In 2012, the Navajo Nation and several of its individual members (collectively, the Navajo Nation) sued San Juan County, alleging that the election districts for both the school board and the county commission violated the Equal Protection Clause of the Fourteenth Amendment to the United States Constitution and the Voting Rights Act (VRA) of 1965, 52 U.S.C. §§ 10301 -14 (transferred from 42 U.S.C. §§ 1973 - 1973o ). The district court denied the county's motion to dismiss, found that the election districts violated the Equal Protection Clause, and awarded summary judgment to the Navajo Nation. It later rejected the county's proposed remedial redistricting plan because it concluded the redrawn districts again violated the Equal Protection Clause. The district court then appointed a special master to develop a proposed remedial redistricting plan, directed the county to adopt that remedial *1275 plan, and ordered the county to hold special elections based on that plan in November 2018.

On appeal, the county challenges each of the district court's decisions. For the reasons explained below, we affirm.

Background

San Juan County occupies the southeastern corner of Utah. Geographically, it's the largest county in the state. The county seat is Monticello, and the county's largest city is Blanding. Approximately 52% of the county's population is Native American. Most Native American residents live in the southern portion of the county, on the Navajo Nation Reservation; the northern portion of the county contains "most of the non-Hispanic [w]hite population." App. vol. 43, 8420. As of the 2010 census, 14,746 people lived in the county.

A three-member county commission governs the county. Until the early 1980s, the county elected its commissioners in at-large elections. But in 1983, the United States sued the county, alleging that the at-large elections violated the Constitution and § 2 of the VRA because they denied Native American residents "an equal opportunity to participate in the [c]ounty political process and to elect candidates of their choice." 1 App. vol. 2, 277. In its complaint, the United States pointed out that although the county had a substantial Native American population, it had never elected a Native American representative to the county commission. 2

Rather than going to trial, the county entered into a consent decree with the United States. The district court accepted the parties' agreement and entered a settlement order. The settlement order acknowledged that the county's at-large election system "fail[ed] to comply fully with the requirements of [§] 2 of the [VRA]." Id. at 223 . As such, it provided that the county would adopt "fairly drawn single[-]member districts." Id.

Accordingly, the county established three single-member county-commission districts: Districts 1, 2, and 3. A former county official who helped design the districts "testified that she understood that District [3] 'was to be heavily loaded with Navajo voters.' " App. vol. 49, 9836 (quoting App. vol. 30, 5580). Thus, when the county first created the single-member districts in the 1980s, Native Americans made up 88% of District 3's population. And in the years since the county adopted single-member districts, "the commissioners elected from Districts [1] and [2] have been white, and the commissioners elected from District [3] have been Native American." App. vol. 3, 450. As the district court put it, "[i]n this way, the [c]ounty moved from a system that historically denied representation to a minority group to one that allowed th[e] group greater participation in the political process." App. vol. 49, 9834.

In 2011, the Navajo Nation asked the county to redraw the county-commission districts in response to the 2010 census. It argued that District 3, which was now over *1276 92% Native American, had "an inordinately large population of Native Americans." App. vol. 3, 450. The county declined to change District 3's boundaries, and it made only a few small changes to Districts 1 and 2 to equalize the population of those districts. In response, the Navajo Nation brought this action against the county in federal district court. The Navajo Nation alleged that the boundaries of the county-commission districts, specifically the boundaries of District 3, were unconstitutionally based on race in violation of the Equal Protection Clause.

Additionally, the Navajo Nation challenged the constitutionality of the county's school-board districts. By state law, the school board has "five members, each elected from a single[-]member district." App. vol. 43, 8344; see also Utah Code Ann. § 20A-14-202(1)(a), (h). As of the 2010 census, the total population deviation among the five school-board districts was around 38%. In other words, the districts weren't equally populated; some districts contained substantially more voters than other districts. The Navajo Nation contended that this high population deviation resulted in vote dilution in violation of the Equal Protection Clause's guarantee of one person, one vote. See Avery v. Midland Cty. , 390 U.S. 474 , 478, 88 S.Ct. 1114 , 20 L.Ed.2d 45 (1968) (explaining that right to vote "is infringed when legislators are elected from districts of substantially unequal population"). 3

The county moved to dismiss the Navajo Nation's claim related to the county-commission districts, arguing that it was an impermissible collateral attack on the 1984 consent decree and settlement order. It also argued that the United States was an indispensable party to any litigation related to the consent decree and settlement order. The district court disagreed and denied the county's motion. 4

The parties then cross-moved for summary judgment on the Navajo Nation's two Equal Protection claims: (1) that District 3 of the county-commission districts was unconstitutionally based on race; and (2) that the population deviation among the school-board districts caused unconstitutional (although not race-related) vote dilution. The district court found that both the school-board and county-commission districts violated the Equal Protection Clause and were therefore unconstitutional. Accordingly, it awarded summary judgment to the Navajo Nation on both claims.

The district court then ordered the county to develop a remedial redistricting plan. 5 See Large v. Fremont Cty. , 670 F.3d 1133 , 1138 (10th Cir. 2012) (noting that when court declares election districts unconstitutional, it should allow legislature to create new plan). It stated that "it would adopt [the county]'s proposed remedial plan[ ] if [the plan] cured the identified violations and [was] otherwise legally sound." App. vol. 54, 10825-26; see also Large , 670 F.3d at 1138 (noting that district *1277 court should adopt legislature's remedial plan "unless it, too, is challenged and found to violate the Constitution" (quoting Wise v. Lipscomb , 437 U.S. 535 , 540, 98 S.Ct. 2493 , 57 L.Ed.2d 411 (1978) (plurality opinion))).

After the county submitted its proposed remedial redistricting plan, the district court found that the remedial plan was also unconstitutional. Specifically, it found that some districts in the remedial plan were based on race and didn't survive strict scrutiny. So the district court rejected the county's plan and appointed a special master. After the special master created several proposed redistricting plans, the district court conducted two public hearings and accepted input from the parties. The special master then drafted a final plan, and the district court ordered the county to adopt it. The district court further ordered the county to hold special elections utilizing the new districts in November 2018.

The county appeals.

Analysis

The county raises five challenges; we examine each in turn. We first address the county's argument that the district court erred in denying its motion to dismiss the Navajo Nation's claim relating to the 2011 county-commission districts. Second, we consider whether the district court erred when it ruled that the county lacked a compelling interest to justify the racially drawn boundaries of county-commission District 3. Third, we review the county's assertion that the district court erred in rejecting the county's justifications for the population deviation in the 1992 school-board districts. Fourth, we ask whether the district court erred in finding that the county's proposed remedial redistricting plan was predominantly based on race and failed to satisfy strict scrutiny. Last, we consider whether, as the county contends, the district court erroneously ordered the county to adopt the special master's remedial redistricting plan. 6

I. The County's Motion to Dismiss

The county first argues that the district court erred in denying its motion to dismiss. In particular, it argues that (1) the 1984 consent decree and settlement order bar the claim related to the county-commission districts because the court that approved the consent decree and entered the settlement order retained jurisdiction over the matter; and (2) the United States, as a party to the consent decree and settlement order, is indispensable to this action. Neither argument succeeds.

