JUSTICE WHEAT
delivered the Opinion of the Court.
¶1 Claimants Gregory Duncan, Sherri Donovan, and Terry Dougherty (collectively Duncan) and Skelton Angus Ranch, Inc. (Skelton) appeal from a decision of the Montana Water Court that amended the Water Master’s (Water) Report and adopted it as amended. We affirm.
¶2 We address the following issues:
¶3 1. Did the ChiefWater Judge properly admit historical documents prepared by Pondera County Canal & Reservoir Company (Pondera) in anticipation of litigation?
¶4 2. Did the ChiefWater Judge correctly reject the Master’s findings regarding certain variables used to determine the historical capacity of a flume on the Thomas ditch?
¶5 3. Did the ChiefWater Judge correctly determine that portions of the claimants’ water rights had been abandoned or never perfected?
¶6 4. Did the Chief Water Judge correctly adopt the Master’s conclusion that the claimants did not acquire any water rights by adverse possession?
FACTUAL AND PROCEDURAL BACKGROUND
¶7 This matter comprises two joined cases, both involving claims to water diverted from the South Fork of Dupuyer Creek in Two Medicine River Basin, into Gansman Coulee, for irrigation in the Teton River Basin.
¶8 In the adjudication process, Duncan and Skelton filed statements of claim for existing rights based on various Notices of Appropriation (NOAs) filed between 1895 and 1913. The claims share a single point of diversion from the South Fork, Dupuyer Creek, through the Thomas ditch, to the claimants’ lands. The Duncan property is positioned where the diverted water enters Gansman Coulee, and the Skelton property adjoins the Duncan property further downstream. Pondera diverts water from South Fork, Dupuyer Creek, downstream of the Thomas ditch. Pondera filed notices of intent to appear in the adjudication of Skelton’s and Duncan’s claims. A Temporary Preliminary Decree for [330]*330Teton River Basin was issued on December 29,2005. The objections at issue here stem from that decree.
¶9 Following a hearing, the Master, in his Report, quantified and assigned priority dates to the claimed water rights. The Master concluded the capacity of a flume through which the water was diverted after the Thomas ditch washed out around 1908 limited the quantity of water that had historically been put to beneficial use and could be claimed. After reviewing the evidence, the Master concluded that the initial flume was built around 1912. Certain water rights were limited in quantity to the capacity of the 1912 flume. The 1912 flume washed out several times and was completely rebuilt in 1931. The 1931 flume was significantly larger than the 1912 flume. Construction of the 1931 flume resulted in creation of new “implied” water rights, with a 1931 priority date, to be distributed among the parties. Because he concluded the flume’s capacity limited the water rights, the Master concluded that claimed quantities of water exceeding the capacity he calculated had been abandoned or never perfected.
¶10 The Master’s Report recognized that determining the flume’s original capacity had been complicated by the number of repairs and improvements that took place over the years. In addition, the Master noted that the original flume had been washed out and rebuilt on many occasions. “In fact,” he wrote, “maintaining this diversion has been a huge problem.” This, he suggested, was in part due to frequent flood events that caused “significant destruction.”
¶11 Information about the flume around the time of its construction primarily derived from documents Pondera compiled in the early 1900s, after it began investigating the water rights in the area to determine the viability of attempting to obtain land under the Federal Carey Land Act. All parties acknowledge that the documents generated by these investigations were created — at least in part — in anticipation of litigation of the water rights on Dupuyer Creek. Prior to the hearing before the Master, the claimants moved to exclude these documents because they were prepared in anticipation of litigation and contained self-serving declarations. The Master determined they were admissible as ancient documents.
