TASHIMA, Circuit Judge:

The Russian River Watershed Protection Committee and Brenda Adelman (“appellants”) appeal from a judgment entered in favor of the City of Santa Rosa (the “City”) after a bench trial in this Clean Water Act case. On appeal, appellants contend that the City’s method of determining compliance with its National Pollutant Discharge Elimination System permits (“NPDES permits” or “permits”) violated the terms of those permits and the Clean Water Act. They also contend that the district court erred in finding that they did not have standing under the Clean Water Act. Because the NPDES permits do not establish a method of determining compliance, we hold that the Executive Officer of the Regional Board had the discretion under California Water Code section 13223(a) to determine a method of compliance with the permits. We also affirm the district court’s decision that appellants lacked standing under the Clean Water Act.

I. BACKGROUND

A. Statutory Framework

The Clean Water Act regulates the discharge of pollutants into navigable waters, prohibiting their discharge unless certain statutory exceptions apply. 33 U.S.C. § 1311(a). One such exception is where the polluter has been issued an NPDES permit. 33 U.S.C. § 1342. The effluent discharge standards or limitations specified in an NPDES permit define the scope of the authorized exception to the prohibition in section 1311(a). Authority to administer the NPDES permit system may be delegated to a state or regional agency where the state or regional regulatory scheme meets certain criteria. 33 U.S.C. § 1342(b). The entity responsible for issuing permits in the Santa Rosa area is the California North Coast Regional Water Quality Control Board (the “Regional Board”). The Regional Board is established pursuant to the Porter-Cologne *1139 Water Quality Control Act, Cal. Water Code § 13000 et seq., and enforces the provisions of that statute as well. Private citizens may bring suit pursuant to 33 U.S.C. § 1365 to enforce effluent standards or limitations, which are defined to include violations of NPDES permits. 33 U.S.C. § 1365(f)(1). Thus, private citizens may bring suit under the Clean Water Act to enforce the standards set out in NPDES permits.

B. Factual Background

The City operates a wastewater treatment system serving various cities and areas in Sonoma County, California. The system’s main plant is the Laguna Subregional Waste-water Treatment Plant (the “Laguna Plant”) which is located outside the City adjacent to the Laguna de Santa Rosa, a tributary of the Russian River. This plant is considered a major discharger, as defined by the Environmental Protection Agency. The Laguna Plant is a modern facility which treats sewage from participating communities at an advanced level referred to as the tertiary level. Sewage passes through various basins and clarifiers which provide treatment to the secondary level before achieving tertiary level treatment through filtration and disinfection. The sewage reclaimed through this process is authorized for irrigation of food crops, golf courses and school playfields. During the summer, all reclaimed water is used for irrigation and during the winter it is released into the Russian River.

The Regional Board issued NPDES permits to the City in 1986,1990 and 1995. The 1990 and 1995 NPDES permits are substantially similar, both containing discharge prohibitions, numeric effluent limitations, receiving water limitations, and requirements for solids disposal, pretreatment of industrial waste, and monitoring and reporting.

The permits prohibit the discharge of reclaimed water during the summer irrigation season, from May 15 through September 30. Between October 1 through May 14, the City may discharge one percent of the flow of the Russian River after it initially reaches a flow of 1,000 cubic feet per second (“cfs”). The permits state that the 1,000 cfs is to be measured by the gauge at the Hacienda Bridge. The permits do not specify how the flow is to be measured. The discharge is released into the Laguna de Santa Rosa, eight to twelve miles above the gauge at Hacienda Bridge, and takes from one to five days to reach the bridge.

In accordance with California Water Code section 13223(a), the Regional Board delegated all powers authorized by that statute to its Executive Officer. The Executive Officer devised a method known as the “seven-day averaging method” for measuring allowable discharge which has been used throughout the terms of the 1990 and 1995 NPDES permits. Under this method, discharge rates are determined by basing the daily discharge rate on the previous day’s highest hourly flow (recorded at Hacienda Bridge). Discharge percentages are then calculated as a weekly average.

The permits set limitations on the discharge of effluents, such as chlorine, coliform and hydrogen ions, from the treated water. The permits state a minimum value of 1.5 mg/1 for chlorine residual and requires that this minimum value “be maintained at the end of the chlorine contact chamber.” There are three such chambers at the Laguna Plant. The Executive Officer determined that coliform and chlorine would be monitored by taking samples at the end of one of three chlorine contact chambers.