When a court accepts a consent decree and enters an accompanying order, it often retains jurisdiction over the matter. If it does so, other courts necessarily lack subject-matter jurisdiction over suits seeking to modify that decree. See Culbreath v. Dukakis , 630 F.2d 15 , 22 (1st Cir. 1980) (noting that "only the district court supervising implementation of the decree will have subject[-]matter jurisdiction to modify the decree"). That typically means parties to a consent decree can't collaterally attack the judgment in a separate action. See Floyd v. Ortiz , 300 F.3d 1223 , 1225 n.1 (10th Cir. 2002) (noting that party to consent judgment can't collaterally attack consent judgment); Barfus v. City of Miami , 936 F.2d 1182 , 1185 (11th Cir. 1991) (citing cases "holding that a party or *1278 privy to a consent decree cannot launch a collateral attack upon the decree").

And that's what the county urges happened here: it contends that this case is an impermissible collateral attack on the consent decree and settlement order. The district court rejected this subject-matter-jurisdiction argument for two reasons. First, it found that because the Navajo Nation wasn't a party to the consent decree and settlement order, the rule that "parties to a consent decree may not collaterally attack the judgment in a separate action" didn't apply. App. vol. 3, 452. Second, it determined that "the present suit does not address the subject matter of the 1984 [consent decree and settlement order]." Id. at 453. Specifically, the district court noted that the consent decree and settlement order aimed to eliminate at-large elections but said nothing about specific boundary lines or the makeup of any single-member district. And in this case, the Navajo Nation didn't seek to return to at-large elections; instead, it sought to redraw certain boundaries. Thus, the district court concluded that although the court that approved the consent decree and entered the settlement order "retain[ed] jurisdiction for all purposes," this case didn't implicate "th[e] matter" over which jurisdiction was retained. Id. (quoting App. vol. 2, 224).

On appeal, the county challenges both of these rationales. "We review de novo the district court's denial of a motion to dismiss for lack of subject[-]matter jurisdiction." Opala v. Watt , 454 F.3d 1154 , 1156-57 (10th Cir. 2006). The county first contends that the Navajo Nation was a party to the consent decree and settlement order because the United States brought the 1983 lawsuit "on behalf of [the] Navajo Nation." Aplt. Br. 37. But this argument misses the mark because the 1983 complaint plainly states that it was brought "on behalf of the United States of America," not the Navajo Nation. App. vol. 2, 273.

Resisting this plain language, the county points out that the 1983 complaint identifies 25 U.S.C. § 175 as one of the statutes authorizing the lawsuit. Section 175 provides that "[i]n all [s]tates ... where there are reservations or allotted Indians[,] the United States [A]ttorney shall represent them in all suits at law and in equity." Traditionally, Native American litigants invoke this statute when they wish to have the United States Attorney represent their interests in court. See, e.g. , Oviatt v. Reynolds , 733 F. App'x 929 , 931 (10th Cir. 2018) (unpublished) (exercising discretion to deny § 175 motion to appoint United States Attorney); Siniscal v. United States , 208 F.2d 406 , 410 (9th Cir. 1953) (holding that § 175 "is not mandatory and that its purpose is no more than to [e]nsure [Native Americans] adequate representation in suits to which they might be parties").

But the county cites no authority for the proposition that the United States' mere mention of § 175 somehow made the Navajo Nation a party to the 1983 lawsuit or brought the Navajo Nation into privity with the United States. Nor does it point to any fact tending to show that the Navajo Nation invoked § 175 and asked the United States to bring the 1983 lawsuit on its behalf. Indeed, although the 1983 complaint cites § 175, it plainly relies on the VRA as authority for the action. The VRA permits the United States to institute a civil action whenever "any person has engaged or there are reasonable grounds to believe that any person is about to engage in any act or practice which would deprive any other person of" the right to vote. § 10101(c). In line with this authority, the 1983 complaint alleges violations of the VRA and requests relief under the VRA.

*1279 Accordingly, we conclude that the mere reference to § 175 on the first page of the 1983 complaint doesn't establish that the United States acted as a representative of the Navajo Nation in bringing the 1983 lawsuit. Thus, we reject the county's argument that the Navajo Nation was a party to the earlier lawsuit simply because the United States used the VRA to protect the voting rights of Native Americans living in San Juan County. And because the Navajo Nation wasn't a party to the consent decree and settlement order, the collateral-attack doctrine doesn't bar this action. Cf. Floyd , 300 F.3d at 1225 n.1 (prohibiting collateral attack by party or person in privity with party to settlement order).

The county next argues that the Navajo Nation's suit seeking to redraw the boundaries of the county-commission districts implicates the 1984 consent decree and settlement order because those documents required the United States to approve the county's new single-member districts. The district court found it "reasonable to infer from the record that the [c]ounty presented a final plan to the [United States] before enacting it." App. vol. 49, 9834 n.34. We accept this inference. But the county takes this inference too far, insisting that (1) the county created District 3's boundaries "at the insistence of the" United States, and (2) these boundaries are permanent and can't be changed without involvement and permission of the United States. Aplt. Br. 37.

Yet we see no support for these additional inferences in the record. Even if the United States approved the original single-member district boundaries, nothing in the 1984 consent decree and settlement order prohibits the county from altering those boundaries over time. The record does indicate that one county official testified that she believed the consent decree and settlement order (1) established District 3's boundaries such that the district would be packed with Navajo voters and (2) locked those boundaries. But those documents simply do not support the official's testimony. Indeed, the county's opening brief at one point concedes that the consent decree and settlement order "do[ ] not dictate that the boundaries of District[ ]3 remain unchanged or even that District[ ]3 be a Navajo-majority" district. Aplt. Br. 42. Accordingly, the district court correctly determined that the Navajo Nation's claim seeking to modify District 3's boundaries doesn't implicate the consent decree and settlement order. As a result, we conclude that although the district court retained jurisdiction over the 1983 consent decree and settlement order, that fact didn't deprive the district court in this case of jurisdiction over the Navajo Nation's claim related to the county-commission districts. 7

Relatedly, and briefly, the county argues that the district court abused its discretion when it ruled that the United States wasn't an indispensable party under Federal Rule of Civil Procedure 19. See N. Arapaho Tribe v. Harnsberger , 697 F.3d 1272 , 1277 (10th Cir. 2012) (noting that we review Rule 19 determinations for abuse of discretion). The county suggests that because the consent decree is a contract, all parties to that contract-including the United States-must be involved in any litigation related to that contract. But as we've just detailed, the Navajo Nation's claim related to the county-commission districts doesn't implicate the consent decree and settlement order. As such, the district court didn't abuse its discretion in *1280 concluding that the United States wasn't indispensable to this action.

