Dow v. Az Nsac

IN THE ARIZONA COURT OF APPEALS DIVISION ONE DEFENDERS OF WILDLIFE, et al., Appellants, v. ARIZONA NAVIGABLE STREAM ADJUDICATION COMMISSION, et al., Appellees. No. 1 CA-CV 20-0295 1 CA-CV 20-0296 1 CA-CV 20-0297 (Consolidated) FILED 2-7-2023 Appeal from the Superior Court in Maricopa County Nos. LC2019-000102-001 LC2019-000101-001 LC2019-000100-001 The Honorable Sigmund G. Popko, Judge Pro Tempore AFFIRMED IN PART; REVERSED IN PART COUNSEL Arizona Center for Law in the Public Interest, Phoenix By Jennifer B. Anderson 1, Daniel J. Adelman Counsel for Appellants Defender of Wildlife, Donald Steuter, Jerry Van Gasse, Jim Vaaler 1Jennifer B. Anderson presented at oral argument but has since left the Arizona Center for Law in the Public Interest. Fennemore Craig, P.C., Phoenix By Sean Hood Co-Counsel for Appellee Freeport Minerals Corporation Snell & Wilmer L.L.P., Phoenix By L. William Staudenmaier Co-Counsel for Appellee Freeport Minerals Corporation Salmon, Lewis & Weldon, P.L.C., Phoenix By John B. Weldon, Jr., Mark A. McGinnis Counsel for Appellees Salt River Project Agricultural Improvement and Power District and Salt River Valley Water Users’ Association Engelman Berger, P.C., Phoenix By William H. Anger Co-Counsel for Appellee City of Mesa Mesa City Attorney’s Office By Wilbert J. Taebel Co-Counsel for Appellee City of Mesa City of Phoenix, Office of the City Attorney, Phoenix By Cris Meyer, Charles L. Cahoy Counsel for Appellee City of Phoenix Tempe City Attorney’s Office, Tempe By Judith R. Baumann, Megan H. Tracy Counsel for the Appellee City of Tempe Morisset, Schlosser, Jozwiak & Somerville APC, Seattle, Washington By Thane D. Somerville Counsel for Appellee Salt River Pima-Maricopa Indian Community Montgomery & Interpreter, PLC, Phoenix By Susan B. Montgomery, Robyn L. Interpreter, Jay M. Tomkus Counsel for Appellees Yavapai-Apache Nation and the Fort McDowell Yavapai Nation Office of General Counsel Gila River Indian Community, Sacaton By Linus Everling, Thomas L. Murphy Counsel for Appellee Gila River Indian Community 2 Perkins Coie LLP, Phoenix By Matthew L. Rojas, Andrea J. Driggs, Karl J. Worsham Counsel for Appellee Arizona Navigable Stream Adjudication Commission The Sparks Law Firm, P.C., Scottsdale By Joe P. Sparks, Laurel A. Herrmann Counsel for Appellee San Carlos Apache Tribe Gust Rosenfeld P.L.C., Phoenix By Scott A. Malm Counsel for Amicus Curiae Land Title Association of Arizona OPINION Presiding Judge David B. Gass delivered the opinion of the court, in which Judge Michael J. Brown and Judge David D. Weinzweig joined. G A S S, Judge: ¶1 The Arizona Navigable Stream Adjudication Commission (ANSAC) has resolved title for most of Arizona’s rivers based on navigability, finding all but the mighty Colorado to be nonnavigable. Three rivers remain, the Verde, the Salt, and the Gila. This appeal promises to be the last step in the streambed litigation, which began in the 1980s. As relevant to this case, ANSAC found 20 segments of the 3 rivers nonnavigable (6 segments of the Verde, 6 segments of the Salt, and 8 segments of the Gila). Of the 20 segments, 17 are the subject of this appeal. ¶2 Long ago, Mikhail Lermontov wrote, “Many a calm river begins as a turbulent waterfall, yet none hurtles and foams all the way to the sea.” 2 We must decide whether the law and facts support ANSAC’s determination as affirmed by the superior court that no segment of three Arizona rivers ran sufficiently calm and deep to be navigable when Arizona became a state in 1912. 3 Ultimately, we affirm ANSAC’s finding of nonnavigability on 16 of the 17 segments challenged in this appeal, the only 2 Mikhail Lermontov, A Hero of Our Time 117 (Dimitri Nabokov et al. trans., A.A. Knopf 1992) (1840). 3 Absent material revisions after the relevant dates, statutes and rules cited refer to the current version unless otherwise indicated. 3 DOW, et al. v. AZ NSAC, et al. Opinion of the Court exception being 1 segment of the Gila. We conclude that segment was navigable as a matter of law on February 14, 1912. FACTUAL AND PROCEDURAL HISTORY I. History of the Rivers and Navigability Adjudications in Arizona ¶3 For millennia, the first peoples to live in Arizona—including the ancestors and current members of the sovereign tribal parties in this case—settled alongside Arizona’s rivers, relying on them to support their communities. And for hundreds of years, newcomers—like the Spanish and early American settlers—came to rely heavily on these rivers as well. ¶4 On February 14, 1912, Arizona joined the United States on equal footing with all states coming before it. Defenders of Wildlife v. Hull, 199 Ariz. 411, 415, ¶ 2 (App. 2001). Arizona, thus, took title to the beds under all navigable waters in the state. See id. ¶5 In 1987, Arizona’s legislature tried to resolve conflicting riverbed title claims by “relinquish[ing] most of the state’s interest in Arizona’s watercourse bedlands[,]” including the Verde, Salt, and Gila riverbeds. State ex rel. Winkleman v. Ariz. Navigable Stream Adjudication Comm’n, 224 Ariz. 230, 234–36, ¶¶ 3–7 (App. 2010). In 1992, after this court ruled the 1987 changes were unconstitutional, the legislature adopted the framework for resolving riverbed ownership by creating ANSAC and “charg[ing] it with the responsibility for determining which watercourses were navigable at statehood by hearing evidence presented by the Arizona State Land Department [(ASLD)] and the public.” Id. at 235, ¶ 5. In 1994 and 1998, the Arizona legislature made further changes to ANSAC’s charge. Id. at 236, ¶ 7. Those later changes failed, in part, because they circumvented the federal legal standard for navigability. Id. at ¶ 8. Since then, ANSAC adjudicated the navigability of around 39,000 streams—finding none to be navigable. See Ariz. State Libr., Archives, & Pub. Recs., Arizona Navigable Stream Adjudication Commission (ANSAC) (July 15, 2019), https://azlibrary.gov/sla/agency_histories/arizona-navigable-stream- adjudication-commission-ansac. II. Adjudicating the Navigability of the Verde, Salt, and Gila ¶6 After a series of hearings from 2003 to 2006, ANSAC determined all three rivers were nonnavigable, considering the upper and lower Salt separately and the Verde and the Gila in whole. The superior court vacated those determinations, concluding ANSAC failed to account for the effect of diversions and other human impacts. See Winkleman, 224 4 DOW, et al. v. AZ NSAC, et al. Opinion of the Court Ariz. at 241–42, ¶ 28 (holding ANSAC failed to evaluate lower Salt River in its “natural” condition). ¶7 On remand, ANSAC reopened the record and followed the United States Supreme Court’s decision in PPL Montana, LLC v. Montana, which said tribunals must evaluate a river’s navigability segment-by- segment. 565 U.S. 576, 594 (2012). ANSAC held another series of hearings between 2014 and 2016, during which the parties’ expert witnesses testified about the natural state of the waterways. The experts also detailed human usage of the three rivers—including trade and travel—long before, around, and after statehood. ¶8 Without objection, ANSAC adopted ASLD’s proposed segmentation of the rivers. To begin, ANSAC divided the Verde into six segments—beginning with segment 0 in the Verde’s headwaters at Sullivan Lake and ending with segment 5 in the southernmost portion of the river where the Verde joins the Salt. Next, ANSAC divided the Salt into six segments—beginning with segment 1 in the northernmost portion of the river and ending with segment 6 in the southernmost portion of the river where the Salt joins the Gila. Last, ANSAC divided the Gila into eight segments—beginning with segment 1 in the easternmost part of the Gila, where it enters Arizona from New Mexico, and ending with segment 8 in the westernmost part where the Gila joins the Colorado River in Yuma. The parties do not dispute the segmentation on appeal. ¶9 In 2018, a majority of ANSAC’s five-member board determined all the segments in the Verde, Salt, and Gila were nonnavigable at statehood. One member dissented. ¶10 Defenders of Wildlife (DOW)—the navigability