C. Procedural Background

In 1994, after examining the self-monitoring reports of the City, Adelman believed that the City was violating its NPDES permit by (1) discharging treated wastewater in excess of the limits stated in the permit, and (2) not monitoring coliform or chlorine as required by the permit. She notified the Regional Board of her concerns and learned about the City’s method for calculating discharge which she alleges had not been discussed at any public hearing.

On February 21, 1995, appellants sent the City a 60-day notice of violations and intent to file suit, as required by the Clean Water Act. See 33 U.S.C. § 1365(b)(1)(A). They filed suit in district court on May 9, 1995, alleging that the City was in violation of its *1140 NPDES permit. Appellants sought injunc-tive relief, civil penalties, costs, and reasonable attorney’s fees under 33 U.S.C. § 1365(d). The City filed a motion to dismiss or in the alternative a motion for summary judgment based on appellants’ lack of standing to sue because of their inability to allege continuing or recurrent violations of the Clean Water Act. The district court denied both motions. It later granted a motion by appellants to bifurcate the trial into a liability phase and a penalty phase.

The district court conducted two bench trials. The first was limited to determining the proper method for measuring compliance with the NPDES permit. After that trial, the district court made findings of fact and conclusions of law which determined that neither the 1990 nor the 1995 NPDES permit specified the manner in which compliance with the discharge requirements were to be calculated. The court further found that no direct temporal correlation existed between the amount discharged and the flow of the river at the Hacienda Bridge. Further, the court found that the gauge at the Hacienda Bridge specified in the permits only measures the height of the river and not the flow volume. Based on these and other findings of fact, the district court concluded that the Executive Officer of the Regional Board has authority under the Porter-Cologne Water Quality Control Act to establish reasonable procedures for measuring compliance with NPDES permits. The court concluded that the procedure established by the Executive Officer was reasonable.

.Appellants argued that the City violated their permit by discharging in excess of one percent when the variance of the current storage was below the operations curve established to determine the capacity of the Laguna Plant’s storage ponds. However, the district court concluded that “variance above the operations curve” as set out in paragraph B(3) of the permits, is merely one of the factors to be considered by the Executive Officer in deciding whether or not to grant permission to the City to discharge in excess of one percent of the flow.

Regarding the location where numeric effluent limitations are to be measured, the court found that the permits state that the chlorine residual shall be measured “at the end of the chlorine contact chamber.” The court also found that the Laguna Plant uses three chlorine contact chambers and the permits do not specify multiple tests at the end of each chamber. The court further found that the sample taken at the end of one of the chambers is representative of all three. Based on these and other findings of fact, the court concluded that the grab samples for chlorine and coliform are properly taken at the end of one of the three chlorine contact chambers.

At the second trial, the court considered appellants’ evidence of alleged violations in light of its earlier ruling. The court issued Findings of Fact and Conclusion of Law Re Specific Alleged Violations which found that the City had not violated its NPDES permit regarding the discharge limitations. The court did find six coliform violations between January 31 and April 12, 1992. It further found that coliform violations occurred on September 7, 1992 and June 20, 1994. It also found that five coliform violations occurred between March 9 and 13, 1995. The district court found hydrogen ion (pH) samples which violated the 1990 permit but would be in compliance with the 1995 permit. The court found that these violations were not likely to recur. Consequently, the district court found that appellants failed to establish intermittent or continuing violations of the NPDES permit which would give them standing to bring a Clean Water Act citizen suit against the City. The district court entered judgment in favor of the City and appellants timely appealed.

II. JURISDICTION

The district court had jurisdiction pursuant to section 505 of the Federal Water Pollution Control Act (the “Clean Water Act”), 33 U.S.C. § 1365(a). We have jurisdiction over this timely appeal pursuant to 28 U.S.C. § 1291.

III. DISCUSSION

A. Compliance with NPDES Permits

1. Standard of Review

We review the district court’s factual findings under the clear error standard. *1141 Tonry v. Security Experts, Inc., 20 F.Bd 967, 970 (9th Cir.1994). The district court’s adoption of much of the City’s proposed findings does not alter this standard. See Saltarelli v. Bob Baker Group Med. Trust, 35 F.3d 382, 384 (9th Cir.1994); Hagans v. Andrus, 651 F.2d 622, 626 (9th Cir.1981). We review the district court’s conclusions of law de novo. Northwest Envtl. Advocates v. City of Port land, 56 F.3d 979, 982 (9th Cir.1995), cert. denied, 518 U.S. 1018, 116 S.Ct. 2550, 135 L.Ed.2d 1069 (1996); Saltarelli 35 F.3d at 385. The district court’s interpretation of unambiguous terms of the NPDES permit is subject to de novo review. Where the terms of the permit are ambiguous and the district court looks to extrinsic evidence, such findings are reviewed for clear error. Northwest EnvtLAdvocates, 56 F.3d at 982. We do not review issues raised only by an amicus curiae. Swan v. Peterson, 6 F.3d 1373, 1383 (9th Cir.1993). 1