In sum, because this case doesn't implicate the 1984 consent decree and settlement order, we affirm the district court's order denying the county's motion to dismiss.

II. The 2011 County-Commission Districts

Next, the county argues that the district court erred in granting summary judgment to the Navajo Nation on the claim that the boundaries of District 3 of the 2011 county-commission districts were unconstitutionally based on race in violation of the Equal Protection Clause. "We review a grant of a motion for summary judgment de novo, applying the same legal standard used by the district court." Adarand Constructors, Inc. v. Slater , 228 F.3d 1147 , 1161 (10th Cir. 2000) (quoting McGarry v. Bd. of Cty. Comm'rs , 175 F.3d 1193 , 1198 (10th Cir. 1999) ). Under that legal standard, summary judgment is appropriate if there are no genuine issues of material fact and the moving party is entitled to judgment as a matter of law. Fed. R. Civ. P. 56(a).

The Fourteenth Amendment's Equal Protection Clause "limits racial gerrymanders in legislative districting plans." Cooper v. Harris , --- U.S. ----, 137 S. Ct. 1455 , 1463, 197 L.Ed.2d 837 (2017). That means a state cannot, "in the absence of 'sufficient justification,' ... 'separat[e] its citizens into different voting districts on the basis of race.' " Id. (quoting Bethune-Hill v. Va. State Bd. of Elections , --- U.S. ----, 137 S. Ct. 788 , 797, 197 L.Ed.2d 85 (2017) ). In racial gerrymandering cases, the initial burden is on the plaintiff to "prove that 'race was the predominant factor motivating the legislature's decision to place a significant number of voters within or without a particular district.' " Id. (quoting Miller v. Johnson , 515 U.S. 900 , 916, 115 S.Ct. 2475 , 132 L.Ed.2d 762 (1995) ). It makes no difference whether "the reason for the racial classification is benign or the purpose remedial." Shaw v. Hunt , 517 U.S. 899 , 904-05, 116 S.Ct. 1894 , 135 L.Ed.2d 207 (1996). If a plaintiff proves that race was the predominant factor, "the design of the district must withstand strict scrutiny." Cooper , 137 S. Ct. at 1464. Under strict scrutiny, "[t]he burden ... shifts to the [government] to prove that its race-based sorting of voters serves a 'compelling interest' and is 'narrowly tailored' to that end." Id. (quoting Bethune-Hill , 137 S. Ct. at 800 ).

Here, the parties largely agree on the underlying material facts. Most critically, the county "admit[s]" that it drew District 3's boundaries "based on race." Aplt. Br. 41. Thus, we need not consider whether the Navajo Nation sufficiently proved that race predominated in the creation of District 3; the county has conceded as much. Instead, we need only determine whether, as the county argues, the district court erred in concluding that it couldn't satisfy the strict scrutiny applied to that race-based decision.

In reaching that conclusion, the district court rejected the county's argument that it had a compelling interest in adhering to the terms of the 1984 consent decree and settlement order. Specifically, the district court acknowledged that county officials subjectively believed that the consent decree and settlement order required packing Navajo voters into District 3. But it found that this subjective belief wasn't enough to establish a compelling interest. In particular, the district court stated that "subjective beliefs cannot supply legal requirements ... that are not found in the documents themselves." App. vol. 49, 9855. It further noted that nothing in the consent decree or settlement order established specific district lines or provided *1281 that such lines must exist, unchanged, in perpetuity; nor did those documents refer to packing Native American voters into a single district. As such, the district court concluded that complying with the consent decree and settlement order did not provide a compelling interest to justify District 3's race-based boundaries. Thus, the district court granted summary judgment in favor of the Navajo Nation on its claim that the boundaries of District 3 of the 2011 county-commission districts were unconstitutionally based on race in violation of the Equal Protection Clause. We review that ruling de novo. See Concrete Works of Colo., Inc. v. City & Cty. of Denver , 321 F.3d 950 , 958 (10th Cir. 2003) (noting that we review de novo whether party satisfies strict scrutiny).

In seeking reversal of this ruling, the county first contends, as it did below, that the 1984 consent decree and settlement order required it to maintain District 3's race-based boundaries. But the county fails to "explain ... why the district court's decision was wrong." Nixon v. City & Cty. of Denver , 784 F.3d 1364 , 1366 (10th Cir. 2015). In fact, the county advances no argument on this front, and it goes as far as to admit that the consent decree and settlement order don't "dictate that the boundaries of District[ ]3 remain unchanged or even that District[ ]3 be a Navajo-majority" district. Aplt. Br. 42. Thus, we find this argument waived through inadequate briefing and decline to address it. See Fed. R. App. P. 28(a)(8) (requiring argument section of appellant's opening brief to contain "appellant's contentions and the reasons for them, with citations to the authorities ... on which the appellant relies"); Nixon , 784 F.3d at 1369 (declining to consider argument inadequately briefed on appeal).

Rather than challenging any aspect of the district court's sound analysis of the language of the consent decree and settlement order, the county instead advances a new compelling-interest argument: it asserts that it has a compelling interest in complying with § 2 of the VRA. Although we assume that "compl[ying] with § 2 of the VRA constitutes a compelling government interest," we nevertheless reject the county's argument. Sanchez v. Colorado , 97 F.3d 1303 , 1328 (10th Cir. 1996) ; see also Cooper , 137 S. Ct. at 1464 (assuming that complying with VRA is compelling interest). That's because the compelling interest is only half the battle of strict scrutiny. And the county doesn't survive the other half because, for the reasons discussed below, it can't show that District 3's race-based boundaries were narrowly tailored to further that interest. 8

Narrow tailoring in the context of VRA compliance means that the county must show "it had 'a strong basis in evidence' for concluding that the [VRA] required its action," or, stated differently, "that it had 'good reasons' to think that it would transgress the [VRA] if it did not draw race-based district lines." Cooper , 137 S. Ct. at 1464 (quoting Ala. Leg. Black Caucus v. Alabama , --- U.S. ----, 135 S. Ct. 1257 , 1274, 191 L.Ed.2d 314 (2015) ). On appeal, the county asserts that the consent decree and settlement order provide the required strong evidence and good reasons for its race-based line drawing. But this argument ignores the context and language of the consent decree and settlement order. As discussed, the county entered into the consent decree to remedy potential VRA violations resulting from its at-large elections. But the consent decree and settlement order themselves didn't *1282 mandate the composition of the single-member districts. Nor did they set the boundaries of those districts. So complying with the 1984 consent decree and settlement order didn't supply the county with a good reason to pack Native American voters into District 3 in 2011.

