, petitioners may challenge the assay methodology as modified by the Order. See Eastern Shore Natural Gas Co., 43 F.E.R.C. ¶ 61,489 at 62,212 (1988) (“If [petitioner] does not want to accept the settlement with the modifications that the Commission found were necessary ... [it] can reject the modified settlement and litigate the issues.”).
2. Merits of the change in methodology
a. Unforeseeability
Petitioners’ initial challenge to FERC’s decision to replace the gravity meth*690odology with the assay methodology is premised on the assumption that the proponents of change bore the burden of proving not only changes in circumstances, but also that the changes were “unforeseen or not reasonably foreseeable at the time of the prior Commission decision.” Brief for Petitioners at 10. While they concede that there has been an increase in the amount of NGLs injected into the common stream since FERC approved the gravity methodology in 1984, they claim that the record is devoid of evidence that this increase was not foreseeable and, as a consequence, that the Commission erred as a matter of law in approving the change.
Petitioners’ premise misstates the law. The ICA requires that all rates charged be “just and reasonable,” 49 U.S.C.App. § 1(5), and empowers FERC to prescribe “just and reasonable” rates or charges when it determines that any rate or practice of a carrier is “unjust or unreasonable,” 49 U.S.C.App. § 15(1), without regard to foreseeability. “FERC has a continuing obligation to ensure that pipeline rates are just and reasonable .... The fact that a rate was once found reasonable does not preclude a finding of unreasonableness in a subsequent proceeding.” Texas Eastern Transmission Corp. v. FERC, 893 F.2d 767, 774 (5th Cir.1990); see also Norfolk & Western Ry. v. United States, 768 F.2d 373, 378 (D.C.Cir.1985) (“rate orders are generally not res judicata because every rate order made may be superseded by another”) (emphasis in original, internal quotation marks omitted). Petitioners refer to cases in which the Commission has cited knowledge of future conditions as a factor where it has refused to modify a carrier’s terms of service, see, e.g., Trailblazer Pipeline Co., 50 F.E.R.C. ¶ 61,-188 at 61,608 (1990), but these decisions in no way suggest that a finding of unforeseeability is required before the Commission may reach the conclusion that a rate that was previously just and reasonable is no longer so.
b. Substantial evidence/arbitrary and capricious
In the alternative, petitioners contend that even if a foreseeable change in circumstances can serve as a basis for FERC’s reevaluation of the Quality Bank’s valuation methodology, the ALJ decision that served as the basis for the Commission’s Order was not supported by substantial evidence; and, because it did not adequately explain the shift in policy, it was also arbitrary and capricious. We disagree.
Petitioners are correct that FERC’s decision to approve a portion of a contested settlement must be supported by substantial evidence, 18 C.F.R. § 385.602(h)(l)(i) (1995), and that we must set aside agency actions that are “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706(2)(A). Our inquiry under the arbitrary and capricious test, of course, is “narrow and a court is not to substitute its judgment for that of the agency.” Motor Vehicle Mfrs. Ass’n of the United States, Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43, 103 S.Ct. 2856, 2860, 77 L.Ed.2d 443 (1983) ("State Farm ”). Where the necessary analysis “requires a high level of technical expertise, we must defer to the informed discretion of the responsible federal agencies.” Marsh v. Oregon Natural Resources Council, 490 U.S. 360, 377, 109 S.Ct. 1851, 1861, 104 L.Ed.2d 377 (1989) (internal quotation marks omitted). Nevertheless, we require the Commission to engage in rational decision-making, see, e.g., State Farm, 463 U.S. at 43, 103 S.Ct. at 2866-67; Laclede Gas Co. v. FERC, 997 F.2d 936, 945 (D.C.Cir.1993), and, when changing course, it “must supply a reasoned analysis indicating that prior policies and standards are being deliberately changed.” Michigan Consol. Gas v. FERC, 883 F.2d 117, 122 (D.C.Cir.1989) (quoting Hall v. McLaughlin, 864 F.2d 868, 872 (D.C.Cir.1989)). For the reasons described below, we find that the ALJ’s recommendation to abandon the gravity methodology, which the Commission adopted, satisfies these requirements.
FERC approved the gravity methodology as just and reasonable in 1984 over Tesoro’s objection not because it believed that methodology precisely valued the petroleum delivered to TAPS, but because the impact on the common stream of materials not accurately *691valued by the gravity method — i.e., NGLs— did not have a “measurable impact_” 1984 Order, 29 F.E.R.C. ¶ 61,123 at 61,239. In 1991, the ALJ found that “[c]ireumstances on TAPS have changed significantly since 1984.” ALJ Decision, 57 F.E.R.C. ¶ 63,010 at 65,050. First, shippers were injecting substantially larger quantities of NGLs into the petroleum delivered at Pump Station No. 1. Id. Second, MAPCO expanded its refinery at GVEA and Petro Star constructed its own refinery there, thus increasing the amount of valuable mid-weight products that was stripped from the common stream. As a consequence, the petroleum returned to TAPS at GVEA contained a still higher percentage of NGLs. Id. at 65,053. These changes, which increased the relative volume of NGLs injected into the common stream at both Pump Station No. 1 and GVEA, were material because “the current gravity-based methodology does not accurately value NGLs.” Id. at 65,052. Together, they provide adequate factual support for the Commission’s conclusion that “due to changed circumstances, the existing methodology is no longer just and reasonable, and that a new methodology is required.” 1998 Order, 65 F.E.R.C. ¶ 61,277 at 62,286 (footnote omitted).