¶12 The available information revealed that, in 1912, H.A. Bestor (Bestor), apparently a surveyor or engineer for Pondera, estimated the flume was 24 inches wide and 8 inches deep, with a slope of .003. Bestor calculated the flume’s capacity at 4.22 cubic feet per second (cfs); the claimants’ expert witness Ryan Casne (Casne) opined that calculation was based on water flowing through the flume at that time [331]*331rather than the flume’s maximum capacity. In 1918, a diary entry provided by Pondera documented the size of the flume at approximately 23 inches wide, 10 inches deep, and 400 feet long. Various interviews with local residents conducted in the early 1920s indicated that the flume washed out in 1916, was rebuilt by 1918, and that part of the flume was rebuilt in 1920. One of the interviewees anecdotally opined that the rebuilt flume was “the same size” as the previous structure — but actual measurements did not corroborate this estimate. Notes taken by a Pondera employee named Mattison, dated August 26,1920, measured the flume at 24 inches wide and 11 inches deep. An Exhibit from Pondera based on 1921 and 1922 field work calculated the flow rate of the flume at 3.49 cfs. An affidavit from Woodrow W. Coffins mentioned a dam, which the Master concluded was “a significant system improvement” in the 1930s. In 1936, a memorandum drafted for Pondera’s predecessor documented the flume was 30 inches wide, 22 inches high and 200 feet long; and estimated the flow rate at 12-15 cfs. From 1947 to 1999, the claimants’ predecessors testified, the flume was approximately 36 inches wide, 24 inches deep and 300 feet long.
¶13 The parties, recognizing the significance of flume capacity, submitted expert testimony regarding that capacity. Experts used two different formulas to calculate the flume’s capacity or flow rate: Manning’s formula for an outlet-controlled structure, and the formula for an inlet-controlled structure. Testimony explained that the flume’s length would determine whether the structure was outlet- or inlet-controlled; if the structure was 300 feet long or greater, experts agreed it was likely outlet-controlled, and its capacity best determined using Manning’s formula.
¶14 Everyone who calculated the original 1912 structure’s capacity assumed that its slope was .003 and that the roughness coefficient was .012, figures taken from Bestor’s 1912 Report. Expert Casne, assuming the flume was 24 inches wide and 8 inches deep, calculated the flume’s capacity at 4.6 cfs. Pondera’s expert Bruce Anderson (Anderson) relied on Bestor’s 1912 calculation and concluded that 4.22 cfs was a reasonable flow rate for the original flume. The Master performed an independent calculation using Manning’s formula for an outlet-controlled structure using an online calculator. Assuming the flume was 24 inches wide by 11 inches deep, and the slope was .003, he calculated the flow rate was 7.6 cfs, almost double the rate any expert witness calculated.
¶15 As for the 1931 flume, assuming that the flume was an inlet-[332]*332controlled structure, expert testimony estimates of its flow rate ranged from 20.0 cfs to 21.97 cfs. Assuming it was an outlet-controlled structure, estimates varied almost 300 percent, due to use of different slopes and roughness coefficients. There was no reliable slope information for the flume after it was enlarged around 1931. Again, the Master performed an independent calculation and arrived at a much larger amount of water using Manning’s formula than the experts did using the formula for inlet-controlled structure. Critical, here, is that Manning’s formula, the formula for an outlet-controlled structure, relies heavily on slope. On the other hand, in the calculation for an inlet-controlled structure, slope is not relevant. In calculating the flow rate of the reconstructed 1931 flume, the Master used the slope coefficient from the 1912 flume, in Manning’s formula.
¶16 The Chief Water Judge rejected the Master’s findings on the capacities of the two flumes and associated water rights because he concluded that the Master “misapprehended the effect of the evidence.” The Chief Water Judge replaced the Master’s findings with factual findings he considered better supported by the evidence in the record, pursuant to his authority under M. R. Civ. P. 53(e)(2) and Rule 23, W.R.Adj.R.
¶17 With regards to the capacity of the original 1912 flume, the Chief Water Judge explained that the dimensions the Master used in conducting his calculations were not supported by the record and were different from those used by the experts who testified. Specifically, the Chief Water Judge found that the flume dimensions the Master used in his calculations were based on the 1920 flume structure, rather than the original 1912 flume — even though the Master used the slope and roughness coefficients from the 1912 flume. The Chief Water Judge pointed out that using a later measurement was problematic because the purpose of determining the original flume’s capacity was to quantify the amount of water originally put to beneficial use, and since the slope of the flume in the 1920s was not known. By using dimensions from a later structure, the Master affected the quantification and priority date of the water right, since those determinations were based in part on the flume capacity.