Indeed, there's no evidence that when redistricting in 2011, the county ever attempted to determine what § 2 required. Instead, the evidence shows that (1) the county mistakenly believed it wasn't permitted to modify District 3's boundaries, and (2) it maintained this belief despite being unsure of what the consent decree and settlement order required. The county's mistaken understanding simply doesn't provide "a strong basis in evidence" or a "good reason[ ]" to believe that § 2 required this racially drawn boundary. Cooper , 137 S. Ct. at 1464 (quoting Ala. Leg. Black Caucus , 135 S. Ct. at 1274 ). Thus, even assuming VRA compliance provides the county with a compelling interest, the county fails to establish-for purposes of narrow tailoring-that it had good reasons or a strong basis in evidence to believe that the VRA required maintaining District 3's race-based boundaries.

Accordingly, we affirm the district court's order granting summary judgment to the Navajo Nation on its claim that the county created county-commission District 3 based on race in violation of the Equal Protection Clause.

III. The 1992 School-Board Districts

The county next challenges the district court's order granting summary judgment to the Navajo Nation on its claim that the school-board districts violate the Equal Protection Clause because the districts contain substantially unequal populations, in violation of the one-person, one-vote principle. The same summary-judgment standard of review applies to this issue, and the parties again agree on the underlying facts.

In addition to prohibiting race-based classifications that fail to survive strict scrutiny, the Equal Protection Clause of the Fourteenth Amendment "prohibits states from restricting or diluting votes in violation of the 'one person, one vote' principle." City of Herriman v. Bell , 590 F.3d 1176 , 1185 (10th Cir. 2010). The Supreme Court first described this principle in Reynolds v. Sims , 377 U.S. 533 , 84 S.Ct. 1362 , 12 L.Ed.2d 506 (1964), and extended it to local governments in Avery , 390 U.S. 474 , 88 S.Ct. 1114 . Under the one-person, one-vote principle, the right to vote "is infringed when legislators are elected from districts of substantially unequal population." Avery , 390 U.S. at 478 , 88 S.Ct. 1114 . Or, as the district court in this case put it, "election districts must be of substantially equal population, to avoid over[ ]representing members of one district (whose votes count proportionally more) or under[ ]representing members of another (whose votes count proportionally less)." App. vol. 43, 8351.

At the same time, "it is a practical impossibility to arrange legislative districts so that each one has an identical number of residents, or citizens, or voters." Reynolds , 377 U.S. at 577 , 84 S.Ct. 1362 . Recognizing this reality, the Court has held that "a maximum population deviation under 10%" is a "minor deviation[ ]" that doesn't "make out a prima facie case of invidious discrimination under the Fourteenth Amendment." Brown v. Thomson , 462 U.S. 835 , 842, 103 S.Ct. 2690 , 77 L.Ed.2d 214 (1983). Conversely, a plan with population disparities over 10% "creates a prima facie case of discrimination and therefore must be justified by the [redistricting entity]." Id. at 842-43 , 103 S.Ct. 2690 . Here, the deviation is around 38%-well over the 10% threshold. Indeed, the county didn't *1283 argue below and doesn't argue on appeal that the Navajo Nation failed to make out a prima facie case of unconstitutional vote dilution. Instead, the question is whether the county can adequately justify the disparity.

Initially, the parties disagree about the level of scrutiny we should apply to the county's justifications for deviating from the one-person, one-vote principle. The Navajo Nation argues that we should apply strict scrutiny, and the county suggests that we should instead follow the district court's lead and use the intermediate balancing test the Court articulated in Anderson v. Celebrezze , 460 U.S. 780 , 103 S.Ct. 1564 , 75 L.Ed.2d 547 (1983), and Burdick v. Takushi , 504 U.S. 428 , 112 S.Ct. 2059 , 119 L.Ed.2d 245 (1992) (the Anderson - Burdick test). 9 Compare Carlson v. Wiggins , 675 F.3d 1134 , 1139 (8th Cir. 2012) ("[D]eviations from [one person, one vote] are subject to strict scrutiny."), with Daly v. Hunt , 93 F.3d 1212 , 1218 n.8 (4th Cir. 1996) ("The one[-]person, one[-]vote principle is apparently not subject to strict scrutiny ...."). But we need not decide, in this case, which level of scrutiny applies to all one-person, one-vote challenges. That's because even under the lower standard of the intermediate Anderson - Burdick test, the county's justifications fail.

Under the Anderson - Burdick test, we "weigh 'the character and magnitude of the asserted injury' ... against 'the precise interests put forward by the [government] as justifications for the burden imposed by its rule,' taking into consideration 'the extent to which those interests make it necessary to burden the plaintiff's rights.' " Burdick , 504 U.S. at 434 , 112 S.Ct. 2059 (quoting Anderson , 460 U.S. at 789 , 103 S.Ct. 1564 ). This test provides a sliding scale under which "the rigorousness of [the] inquiry into the propriety of a state election law depends upon the extent to which a challenged regulation burdens" individual rights. Id.

We begin with "the character and magnitude of the asserted injury." Anderson , 460 U.S. at 789 , 103 S.Ct. 1564 . On this point, the county takes no issue with the district court's conclusion that "[t]wo decades of significant non[ ]compliance with the one-person, one-vote principle on its own strongly suggests an arbitrary abdication of constitutional responsibilities." App. vol. 43, 8363. Indeed, when the county first drew the districts in 1992, the deviation was 18.7%. And the county never attempted to correct this deviation. In fact, it failed to redistrict every 10 years as required by state law. See Utah Code Ann. § 20A-14-201(2)(a)(i). This "stark and dramatic population deviation ... render[s] the apportionment more constitutionally suspect than a deviation closer to the 10% safe harbor." App. vol. 43, 8362; see also Mahan v. Howell , 410 U.S. 315 , 329, 93 S.Ct. 979 , 35 L.Ed.2d 320 (1973) (noting that deviation of 16.4% "may well approach tolerable limits"); cf. Daly , 93 F.3d at 1218 ("[T]here is a level of population disparity beyond which a [districting entity] can offer no possible justification.").

Likewise, the county doesn't dispute that the seriousness of the injury increases "the rigorousness of [the] inquiry into the propriety" of the county's justifications.