Petitioners argue that the AL J’s reasoning that the gravity methodology is unjust because it fails to properly value NGLs is nonsensical in the context of the TAPS Quality Bank. They point out that NGLs are not delivered separately to TAPS, but merely constitute a portion of various petroleum streams, and argue that how a stream acquires its “quality” is “irrelevant to an assessment of that quality in the market place.” Brief for Petitioners at 19. The ALJ recognized that the issue was not whether the gravity methodology accurately valued NGLs per se, but whether it placed a proper value on petroleum whose gravity had been increased as a result of the injection of substantial quantities of NGLs. He examined the evidence and concluded that “the current straightline gravity basis for valuing crude oil does not assign an accurate value for NGLs.” ALJ Decision, 57 F.E.R.C. ¶ 63,010 at 65,050. It follows from this conclusion that shippers delivering oil with a high NGL content were either being overcompensated or undercompensated under the gravity methodology for their contribution to the value of the common stream.
Petitioners’ arguments that the ALJ’s reasoning was arbitrary or illogical must fail because what is really at issue is a disagreement between petitioners and the ALJ over whether the weight of the evidence indicates that the gravity methodology fairly values petroleum with a high NGL content. Petitioners strenuously argued to the ALJ that the high gravity of NGLs eommensurately increases the value of the stream into which it is injected, but the ALJ considered their “evidence on this point ... not convincing,” id., and, in fact, was of the view that “the more substantial evidence indicate[d] that NGLs devalue the common stream....” Id.
Determining the effect of NGLs on a stream’s value is a question of fact. It calls for “a high level of technical expertise,” requiring us to “defer to the informed discretion” of the agency. Oregon Natural Resources Council, 490 U.S. at 377, 109 S.Ct. at 1861; see also Board of Trade of Kansas City v. United States, 314 U.S. 534, 546, 62 S.Ct. 366, 372, 86 L.Ed. 432 (1942) (observing that “[t]he process of rate making is essentially empiric” and “Congress has therefore delegated the enforcement of transportation policy to a permanent expert body_”). Here, petitioners point to nothing in the record undermining the evidence — and the Commission’s conclusion — that the gravity method is an inaccurate method for valuing a common stream with NGLs in as high a proportion as prevail at Pump Station No. 1 and GVEA.
Finally, shifting gears, petitioners contend that it was irrational for the Commission to approve a change in methodology at Pump Station No. 1 and GVEA based on the finding that the gravity methodology is no longer just and reasonable at those locations while retaining that methodology at Valdez where NGLs unarguably are still present in the common stream. Again we disagree. It is no doubt true that if the assay methodology provides a more accurate valuation at Pump Station No. 1 and GVEA, *692it will provide a more accurate valuation at Valdez as well. But the Commission may approve the methodology proposed in the settlement agreement if it is “just and reasonable”; it need not be the only reasonable methodology, or even the most accurate. Cf. City of Bethany v. FERC, 727 F.2d 1131, 1136 (D.C.Cir.1984) (when determining whether proposed rate was “just and reasonable,” as required by Federal Power Act, FERC properly did not consider “whether a proposed rate schedule is more or less reasonable than alternative rate designs”).
At Valdez, the Quality Bank adjusts for the differences in the quality of the common stream from one day to the next, rather than the differences in quality between the petroleum streams delivered to TAPS by various shippers or refiners. This difference is important for two reasons. First, as the record indicates, because the daily variations in the gravity of the common stream at Valdez are slight, Quality Bank adjustments there are “relatively inconsequential.” Testimony of William H. Clifton, reprinted in Joint Appendix (“J.A.”) at 2061-62. Second, any overvaluation of NGLs will not systematically adversely affect certain shippers and advantage others, because whether a particular shipper obtains a tanker of petroleum with a marginally higher or lower NGL content is a matter of random chance. Because the precision of the valuation methodology is so much less important at Valdez than at the other Quality Bank locations, it is not surprising that no shipper presented evidence to the Commission suggesting that the gravity methodology produced unjust or unreasonable results at Valdez, see ALJ Decision, 57 F.E.R.C. ¶ 63,-010 at 65,053, or that the proposed settlement did not seek a change in the valuation method there. On the basis of this record, it was entirely reasonable for the ALJ to recommend, and FERC to approve, the maintenance of the status quo at the TAPS’ terminus.
B. The Distillate Cut
Petitioners MAPCO, Petro Star and Alaska challenge, as arbitrary and capricious, FERC’s decision to value the distillate cut by splitting it into “light” and “heavy” distillates and valuing the former at the price of jet fuel and the latter at the price of No. 2 fuel oil. Intervenors ARCO, BP, OXY, and Tesoro assert that this challenge is not properly before the court because none of the petitioners raised this objection in the original petitions for rehearing filed subsequent to the 1993 Order. We find the intervenors’ procedural challenge meritless and agree with petitioners that FERC’s valuation of distillate is arbitrary.
1. Exhaustion
The intervenors’ procedural argument appears to be a claim that petitioners did not exhaust their administrative remedies because they failed to raise their objection to FERC’s distillation cut methodology in a petition for rehearing. Their basic complaint is that, although the 1993 Order described the Commission’s decision to split the distillate cut and value light and heavy distillates separately, petitioners did not object to this decision until after the Commission had issued its Rehearing Order. The intervenors urge us to refuse to sanction this delay, which they call a “tactical ruse.” Brief for Intervenors at 12.