2. Method for Determining Compliance with Permits

a. The Proper Method for Measuring Allowable Discharge

Appellants contend that the City’s method modifies the permit and violates the Clean Water Act. In its first Findings of Fact and Conclusions of Law, the district court found that neither the 1990 nor the 1995 NPDES permit specifies the manner in which compliance with the discharge requirements is to be calculated. This finding is not clearly erroneous. A review of both the 1990 and the 1995 permits indicates that they do not specify any manner in which the discharge requirements are to be calculated. This fact was confirmed by the testimony of the Executive Officer and was not contested by any other testimony. In fact, appellants’ own expert testified that he had personally examined fifteen different methods for measuring compliance with the permits.

The district court did not err in finding that the procedure established by the Executive Officer, based on the seven-day averaging of each previous day’s highest hourly flow as measured at the Hacienda Bridge, is reasonable. Tuck Vath, a staff member of the Regional Board, testified that the seven-day method is used in order to account for irregularities in the system. The Senior Engineer with the Regional Board, Mr. Tancreto, testified that the seven-day method was chosen because it was workable and was based on a short number of days, similar to the other time periods contained in the permit. Miles Ferris, the Director of Utilities Engineer for the City, testified that the seven-day averaging method was chosen after comparing numerous other methods. Mr. Ferris explained that he searched for methods that would approximate the real flows of the Russian River. He examined a whole batch of flows and attempted to predict them and the closest he came was by selecting the highest hour of the previous day. This method— which was incorporated into the seven-day averaging method — was the closest approximation to the actual flow of the Russian River. Mr. Ferris further testified that this method was the most logical method that people could understand.

The district court properly gave substantial deference to the Executive Officer’s selection of the seven-day averaging method because it was a reasonable interpretation of the NPDES permit. Cf. Arkansas v. Oklahoma, 503 U.S. 91, 110, 112 S.Ct. 1046, 1058-59, 117 L.Ed.2d 239 (1992) (holding that substantial deference should be given to an agency’s reasonable and consistently held interpretation of its own regulations).

The district court further found that no direct temporal correlation exists between the amount discharged and the flow of the river at the Hacienda Bridge as the dis *1142 charge points are eight to twelve miles upstream from the Hacienda gauging station. It found that when the City operator sets the discharge rate each morning, the only information available to the operator is the previ-mus day’s flow measurements. These findings are not clearly erroneous. Messrs. Tancreto and Ferris testified that it can take up to several days for discharged water to reach the Hacienda Bridge. Mr. Ferris further testified that there is no direct relationship between the discharged amount and the flow of the Russian River at Hacienda Bridge.

Appellants contend that the City can determine the flow of the Russian River from the gauge at the Hacienda Bridge. However, the district court found that the gauge at the Hacienda Bridge only measures the height of the river. In order to calculate flow volume, the bottom contour of the river must be measured, which the U.S. Geological Survey (USGS) does periodically. Such measurements are passed on after a delay to the City. Therefore, the district court found that the true flow at the Hacienda Bridge cannot be determined on a day-to-day basis. Messrs. Ferris and Tancreto both testified that the gauge at the Hacienda Bridge yields only elevation data and not accurate flow information. Mr. Ferris explained that in order to determine the true flow of the River, one needs stream gauge measurements from the USGS which measures the stream bottom periodically. Mr. Ferris further testified that the City only receives this information from the USGS after some delay. The district court’s findings were not clearly erroneous.

Appellants contend that the seven-day averaging method is improper because it was never presented to the Regional Board. However, the district court further found that the seven-day averaging method was presented to the Regional Board during the 1995 permit hearings. This finding is not clearly erroneous. Mr. Vath testified that the method was discussed in the context of an attempt to move to a thirty-day averaging method. Ms. Adelman did not dispute that Mr. Vath explained the seven-day averaging method at a Regional Board meeting in 1995 which she attended. The transcripts of that meeting which Ms. Adelman read in court explicitly discuss the seven-day averaging method.