*1284 Burdick , 504 U.S. at 434 , 112 S.Ct. 2059 . Instead, the county primarily challenges the manner in which the district court analyzed the county's proffered justifications for the population disparity. See id. (directing court to review "precise interests put forward by the [government] as justifications for the burden imposed" (quoting Anderson , 460 U.S. at 789 , 103 S.Ct. 1564 )). In particular, the county asserts that the district court erred by considering its proffered justifications individually, rather than collectively. Yet it provides no authority for the proposition that a court must consider the justifications as a whole. Cf. Chapman v. Meier , 420 U.S. 1 , 24-25, 95 S.Ct. 751 , 42 L.Ed.2d 766 (1975) (considering and rejecting proffered justifications one by one). Nor is it clear that the district court in fact failed to consider the factors together; it merely discussed them one at a time. See id. But even if we accept both the county's premises-that the district court must consider the justifications collectively and failed to do so here-we nevertheless reject the county's argument and affirm the district court. That's because even considered collectively, the county's justifications don't satisfy the intermediate Anderson - Burdick test.

The county's primary justification on appeal is its school-community philosophy. Under this philosophy, the school-board "districts are essentially drawn around the[ ] schools, creating communities of interest" to "ensur[e] that each school ... has specific representation on the [s]chool [b]oard by placing voters who are within the school's boundaries in the voting district for that school's board member." Aplt. Br. 48. The county argues that this philosophy, in combination with the county's geography and sparse population, justifies the 38% population deviation.

The county's argument suffers from several flaws. First, "sparse population is not a legitimate basis for a departure from the goal of equality." Chapman , 420 U.S. at 24 , 95 S.Ct. 751 . In fact, the Court indicated in Chapman that sparse population actually cuts the other way because "in a [s]tate with a small population, each individual vote may be more important to the result of an election than in a highly populated [s]tate." Id. Nor do geographic considerations typically justify significant deviations: "Modern developments and improvements in transportation and communications make rather hollow ... most claims that deviations from population-based representation can validly be based solely on geographical considerations." Reynolds , 377 U.S. at 580 , 84 S.Ct. 1362 . One example of such improvements in communications is the county's own vote-by-mail system, which negates concerns about the location of polling places in the county. As such, we conclude that the county's geography and sparse population aren't persuasive justifications, even considered collectively, for the 38% population deviation.

Nor does the addition of the school-community philosophy change this analysis. Like the district court, we accept that the school-community philosophy could potentially justify the population deviation. 10 But the county's method of implementing this philosophy runs contrary to state law. By statute, Utah mandates three considerations for drawing school-board districts, one of which is that the districts must be "substantially equal in population."

*1285 § 20A-14-201(1)(b). State law says nothing about a school-community philosophy. Thus, the county's goal of prioritizing the school-community philosophy over a state-mandated factor that mirrors the Fourteenth Amendment's one-person, one-vote guarantee fails to justify vote dilution, even when considered in combination with the county's sparse population and geography.

In any event, even assuming these justifications carry some weight and could explain part of the population disparity, the county failed to demonstrate below and likewise fails to demonstrate on appeal the extent to which the unequal districts are a necessary result of its school-community philosophy, sparse population, and geography. In the absence of such a showing, we can't determine the "extent to which" the school-community philosophy "make[s] it necessary to burden" the voting rights of the county's residents. Burdick , 504 U.S. at 434 , 112 S.Ct. 2059 (quoting Anderson , 460 U.S. at 789 , 103 S.Ct. 1564 ). In other words, even if we accept that the county drew unequal districts in an effort to implement its school-community philosophy, the county fails to show that the substantial 38% deviation was an unavoidable consequence of this goal. Of course, the county need not demonstrate that the unequal districts were necessary to implement its school-community philosophy. See Mahan , 410 U.S. at 326 , 93 S.Ct. 979 (noting that in one-person, one-vote cases, "the proper equal[-]protection test is not framed in terms of 'governmental necessity' " (quoting Reynolds , 377 U.S. at 580-81 , 84 S.Ct. 1362 )). But even under the intermediate Anderson - Burdick test, the county must show something more than a rational motivation for drawing unequal districts; it must provide a reason or reasons that are strong enough to justify the significant dilution of its citizens' right to vote.

In a final attempt to persuade the court that it has made this showing, the county relies on Frank , 336 F.3d 570 . There, two of the 21 county-supervisor districts in a sparsely populated Wisconsin county varied in population by 18%. Frank , 336 F.3d at 572 . In a vote-dilution challenge arising under the one-person, one-vote principle, the Seventh Circuit affirmed summary judgment for the county because the high number of districts combined with the low population made it impossible to draw more equal districts without creating odd shapes and crossing local government boundaries. See id. at 573 .

This case is distinguishable from Frank based solely on the number of districts involved-five here, compared to 21 in Frank . Critically, it's far more difficult to equally divide a county's residents into 21 districts than it is to equally divide them into only five districts. See id. at 571, 573 (noting that "the more numerous the districts," the "harder [it is] to create districts of equal population"). Thus, the high number of districts at issue in Frank gave the county a sufficiently strong justification for the population variance. And that justification is entirely lacking here. Further, the 18% deviation in Frank is far less than the 38% deviation at issue in this case. See id. at 572 . Thus, we find Frank unpersuasive.

In summary, the county's reasons for its substantially unequal school-board districts fail to adequately justify the heavy and longstanding burden on its residents' right to vote. We therefore affirm the district court's order granting summary judgment to the Navajo Nation on its claim that the school-board districts violate the Equal Protection Clause.

IV. The County's Proposed Remedial Redistricting Plan

The county argues that the district court erred when it determined that the county's proposed remedial redistricting plan was *1286 predominantly based on race (a factual finding we review for clear error) and that the plan didn't survive strict scrutiny (an analysis we review de novo). See Cooper , 137 S. Ct. at 1465 (noting that we review racial-predomination factual finding for clear error); Concrete Works of Colo. , 321 F.3d at 958 (noting that we review de novo whether party can satisfy strict scrutiny).

Because "[f]ederal-court review of districting legislation represents a serious intrusion on the most vital of local functions," Miller , 515 U.S. at 915 , 115 S.Ct. 2475 , a plaintiff alleging racial discrimination in this context faces a "demanding" burden of proof, Cooper , 137 S. Ct. at 1479 (quoting Easley v. Cromartie , 532 U.S. 234 , 241, 121 S.Ct. 1452 , 149 L.Ed.2d 430 (2001) ). Specifically, "[t]he plaintiff's burden is to show, either through circumstantial evidence of a district's shape and demographics or more direct evidence going to legislative purpose, that race was the predominant factor motivating the legislature's decision to place a significant number of voters within or without a particular district." Miller , 515 U.S. at 916 , 115 S.Ct. 2475 (emphasis added); accord Cooper , 137 S. Ct. at 1479.