We need not investigate the propriety of petitioners’ delay in seeking rehearing. The intervenors’ argument must fail because there is no statutory or regulatory requirement that petitioners seek rehearing of an order issued by FERC under the ICA prior to seeking judicial review, 18 C.F.R. § 385.1902(b) (1995); and we will not require an aggrieved party to seek optional administrative appeals prior to petitioning for our review. Darby v. Cisneros, — U.S. -, -, 113 S.Ct. 2539, 2544-45, 125 L.Ed.2d 113 (1993) (exhaustion of administrative remedies doctrine limited to requirements “the statute or rule clearly mandates”); compare Tennessee Gas Pipeline Co. v. FERC, 9 F.3d 980, 981 (D.C.Cir.1993) (noting that a request for rehearing is a prerequisite for judicial review of FERC orders issued under the statutory authority of the Natural Gas Act, 15 U.S.C. § 717r).
*693 2. FERC’s methodology
The Commission’s method of valuing distillate in the TAPS streams deviates from the proposed settlement in two material respects. First, FERC determined that heavy and light distillates are “distinctly different and are marketed separately,” and thus should be valued separately. 1993 Order, 65 F.E.R.C. ¶ 61,277 at 62,288. Second, because of its determination that all cuts should be valued at unaltered market prices, FERC rejected the proposed settlement’s concept of arriving at valuations by adjusting market prices of similar but not identical products. Instead, it ordered the distillate cuts valued at the precise spot market prices of jet fuel and No. 2 fuel oil. Petitioners assert that both departures from the proposed settlement were arbitrary and capricious and/or not supported by substantial evidence.
Petitioners contest FERC’s decision to split the single distillate cut into heavy and light cuts by arguing that the record is devoid of evidence that the two cuts are, in fact, marketed separately. This objection reads FERC’s rationale for splitting the cut too narrowly. The Commission subdivided the distillate cut into light and heavy components because lighter distillate is often refined into jet fuel and then marketed as such, whereas the heavier distillate is incompatible with this use and is therefore processed into and marketed as less valuable products. The record amply supports this conclusion. Given the differences in uses, the Commission reasonably concluded that the assay methodology should place a higher value on light distillate than on heavy distillate.
Petitioners’ more compelling argument is that FERC’s decision to value light distillate at the market price of jet fuel and heavy distillate at the market price of No. 2 fuel oil is flawed because these market proxies are the prices of finished products rather than of raw materials. They claim that considerable processing is necessary before light distillate can be sold as jet fuel, or heavy distillate as No. 2 fuel oil, and that FERC’s valuation methodology thus overvalues both light and heavy distillates relative to other cuts in the common stream by not taking these costs into account.
The Commission does not dispute petitioners’ claim of overvaluation; rather, it defends its methodology on the ground that jet fuel and No. 2 fuel oil spot prices are “reasonable proxies” for light and heavy distillate and explains that it chose to avoid attaching adjustment formulas to the finished products’ market prices to protect the objectivity of the Quality Bank’s valuation methodology. Thus, it contends, the pricing decision fell within its proper discretion.
We cannot agree. The goal of the Quality Bank valuation methodology, as all parties agree, is to assign accurate relative values to • the petroleum that is delivered to TAPS and becomes part of the common stream. In order to achieve this goal, FERC must accurately value all cuts—not merely some or most of them—or it must overvalue or undervalue all cuts to approximately the same degree. If light and heavy distillates are overvalued and other cuts are not, streams rich in these distillates will be overvalued relative to other streams and their owners will receive a windfall in the form of Quality Bank credits. FERC’s position appears to be that because jet fuel bears some relation to light distillate and No. 2 fuel oil bears some relation to heavy distillate, the prices of the finished products are close enough to the values of the raw materials to serve as their proxies, although it presents no data to indicate how close the values are in fact. We find this reasoning arbitrary and capricious and thus conclude that, absent a more persuasive justification, FERC’s method of valuing distillates violates the APA. 5 U.S.C. § 706(2)(A).
The intervenors who support the Commission present a stronger argument. Focusing on light distillate, they concede that converting the raw material into jet fuel requires some processing but contend that this treatment is minimal and not unlike the minor processing required to bring other cuts of the common stream up to the specifications assumed by the spot market prices used to value them. This premise, if true, might support the reasonableness of FERC’s light distillate valuation method. It is not the reason that the Commission adopted, *694however, and we may affirm the agency only on the grounds upon which it relied in exercising its power. See SEC v. Chenery Corp., 318 U.S. 80, 95, 63 S.Ct. 454, 462-63, 87 L.Ed. 626 (1943); Puerto Rico Higher Educ. Assistance Corp. v. Riley, 10 F.3d 847, 850 (D.C.Cir.1993). It is also not supported by substantial evidence in the record. Although one witness characterized the processing required to turn light distillate into jet fuel and heavy distillate into No. 2 fuel oil as “minor,” Testimony of William Stancil, reprinted in J.A. at 1232, there is no attempt in the record that we are aware of to quantify these “minor” costs or to project the extent to which ignoring them would result in the overvaluation of TAPS streams heavily laden with distillate.
We are conscious of the difficult and necessarily imprecise task FERC faces when it must split petroleum streams into component parts and then place a value on each of them. We agree with the Commission that there is no “perfect way” to value the different quality oils shipped on TAPS, 1993 Order, 65 F.E.R.C. ¶ 61,277 at 62,286, especially in the case of products without a readily ascertainable market price; and we will not hold the Commission to an impossibly high standard. But if the agency chooses to value some cuts of petroleum at the prices they command in the market without the benefit of processing, as it appears to have done, it must attempt, to the extent possible, to value all cuts at the price they would command without processing. It cannot, consistent with the requirement of reasoned decisionmaking, value some cuts precisely and others haphazardly. Accordingly, we must remand its distillate valuation methodology for further consideration.