proponent— unsuccessfully challenged ANSAC’s determinations in the superior court. This court has jurisdiction under article VI, section 9, of the Arizona Constitution, and A.R.S. §§ 12-120.21.A.1 and -2101.A.1. ANALYSIS ¶11 DOW challenges the sufficiency of the evidence to support the determinations. DOW also contends ANSAC failed to apply the correct legal standards when it determined the three rivers were nonnavigable. DOW and other parties also raise other issues related to the riverbeds’ title. We address these issues in turn. 5 DOW, et al. v. AZ NSAC, et al. Opinion of the Court I. Standard of Review and Burden of Proof ¶12 “On appeal from a superior court’s review of an administrative decision, [this court] must determine . . . whether the administrative action was illegal, arbitrary, capricious or involved an abuse of discretion.” Eaton v. Ariz. Health Care Cost Containment Sys., 206 Ariz. 430, 432, ¶ 7 (App. 2003); see also A.R.S. § 12-910.F. This court, however, reviews questions of law de novo, including whether ANSAC applied the appropriate legal test to the facts. See Eaton, 206 Ariz. at 432, ¶ 7. ¶13 If ANSAC applied the correct legal tests, navigability is usually a question of fact. Winkleman, 224 Ariz. at 238, ¶ 14 (citing Ariz. Ctr. for Law in the Pub. Int. v. Hassell, 172 Ariz. 356, 363 n.10 (App. 1991)). Because ANSAC is Arizona’s fact-finding agency for stream adjudications, this court will not substitute its judgment on matters of fact unless ANSAC’s findings are clearly erroneous. See Winkleman, 224 Ariz. at 238, ¶ 14. Instead, this court reviews “the record to determine whether substantial evidence supports [ANSAC’s] decision and whether [ANSAC] exercised its discretion reasonably and with due consideration.” Id. (citing Callen v. Rogers, 216 Ariz. 499, 502, ¶ 9 (App. 2007); Siegel v. Ariz. State Liquor Bd., 167 Ariz. 400, 401 (App. 1991)). But “[b]oth the standards and the ultimate conclusion [in navigability cases] invo[l]ve questions of law inseparable from the particular facts to which they are applied.” United States v. Appalachian Elec. Power Co., 311 U.S. 377, 404 (1940), superseded in part by statute as recognized in Rapanos v. United States, 547 U.S. 715, 723–34 (2006). This court, thus, “may draw its own legal conclusions from facts found or inferred in the judgment and is not bound by findings of fact on mixed questions of law and fact.” Huskie v. Ames Bros. Motor & Supply Co., 139 Ariz. 396, 401 (App. 1984). ¶14 The legal standard does not impose a presumption for or against navigability. “ANSAC’s approach and analysis must be wholly impartial.” Winkleman, 224 Ariz. at 239, ¶ 18. But the proponent of navigability, DOW, bears the burden of proving a river segment navigable by a preponderance of the evidence. See id. at 238–39, ¶ 17. II. Evidentiary Disputes ¶15 DOW argues ANSAC abused its discretion by improperly considering certain evidence while failing to consider other evidence in arriving at its decision on the navigability of the Verde, Salt, and Gila. Moving through DOW’s evidentiary challenges, we address four issues: (1) we first address if ANSAC erred by considering expert testimony about 6 DOW, et al. v. AZ NSAC, et al. Opinion of the Court boating from individuals not qualified to testify on that subject matter; (2) next, we decide whether ANSAC improperly relied on alleged “irrelevant evidence,” such as evidence of non-boating transportation, land grants, and land maps; (3) we then address the evidentiary weight of Arizona’s long delay in asserting title to the riverbeds of the Gila, Verde, and Salt on ANSAC’s navigability determinations; and (4) last, we decide whether ANSAC erred by “failing to consider” certain evidence relevant to the navigability of the three rivers, including ferry usage and the best evidence of the rivers’ ordinary and natural condition. A. Expert Testimony and Qualifications ¶16 DOW argues ANSAC erred in considering certain expert testimony from persons not qualified as boating experts because their boating testimony was speculative. DOW adds the superior court erred in finding DOW waived these arguments by not objecting before ANSAC. ¶17 ANSAC must review and consider “all relevant historical and other evidence.” A.R.S. § 37-1123.A; see also Winkleman, 224 Ariz. at 243, ¶ 31 (“[T]he determination of the relevance and weight to be afforded the evidence is generally for ANSAC to make.”); Hull, 199 Ariz. at 425, ¶ 52. An agency, board, or commission conducting a hearing is not limited by evidentiary rules and has “exceptional discretion” to determine which witnesses may provide expert testimony. Lathrop v. Ariz. Bd. of Chiropractic Examiners, 182 Ariz. 172, 181 (App. 1995). ¶18 Even though the superior court found DOW may have waived the argument, it reached the merits. And we agree with the superior court’s merits analysis. To begin, DOW contended the testimony was unqualified because “opponent’s expert witnesses had [n]ever boated or tried to boat the Verde.” All parties agree all three rivers are not currently in the same ordinary and natural condition they were at statehood. It follows, then, whether an expert witness tried to boat one of the rivers today is immaterial to their expert qualifications and to navigability. ¶19 Moreover, DOW points to a portion of one expert’s testimony in which he did not explain the specific boats he envisioned in considering navigability on Verde segments. That same expert was a hydraulic engineering manager with 35-years’ experience consulting in hydrology and rivers. He studied the depth, dynamic geomorphic character, longitudinal variability, likelihood of periodic changes, vegetation, and ranges of flows of the Verde. And, in his third day of testifying, he detailed the types of boats in existence at statehood and their physical compositions. 7 DOW, et al. v. AZ NSAC, et al. Opinion of the Court ¶20 We cannot say this expert was unqualified to discuss the rivers’ susceptibility to use by watercraft just because he lacked the moniker “boating expert,” and DOW cites no authority to the contrary. To the extent DOW asks us to reweigh evidence on appeal, we decline to do so. See Winkleman, 224 Ariz. at 242, ¶ 29. B. Reliance on Federal Surveys, Land Patents, Land Grants, Maps, and Other Documents ¶21 DOW also contends ANSAC erred in relying on “irrelevant evidence,” including non-boating transportation, land grants, and land patents. ¶22 DOW argues this court in Hull precluded the use of non- boating transportation when determining nonnavigability. See 199 Ariz. at 424–25, ¶¶ 47–48. DOW overstates Hull’s holding. Hull ruled ANSAC could not presume nonnavigability based on non-boating transportation in proximity to a watercourse, and ANSAC could not consider such evidence “highly probative.” Id.; see A.R.S. § 37-1128.D.8 (2000) (repealed). Though Hull rejected the “highly probative” standard, Hull acknowledged other courts found evidence of non-boating transportation relevant. 199 Ariz. at 424–25, ¶¶ 47–48 (citing Monroe v. State, 111 Utah 1 (Utah 1946); Lykes Bros. Inc. v. U.S. Army Corps of Eng’rs, 821 F. Supp. 1457 (M.D. Fla. 1993), aff’d, 64 F.3d 630 (11th Cir. 1995)). The Hull court did not say the evidence was irrelevant, and we decline to do so here. ¶23 DOW also argues ANSAC erred by relying on federal surveys, land patents, land grants, maps, and other documents because “there is no indication that the surveyors understood, let alone applied, the term ‘navigability’ as defined in The Daniel Ball and interpreted by federal and state courts.” 77 U.S. (10 Wall.) 557, 563 (1870). ¶24 True enough, surveyors “not clothed with power to settle questions of navigability” drew the land patents, land grants, and maps. Oklahoma v. Texas, 258 U.S. 574, 585 (1922). But the evidence is still relevant to ANSAC’s navigability assessment. See generally Ariz. R. Evid. 401 (evidence is relevant if it tends to make a fact of consequence “more or less probable”). Other courts considering navigability-for-title have looked to comparable evidence from early surveyors to inform their factual analysis. See State v. Adams, 89 N.W.2d 661, 668–72 (Minn. 1957); Ryals v. Pigott, 580 So.2d 1140, 1171 (Miss. 1990) (Blass, J., dissenting); see also Lykes Bros., 64 F.3d at 636 n.5. 