"To make this showing, a plaintiff must prove that the legislature subordinated traditional race-neutral districting principles, including but not limited to compactness, contiguity, and respect for political subdivisions or communities defined by actual shared interests, to racial considerations." 11 Miller , 515 U.S. at 916 , 115 S.Ct. 2475 (emphasis added). That is, it's not enough to show that a redistricting entity is "aware of racial demographics." Id. Nor is it enough to show that the redistricting entity had a "policy of prioritizing mechanical racial targets above all other districting criteria (save one-person, one-vote)." Ala. Leg. Black Caucus , 135 S. Ct. at 1267 . Such a policy is "evidence that race motivated the drawing of particular lines," but it isn't conclusive proof that race predominated over traditional redistricting principles in any particular district. Id.

If a plaintiff shows that race predominated the districting process, then the district court will subject the racially drawn lines to strict scrutiny. See Miller , 515 U.S. at 920 , 115 S.Ct. 2475 . And as discussed earlier, "[t]o satisfy strict scrutiny," the county "must demonstrate that its districting legislation is narrowly tailored to achieve a compelling interest." Id.

Here, the district court found that racial considerations predominated over traditional districting principles in the county's redistricting of both the school-board and county-commission districts. To reach this conclusion, the district court first noted that the county's primary goal, after adhering to the one-person, one-vote principle, was complying with § 2 of the VRA. But rather than rely on any particularized § 2 analysis, the county aimed for overall racial proportionality "as a proxy for compliance with [§] 2 of the [VRA]." App. vol. 54, 10852. That proportionality goal effectively created one county-commission district with a "safe [w]hite majority," one county-commission district with a "safe Native American majority," and one county-commission district with "a slim Native American majority." Id. at 10853-54. For the school-board districts, the county created two districts "with a safe majority of [w]hite voters," two districts "with a safe *1287 majority of Native American voters," and one district with "a slim Native American majority." Id. at 10853. The county's proposed remedial redistricting plan was meant to mirror the overall racial demographics of the county, which is just over half Native American. Based on this information, the district court concluded that the county "adopted a countywide policy of prioritizing racial targets above all other redistricting criteria." Id. at 10850.

The district court next considered whether and how this proportionality policy impacted the individual districts. See Ala. Leg. Black Caucus , 135 S. Ct. at 1267 . It concluded that race predominated in the drawing of school-board District 3 (but not in any of the other school-board districts). In support of its conclusion that race predominated in that district, the district court pointed to (1) the racial demographics showing a narrow racial split that mirrored the county as a whole; (2) the odd and noncompact "horseshoe[-]like" shape of the district; 12 and (3) the four precinct splits 13 in this district (out of eight such splits in the county as a whole). App. vol. 54, 10855.

It likewise concluded that race predominated in the drawing of county-commission Districts 1 and 2 (but not District 3). For District 2, it pointed to (1) testimony from the county's expert that he specifically moved voters from District 1 into District 2 in order to increase the percentage of Native Americans in District 2; (2) testimony from the county's expert that he split a precinct in order to complete that increase; and (3) the racial demographics showing a narrow racial split that mirrored the county as a whole. For District 1, the district court noted that the same expert testimony supported a finding that race predominated in creating its boundary, as did evidence showing that the county's expert grouped disparate communities from the far north and the far south of the county into District 1.

Having concluded that race predominated in creating these districts, the district court then determined that the county could not satisfy strict scrutiny. It did so primarily because the county chose "to stand on its argument that its consideration of race in redistricting was permissible" rather than attempting to meet strict scrutiny. Id. at 10859. But the district court nevertheless interpreted the county's purported attempts to comply with § 2 of the VRA as a compelling interest. See Sanchez , 97 F.3d at 1328 (holding that VRA compliance is compelling interest). It then applied the "relaxed version of narrow tailoring" that applies when a redistricting entity asserts VRA compliance as a compelling interest. App. vol. 54, 10860; see also Ala. Black Leg. Caucus , 135 S. Ct. at 1274 (noting that redistricting entity may have "strong basis in evidence to use racial classifications in order to comply with a statute when they have good reasons to believe such use is required"). In so doing, the district court determined that "the [c]ounty provided no evidence that there was a potential [§] 2 violation to remedy as to Native American voters." App. vol. 54, 10861. Accordingly, the district court found that the county lacked "a strong basis in evidence" or "good reasons"

*1288 for its race-based choices. Cooper , 137 S. Ct. at 1464 (quoting Ala. Leg. Black Caucus , 135 S. Ct. at 1274 ). Having concluded that racial considerations predominated in creating certain districts and that those race-based decisions didn't withstand strict scrutiny, the district court rejected the county's proposed remedial redistricting plan.

On appeal, the county contends that the district court's factual finding-that race predominated the drawing of these districts-was clearly erroneous. See id. at 1465 (noting that we review factual finding about "whether racial considerations predominated in drawing district lines" for clear error). But it fails to point to any part of the district court's analysis as unsupported by the record. See id. at 1474 (noting that under clear-error review, we affirm court's finding so long as it's "plausible," and "we reverse only when 'left with the definite and firm conviction that a mistake has been committed' " (quoting Anderson v. City of Bessemer City , 470 U.S. 564 , 573-74, 105 S.Ct. 1504 , 84 L.Ed.2d 518 (1985) )). Instead, the county asserts that it drew these boundaries to meet the goals of adhering to the one-person, one-vote principle; complying with the VRA; and maintaining precinct boundaries. None of these assertions convince us that the district court clearly erred in concluding that racial considerations predominated in creating the county's proposed remedial redistricting plan.

First, the one-person, one-vote principle is "taken as a given[ ] when determining whether race, or other factors, predominate." Ala. Leg. Black Caucus , 135 S. Ct. at 1270 . That the county was following this principle isn't relevant to the question of whether race predominated in creating these districts. Second, it's true that complying with the VRA can be a compelling interest that justifies the race-based boundaries of an election district. See Sanchez , 97 F.3d at 1328 . But this, too, is irrelevant to the question at hand. That's because we aren't concerned, at this point in the analysis, with the reasons for the county's race-based redistricting; instead, we are concerned with whether race-based redistricting in fact occurred.

Third, as to maintaining precinct boundaries, the district court specifically and correctly found that race predominated over this goal. The racially proportionate school-board District 3 included four of eight total precinct splits, and the county offered no other justification for these splits. 14 Also, the county's expert testified that he intentionally split a precinct between county-commission Districts 1 and 2 to increase the Native American population in District 2. In short, the county's challenge to the district court's conclusion that race predominated the county's redistricting process doesn't leave us with "the definite and firm conviction that a mistake has been committed." Cooper , 137 S. Ct. at 1474 (quoting City of Bessemer City , 470 U.S. at 573 , 105 S.Ct. 1504 ). We thus hold that the district court did not clearly err when it ruled that race predominated the creation of these districts.