C. The Resid Cut
Recall that the settlement agreement proposed that the TAPS Quality Bank value resid — the heaviest portion of the petroleum stream — at the price of FO-380 less the cost of converting it into FO-380. Under this approach, all resid would not be valued equivalently: its value would vary based in part on its viscosity, as more viscous resid would require more of a blending agent to be converted into FO-380.
The Commission found this “blending” valuation methodology “arbitrary, unreasonable, and impractical,” 1993 Order, 65 F.E.R.C. ¶ 61,277 at 62,288, due to its determination that “[rjesid is primarily used as a [refinery] feedstock [to manufacture other products] in processes where its viscosity is irrelevant to its value_” Trans Alaska Pipeline System, 67 F.E.R.C. ¶ 61,175 at 61,531 (1994). The Commission decided it was appropriate to value all resid with a boiling point higher than 1050 degrees identically because such a valuation would be consistent with “the use to which the overwhelming majority of North Slope resid is put.” Id. Notwithstanding this position, in an attempt to accommodate shippers who feared that a departure from the blending methodology would result in the overvaluation of resid, the Commission expanded the resid cut to include petroleum with a boiling point as low as 1000 degrees, valuing resid with a boiling point higher than 1050 (“1050 + resid”) at the price of FO-380 and petroleum with a boiling point between 1000 and 1050 degrees (“lighter resid”) at the price of No. 6 fuel oil.
As is true of distillate, there is no published market price for resid, which makes valuing it difficult. In spite of the vexing nature of the problem and the deference we owe the Commission’s judgments, we find that its approach to resid valuation fails to satisfy the APA’s basic requirement of reasoned deci-sionmaking. See, e.g., State Farm, 463 U.S. at 43, 103 S.Ct. at 2866-67; Laclede Gas Co., 997 F.2d at 945. Consequently, we remand this portion of the assay methodology to the Commission for further consideration.
1. The 1050+ resid
The Commission searched for a proxy price for the 1050+ resid that would be consistent with its belief that objectivity was best served by valuing petroleum cuts only at unadjusted market prices. It offered the following justification for its selection of the published price of FO-380 as an appropriate proxy:
We are cognizant that [FO-380] does not correlate directly with the specifications of [resid]; however, it is the best commonly available published pricing indicator which *695most closely approximates the specifications of [resid]. This reference price represents the least desirable, heaviest gravity product still traded on major markets. Moreover, since the purpose of the Quality Bank is only to establish relative values for the TAPS crude streams ... our purpose is adequately served by this reference pricing scheme.
Rehearing Order, 66 F.E.R.C. ¶ 61,188 at 61,420.
Petitioners Exxon, Alaska, and Tesoro challenge FERC’s valuation of the 1050 + resid and in doing so reiterate their support for the settlement proposal’s methodological approach. They believe that FERC’s methodology significantly overvalues resid, thus penalizing shippers of petroleum having a low content of the 1050+ resid.
The Commission’s resid valuation methodology suffers from the same conceptual flaw that plagues its distillate methodology: the record demonstrates no more than that the price of FO-380 bears some remote relationship to the value of 1050+ resid as a feedstock. FERC offers two arguments in defense of its use of FO-380 as a proxy, neither of which is convincing. First, relying on expert testimony, the Commission claims that FO-380 can substitute for the 1050+ resid as a feedstock. Notably, neither the witness who so testified nor any other stated that it was a common industry practice to use FO-380 as a feedstock when resid would do the job. Consequently, although the cited testimony supports the conclusion that FO-380 and the 1050+ resid share some physical properties, it in no way suggests the two materials have equal or even near-equal market values. As petitioners aptly note, FERC’s reasoning is akin to suggesting that if diamonds can be substituted for coal as a source of carbon, the price of diamonds would be an appropriate proxy for the price of coal. The Commission’s conclusion simply does not follow from its premise.
The Commission’s alternate justification is that it has assigned, as a proxy for this least valuable component of the common stream, the petroleum product having the lowest published price. The fact that FO-380 is cheaper than other petroleum products with active markets, however, in no way demonstrates that its value is even remotely commensurate with that of resid. If the Commission values other cuts in the TAPS streams using relevant market prices but significantly overvalues resid, the Quality Bank will consequently overvalue resid-laden petroleum streams relative to those with a significantly lower resid content. We therefore find the 1050+ resid portion of the assay methodology arbitrary and capricious and remand it to the Commission for further consideration.
Petitioners’ criticism of the Commission’s 1050+ resid valuation methodology, however, runs deeper than the methodology’s most obvious shortcoming. Petitioners object to the Commission’s emphasis on resid’s value as a feedstock. They argue that, according to basic economic principles, a product’s marginal use — not its most prevalent use — dictates its market price. While conceding that most North Slope resid is used as a feedstock, petitioners contend, and FERC does not dispute, that the 1050 + resid’s marginal use is as an ingredient of FO-380 or similar fuel oils. To convert this resid into the less viscous FO-380, lighter oils more expensive than FO-380 must be used as a blending agent. This leads petitioners to their ultimate conclusion that 1050+ resid must be less valuable than FO-380. The Commission, in its various orders, has failed to respond to this argument.
It is true that a competitive market will set a product’s price at its marginal use value, see Paul A. Samuelson and Anthony Scott, Economics: An Introductory Analysis 471 (1966), but this hardly proves that a market would price resid precisely at its value as a blending agent to the refiners that use it for that purpose. If there are market imperfections, petitioners’ position might not be correct. Thus, their suggested resid valuation methodology is not necessarily the only reasonable one. Nonetheless, petitioners’ argument has sufficient analytical force. On remand, the Commission should explicitly address whether the marginal use of 1050 + resid should be taken into account in that cut’s valuation methodology.