¶18 As to the second, 1931 flume, the Chief Water Judge concluded the Master committed clear error by relying on a formula that depended heavily on slope when the factual record contained no slope measurement for the second flume. The Master simply used Bestor’s slope measurement from the 1912 flume, even though the flume had, by all accounts, been completely rebuilt by 1931. Further, the slope [333]*333measurement the Master used was almost twice as steep as post-reconstruction slope information. The Chief Water Judge also reversed the Master’s determinations with regards to the validity of some of the water rights at issue.
¶19 The 1895 B.P. Clark NOA claimed 25 cfs. Dimcan and Skelton both claim the full 25 cfs. For water appropriated pursuant to the B.P. Clark NOA, the Master ultimately recommended that the right was abandoned except what could be carried in the 7.6 cfs flume. Due to a subsequent land transfer, he found that Duncan was entitled to 0.35 cfs of this right and Skelton was entitled to 7.25 cfs. The Chief Water Judge determined that the evidence was insufficient to establish perfection of rights based on the B.P. Clark NOA on lands owned by Skelton. The Chief Water Judge determined that the B.P. Clark appropriation right was severed from lands owned by Duncan because it was reserved when Armedia Clark, in 1903, specified the rights conveyed in a deed to Duncan’s predecessor. Consequently, the Chief Water Judge did not allow the claimants any water based on this NOA.
¶20 Both Duncan and Skelton also claim water based on a 1902 NOA for 50 cfs filed on behalf of Armedia Clark. Both claimants sought the full 50 cfs. The Master held that, of the original 50 cfs, only 4.5 cfs were acquired by Duncan’s predecessors through the 1903 Armedia Clark deed and should pass on to Duncan. The Master held that Skelton’s predecessor abandoned the right to appropriate water under the Armedia Clark NOA and the right could not pass to Skelton. The Chief Water Judge adopted the Master’s recommendations for the Armedia Clark NOA.
¶21 Skelton claims water based on a NOA filed by Mustard, Roberts, and Deschenau in 1904, and another by Walter Clark in 1906. The Master and the Chief Water Judge agreed that the claims based on these NOAs were invalid because they were never perfected.
¶22 Dimcan claimed water based on a 1912 or 1913 NOA for 10 cfs filed on behalf of Theresa Flacker. That right was diverted through the Flacker ditch by 1913, but the Flacker ditch was tom out soon afterwards. The Master found that Flacker conveyed this appropriation to one of Duncan’s predecessors in a land sale. The new owner, Duncan claimed, diverted the right through the Thomas flume, beginning in 1915. The Master accepted this proposition, but held that the owner lost all but 3.1 cfs due to the capacity of the flume. The Master recommended that Duncan acquired this 3.1 cfs as a “use right,” because of discrepancies with the filing date. The Chief Water Judge, having determined the flume’s capacity was significantly smaller than [334]*334what the Master calculated, concluded that the flume did not have sufficient capacity to convey both the Flacker and the Armedia Clark appropriations. He further noted that the only real evidence in the record of the Flacker right was that the ditch had been destroyed and the right never used again. The Chief Water Judge effectively concluded that any claim based on the Flacker NOA had been abandoned.
¶23 Finally, both Duncan and Skelton claim water based on the increased flow rate that resulted from the installation of a larger flume in 1931. The Master recommended creating implied claims for Duncan and Skelton for the difference between the capacity of the post-1931 flume, which he determined to be 36.32 cfs, and the amount of water allocated to each claimant based on the 1912 flume. The Master’s reason for creating an implied claim was that the 1931 appropriation was based on several existing claims which appeared to be overstated, creating a need to distill these into separate water right claims with a 1931 priority date. See W.R.C.E.R. 35. For the implied claim rights, the Master determined that Duncan was entitled to 28.37 cfs and Skelton was entitled to 29.07 cfs — both with a December 31, 1931, priority date. On review of the Master’s report, the Chief Water Judge held: “Other than the amounts of the implied claims, the Master’s rationale is sound.” Because the Chief Water Judge determined that the capacity of the post-1931 flume was only 20 cfs, and that Duncan was entitled to use 4.5 cfs pursuant to the Armedia Clark NOA, the court concluded that the claimants’ implied rights with a 1931 priority date only amounted to 15.5 cfs each.