8 DOW, et al. v. AZ NSAC, et al. Opinion of the Court ¶25 The record shows no instance in which ANSAC relied on this evidence or considered it highly probative in making its determination, and we cannot say it is irrelevant to determining navigability. ANSAC did not abuse its discretion in considering this evidence. C. Delay in Asserting Title to Riverbeds ¶26 Salt River Pima-Maricopa Indian Community argues Arizona’s delay in asserting title to the Verde, Salt, and Gila riverbeds is evidence of nonnavigability. But ANSAC made no finding the delay was at all probative. PPL Montana did note a state’s “long failure to assert title is some evidence to support the conclusion that the river segments were nonnavigable for purposes of the equal-footing doctrine.” 565 U.S. at 604. But as Hassell said, “[t]hat generations of trustees have slept on public rights does not foreclose their successors from awakening.” 172 Ariz. at 369. Arizona’s decades-long failure to assert title simply is not probative of navigability, and we find no error in ANSAC treating it as such. D. Ignoring DOW’s Evidence ¶27 DOW argues ANSAC erred by disregarding evidence DOW brought forward to prove the rivers’ navigability. Under A.R.S. §§ 37- 1123.A and -1128.A, ANSAC must consider all relevant evidence before it. See Winkleman, 224 Ariz. at 243, ¶ 31. This court will defer to ANSAC’s “determination of the relevance and weight [of] the evidence” when “the evidence has indicia of reliability.” Id. III. The Navigability Test ¶28 Arizona holds title to all riverbeds if the river was navigable at the time of statehood. Hassell, 172 Ariz. at 359–60. Arizona must apply the federal Daniel Ball test to determine whether a river was navigable under the equal-footing doctrine. Hull, 199 Ariz. at 419, ¶¶ 16–17 (citing Hassell, 172 Ariz. at 362; The Daniel Ball, 77 U.S. (10 Wall.) 557); PPL Montana, 565 U.S. at 592. The test is whether the river was “navigable in fact” at statehood. Hull, 199 Ariz. at 419, ¶¶ 16–17 (quoting The Daniel Ball, 77 U.S. (10 Wall.) at 563). Rivers are “navigable in fact when they are used, or are susceptible of being used, in their natural and ordinary condition, as highways for commerce, over which trade and travel are or may be conducted in the customary modes of trade and travel on water.” See United States v. Utah (Utah I), 283 U.S. 64, 76 (1931); see also The Daniel Ball, 77 U.S. (10 Wall.) at 563. 9 DOW, et al. v. AZ NSAC, et al. Opinion of the Court ¶29 Arizona state law reflects the Daniel Ball test, though its wording is slightly different. Under A.R.S. § 37-1101.5, a navigable river for purposes of title means a river “in existence on February 14, 1912, and at that time was used or was susceptible to being used, in its ordinary and natural condition, as a highway for commerce, over which trade and travel were or could have been conducted in the customary modes of trade on water.” ¶30 Even if a parsing of the words in Arizona’s statute might yield a different result than under the federal Daniel Ball test, the federal test controls. See Hull, 199 Ariz. at 418–20, ¶¶ 14–16, 21. Arizona cannot expand the federal navigability test because doing so would allow Arizona to take title to rivers not navigable at statehood. See PPL Montana, 565 U.S. at 603 (explaining “federal law determines riverbed title under the equal-footing doctrine”). Arizona also cannot narrow the navigability test because doing so would have Arizona abandon its title to rivers in violation of Arizona’s trust obligations. See Winkleman, 224 Ariz. at 234–35, ¶¶ 3–4 (citing Hassell, 172 Ariz. at 359–60, 364–65, 369–72). IV. Applying the Navigability Test ¶31 DOW argues ANSAC misapplied navigability throughout these proceedings. To determine whether ANSAC “applied the proper legal tests to the facts presented” and “exercised its discretion reasonably and with due consideration,” we consider if ANSAC meets each essential component of the navigability test in turn. See Winkleman, 224 Ariz. at 238, ¶ 14–15. ¶32 From the words of the test and navigability precedent, we discern five essential components: 1. The river’s ordinary and natural condition at the time of statehood. PPL Montana, 565 U.S. at 592; Winkleman, 224 Ariz. at 241–42, ¶ 28. 2. The types of commerce, in terms of both trade and travel, contemplated at statehood. PPL Montana, 565 U.S. at 600. 3. The customary modes of trade and travel on water at statehood. Id. at 601. 4. Actual navigation of the river, before and after statehood. Utah I, 283 U.S. at 11–12. 10 DOW, et al. v. AZ NSAC, et al. Opinion of the Court 5. The river’s susceptibility to use as a highway for commerce at the time of statehood, assuming the river had been in its ordinary and natural condition. PPL Montana, 565 U.S. at 600– 01. ¶33 ANSAC must determine navigability for each river segment, not the entire river. See id. at 593. Because of geographical variations in river conditions, commerce, modes of transportation, and actual navigation, ANSAC must consider each of the essential components on a segment-by- segment basis. See id. at 601. ANSAC, however, may conclude certain components—for instance, customary craft—do not significantly vary along a river. See id. ANSAC need not necessarily make express findings on each component, for each segment, but it must at least demonstrate it has considered all the components and the evidence material to them. See Winkleman, 224 Ariz. at 240, ¶ 22. ¶34 DOW argues ANSAC erred as a matter of law in applying each component. We address DOW’s arguments as to each component in turn, combining our review of actual use and susceptibility because DOW’s arguments on those components are interrelated. A. Ordinary and Natural Condition ¶35 ANSAC had to consider what each river was like or would have been like at statehood in its ordinary and natural condition. Winkleman, 224 Ariz. at 241, ¶ 28; cf. United States. v. Holt State Bank, 270 U.S. 49, 56–57 (1926). Evidence of a river’s natural flow, obstacles, and channel characteristics—specifically width and depth—is relevant to the analysis. Oklahoma, 258 U.S. at 589; Utah I, 283 U.S. at 84–87. ¶36 The best evidence of a river’s natural condition is from periods without significant man-made diversions and obstructions. Winkleman, 224 Ariz. at 242, ¶ 30 (recognizing the Salt was in its natural condition in the 1800s, after Hohokam diversions largely ceased, until modern settlement began). ANSAC may consider evidence of the river’s condition during other periods—though it may have “minimal probative value”—but ANSAC must account for the effect of human impacts on the river’s characteristics. Id. at 241–43, ¶¶ 28, 31. ¶37 DOW argues ANSAC erred when it disregarded the “best evidence” of the rivers’ ordinary and natural condition from the period after significant Native American diversions ceased but before significant impacts from Euro-American settlers began. Because ANSAC did consider this evidence, we disagree. ANSAC generally considered historical 11 DOW, et al. v. AZ NSAC, et al. Opinion of the Court accounts and streamflow measurements from that period, and it considered how natural forces (namely floods) and human impacts had changed the rivers by the time of statehood and afterwards. ANSAC also generally considered expert estimates of what streamflow would have been at the time of statehood absent human impacts. ¶38 DOW admits ANSAC considered “extensive historical evidence” from the 1800s but argues ANSAC “trivializ[ed] and disregard[ed] this evidence.” To the extent DOW contends ANSAC erred in weighing the evidence or considering other evidence, the law required ANSAC to consider all relevant evidence. See Winkleman, 224 Ariz. at 243, ¶ 31; see also A.R.S. §§ 37-1123.A (“all evidence”), -1128.A (“all relevant evidence”). And this court must defer to ANSAC’s determination of the evidence’s relevancy when it has indicia of reliability. See id.; see also A.R.S. §§ 37-1123.A, –1128.A. Except