*696 2. The lighter resid
Petitioners also challenge FERC’s decision to expand the resid cut to include petroleum with a boiling point between 1000 and 1050 degrees, and to value that portion of the resid cut at the price of low-sulfur No. 6 fuel oil. They claim that there is no evidence in the record that supports either of these decisions.
We believe there is sufficient evidence to support the Commission’s decision to expand the range of the components of the petroleum streams that it categorizes as “resid.” In the rehearing proceedings, an Exxon consultant testified that the Commission’s previous selection of 1050 degrees as the cutoff point for the resid cut was arbitrary and that 1000 degrees was more consistent with what West Coast refineries tend to characterize as residual fuel oil. Aff. of Joe F. Moore, reprinted in J.A. at 2162-63. In light of the fact that there is no published market price for resid and, consequently, no precise specifications for it, we find it was well within FERC’s discretion to rely on this testimony to reach its decision to expand the resid cut.
Petitioners are correct, however, that there is no evidence in the record to justify the Commission’s decision to value the lighter resid at the price of No. 6 fuel oil. FERC defends this decision on the ground that it has a lower viscosity than the 1050 + resid and thus is more valuable. This response is inconsistent with the Commission’s position, discussed above, that the viscosity of resid has no bearing on its value because most resid is used as a feedstock. In addition, as is the problem with the Commission’s valuation of the 1050+ resid and the light and heavy distillates, we know of no evidence in the record that suggests that the price of No. 6 fuel oil bears a close relationship to the value of North Slope petroleum with a boiling point between 1000 and 1050 degrees. As a result, we must remand the Commission’s method of valuing the fighter resid as well.
D. Refunds
Petitioner OXY challenges FERC’s determination that changes in the Quality Bank valuation methodology would be prospective only. Intervenors ARCO, Exxon, MAPCO, and the TAPS Carriers collectively counter with a challenge to OXY’s standing to contest the Commission’s ruling. We find that OXY demonstrated that it was “aggrieved” by FERC’s order and thus has standing, but that the Commission properly declined to order refunds.
1. Standing
As noted above, only “aggrieved” parties may seek judicial review of a final FERC order issued under the ICA. 28 U.S.C. § 2344; Shell Oil, 47 F.3d at 1200. We evaluate aggrievement by reference to traditional principles of standing. Water Transport Ass’n v. ICC, 819 F.2d 1189, 1193 (D.C.Cir.1987). To be aggrieved, then, OXY must have suffered an “injury in fact” traceable to FERC’s action, a decision in its favor must be capable of redressing that injury, and its interest must be “arguably within the zone protected or regulated by the ... statutory [provision] ... in question.” Id. OXY bears the burden of proof on all of these elements. Lujan v. Defenders of Wildlife, 504 U.S. 555, 559-61, 112 S.Ct. 2130, 2136, 119 L.Ed.2d 351 (1992).
Intervenors first argue that OXY has not satisfied its burden of proving injury in fact because there is no evidence that it would have benefited from a refund of past Quality Bank payments. It is undisputed that OXY is not a TAPS shipper; it sells petroleum to Conoco, Inc., which in turn sells petroleum to TAPS shippers. The intervenors concede that OXY may have an interest in the methodology employed by the Quality Bank in the future because that methodology might affect the prices it can charge for its petroleum, but they argue that there is no evidence that OXY would be contractually entitled to share in any rebate the Commission might order.
We disagree. In a 1990 TAPS hearing, an expert witness testifying on behalf of OXY and Conoco concerning a potential change in methodology calculated the amount of money that would be due to OXY for 1989 and 1990 *697if the Commission ordered refunds. In the absence of any counter-evidence suggesting OXY would not, in fact, participate in any refund, we find in this testimony a sufficient demonstration of injury to satisfy the Supreme Court’s requirement that the injury be “actual ... not conjectural or hypothetical.” Id. (internal quotation marks omitted).
Intervenors also contend that OXY does not fall within the ICA’s zone of interest because the primary purpose of the ICA is to protect shippers. They have far too narrow a view of the statute’s objectives. The ICA grants the Commission broad authority to respond to complaints concerning “anything done or omitted to be done by any common carrier” subject to the statute lodged by
[a]ny person, firm, corporation, company, or association, or any mercantile, agricultural, or manufacturing society or other organization, or any body politic or municipal organization, or any common carrier....
49 U.S.C.App. § 13(1) (1988). The statutory scheme indicates a congressional intent to protect the interests of a broad category of entities affected by the practices of common carriers. Conversely, nothing in the statute suggests that concern is limited to parties in privity with common carriers. We believe it would be impossible to say that companies like OXY, ie., those that produce and sell petroleum that is subsequently shipped through TAPS, are not at least “arguably within the zone protected or regulated by the [ICA].... ” Water Transport, 819 F.2d at 1193. Therefore, we conclude that OXY was “aggrieved” by FERC’s decision to not order refunds and may challenge that decision here.
2. Tariff violations
FERC has consistently maintained that shippers would be entitled to refunds from July 1, 1989, the effective date of the updated tariff filings, only if it were found that the NGLs were being shipped in violation of the tariffs or that Quality Bank debits and credits were being incorrectly calculated. 1993 Order, 65 F.E.R.C. ¶ 61,277 at 62,292; see also 1990 Order, 51 F.E.R.C. ¶ 61,062 at 61,137; 1989 Order, 49 F.E.R.C. ¶ 61,349 at 62,264-65. The ALJ found no tariff violations or calculation errors. ALJ Decision, 57 F.E.R.C. ¶ 63,010 at 65,041-45. Consistent with this finding, the proposed settlement did not call for refunds.. FERC’s approval of the settlement reiterated that the change in the Quality Bank methodology would be prospective only, presumably agreeing with the ALJ that no tariff violations had occurred. See 1993 Order, 65 F.E.R.C. ¶ 61,277 at 62,291-92. OXY now challenges that determination.