¶24 Duncan and Skelton now appeal from the Water Court’s order.
STANDARD OF REVIEW
¶25 When reviewing the Water Court’s decision, two standards are relevant: The standard the water judge applies to the Master’s report and the standard we apply to the Water Court’s opinion. Heavirland v. State, 2013 MT 313, ¶ 13, 372 Mont. 300, 311 P.3d 813. The Water Court reviews the Master’s findings of fact for clear error and the Master’s conclusions of law to determine whether they are correct. Heavirland, ¶ 14. Applyingthese standards of review, “the water judge may adopt, modify, or reject the [Master’s] report, in whole or in part, or may receive further evidence or recommit it with instructions.” W. R. Adj. R. 23; M. R. Civ. P. 53(e)(2).
¶26 We review the Water Court’s order de novo, to determine whether it correctly applied the clear error standard of review to the Master’s [335]*335findings of fact and whether its conclusions of law were correct. Heavirland, ¶ 15. “[W]hether the standard of review was applied correctly is a question of law.” Heavirland, ¶ 15 (citing Milliken Research Corp. v. Dan River, Inc., 739 F.2d 587, 593 (Fed. Cir. 1984)). We review the Water Court’s findings to determine whether they are clearly erroneous. Weinheimer Ranch v. Pospisil, 2013 MT 87, ¶ 19, 369 Mont. 419, 299 P.3d 327; see Milliken Research Corp., 739 F.2d at 593 (“we must first review, as a matter of law, the correctness of the district court’s setting aside any factual finding by the master and, if that is upheld, review any substitute or additional findings of the district court under the ‘clearly erroneous’ standard....”).
¶27 The standard we use to determine whether a finding of fact by a court sitting without a jury is clearly erroneous is set forth in Interstate Prod. Credit Ass’n v. Desaye, 250 Mont. 320, 323, 820 P.2d 1285, 1287 (1991); see also e.g. In re the Existing Rights within the Jefferson River Drainage Area, Nos. 41G-137, 41G-W-182145-00, 1999 Mont. Water LEXIS 1 at **3-4 (Dec. 27, 1999) (setting forth the test the Water Court uses when reviewing objections to a Master’s Findings of Fact; citing Desaye). First, the reviewing court reviews the record to see whether the findings are supported by substantial evidence. Desaye, 250 Mont. at 323, 820 P.2d at 1287. Second, even if the findings are supported by substantial evidence, the reviewing court may determine a finding is clearly erroneous if the trial court misapprehended the effect of the evidence. Desaye, 250 Mont. at 323, 820 P.2d at 1287. Third, even where the findings are supported by substantial evidence and the court has not misapprehended the evidence’s effect, the reviewing court may determine a finding is clearly erroneous when “ ‘although there is evidence to support it, the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed.’ ” Heavirland, ¶ 16 (quoting U.S. v. U.S. Gypsum Co., 333 U.S. 364, 395, 68 S. Ct. 525, 542 (1948)). “Substantial evidence is evidence which a reasonable mind might accept as adequate to support a conclusion, even if the evidence is weak or conflicting.” Arnold v. Boise Cascade Corp., 259 Mont. 259, 265, 856 P.2d 217, 220 (1993). It need not amount to a preponderance of the evidence, but it must be more than a scintilla. State v. Shodair, 273 Mont. 155, 163, 902 P.2d 21, 26 (1995). Although the standard is a deferential one, substantial evidence is not synonymous with clearly erroneous and a reviewing court may find that a finding is clearly erroneous even though there is evidence to support it. Heavirland, ¶ 16.