for segment 8 of the Gila, we generally find no error in ANSAC’s application of the ordinary-and-natural-condition component of the navigability test. B. Highway for Commerce ¶39 In determining navigability, ANSAC must consider “the kinds of commercial use,” in terms of both trade and travel, “that, as a realistic matter, might have occurred at the time of statehood.” PPL Montana, 565 U.S. at 600. Relevant evidence includes the types of industries and trades in the vicinity of the river, the demand for personal transportation, and the geographical context, such as where goods and people were going and how they got there. See, e.g., Holt State Bank, 270 U.S. at 57 (noting settlements used a lake connected to navigable rivers “in sending for and bringing in their supplies”); Oklahoma, 258 U.S. at 583, 589– 90 (noting shipping boats ceased operating after railroads were constructed in the vicinity about 30 years before statehood). But see Utah I, 283 U.S. at 83 (explaining potential future profitable use may be relevant when exploration and settlement explain a lack of commercial demand at statehood). ¶40 ANSAC made findings about the types of commercial trade and travel contemplated at statehood for the Verde and Gila, but not the Salt. For both the Verde and the Gila, ANSAC listed transport of military supplies, mining materials, and agricultural goods as well as travel and transport of people as contemplated forms of commerce. For the Gila, ANSAC also found people contemplated trapping and hunting. ANSAC, however, made no findings about the types of commerce contemplated for the Salt. Instead, ANSAC said the Salt “was not actually used as a ‘highway 12 DOW, et al. v. AZ NSAC, et al. Opinion of the Court of commerce.’” ANSAC’s lack of findings regarding commerce contemplated on the Salt takes away from its findings about customary craft used on the Salt, which in turn gives DOW ammunition for its argument. ¶41 DOW contends ANSAC “misconstrued the ‘highway for commerce’ requirement” by focusing on whether the rivers were or could have been used for “commercial” purposes, and if such purposes were profitable. DOW posits “highway for commerce” encompasses either trade or noncommercial travel. We begin with the commerciality analysis, then turn to profitability. 1. Commerciality ¶42 Contrary to DOW’s argument, the navigability test does not contemplate noncommercial travel. Though the United States Supreme Court did not grant certiorari in PPL Montana to address the “highway for commerce” component of the federal navigability test, it did clarify a “highway for commerce” must have an element of “commercial reality.” 565 U.S. at 602–03. We agree with the Utah Supreme Court’s formulation: “the touchstone of navigability is commercial utility.” Utah Stream Access Coal. v. Orange St. Dev., 416 P.3d 553, 560, ¶ 29 (Utah 2017); see also Utah I, 283 U.S. at 81; The Daniel Ball, 77 U.S. (10 Wall.) at 560 (noting a river can be “a highway for commerce” and navigable). When a navigability determination is based on actual travel or susceptibility to travel, the travel must be for a commercial purpose, not just for recreation or other personal reasons. See Utah Stream Access Coal., 416 P.3d at 560, ¶ 29; see also Alaska v. Ahtna, Inc., 891 F.2d 1401, 1404 (9th Cir. 1989). Even so, courts have recognized “commercial activity” can be broad enough to encompass the “recreation industry.” Ahtna, 891 F.2d at 1405. But “commercial activity” cannot be so broad as to include noncommercial travel. PPL Montana, 565 U.S. at 602–03. For example, self-guided recreational boating trips are not commercial use, but the Forty-Niner trips down the Gila to get to California (discussed in infra ¶ 91) were. See Utah v. United States (Utah II), 403 U.S. 9, 11 (1971). ¶43 To the extent Hull suggested noncommercial travel could suffice, we disagree. See 199 Ariz. at 421, ¶ 28. Such a holding would depart from Utah II. In Utah II, the United States Supreme Court upheld a navigability determination for the Great Salt Lake because ranchers used the lake to haul cattle and sheep. 430 U.S. at 11. Because livestock hauling was “done by owners and ‘not by a carrier for the purpose of making 13 DOW, et al. v. AZ NSAC, et al. Opinion of the Court money,’” the Hull court determined the federal test did not turn on commercial use. 199 Ariz. at 421, ¶ 28 (quoting Utah II, 430 U.S. at 11). ¶44 Contrary to Hull’s suggestion, Utah II did not find the livestock owners used the lake for noncommercial purposes. Indeed, the United States Supreme Court said “people who performed ranching operations” used the lake. Utah II, 430 U.S. at 11. And the Court did not say commercial use was irrelevant. Id. Instead, the Court said it was irrelevant “the business of the boats was ranching and not carrying water-borne freight.” Id. ¶45 DOW misplaces its reliance on Utah I and Utah II to support its noncommercial travel argument. Neither Utah I nor Utah II say susceptibility to noncommercial use was enough to establish navigability. Utah I simply says limited, noncommercial use is relevant evidence of whether “rivers are . . . capable of commercial use.” 283 U.S. at 82. And, as explained in the previous two paragraphs, Utah II simply says any commercial purpose will suffice. See 403 U.S. at 11. Put simply, neither Utah I nor Utah II articulated a bright-line rule that noncommercial travel establishes navigability per se. Instead, both cases highlight how limited commercial and noncommercial uses can factor into a river’s susceptibility to commercial use. ¶46 The commerciality requirement tracks both the federal test and Arizona’s statutory corollary. Both the federal test and Arizona’s test use the conjunction “and.” The federal navigability-for-title test looks for use or susceptibility “as highways for commerce, over which trade and travel are or may be conducted.” The Daniel Ball, 77 U.S. (10 Wall.) at 563 (emphasis added). In Arizona, a “‘navigable watercourse’ means a watercourse” used or susceptible to use “as a highway for commerce, over which trade and travel were or could have been conducted.” A.R.S. § 37- 1101(5) (emphasis added). Though some courts have interpreted the phrase “trade and travel” as disjunctive, rather than conjunctive, we need not reach that issue. See Utah Stream Access Coal., 416 P.3d at 560, ¶ 32. Indeed, the conjunctive-disjunctive dichotomy is a red herring in the navigability-for- title context, as demonstrated by a close review of Arizona’s definition of “highway for commerce.” ¶47 Arizona defines a “highway for commerce” as “a corridor or conduit within which the exchange of goods, commodities or property or the transportation of persons may be conducted.” A.R.S. § 37-1101(3) (emphasis added). Arizona cannot enlarge the federal navigability-for-title test by using noncommercial travel to establish navigability because doing 14 DOW, et al. v. AZ NSAC, et al. Opinion of the Court so would vest title to Arizona in rivers that were not navigable at statehood. See supra ¶ 29. As a result, to accord with the federal test, a highway for commerce, as defined in A.R.S. § 37-1101(3), must be based on “the transportation of persons” for commercial reasons. See Blake v. Schwartz, 202 Ariz. 120, 122, ¶ 10 (App. 2002) (citing State v. McDonald, 191 Ariz. 118, 120 (App. 1997)) (“If possible, this court has a duty to construe a statute so that it will be constitutional.”). ¶48 Interpreting “travel,” under federal law, and “the transportation of persons,” under Arizona law, to require a commercial purpose obviates the dichotomy between “trade and travel” and “trade or travel.” Once we view the navigability-for-title test through the lens of “commercial utility,” the conjunctive-disjunctive debate becomes a distinction without a difference. ¶49 In short, travel in the navigability-for-title context encompasses commercial travel. ANSAC did not legally err by requiring use or susceptibility for commercial purposes. 