Appellants further contend that the City violated the permit by discharging in excess of one percent when the variance in the storage ponds was below the operations curve at the time of discharge. The district court found that “variance above the operations curve” is merely one of the factors to be considered by the Executive Officer in deciding whether or not to grant permission to the City to discharge in excess of one percent of the flow. Reviewing the NPDES permits de novo, we conclude that the district court was correct. Northwest Envtl. Advocates, 56 F.3d at 982 (district court’s interpretation of NPDES permits reviewed de novo). Messrs. Kor and Vath testified that “variance above the operations curve” is one of the factors that they did consider in deciding whether or not to grant the City permission to discharge in excess of one percent of the flow.

In sum, appellants’ argument boils down to their dissatisfaction with the method used by the City to calculate compliance with the NPDES permits. Appellants emphasize that the permits state, “In no case shall any discharge of wastewater exceed five percent of the flow of the Russian River.” They seek as close to real-time measurement of the discharge as possible. They prefer a daily monitoring system to the seven-day averaging method selected by the Executive Officer. Appellants successfully opposed the use of a different method at the hearings preceding adoption of the 1995 permit, but did not succeed in opposing the seven-day averaging method. Appellants’ preference, however, is unsupported by any showing that the seven-day averaging method is improper. The district court’s factual findings are supported by the evidence and are not clearly erroneous.

b. Location for Measuring Effluent Limitations

Appellants contend that the City failed to monitor final effluent chlorine and coliform organisms as required by the permits. The district court found that the permits require measurement of the numeric *1143 effluent limitations “at the end of the chlorine contact chamber,” and that the permits do not specify multiple tests at the end of each of the three chambers used at the Lagu-na Plant. Our review of the language of the NPDES permits indicates that the district court’s findings are not clearly erroneous.

The finding that the sample taken at the end of one of the chlorine contact chambers is representative of all three chambers is not clearly erroneous. Mr. William Stinebaugh, the Deputy Director of Operations for the City, testified that he oversees the Laguna Plant and that there is no differential between taking samples at the end of one of the chambers compared to taking them at the end of all of the chambers. Mr. Adam Russell, the clerk technician with the City who actually takes the samples, testified that taking samples from one chamber is representative of all three. The district court therefore properly concluded that the grab samples for chlorine and coliform are properly taken at the end of one of the three chlorine contact chambers.

3. Discretion of the Executive Officer

Appellants contend that the district court erred in concluding that California Water Code section 13223(a) grants the Executive Officer the authority to determine the method of compliance with the permit. They further contend that the Executive Officer’s interpretation constituted de facto modifications of the permits in violation of the statute.

California Water Code section 13223(a) gives a Regional Board the authority to delegate any of its powers and duties, with limited exceptions, to its Executive Officer. 2 In this case, the Regional Board so delegated its authority. The Executive Officer routinely established the method for measuring compliance with NPDES penhits including those at issue here.

There is no caselaw interpreting section 13223(a), but under federal law, the establishment of a method of compliance with an NPDES permit does not constitute a modification of the permit. Cf. Citizens for a Better Env’t v. Union Oil Co., 83 F.3d 1111, 1119-20 (9th Cir.1996) (holding that the Regional Board’s cease and desist order deferring compliance with a permit condition did not constitute a modification of the terms of the NPDES permit). We therefore conclude that the Executive Officer’s adoption, of a method of compliance was not a “modification” of the permit in violation of section 13223.

4. Standing

The Clean Water Act does not permit citizen suits for wholly past violations. Gwaltney of Smithfield, Ltd. v. Chesapeake Bay Found., Inc., 484 U.S. 49, 64, 108 S.Ct. 376, 385, 98 L.Ed.2d 306 (1987). To have standing under the Clean Water Act, appellants must prove the existence of ongoing violations or the reasonable likelihood of continuing future violations. Id. at 57, 108 S.Ct. at 381. The district court found that numerous allegations of violations could not be substantiated given the validity of the seven-day averaging method. The court did find seven coliform violations in 1992, one in 1994, and five in 1995. It also found six pH violations in 1992 and one in 1993. The district court correctly held that appellants did not prove the existence of ongoing violations or the reasonable likelihood of continuing future violations. The coliform violations resulted from changes in the treatment process requested by the Regional Board, faulty testing, out of service tanks, and a major flood. The pH violations that occurred would not have violated the 1995 permit. The 1995 permit lowered the minimum pH level from 6.5 to 6.0. Because the pH level has never fallen below 6.0, the likelihood of future violations is minimal. Under Gwaltney, appellants do not have standing under the Clean Water Act.