[336]*336¶28 We review a District Court’s evidentiary rulings for an abuse of discretion. Stevens v. Novartis Pharms. Corp., 2010 MT 282, ¶ 23, 358 Mont. 474, 247 P.3d 244.
DISCUSSION
¶29 1. Did the Chief Water Judge properly admit historical documents prepared by Pondera in anticipation of litigation?
¶30 Duncan and Skelton argue that Pondera should not have been allowed to introduce its historical documentation of the water rights into evidence because the documents constitute self-serving hearsay evidence that should have been admissible only if offered against Pondera.
¶31 “Hearsay is a statement, other than one made by the declarant while testifying at the trial or hearing, offered in evidence to prove the truth of the matter asserted.” M. R. Evid. 801(c). “Hearsay is not admissible except as otherwise provided by statute, [the Montana Rules of Evidence], or other rules applicable in the courts of this state.” M. R. Evid. 802. Statements in “ancient documents” — defined as “a document in existence for twenty years or more, the authenticity of which is established” — are exceptions to the hearsay rule. M. R. Evid. 803(16).
¶32 The authenticity requirement of the ancient documents exception is a condition precedent to admissibility and “is satisfied by evidence sufficient to support a finding that the matter in question is what its proponent claims.” M. R. Evid. 901(a). Rule 901(b) provides an illustration of criteria that, if met, would fulfill the authenticity requirement for ancient documents:
Evidence that a document or data compilation, in any form, (A) is in such condition as to create no suspicion concerning its authenticity, (B) was in a place where it, if authentic, would likely be, and (C) has been in existence 20 years or more at the time it is offered.
M. R. Evid. 901(b)(8).
¶33 The Master explained that Pondera made extensive studies of the entire area and that, while “it is clear that a certain amount of that data was acquired in anticipation of potential water right litigation ... it is apparent this was not the only reason for the work.” The documents in question were prepared by Pondera employees and consist of reports, surveys, maps, hydraulic data, memoranda, and interview notes. Pondera documented visits to the diversion site by various employees and recorded numerous measurements of the flume [337]*337over the years. The Master held that “[w]hile there may have been a tendency to minimize competing appropriations, it is not appropriate to dismiss all of this documentation as inadmissibly self-serving.” The Master noted that the data could help establish an approximate size of the flume.
¶34 The claimants concede that the documents have existed for well over twenty years. Pondera’s documents contain a significant amount of historical information about the diversion dating generally from 1900 to 1940. Claimants also concede that the documents are what Pondera claims them to be and that the documents were stored in a place where such documents would be kept. Authenticity, accordingly, is not in dispute.
¶35 Claimants argue instead that the documents are inadmissible because the Pondera documents do not antedate the present controversy and a motive for misrepresentation already existed when the documents were created. Claimants cite Palmer v. Hoffman, 318 U.S. 109, 63 S. Ct. 477 (1943), in which the U. S. Supreme Court affirmed the trial court’s determination that an accident report prepared in contemplation of litigation lacked the trustworthiness required to satisfy the business records exception to the hearsay rule. Palmer, 318 U.S. at 114, 63 S. Ct. at 481. The generally accepted view of Palmer is that “documents prepared for litigation are excluded, not on a per se basis, but rather upon an inquiry into whether such documents bear circumstantial indicia of lack of trustworthiness.” Jefferson Garden Assocs. v. Greene, 520 A.2d 173, 181 (Conn. 1987) (noting that courts may exclude documents prepared in anticipation of litigation under the Palmer test, but that the decision whether to exclude such documents requires “the exercise of appropriate discretion”). While we have previously addressed documents containing self-serving declarations, see King v. Schultz, 141 Mont. 94, 99, 375 P.2d 108, 111 (1962) and Osnes Livestock Co. v. Warren, 103 Mont. 284, 296, 62 P.2d 206, 211 (1936), we have not created a per se rule that all documents created in anticipation of litigation must be excluded.