2. Profitability ¶50 DOW correctly argues the navigability-for-title analysis does not require a “profitable” commercial enterprise. On this account, we agree with Hull and need not revisit the issue here. See 199 Ariz. at 422, ¶ 34. But Hull dealt with legislative presumptions and held navigability cannot hinge on profitability. Id.; see A.R.S. § 37-1128.D.2 (2000) (repealed). ANSAC did not err in considering profitability. Indeed, ANSAC rarely discussed profitability and most references were direct quotes from DOW’s expert. ¶51 Though Hull rejects a profitability requirement, it does not prohibit ANSAC from considering profitability when determining what constitutes commerce. Profitability may be probative of whether navigability is a “commercial reality.” See generally Ariz. R. Evid. 401, 402; A.R.S. § 37-1123.A (directing ANSAC to receive, review, and consider all relevant evidence). And considering DOW offered expert testimony on profitability, DOW seems to believe it is probative as well. ANSAC did not err in admitting and considering testimony on profitability as part of its navigability analysis. C. Customary Modes of Trade and Travel on Water ¶52 ANSAC had to consider the customary modes of trade and travel used or that could have realistically been used for the types of commerce contemplated on the river at statehood. See PPL Montana, 565 15 DOW, et al. v. AZ NSAC, et al. Opinion of the Court U.S. at 602–03; Ahtna, Inc., 891 F.2d at 1405. Customary modes may differ for trade and travel. Utah I, 283 U.S. at 82. For instance, determining steamboats could not have navigated a river to ship ore is not probative of whether rowboats might have transported people. Cf. Oregon v. Riverfront Prot. Ass’n, 672 F.2d 792, 795–96 (9th Cir. 1982) (considering whether river was navigable for “log drives”). ¶53 DOW asserts ANSAC erred because it did not find the rivers were navigable by small, low-draft watercraft, and in so doing, erroneously required the rivers be navigable by large commercial vessels. We agree ANSAC erred in this respect for segment 8 of the Gila (see infra § V-D), but not for the Verde, the Salt, or segments 1 through 7 of the Gila. ¶54 ANSAC did not treat segments as “presumptively non- navigable” based on the lack of large, commercial boating. See Hull, 199 Ariz. at 422–23, ¶¶ 35–37 (quoting A.R.S. § 37-1128.D.3 (2000) (repealed)). For the Verde and Salt, ANSAC made findings about the watercraft customarily used at the time of statehood. For the Verde, ANSAC found various small boats and larger steamboats would have been customary, but those boat types “would need a dependable and reliable draft of around two feet.” ANSAC found there was not enough water in the river for reliable navigation by rowboats and the like. The record reasonably supports those findings. ¶55 For the Salt, ANSAC found at statehood that larger boats such as “keelboats, steamboats, and mountain boats” would have been typical for commercial trade and travel. ANSAC based its finding on expert testimony about how canoes and other smaller boats were no longer customary for commercial cargo hauling or passenger transportation by the time of statehood. DOW correctly points out the record does not indicate whether ANSAC considered if small boats might have been used on the Salt for other commercial purposes, namely hunting and trapping, despite evidence DOW cites showing multiple trappers used segment 6 in the 1890s. ANSAC’s analysis also does not discuss the types of commerce contemplated on the Salt at statehood despite finding that hunting and trapping were contemplated on the Gila at statehood. See infra § V-B. We, however, conclude this gap in ANSAC’s analysis of the Salt does not merit reversal because of our standard of review. See infra § V-B. ¶56 As for the Gila, ANSAC made no express finding about whether small boats were customarily used at the time of statehood. We address the oversight below. See infra § V-D. Ultimately, we conclude the viability of small boats is not significant for segments 1 through 7 of the 16 DOW, et al. v. AZ NSAC, et al. Opinion of the Court Gila, but it is a significant component in our conclusion segment 8 was navigable in fact at statehood. See infra § V-D. D. Actual Use ¶57 ANSAC must consider evidence of actual use, before, around, or after the time of statehood. See, e.g., Utah II, 403 U.S. at 12; Utah I, 283 U.S. at 82; PPL Montana, 565 U.S. at 602–03. The river’s use as a highway for commerce at those times is persuasive evidence. Utah I, 283 U.S. at 82. Evidence of other types of use still may prove susceptibility. Id. The analysis, thus, focuses on the characteristics of the boats used and the river’s condition. See PPL Montana, 565 U.S. at 602–03; see, e.g., United States v. State of Oregon, 295 U.S. 1, 20–21 (1935). ¶58 Here, ANSAC considered evidence of historical and modern boating for all three rivers. For the Verde and Gila, ANSAC made findings about how watercraft and river conditions compared to customary craft and ordinary and natural river conditions at the time of statehood. For the Salt, ANSAC made similar findings about modern boating. As to historical boating, ANSAC said only “there [w]ere instances of historic navigation under unique circumstances or within brief windows of time” and the Salt “was never used for any type of regular trade or travel.” Still, DOW correctly points out the record contains a handful of accounts about successful downstream trips between the late 1800s and 1919 on segments 3 through 5 and more than a dozen such accounts on segment 6. For segment 6, those trips included navigation throughout the year. We agree ANSAC’s characterization of the trips as occurring “within brief windows of time” is not quite accurate. But we find no abuse of discretion because the record supports ANSAC’s finding historical use was not “regular.” E. Susceptibility to Use ¶59 The Daniel Ball test required ANSAC to consider whether a river in its ordinary and natural condition at the time of statehood was susceptible for use as a highway for commerce. Utah I, 283 U.S. at 82. ANSAC had to address whether it would have been commercially realistic for customary craft to navigate the river. See PPL Montana, 565 U.S. at 601. ANSAC could base its determination on a comparison of the river’s condition and the operating requirements of customary craft. See, e.g., Utah I, 283 U.S. at 85; see also, e.g., Utah II, 403 U.S. at 12. In the alternative, or in addition, ANSAC could base its determination on evidence of actual navigation. See PPL Montana, 565 U.S. at 601; see, e.g., Utah I, 283 U.S. at 82; see also, e.g., Ahtna, Inc., 891 F.2d at 1405. Because the proper application of 17 DOW, et al. v. AZ NSAC, et al. Opinion of the Court the Daniel Ball test is a question of law, this court reviews the issue de novo. See Winkleman, 224 Ariz. at 238, ¶ 15. ¶60 DOW contends ANSAC erred by failing to treat evidence of irregular historic use as evidence of navigability and by misapplying the law with respect to seasonal navigability. We address these arguments in turn. 1. Irregular Historical Use ¶61 DOW contends ANSAC erred by overvaluing evidence the rivers were not actually used as highways for commerce by implicitly requiring that actual use be “commercially realistic” to support a navigability determination. As to the Verde, DOW’s argument misconstrues ANSAC’s decision. ANSAC’s analysis of the Salt and Gila are closer calls. ¶62 For the Verde, ANSAC considered both actual use and susceptibility, recognizing the distinction. ANSAC did not conclude the Verde was nonnavigable solely because historic navigation was irregular or commercially unrealistic. Instead, ANSAC concluded navigability would have been too sporadic to be a “commercial reality” at the time of statehood. ANSAC based its conclusion not just on the scarcity and irregularity of historic navigation, but also on the Verde’s ordinary and natural condition at statehood, the commercial context, and the customary craft operating at the time of statehood. Because commercial use is relevant, we cannot conclude ANSAC erred by finding a lack of commercial use weighed against a navigability finding. See Utah I, 283 U.S. at 82. Though DOW disagrees with the weight