*1144 B. Costs

1. Standard of Review

We review a district court’s award of costs for an abuse of discretion. Schwarz v. Secretary of Health & Human Servs., 73 F.3d 895, 900 (9th Cir.1995); National Info. Servs., Inc. v. TRW, Inc., 51 F.3d 1470, 1471 (9th Cir.1995). Whether the district court has the authority to award costs, however, is a question of law reviewed de novo. United States ex rel. Lindenthal v. General Dynamics Corp., 61. F.3d 1402, 1412 n. 13 (9th Cir.1995), cert. denied, 517 U.S. 1104, 116 S.Ct. 1319, 134 L.Ed.2d 472 (1996).

2. Award of Costs to Defendant

Appellants contend that the district court erred in awarding costs to the City. They assert that under the Clean Water Act, the district court may only award costs to the prevailing defendant where the plaintiffs’ claim was frivolous, meritless, or vexatious. 3 Appellants’ contention is incorrect.

The City correctly argues that the district court was empowered by Federal Rule of Civil Procedure 54(d) 4 to make an award of taxable costs to the prevailing party. The cases on which appellants rely discuss only the question of attorney’s fees which were not taxed as costs here. 5 This circuit has explicitly rejected the position proffered by appellants. See National Org. for Women v. Bank of Cal., NA, 680 F.2d 1291, 1294 (9th Cir.1982) (refusing to apply the Christiansburg standard to cost awards). Rule 54(d)(1) creates a presumption in favor of awarding costs to the prevailing party which may only be overcome by pointing to some impropriety on the part of the prevailing party that would justify a denial of costs. National Info. Servs., 51 F.3d at 1471-72 (citations omitted). Appellants have alleged no such improprieties. Accordingly, the district court properly awarded costs to the City under Rule 54(d)(1).

IV. CONCLUSION

Because the NPDES permits did not set out a method of determining compliance, the Executive Officer of the Regional Board had the discretion under Cal. Water Code § 13223(a) to determine a reasonable method of compliance with the permits. The selection of the seven-day averaging period was reasonable, as was the decision to sample chlorine and eoliform at the end of one of the three chlorine contact chambers. Finally, the district court did not err in awarding costs to the City. 6

AFFIRMED.

. Amicus California State Water Resources Control Board (the "State Board”) contends that the doctrines of administrative res judicata and collateral estoppel bar appellants' challenge to the Regional Board's method for measuring compliance with the NPDES permits. We decline to address this argument because, as the State Board candidly acknowledges, it is raised for the first time on appeal, and not by any party. See Swan, 6 F.3d at 1383. Generally, we will not consider on appeal an issue raised only by an amicus. Id. (citing Sanchez-Trujillo v. INS, 801 F.2d 1571, 1581 n. 9 (9th Cir.1986); Preservation Coalition, Inc. v. Pierce, 667 F.2d 851, 862 (9th cir.1982)). Appellee did not adopt amicus the State Board’s argument in its brief; therefore, the issue has been waived. See Swan, 6 F.3d at 1383.

. Section 13223(a) provides, inter alia:

Each regional board may delegate any of its powers and duties vested in it by this division to its executive officer excepting only the following:

... (2) the issuance, modification, or revocation of any water quality control plan, water quality objectives, or waste discharge requirement---Cal. Water Code § 13223(a).

. Appellants rely on Ruckelshaus v. Sierra Club, 463 U.S. 680, 682 n. 1, 103 S.Ct. 3274, 3276 n. 1, 77 L.Ed.2d 938 (1983) and Christiansburg Garment Co. v. EEOC, 434 U.S. 412, 98 S.Ct. 694, 54 L.Ed.2d 648 (1978).

. The Rule states:

"(1) Costs Other than Attorneys’ Fees. Except when express provision is made either in a statute of the United States or in these rules, costs other than attorneys' fees shall be allowed as of course to the prevailing party unless the court otherwise directs ... Such costs may be taxed by the clerk on one day's notice...."

Fed.R.Civ.P. 54(d)(1).

. Costs were taxed in the amount of $14,275.53 and did not include any attorney’s fees. The clerk reduced the claim for expert witness fees to "ordinary" witness fees.

.We have also reviewed the evidentiary rulings of the district court challenged by appellants. Our review is governed by the abuse of discretion standard. See Masson v. New Yorker Magazine, Inc., 85 F.3d 1394, 1399 (9th Cir.1996) (admission or exclusion of evidence); United States v. Easter, 66 F.3d 1018, 1020 (9th Cir.1995) (relevance of evidence), cert. denied, 516 U.S. 1150, 116 S.Ct. 1026, 134 L.Ed.2d 104 (1996); United States v. VonWillie, 59 F.3d 922, 929 (9th Cir. 1995) (admission of lay opinion testimony). After a careful review of the record, we conclude that the district court did not abuse its discretion in making any of the challenged evidentiary rulings.