¶36 The trial court has broad discretion to determine the admissibility of evidence. Novartis Pharms. Co., ¶ 24. The Master was convinced that the documents had sufficient circumstantial indicia of trustworthiness for admission and permissibly allowed them into evidence. Although the admissibility of Pondera’s documents was a point of contention between the parties, once the Master made his ruling regarding admissibility, both parties moved to admit Pondera [338]*338documents and relied on data in the documents at the hearing. The Master and the Chief Water Judge both recognized that purely objective data was scarce and that the Pondera documents would shed some historic light on the questions before the court. The Master did not abuse his discretion in admitting the documents under the ancient document exception to the hearsay rule. The Chief Water Judge correctly did not disturb the Master’s admission of the evidence.
¶37 2. Did the ChiefWater Judge correctly reject the Master’s findings regarding certain variables used to determine the historical capacity of a flume on the Thomas ditch ?
¶38 Our de novo review begins by determining whether the Water Court correctly applied the clear error standard of review to the Master’s findings. Pursuant to Desaye and Heavirland, even when the Water Court, upon a review of the record, finds the Master’s findings supported by substantial evidence, the Water Court may still overturn those findings. If the Water Court properly rejected the Master’s findings, we will review the Water Court’s replacement findings to determine whether they are clearly erroneous. See Milliken Research Corp., 739 F.2d at 593.
¶39 The ChiefWater Judge, after an exhaustive review of the record supporting the Master’s factual determinations regarding historic flow rates, determined that the Master had misapprehended the effect of the evidence. He reasoned, in part, that the evidence did not support the Master’s calculations of the flumes’ flow rates in 1912 and 1931. He noted that the Master, in calculating the 1912 flume’s capacity, used flume dimensions in Manning’s formula that were not supported by the record or relied on by the experts. Given the importance of calculating the flume’s capacity at a given date to this proceeding, the ChiefWater Judge reasoned that the Master’s reliance on 1920 dimensions to calculate the flow rate of a 1912 flume was problematic. He also concluded that the Master’s use of a slope variable from 1912 in calculating the rebuilt 1931 flume’s capacity was not supported by the record since the 1912 flume had been completely rebuilt and the record contained no reliable slope information for the post reconstruction flume.
¶40 Our review of the record supports the Chief Water Judge’s determination. The evidence shows that the flume’s capacity was expanded between 1912 and 1920. The flume washed out in 1916 and was rebuilt by 1918; and part of the flume was rebuilt in 1920. The flume dimensions from measurements taken at those times confirm that the flume was enlarged with each reconstruction. The Master’s [339]*339Report acknowledged that the flume’s height may have increased by as much as 3 inches between 1912 and 1922. In 1912, the flume was measured to be 24 inches wide and 8 inches deep. In 1918, the flume was measured to be about 24 inches wide and 10 inches deep. In 1920, the flume was measured to be 24 inches wide and 11 inches deep. By using dimensions from the 1920 report, the Master introduced error into his calculation, because the flume in place in 1920 was 3 inches taller than the one in place in 1912. None of the evidence indicated that the flume was 24 inches wide and 11 inches deep, in 1912, or that the slope was .003 in 1920.
¶41 Similarly, the evidence did not support the Master’s use of the slope of the 1912 flume in Manning’s equation to calculate the flow rate of the rebuilt 1931 flume. Expert Casne testified that “slope is a big factor” in determining flow using Manning’s formula. This variable is responsible, in part, for the wide range of results in the experts’ calculations when they used Manning’s formula to calculate the capacity of the 1931 flume. The Master used the 1912 flume’s slope, “absent a more compelling figure,” to calculate the 1931 flume’s capacity using Manning’s formula. The Master and Chief Water Judge agreed, however, that the flume had been expanded and completely rebuilt by 1931. No evidence of the flume’s slope in 1931 existed at all. The Chief Water Judge pointed out, “[t]he slope selected by the Master was taken from the earliest data available and was almost twice as steep as later slope information.” We agree with the Chief Water Judge’s determination that selecting a slope number for use in a slope-dependent formula, where no evidence of slope exists, is clearly erroneous in this context.