ANSAC assigned to the historical boating evidence, this court cannot reweigh the evidence or substitute its judgment for ANSAC’s. See Winkleman, 224 Ariz. at 242, ¶ 29. ¶63 For the Salt, ANSAC misconstrued the susceptibility component of the Daniel Ball test. In analyzing the Salt’s susceptibility to commercial navigation, ANSAC said people “simply failed to comprehend the potential usefulness of the [r]iver as an avenue for navigation.” ¶64 ANSAC’s analysis focused on actual and regular commercial use, implying it might be a prerequisite to navigability, at least in the presence of sufficient commercial demand. DOW argues ANSAC erred because well-established jurisprudence illustrates people need not have used a river as a highway for commerce for the river to be navigable. A river segment is navigable either if it was used or if it “[was] susceptible of being used” as a highway of commerce at the time of statehood. PPL Montana, 565 18 DOW, et al. v. AZ NSAC, et al. Opinion of the Court U.S. at 600 (quoting Utah I, 283 U.S. at 76 (quoting The Daniel Ball, 77 U.S. (10 Wall.) at 563)); see also Nw. Steelheaders Ass’n v. Simantel, 112 P.3d 383, 389 (Or. Ct. App. 2005). This point is true regardless of the commercial demands in the area before statehood because the relevant period for analyzing commercial demands is the time of statehood. See PPL Montana, 565 U.S. at 600; see also Utah I, 283 U.S. at 82. True, ANSAC could not presume actual and regular commercial use was a prerequisite to navigability. Still the record reasonably supports ANSAC’s nonnavigability determination. ANSAC weighed the evidence, including the dearth of evidence showing actual use in a populated area. ¶65 ANSAC found the Gila would have been used for commerce if it had been susceptible for such use. ANSAC in fact rejected ASLD’s arguments about people preferring to travel by wagon or train, for cost, convenience, and capacity. But the record does not fully explain why. First, ANSAC was particularly concerned with the Gila’s non-use for mining purposes. That lack of mining use was not particularly probative of segment 8’s navigability, because the mines would either need to transport ore on segments 2 through 7 (which ANSAC found were not navigable) or haul ore hundreds of miles by wagon to Dome (only to boat 15 miles or so to Yuma). And just because the Gila was not a highway for shipping “tons of ore” is not particularly probative of whether trappers and hunters might have used the Gila to transport themselves and their pelts to and from the hunting and trapping grounds. ¶66 Second, ANSAC’s susceptibility analysis was somewhat flawed. Navigability precedent does not stand for the proposition that if commercial use is not found, susceptibility also will not be found. Utah I stands for the inverse proposition—commercial use may be the most persuasive evidence of susceptibility. See Utah I, 283 U.S. at 82. In fact, Utah I expressly recognizes commercial use may not develop before the time of statehood, so the absence of commercial use does not preclude a susceptibility finding. See id. As explained above, Arizona’s territory before statehood fits within that scenario. Though the error is problematic for all three rivers, the effect of this error on Gila’s segment 8 requires reversal because of segment 8’s unique situation. See infra § V-D. As to the other segments, sufficient evidence supports ANSAC’s findings such that it did not abuse its discretion. 2. Actual Use and Nonuse ¶67 DOW argues ANSAC’s focus on actual use rendered the susceptibility prong a nullity. We disagree. 19 DOW, et al. v. AZ NSAC, et al. Opinion of the Court ¶68 Though actual use cannot subsume susceptibility, the United States Supreme Court in PPL Montana recognized actual use, including “[e]vidence of recreational use . . . may bear upon susceptibility of commercial use at the time of statehood.” 565 U.S. at 600–01; see also Utah I, 283 U.S. at 82. Actual use is not dispositive. See Utah I, 283 U.S. at 82. ¶69 Though DOW highlights several sections of ANSAC’s decisions discussing actual use in determining susceptibility, DOW fails to recognize susceptibility is not severable from the greater navigability context. Whether a river segment is susceptible to navigation depends on the rest of the Daniel Ball test—the natural and ordinary condition of the river segment and what constituted commerce at and around statehood. ¶70 Consistent with this framework, ANSAC did not rely solely on actual use when making its susceptibility analysis. Instead, ANSAC considered the voluminous record to determine: the types of commerce at statehood; the rivers’ physical characteristics, geomorphology, and hydrology; and actual use during various historical periods—all of which bore on susceptibility. And ANSAC discussed actual use in the context of exploration, settlement, and commercial needs, seeking an explanation as to why people had not actually used a segment for commerce if that segment was susceptible to commercial use. ¶71 Because actual use may be the “most persuasive” component in the susceptibility analysis, we cannot say ANSAC abused its discretion by relying heavily—but not exclusively—on nonuse. See Winkleman, 224 Ariz. at 243, ¶ 31 (“[T]he relevance and weight to be afforded the evidence is generally for ANSAC to make.”). Though DOW brought forth evidence to explain nonuse, ANSAC considered DOW’s explanation unconvincing. The weight ANSAC places on evidence influencing susceptibility goes to whether ANSAC properly applied the navigability test. Except as to segment 8 of the Gila, it did. 3. Seasonal Navigability ¶72 A river may be navigable even if it is not susceptible to commercial use all the time or even most of the time. Susceptibility need only be persistent and regular enough to make navigation “a commercial reality.” PPL Montana, 565 U.S. at 602–03; Utah I, 283 U.S. at 87 n.12. Seasonal navigability is sufficient to establish navigability in fact, even if the season is short but still ordinary and long enough to make commercial use realistic. Compare Oklahoma, 258 U.S. at 591 (explaining navigability cannot be “confined to the irregular and short periods of temporary high 20 DOW, et al. v. AZ NSAC, et al. Opinion of the Court water”) with Econ. Light & Power Co. v. United States, 256 U.S. 113, 122 (1921) (explaining navigation need not “be open at all seasons of the year, or at all stages of the water”); see also Riverfront Prot. Ass’n, 672 F.2d at 795. ¶73 DOW argues ANSAC failed to recognize periodic navigation can support navigability. Contrary to DOW’s argument, ANSAC did not impose a requirement of constant navigability. But it did reject navigation “so brief that it . . . is not a commercial reality, requiring more than [o]ccasional use in exceptional times.” ANSAC characterized the three rivers’ flows as “erratic,” “highly variable,” and fluctuating between flood, even flow, and dryness. Those findings were relevant to whether boating would be commercially realistic. DOW also faults ANSAC for considering whether historical trips took place during a period of “high flows.” But river conditions during boat trips are relevant and necessary to determine whether actual use supports the susceptibility determination. See PPL Montana, 565 U.S. at 601. Because ANSAC received competing evidence on the natural and ordinary condition of the rivers and what constituted realistic commerce at statehood, we cannot say ANSAC misapplied the law with respect to seasonal navigability. V. Sufficiency of the Evidence ¶74 DOW argues substantial evidence does not support ANSAC’s nonnavigability findings for segments 1 through 5 of the Verde, segments 3 through 6 of the Salt, and all segments of the Gila. “[T]his court considers the river on a segment-by-segment basis to assess whether the segment of the river, under which the riverbed in dispute lies, is navigable or not.” PPL Montana, 565 U.S. at 577. Here, no party disputes ANSAC’s segmentation. ¶75 “In reviewing the superior court’s ruling affirming an agency’s order, [this court] ‘independently examines the record to determine whether the evidence supports the judgment,’ under a preponderance of the evidence standard.” Parsons v. Ariz. Dep’t of Health Servs., 242 Ariz. 320, 322, ¶ 10 (App. 2017) (quoting Webb v. State ex rel. Ariz. Bd. of Med. Exam'rs, 202 Ariz. 555, 557, ¶ 7 (App. 2002)). Substantial evidence exists when the record supports the agency’s decision, “even if the record also supports a different conclusion.” Gaveck, Ariz. State Bd. of Podiatry Exam’rs, 222 Ariz. 433, 436, ¶ 11 (App. 2009). This court views the evidence in a “light most favorable to upholding the [agency’s] decision and will affirm if any reasonable interpretation of the record supports the decision.” Lewis v. Ariz. State Pers. Bd., 240 Ariz. 330, 334, ¶ 15 (App. 2016) (citing Baca v. Ariz. Dep’t of Econ. Sec., 191 Ariz. 43, 46, ¶ 6 (App. 1997)). 