Recognizing the Commission’s technical expertise in the area of pipeline ratemaking, we owe “substantial deference” to its interpretation of the TAPS Carriers’ tariff provisions. Natural Gas Clearinghouse v. FERC, 965 F.2d 1066, 1070 (D.C.Cir.1992). Deference does not imply abdication of our obligation to review the Commission’s orders, however. We inquire whether the Commission’s interpretations are “amply supported both factually and legally” and accept them only if they are “the result of reasoned and principled decision-making that can be ascertained from the record.” Tarpon Transmission Co. v. FERC, 860 F.2d 439, 442 (D.C.Cir.1988) (internal quotation marks omitted). The ALJ’s interpretation of the TAPS Carriers’ tariffs easily meets this standard.
OXY argues that the TAPS Carriers violated three provisions of their tariffs by transporting NGL-laden petroleum streams which increased the gravity of the common stream and thus increased the payments owed the Quality Bank by the owners of streams relatively low in NGLs. First, OXY claims the Carriers violated tariff provisions permitting them to transport only petroleum that had a gravity of between 17 and 40 degrees, which it believes effectively prohibits the direct shipment of the high-gravity NGLs. The ALJ pointed out, however, that “[n]othing contained in the tariffs ... requires the carriers to separately test the components of the streams,” ALJ Decision, 57 F.E.R.C. ¶ 63,010 at 65, 043, and that all of the Carriers’ tariffs specifically permit the shipment of “unrefined liquid hydrocarbons including gas liquids” — ie., NGLs. Id. From these facts, the ALJ reasoned that the tariffs’ gravity limitations pertain to petrole*698um streams as delivered to TAPS, not to the individual components of those streams. As OXY concedes that no petroleum shipped on TAPS has ever exceeded the gravity restrictions, the ALJ reasonably found that the inclusion of NGLs in the crude delivered to the pipeline did not violate the tariffs.
Second, OXY claims that the TAPS Carriers violated their tariffs by failing, on some occasions, to demand assays of tendered petroleum streams prior to shipment. It believes that the Carriers should have used the assays to determine that excessive NGL injections were devaluing the common stream and explore ways to address the problem. The ALJ determined that the purpose of the assay provisions is to help the Carriers transport petroleum safely and efficiently, not to benefit the shippers in any way. In support of this interpretation, he pointed out that some of the tariffs permit the Carriers to demand an assay but do not require that they do so. One tariff, for example, provides that the “Carrier may require ... a suitable assay of the tendered Petroleum,” while another permits a Carrier to reject any petroleum tendered without such an assay. Id. at 65,041. The ALJ was clearly justified in concluding that any failure to conduct assays prior to shipment on the part of the TAPS Carriers did not constitute a violation of their tariffs or harm the shippers.
Third, OXY contends that the shipment of NGLs violated tariff provisions that require the TAPS Carriers to reject petroleum whose characteristics will materially affect the quality or value of other shippers’ petroleum unless the Quality Bank makes reasonable monetary adjustments. OXY argues that the ALJ’s findings that the gravity valuation methodology no longer produces just and reasonable results demonstrates that the Quality Bank failed to compensate it for the loss it incurred as a result of the commingling of its petroleum with streams that had a high NGL content. The ALJ viewed this argument as essentially addressed to the propriety of the gravity methodology, which FERC had approved as just and reasonable in 1984, rather than to the manner in which the Carriers had implemented the Commission tariffs. He succinctly and appropriately reasoned that
[t]he concept of the transport of NGLs was anticipated at the time of the [1984] Settlement. The Quality Bank was established to provide reasonable monetary adjustments for the varying quality of shipments. The TAPS Carriers have the right to rely upon the Quality Bank to do so. I find no violation of this tariff provision.
ALJ Decision, 57 F.E.R.C. ¶ 63,010 at 65,-042. In other words, when FERC approved the gravity valuation methodology in 1984, the Quality Bank adjustments made pursuant to that methodology de facto became “reasonable monetary adjustments.” Cf. Montana-Dakota Utilities Co. v. Northwestern Public Service Co., 341 U.S. 246, 251, 71 S.Ct. 692, 695, 95 L.Ed. 912 (1951) (“[T]he right to a reasonable rate is the right to the rate which the Commission files or fixes....”)
3. Retroactive application of the new methodology
The Commission has consistently defended its decision not to apply the assay methodology retroactively on two related grounds. First, it points out that it launched its investigation into the methodology pursuant to section 13(2) of the ICA, which does not permit the Commission to issue “orders for the payment of money.” 49 U.S.CApp. § 13(2). See 1990 Order, 51 F.E.R.C. ¶ 61,-062 at 61,137. Second, it contends, essentially, that because the Quality Bank valuation methodology and the resulting cash adjustments among shippers are integral parts of the TAPS tariff structure, changes must conform to the filed rate doctrine, which protects the ability of shippers and common carriers to rely on filed rates until they are formally changed. 1993 Order, 65 F.E.R.C. ¶ 61,277 at 62,292,
OXY counters that FERC suspended the TAPS Carriers’ 1989 rate filings pursuant to its authority under section 15(7) of the ICA, which gives it the authority to order refunds of “increased rates or charges as by its decision shall be found not justified.” 49 U.S.C.App. § 15(7). Because the Commission has found that the gravity methodology *699is no longer just and reasonable, OXY believes FERC has the authority to retroactively implement the new methodology and the responsibility to provide a reasoned opinion for its failure to exercise that authority.