The county next argues that assuming race did predominate, the districts nevertheless survive strict scrutiny. We review the application of strict scrutiny de novo. See Concrete Works of Colo. , 321 F.3d at 958 . As we did earlier, we assume that complying with § 2 of the VRA is a compelling interest. See Cooper , 137 S. Ct. at 1469 (assuming that VRA compliance can be compelling interest); Sanchez , 97 F.3d at 1328 (holding that VRA compliance *1289 can be compelling interest). But the county must also meet the narrow-tailoring requirement. See Cooper , 137 S. Ct. at 1464. On that front, the county asserts that (1) the 1984 consent decree and settlement order and (2) the district court's redistricting instructions provide the required "good reasons" or "strong bas[e]s in evidence" for the race-based boundaries. Id. (quoting Ala. Leg. Black Caucus , 135 S. Ct. at 1274 ).

Yet as we previously discussed at length, this case doesn't implicate the consent decree and settlement order. That the county entered into the agreement to remedy an alleged § 2 violation doesn't change the analysis; the violation alleged in the 1983 lawsuit resulted from an at-large election system that's no longer in place. So the § 2 concerns that led to the 1983 lawsuit and the consent decree and settlement order aren't relevant to whether the remedial single-member election districts create any § 2 issues.

Next, we reject the county's contention that the district court's instructions in this case provided the necessary good reasons to believe that the new single-member districts might violate § 2. The district court mentioned in passing, from the bench, that the county's remedial redistricting plan should comply with § 2: "The county is going to develop a plan, and it can be constitutionally infirm if the county chooses not to take into account the [§ 2] issues that might arise ...." App. vol. 58, 11273. But as the Navajo Nation points out, acknowledging that § 2 issues "might arise" isn't the same as advising the county that its plan would or did violate § 2. Id. This casual comment about a source of law relevant to redistricting simply isn't " 'a strong basis in evidence' for concluding that the [VRA] required" strict proportionality and the resulting race-based boundary lines. Cooper , 137 S. Ct. at 1464 (quoting Ala. Leg. Black Caucus , 135 S. Ct. at 1274 ).

Moreover, the county acknowledges that its expert did not conduct a § 2 analysis. And in the absence of any investigation into what § 2 might require, the county lacks any basis to argue that it had good reasons to believe § 2 of the VRA required strict proportionality and race-based boundaries. Indeed, the Supreme Court tends to reject strict-scrutiny arguments based on § 2 compliance when "the [redistricting entity] c[an] point to 'no meaningful legislative inquiry' into 'whether a new, enlarged' district, 'created without a focus on race, ... could lead to § 2 liability.' " Abbott v. Perez , --- U.S. ----, 138 S. Ct. 2305 , 2334-35, 201 L.Ed.2d 714 (2018) (omission in original) (quoting Cooper , 137 S. Ct. at 1471 ). 15

Last, the county argues that it used strict proportionality to avoid any potential § 2 violations because "[a] truly proportional plan" is the best way to ensure that "minority voters ... [can] elect the candidates of their choice." Rep. Br. 17. But the Supreme Court has previously rejected strict proportionality as a safe harbor from § 2 of the VRA. See Johnson v. De Grandy , 512 U.S. 997 , 1017-18, 114 S.Ct. 2647 , 129 L.Ed.2d 775 (1994). In Johnson , the Court concluded that such "[a]n inflexible rule would run counter to the textual command of § 2, [which is] that the presence *1290 or absence of a violation be assessed 'based on the totality of circumstances.' " Id. (quoting § 1973(b) ). Accordingly, we reject the county's assertion that it had good reasons for using strict proportionality to avoid potential § 2 liability.

In sum, we conclude that the district court didn't clearly err when it found that race predominated over traditional districting principles in the county's proposed remedial redistricting plan. Further, we agree with the district court that the county's race-based plan doesn't survive strict scrutiny. We therefore affirm the district court's order rejecting the county's proposed remedial redistricting plan.

V. The District Court's Remedial Redistricting Plan

Last, the county argues that the district court erred when it ordered the county to adopt the remedial redistricting plan of the court-appointed special master, Bernard Grofman, a professor of political science at the University of California, Irvine. To more fully address the county's specific arguments, we begin our analysis by reviewing Grofman's process and the details of Grofman's plan.

Grofman took the following approach: he used census blocks to ensure compliance with the one-person, one-vote principle; 16 he chose to draw new boundaries rather than relying on the county's former, unconstitutional districts; he aimed to avoid the use of race as a predominant factor; he didn't consider partisanship; he focused on keeping census places and cities whole; he aimed for contiguity and compactness; and he tried to unpair incumbents. At the final stage, he reviewed the racial makeup of the resulting districts for any potential § 2 problems. He believed that packing minorities into single districts can violate § 2, so he adjusted one school-board district to reduce its Native American population from over 96% to less than 90%. See Voinovich v. Quilter , 507 U.S. 146 , 153, 113 S.Ct. 1149 , 122 L.Ed.2d 500 (1993) (explaining that racial packing can result in vote dilution in violation of VRA if minority group has "sufficient numbers to constitute a majority in three districts" but is "packed into two districts in which it constitutes a super[ ]majority"). Grofman also incorporated the county's objections when possible.

The resulting districts have population deviations of less than 2%. Two of the three county-commission districts and three of the five school-board districts are majority Native American. Nevertheless, Grofman stated that in his plan, the county commission and the school board each contain a true swing district. That is, although the Native American population in those so-called swing districts was around 65%, expert testimony and empirical data on voter turnout indicated that such a percentage didn't make the districts safe Native American seats. Additionally, the county-commission plan "keeps Monticello whole, splits the [c]ity of Blanding in only two parts, splits [the] Navajo Nation [Reservation] in only two parts, and keeps all census places whole." App. vol. 56, 11096. Likewise, the school-board plan "keeps Monticello whole, splits the city of Blanding into only the two mathematically required *1291 parts, splits [the] Navajo Nation [Reservation] into only the three mathematically required parts, and keeps all census places whole." Id.

In the district court, the county argued that Grofman's remedial redistricting plan violated the Equal Protection Clause because it was unconstitutionally based on race, violated § 2 of the VRA, and ignored traditional districting principles. The district court rejected each of the county's arguments and ordered the county to use the new districts and hold special elections in November 2018. On appeal, the county renews two of its three challenges (it doesn't argue that Grofman's remedial plan violates § 2).