21 DOW, et al. v. AZ NSAC, et al. Opinion of the Court A. The Verde ¶76 Citing evidence of historical use, DOW argues ANSAC erred by finding segments 1 through 5 of the Verde nonnavigable. Because substantial evidence supports ANSAC’s findings of nonnavigability, we disagree. ¶77 ANSAC found segments 1 through 5 nonnavigable based on the following factual findings: (1) unpredictable flooding caused persistent channel changes; (2) seasonal periods of high flows; (3) a lack of historical use as a highway for commerce; (4) federal survey evidence and federal patent evidence; (5) historical descriptions of the Verde as well as its hydrology and geomorphology; (6) differences in modern boats; and (7) river obstacles. ¶78 Those findings are sufficient to support ANSAC’s nonnavigability determinations for all five segments. ANSAC cited experts’ estimates of median river depths, based on undepleted flows, of less than two feet for segments 1 through 5. ANSAC found boats readily available at the time of statehood, like flat-bottom fishing boats and rowboats, would need “around two feet.” ANSAC also cited evidence of natural rapids of Class II grade in segments 1, 2, and 5, Class III grade in segment 4, and Class IV grade in segment 3. DOW has not challenged these specific findings, which are plainly sufficient to establish even the smallest boats would not have been able to reliably navigate the Verde. DOW cites other evidence favoring navigability, asserting that evidence was “far more than a preponderance.” But we must affirm ANSAC’s determination because the record supports it, even if sufficient evidence supports a contrary conclusion. B. The Salt ¶79 DOW next argues ANSAC erred when it found segments 3 through 6 of the Salt nonnavigable because evidence about historical boating and customary craft compelled a navigability finding. We disagree. ¶80 ANSAC concluded the Salt was not navigable because natural impediments to navigation “would require portages” because of the irregularity of historic use, and because modern boating evidence was unpersuasive given dissimilarities in boat durability. Even so, ANSAC’s analysis of historical boating on the Salt is not comprehensive. See supra §§ IV-C, IV-E. Because DOW premises its argument on historic boating evidence and customary craft, rather than modern boating, the remaining 22 DOW, et al. v. AZ NSAC, et al. Opinion of the Court issue is whether the record supports ANSAC’s conclusion segments 3 through 6 required portage. ¶81 ANSAC described segments 5, 6, and parts of segment 3 as wide, shallow, and extensively braided. ANSAC also noted “bars” in segment 6. ANSAC further noted hydrological records showing the Salt flowed erratically, “fluctuating between flood, even flow, and dryness.” ANSAC cited evidence of natural Class II rapids in segments 3 through 5 and shallow riffles in segment 6. ANSAC also cited evidence showing, before statehood, federal surveyors described the Salt as shallow and did not note characteristics of navigability in their surveys. And ANSAC noted federal land grants did not depict the Salt as navigable. ¶82 DOW points to evidence that could sustain an opposite conclusion. The record suggests Class II rapids are navigable, “without scouting,” with only “occasional maneuvering.” ANSAC’s analysis did not say sandbars, riffles, rapids, or fluctuating river flows would have served as portage-like barriers to commercial navigation on the Salt. See Utah I, 283 U.S. at 85–87. DOW also argues the navigability conclusions of surveyors and federal agencies are not particularly probative as they are not “clothed with power to settle questions of navigability.” See Oklahoma, 258 U.S. at 585. And their observations of “shallowness” are too vague to be useful in determining whether the segments might be navigable by low-draft boats. Assuming small boats were customary at the time of statehood, the record suggests operating depths of one to two feet would be reasonable, with lower depths necessary in sections without rapids. The record suggests the undepleted river had median depths of at least 1.5 feet in the four segments but contains conflicting evidence about whether depths exceeded 2 feet. ¶83 ANSAC faced an enormous record and had the responsibility of weighing various components with competing, and sometimes contradictory, evidence. ANSAC had the duty to weigh the evidence. See Winkleman, 224 Ariz. at 242, ¶ 29. ANSAC did not specifically indicate how it weighed this conflicting evidence. Even so, “[i]n reviewing factual determinations, we will not substitute our conclusion for that of the administrative agency; instead, we review the record to determine whether substantial evidence supports the agency’s decision and whether the agency exercised its discretion reasonably and with due consideration.” See id. at 238, ¶ 14. We, thus, affirm ANSAC’s conclusion regarding segments 3 through 6 of the Salt. 23 DOW, et al. v. AZ NSAC, et al. Opinion of the Court C. The Gila ¶84 Again, citing historical evidence, DOW argues ANSAC erred when it found all the Gila nonnavigable. As to segments 1 through 7, we disagree. ¶85 ANSAC concluded the Gila was nonnavigable based on the following factual findings: (1) unpredictable flooding caused persistent channel changes and seasonal high-flow periods; (2) a lack of historical or modern use as a highway for commerce; (3) natural Class II and III rapids in segment 4 and natural Class I rapids in segments 1, 2, and 5; (4) “highly braided channel[s]” in segments 1, 3, and 6 at the time of statehood because of flooding around the turn of the century; (5) sandbars, rock outcroppings, beaver dams, marshes, strainers, and other obstacles in “various parts of the river”; (6) improved navigability of modern boats; and (7) almost all observers before and at the time of statehood considered the Gila nonnavigable. ¶86 Those findings—when viewed in the light most favorable to affirming ANSAC—are sufficient to support ANSAC’s nonnavigability determination. Despite other compelling evidence, ANSAC did not err in finding the lack of regular historical use was compelling evidence of the Gila’s nonnavigability. See supra § IV-E-1. ANSAC discussed the evidence of historical boating on the Gila at length. Though the record may also support DOW’s assertion boating was common enough to compel a navigability determination, the record is sufficient to uphold ANSAC’s determination to the contrary. ¶87 ANSAC’s findings regarding rapids, channel changes, sandbars, rock outcroppings, beaver dams, marshes, strainers, and other obstacles do not necessarily compel a nonnavigability determination. See Utah I, 283 U.S. at 82–87. Again, viewing the record in the light most favorable to upholding ANSAC’s nonnavigability determination, the record sufficiently supports ANSAC’s findings. ¶88 Historical observations about a river’s nonnavigability are of limited probative value. See Oklahoma, 258 U.S. at 585. But here, the historical observations were the best evidence of the Gila’s ordinary and natural condition, see Winkleman, 224 Ariz. at 242, ¶ 30, and ANSAC did not indicate whether it found the historical observations or expert estimates more credible. ¶89 DOW argues the evidence compels a navigability determination. We agree the record