We conclude that FERC properly determined that it lacked the authority to apply the new methodology retroactively. Although the Quality Bank valuation methodology is a formula rather than an actual “rate,” we agree with FERC that the methodology has been an integral element of the TAPS Carriers’ tariff structure since it approved the 1984 settlement. 1993 Order, 65 F.E.R.C. ¶ 61,277 at 62,292. That structure establishes the conditions governing the shippers’ access to the pipeline. As we observed in Moss v. C.A.B., 430 F.2d 891, 897 (D.C.Cir.1970), “[a]s a practical matter, the [agency’s] order [of a ratemaking formula] amount[s] to the prescription of rates_” Thus, the filed rate doctrine applies to changes in that methodology.
The filed rate doctrine “forbids a regulated entity to charge rates for its services other than those properly filed with the appropriate federal regulatory authority.” Arkansas Louisiana Gas Co. v. Hall, 453 U.S. 571, 577, 101 S.Ct. 2925, 2930, 69 L.Ed.2d 856 (1981). The doctrine is based on the long-established principles of regulatory law that “the rate of the carrier duly filed is the only lawful charge,” Louisville & Nashville R.R. v. Maxwell, 237 U.S. 94, 97, 35 S.Ct. 494, 495, 59 L.Ed. 853 (1915), and that shippers on common carriers are entitled to rely on filed rates until those rates are changed. Arizona Grocery Co. v. Atchison, Topeka and Santa Fe Ry., 284 U.S. 370, 387-88, 52 S.Ct. 183, 185, 76 L.Ed. 348 (1932) (A carrier “cannot have reparation from the shippers for a rate collected under [an Interstate Commerce Commission] order upon the ground that it was unreasonably low.”)
The doctrine’s corollary, applicable here, is the rule that agencies may not alter rates retroactively. See, e.g., Arizona Grocery, 284 U.S. at 389, 52 S.Ct. at 186; Town of Concord v. FERC, 955 F.2d 67, 71 (D.C.Cir.1992). Together, these principles prevent unjust discrimination and, more relevant to this case, they ensure predictability. See Natural Gas Clearinghouse, 965 F.2d at 1075. The rule against retroactive ratemak-ing, however, “does not extend to cases in which [customers] are on adequate notice that resolution of some specific issue may cause a later adjustment to the rate being collected at the time of service.” Id. The goals of equity and predictability are not undermined when the Commission warns all parties involved that a change in rates is only tentative and might be disallowed.
The provisions of the ICA reflect these general doctrinal rules. Section 13(2), which the Commission invoked in this case, authorizes it to investigate and, in appropriate circumstances, to change an existing rate, with this qualification: FERC may not issue “orders for the payment of money.” Thus the Commission has no authority under that section to apply a change retroactively. On the other hand, section 15(7), on which OXY relies, creates a mechanism by which FERC may allow a challenged rate increase to take conditional effect pending an investigation into its reasonableness. Section 15(7) procedures do not undermine the rule against retroactive ratemaking because all parties are placed on notice that the agency has the authority to order a refund of any part of the increase that it finds to be unjustified. This statutory scheme is similar in structure to other statutes that govern FERC’s ratemak-ing authority. See City of Batavia v. FERC, 672 F.2d 64, 75-76 (D.C.Cir.1982) (distinguishing between section 205 of the Federal Power Act, which provides FERC with authority to order refunds of new schedules found to be unjust, and section 206, which permits FERC to investigate existing rates and change them prospectively); Sea Robin Pipeline Co. v. FERC, 795 F.2d 182, 189 n. 7 (D.C.Cir.1986) (finding FERC may not order a retroactive refund after finding a filed rate unjust subsequent to an investigation launched on its own initiative pursuant to section 5 of the Natural Gas Act).
Section 13(2) governs the change to the assay methodology for two independent reasons. In their 1989 filing, the TAPS Carriers proposed increases in the Quality Bank adjustments; they did not propose a change in the gravity methodology. Thus while it *700was entirely proper for the Commission to consider the proposed adjustments under the provisions of section 15(7) and, if warranted, to order refunds, the gravity methodology could not be subject to those proceedings because it remained the established method of calculating Quality Bank credits and debits. Furthermore, because the filing had placed no one on notice that a change to the assay methodology was in prospect, the change could not have been imposed retroactively without violating the filed rate doctrine. On the other hand, the Commission’s investigation into the gravity methodology pursuant to section 13(2) was clearly appropriate, as was its statement that any change would be prospective only. See 1989 Order, 49 F.E.R.C. ¶ 61,349 at 62,264-65; 1990 Order, 51 F.E.R.C. ¶ 61,062 at 61,137. Because any refund would have constituted impermissible retroactive ratemaking, the Commission quite properly applied the assay methodology prospectively.
E. The Prudhoe Bay Unit Dispute
The final challenge to FERC’s 1998 Order relates only indirectly to the TAPS Quality Bank valuation methodology. The Prudhoe Bay Unit’s petroleum stream (“Unit stream”), which flows into TAPS at Pump Station No. 1, is co-owned by eleven petroleum producers (“Unit shippers”) including petitioner BP. This petroleum is itself a combination of two feeder streams produced from different areas within the Unit, which are commingled at a location some 700 feet upstream of Pump Station No. 1. One of the two feeder streams consists primarily of NGLs; the other, known as the Separator Liquid Production (“SLP”) stream, does not. Under the assay methodology, the petroleum in the SLP stream has a higher value than does that in the NGL stream, Rehearing Order, 66 F.E.R.C. ¶ 61,188 at 61,420, whereas under the gravity methodology the reverse was the case.