The county first argues that, in rejecting the county's argument that Grofman's plan was unconstitutionally based on race in violation of the Equal Protection Clause, the district court relied on a clearly erroneous factual finding. That is, the county asserts that the district court clearly erred when it concluded that race was not the predominant factor in Grofman's plan. See Cooper , 137 S. Ct. at 1463-65 (noting that clear-error review applies to question of whether race predominated in districting plan; explaining that to establish equal-protection violation, party must first show that race predominated in creation of election district). In support, the county baldly states that "[n]o plan[ ] drawn solely on race-neutral districting principles could achieve" the "significant Navajo majorities in two-thirds of the election districts." Aplt. Br. 62. But the county points to no evidence that in redistricting, Grofman was racially motivated or that race predominated his process. The district court specifically found that "[r]ace entered into ... Grofman's line[-]drawing calculation only at the final stage and only with respect to the [s]chool[-b]oard districts." App. vol. 56, 11083. The county points to nothing in the record to refute that finding on appeal. 17

Instead, the county contends that the district court's reasoning was flawed. In particular, it accuses the district court of (1) rejecting the county's remedial plan because the county tried to comply with § 2 but (2) accepting Grofman's plan, which also accounted for § 2 compliance. The problem with this position is that it conflates two very different approaches to § 2 compliance. The county's "efforts to comply with [§] 2 of the [VRA] amounted to nothing more than setting racial targets for each district" and then making additional race-based adjustments to meet those targets. App. vol. 54, 10851. But Grofman drew his lines based on race-neutral districting principles and then, at the end, accounted for a potential § 2 problem in one of the resulting school-board districts. These approaches are diametrically opposed: the county began its redistricting with strict racial proportionality in mind and made changes to districts to achieve that goal, and Grofman ended his process with one individual adjustment based on race. As such, we find no error in the district court's reasoning, and its factual finding that race didn't predominate in Grofman's plan isn't clearly erroneous. The district court therefore properly rejected the county's argument that Grofman's plan violated the Equal Protection Clause.

Next, the county argues that the district court erred in ordering the county to adopt Grofman's redistricting plan because that plan "ignores" certain "traditional districting principles." Aplt. Br. 60. On this point, *1292 our review is for abuse of discretion. See Connor v. Finch , 431 U.S. 407 , 415, 97 S.Ct. 1828 , 52 L.Ed.2d 465 (1977) (noting that when federal court draws election districts "[i]n the wake of a legislature's failure" to do so constitutionally, "the court's task is inevitably an exposed and sensitive one that must be accomplished circumspectly, and in a manner 'free from any taint of arbitrariness or discrimination' " (quoting Roman v. Sincock , 377 U.S. 695 , 710, 84 S.Ct. 1449 , 12 L.Ed.2d 620 (1964) )); Large , 670 F.3d at 1139 (noting that in redistricting cases, we review "appropriateness of the district court's chosen remedy ... for abuse of discretion"). Applying this standard, we reject each of the county's objections in turn.

The county first challenges how Grofman's plan splits the city of Blanding. It acknowledges that Blanding's population is "large enough that in order to meet the equal-population principle of 'one-person, one-vote,' it must be divided between two [c]ounty[-c]ommission districts." Aplt. Br. 60. But the county asserts that Grofman's plan unnecessarily splits Blanding into three county-commission districts, not just two.

The county's position is factually untrue. As the district court noted, Grofman's redistricting plan splits Blanding only into two districts, not three. The county tries to avoid this conclusion by arguing, as it did below, that the city of Blanding includes its suburbs and that Grofman placed those outlying areas in a third district. 18 But as the Navajo Nation points out, the alleged boundaries of these outlying areas don't appear anywhere in the record. As such, the county cannot reasonably argue on appeal that the district court erred by considering only the known boundaries of the city and not the unknown boundaries of the surrounding area.

Relatedly, the county hints at a partisan-gerrymandering argument when it alleges that the city of Blanding is heavily Republican and that Navajo individuals tend to vote for Democratic candidates. But Grofman specifically testified that he didn't consider any partisan factors when drawing his plan. And like it did below, the county fails to develop a partisan-gerrymandering argument on appeal. Any argument it does make appears only in its reply brief and in a letter of supplemental authority filed after oral argument. Accordingly, we decline to consider any such argument. 19 See Bronson v. Swensen , 500 F.3d 1099 , 1104 (10th Cir. 2007) ("[W]e routinely have declined to consider arguments that are not raised, or are inadequately presented, in an appellant's opening brief.").

The county's third point related to Blanding is that one of the school-board districts in the Blanding area contains only one school. The county doesn't explain why this fact weighs in its favor, and we see no indication that it does. There are only so many ways to divide the county's 12 schools among five school-board districts, and it's not patently unreasonable to have one district with only one school in it. We therefore find no abuse of discretion in the decision to accept Grofman's plan despite the Blanding split and the school-board *1293 district with only one school in it. See Large , 670 F.3d at 1139 .

The county next suggests that the district court erred in approving a plan with election districts based on census blocks rather than on county precincts. Grofman explained that even though the county had long relied on voting precincts based on survey lines, he used census blocks because these are "the only units of geography for which we have reliable population estimates." App. vol. 56, 11072. And Grofman and the district court needed an accurate population count and distribution to comply with the one-person, one-vote principle-a principle that's even more important when a court, and not a legislative body, creates election districts. See Connor , 431 U.S. at 414 , 97 S.Ct. 1828 (noting that court-drawn election districts "must ordinarily achieve the goal of population equality with little more than de minimis variation"). The county doesn't seriously dispute the need for equally populated districts; it merely cites its administrative concerns related to switching its system from precincts to census blocks. But the district court didn't abuse its discretion when it prioritized compliance with the constitutionally mandated one-person, one-vote principle over the county's administrative burden.

To conclude, the district court didn't clearly err when it found that Grofman's plan wasn't predominantly based on race. And because the district court approved Grofman's plan "in a manner 'free from any taint of arbitrariness or discrimination,' " we affirm its order directing the county to adopt that plan. Id. at 415 , 97 S.Ct. 1828 (quoting Roman , 377 U.S. at 710, 84 S.Ct. 1449 ).

Conclusion

In summary, we find no error in the district court's well-reasoned rulings. First, the district court correctly determined that the 1984 consent decree and settlement order have nothing to do with this action. It therefore properly denied the county's motion to dismiss the Navajo Nation's county-commission claim. Second, the district court didn't err when it ruled that the county lacked a compelling interest to justify the racially drawn boundaries of county-commission District 3. Third, the district court correctly rejected as inadequate the county's justifications for the longstanding and substantial population deviation among the 1992 school-board districts. Fourth, the record supports the district court's factual finding that several of the districts in the county's remedial redistricting plan were predominantly based on race. And it correctly concluded that the county lacked the required good reasons or a strong basis in evidence to justify this race-based line drawing. Fifth, the district court didn't clearly err when it found that the special master's remedial plan wasn't predominantly based on race; nor did it otherwise abuse its discretion in ordering the county to adopt that plan.

Accordingly, we affirm the district court's decisions in all respects.