could reasonably support such a 24 DOW, et al. v. AZ NSAC, et al. Opinion of the Court finding. Even so, we affirm the nonnavigability determination except for segment 8 because substantial evidence supports ANSAC’s decision and reasonably supports the opposite conclusion. See Winkleman, 224 Ariz. at 238, ¶ 14. D. Segment 8 of the Gila ¶90 ANSAC found “the evidence regarding navigability and nonnavigability [was] evenly weighted” and, thus, concluded that segment 8 was nonnavigable. Because of errors described below, the scales tip in favor of navigability based on ANSAC’s own assessment of the evidence. ¶91 ANSAC cited historical accounts of depths of 9 to 15 feet in segment 8 between the 1850s and the 1890s, from the period during and immediately after the river was in its ordinary and natural condition. ANSAC also noted accounts reported the river was dry during “some seasons” in segment 8, but seasonal lows do not preclude a finding of navigability. See supra § IV-E-3. Small, low-draft boats would have had no trouble navigating at the reported depths, and the record suggests larger craft such as barges, mountain boats, and even steamboats would have been capable as well. Those craft would have been commercially useful for transporting people, military supplies, mining materials, and agricultural goods as well as for hunting and trapping, which ANSAC found were types of commerce “contemplated prior to and at statehood” on the Gila. See § IV-C. As such, based on the ordinary and natural river conditions, the commercial demands, and the customary watercraft, segment 8 would have been susceptible to at least seasonal commercial use. ¶92 Likewise, the historical boating evidence on segment 8 compels a navigability finding. ANSAC cited many reports of successful downstream boat trips on segment 8 before the turn of the century. ANSAC noted: (1) the Howard family boating from Gila Bend to Yuma in 1849; (2) a “few accounts” of Forty-Niners traveling to Yuma in small boats; (3) the Hamilton party taking a homemade skiff from Phoenix to Yuma in 1879; (4) William Eaton “clear[ing] $1,500”on the lower Gila in 1884; (5) the Day brothers trapping along the lower Gila in 1891 to 1892; (6) G.W. Evans and Amos Adams traveling from Phoenix to Yuma in a homemade wooden flat boat in 1895; (7) and Lieutenants Gully and Richardson floating another homemade wooden boat from the Pima Villages to Yuma in 1896. The record also contains evidence of steamboats traveling a handful of miles upstream from the Colorado River confluence in the 1860s. 25 DOW, et al. v. AZ NSAC, et al. Opinion of the Court ¶93 Though a few of those trips happened while the Gila was in its ordinary and natural condition, most were after significant diversions had occurred. Those diversions generally would have made navigability more difficult. The boats used in the trips listed in the paragraph above would have been meaningfully similar to boats at the time of statehood. And steamboats aside, at the very least, the historical evidence of small boats navigating segment 8 shows that segment was susceptible to commercial use for hunting and trapping in small boats at statehood. ¶94 In sum, the evidence of the Gila’s pre-diversion condition and navigation was not, as ANSAC said, “evenly weighted.” Instead, DOW met its burden, by a preponderance of the evidence, of showing segment 8 was at the time of statehood susceptible to use for trade and travel as a highway for commerce. We need not consider ANSAC’s failure to address the effects of human impacts because the human impacts would have reduced, not increased, the flow rate at the time of statehood. See Winkleman, 224 Ariz. at 241–42, ¶ 28. And as noted above, even without considering those effects, segment 8’s flow rate was sufficient to establish navigability. As a result, we reverse ANSAC’s nonnavigability finding on segment 8 of the Gila and hold segment 8 of the Gila was navigable at statehood. VI. Title to Riverbeds A. Pre-Statehood Dams ¶95 SRP argues this court may not reverse a finding of navigability if it determines any segments of the Salt or lower segments of the Gila would be navigable but for the federal government’s construction of pre-statehood dams—primarily the Roosevelt Dam and Granite Reef Dam—or diversions under the Reclamation Act. Specifically, SRP argues because the 1910 Arizona-New Mexico Enabling Act reserved the federal government’s interest in the riverbeds subject to pre-statehood projects pursued under the Reclamation Act, this court should find any such projects have no bearing on the rivers’ ordinary and natural condition. We disagree. ¶96 “States enjoy a presumption of title to submerged lands beneath inland navigable waters within their boundaries.” Alaska v. United States, 545 U.S. 75, 78 (2005). But “[t]he Federal Government can overcome the presumption and defeat a future State’s title to submerged lands by setting them aside before statehood in a way that shows an intent to retain title.” Id. at 79 (citing United States v. Alaska, 521 U.S. 1, 33–34 (1997)). The federal government’s “intent . . . must be definitely declared or otherwise 26 DOW, et al. v. AZ NSAC, et al. Opinion of the Court made very plain.” See id. (citing Alaska, 521 U.S. at 34) (quoting Holt State Bank, 270 U.S. at 55). ¶97 SRP relies on two United States Supreme Court cases. See id; Idaho v. United States, 533 U.S. 262 (2001). But these cases both deal with riverbeds underlying pre-statehood federal reservations. Alaska, 545 U.S. at 102–03; Idaho, 533 U.S. at 265–71. To the extent SRP argues the federal government reserved title to the submerged lands, we do not resolve that issue because the issue before ANSAC was navigability, not riverbed ownership. ¶98 As part of this argument, SRP urges this court to adopt a rule excepting the effects of federal infrastructure made under a federal reservation on a river’s ordinary and natural condition. But the federal test for navigability is whether the rivers would have been navigable in their “natural and ordinary condition” at statehood. PPL Montana, 565 U.S. at 592 (quoting Oklahoma, 258 U.S. at 591). And this court has already considered and rejected an approach allowing ANSAC to ignore the effects of pre- statehood dams and diversions. See Winkleman, 224 Ariz. at 240–41, ¶¶ 23– 28. As a last point, SRP cites no authority to support this proposition, and we find none. B. Riverbed Title on Indian Reservations ¶99 We need not resolve the effect of our ruling on portions of riverbeds lying within the Yavapai-Apache Reservation, Fort McDowell Yavapai Nation, Gila River Indian Reservation, San Carlos Apache Indian Reservation, and Salt River Indian Reservation. See Ariz. Const. art. 20, § 4. Segment 8 of the Gila is the only segment we conclude was in fact navigable at statehood, and no portion of segment 8 lies within a federally recognized Indian reservation. ¶100 This court’s ruling on navigability or nonnavigability, thus, has no consequence on any tribe’s title to lands in Indian country. ATTORNEY FEES ¶101 DOW requests attorney fees under A.R.S. § 12-348 and “the private attorney general doctrine” as laid out in Arnold v. Ariz. Dep’t of Health Servs., 160 Ariz. 593 (1989). Because DOW failed to prevail on most issues, we decline its request for costs and attorney fees under A.R.S. § 12- 348. See Aqua Mgmt., Inc. v. Abdeen, 224 Ariz. 91, 96–97, ¶ 23 (App. 2010) (awarding attorney fees to party who prevailed on majority of issues). And given DOW’s failure to prevail on most issues, we exercise our discretion 27 DOW, et al. v. AZ NSAC, et al. Opinion of the Court to decline its request for attorney fees under the private attorney general doctrine. See Arnold, 160 Ariz. at 609. We decline to award costs under A.R.S. § 12-341 because DOW prevailed only in small part. CONCLUSION ¶102 We affirm ANSAC’s determination for all segments of the Verde and the Salt and for segment 1 through 7 of the Gila. We reverse ANSAC’s determination for segment 8 of the Gila and hold segment 8 of the Gila was navigable at statehood. AMY M. WOOD • Clerk of the Court FILED: AA 28