The Unit shippers own varying interests in the fields from which the two streams are produced. Because BP owns a larger interest in the SLP stream than in the NGL stream, its per barrel contribution to the Unit stream has a higher value under the new methodology than does the per barrel contribution of certain other shippers who own a larger interest in the NGL than in the SLP stream. Nevertheless, although the TAPS Carriers are aware of the Unit shippers’ respective ownerships in the two feeder streams, they make Quality Bank payments to them on a “pro rata” basis; that is, payments are based on the total number of barrels a shipper contributed to the Unit stream rather than on their value.
BP complains that as a result of this payment practice, it does not receive its proper portion of the Quality Bank payments and objects to the Commission’s refusal to require the Carriers to employ some practice that would allocate the payments among the Unit shippers based on the contribution that each of them makes to the value of the Unit stream. While not resolving the merits of BP’s claim, the proposed settlement agreement spoke to it by containing a dispute resolution procedure that the TAPS Carriers would apply when confronted with a dispute among shippers over rights to Quality Bank payments resulting from their co-ownership of a single incoming stream. 1993 Order, 65 F.E.R.C. ¶ 61,277 at 62,291; Rehearing Order, 66 F.E.R.C. ¶ 61,188 at 61,420-21.
The Commission declined to approve the proposed dispute resolution procedure as part of the settlement for the following reasons:
The ownership dispute relates to events that occur before the petroleum is injected into TAPS. We agree that these provisions relate to a private contractual matter that is outside the scope of the Commission’s jurisdiction. Accordingly, our approval of the settlement does not constitute approval of these provisions, and those procedures should not be part of the tariff fifing made by the TAPS Carriers.
1993 Order, 65 F.E.R.C. ¶ 61,277 at 62,291 (footnote omitted). On rehearing, FERC elaborated on the problem but maintained its initial position that “[t]he Commission’s jurisdiction applies to the streams transported on TAPS, and does not cover events that precede[d] it.” Rehearing Order, 66 F.E.R.C. ¶ 61,188 at 61,421. It reasoned that Quality Bank adjustments should be paid based on *701the assays of the streams tendered to TAPS. Id. at 61,421-22. If particular Unit shippers felt they were entitled to more than their pro rata payment, the Commission suggested that they renegotiate their private contracts with their co-owners or litigate against them. Id. at 61,422.
BP now argues that the Commission erred by finding it lacked jurisdiction over the dispute. Section 15(1) of the ICA provides, in pertinent part, that
[w]henever, ... the Commission shall be of [the] opinion that any ... practice whatsoever of ... carriers subject to the provisions of this chapter, is or will be unjust or unreasonable or unjustly discriminatory or unduly preferential or prejudicial, ... the Commission is authorized and empowered to determine and prescribe what ... practice is or will be just, fair, and reasonable. ...
49 U.S.CApp. § 15(1). BP believes that the TAPS Carriers’ method of making Quality Bank payments to Unit shippers based on the quantity of the stream they owned without regard to the quality of what they owned constitutes a “practice” subject to the Commission’s jurisdiction under this section. FERC continues to characterize the issue as a dispute among the Unit shippers as to how to allocate the Quality Bank payments that is “not integral to the operation of the TAPS,” Brief for Respondent at 47, and that the method of resolving it is not a proper subject for the TAPS Carriers’ tariff filings.
The Commission has staked out an internally inconsistent and thus untenable position. On the one hand, it claims to lack jurisdiction over how the Quality Bank payment due the Unit shippers is to be divided. Yet at oral argument the Commission conceded that, as BP charges, it has approved the TAPS Carriers’ current method of dividing payments among the co-owners of the Unit stream on a pro rata basis because that method is embodied in previous filings. FERC thus has acted in the past as if it has jurisdiction over the very issue that it now explicitly maintains it lacks the authority to consider.
Until the Commission articulates a consistent position on the limits of its jurisdiction, we are unable to resolve BP’s specific challenge. We have held that an agency’s interpretation of the limits of its jurisdiction is entitled to “Chevron deference.” Oklahoma Natural Gas Co. v. FERC, 28 F.3d 1281, 1283-84 (D.C.Cir.1994) (referring to deference due an agency’s interpretations of statutes it is charged with administering, as set forth in Chevron U.S.A. Inc. v. NRDC, 467 U.S. 837, 842-44, 104 S.Ct. 2778, 2781-82, 81 L.Ed.2d 694 (1984)). We cannot deferentially review both the Commission’s explicit opinion that it lacks jurisdiction and its implicit opinion that it does not. Therefore, we remand the matter to the Commission with the instruction that it establish a consistent and reasoned position as to whether it has jurisdiction over the method by which the TAPS Carriers distribute Quality Bank payments among co-owners of streams delivered to TAPS.
III. Conclusion
We deny the petitions for review to the extent that they challenge the Commission’s decisions to replace the TAPS Quality Bank’s gravity methodology with an assay methodology and to implement the new methodology prospectively only. We grant the petitions to the extent they challenge the Commission’s methods of valuing the distillate and resid cuts and its decision that it lacked jurisdiction over the method by which the TAPS Carriers compensate the co-owners of petroleum streams shipped on TAPS. The cases are remanded to the Commission for further consideration in accordance with this opinion.
So ordered.