¶42 The Dissent, relying on Weinheimer Ranch, ¶ 19, and Amadeo v. Zant, 486 U.S. 214, 223, 108 S.Ct. 1771, 1777 (1988), suggests that the Chief Water Judge incorrectly substituted his own deductions for those of the Master because the evidence supported more than one conclusion. This case is distinguishable from those cases, however. In Weinheimer Ranch, the claimant-objectors challenged the Water Court’s failure to infer that a water claim inaccurately documented the location of the point of diversion and we declined to adopt a rule that such an inference was required. Weinheimer Ranch, ¶¶ 22-24. In Zant, 486 U.S. at 224, 108 S.Ct. at 1777-78, the U.S. Supreme Court held that an appellate court had erred in replacing a district court’s finding of fact with its own because it did not apply the appropriate standard of review and the district court’s inference was not clearly erroneous in light of the evidence in the record. These cases do not prevent us [340]*340from affirming a water judge’s decision when the judge correctly concludes, applying the proper standard of review, that a Master has made an inference that is not warranted by the record.
¶43 Here, the Water Court applied the correct standard of review and determined the evidence did not support the Master’s conclusion. Although it would have been more accurate for the Water Court to reason that the Master’s findings were not supported by substantial evidence, rather than determining the Master had misapprehended the effect of the evidence, this is not fatal to the court’s determination. “Unless justice requires otherwise, no error in admitting or excluding evidence — or any other error by the court or a party — is ground for granting a new trial, for setting aside a verdict, or for vacating, modifying, or otherwise disturbing a judgment or order.” M. R. Civ. P. 61. Our review of the record supports the Chief Water Judge’s determination that the Master committed clear error because the evidence did not support the variables the Master used in calculating flume capacity. It cannot be said that the Chief Water Judge substituted his view of the evidence for the Master’s when the evidence supported more than one conclusion, because the evidence did not support the Master’s findings.
¶44 Having determined that the Chief Water Judge correctly rejected the Master’s findings as clearly erroneous, we now evaluate the Chief Water Judge’s replacement findings for clear error. The record before the Water Court contained all available evidence regarding the water rights at issue. The Chief Water Judge found, and we agree, that this evidence was sufficiently substantial to adequately determine the flume’s capacity.
¶45 The Chief Water Judge’s quantification of the 1912 flume’s capacity was based squarely on testimony by the parties’ experts. Expert witness Anderson testified that Bestor’s calculation of 4.22 cfs was a reasonable estimate for the flume’s capacity. Expert witness Casne stated that the calculation of the flume’s capacity in the 1912 Bestor Report was not based on the actual capacity of the flume, but rather, on water levels in the flume. Accordingly, Casne, using Bestor’s measurements, calculated the flume’s capacity at 4.6 cfs, assuming the flume was 24 inches wide and 8 inches deep. The Chief Water Judge considered that the flume would not carry its maximum capacity at all times and so, concluded that the right was limited to the portion of the 1902 Armedia Clark appropriation that was used in the Thomas ditch, 4.5 cfs.
¶46 The Chief Water Judge’s quantification of the 1931 flume’s [341]*341capacity relied on expert testimony and avoided the problem with using Manning’s formula where no evidence of slope existed. As the Chief Water Judge pointed out, flowrates calculated for the 1931 flume using Manning’s formula resulted in nearly a 300 percent variation. Manning’s formula relies heavily on slope and the slope of the 1931 flume was not known. In contrast, the variation in expert calculations of the flume’s flow rate using the formula for an inlet-controlled structure, where slope is not relevant, varied by less than 2 cfs. The Chief Water Judge concluded that, because of the huge variation, expert testimony regarding the capacity of the 1931 flume using Manning’s formula was not reliable. Instead, he determined to rely on expert numbers for an inlet-controlled structure to find the flume’s capacity was 20 cfs. The use of the inlet-control equation was consistent with the evidence and expert testimony. Expert Casne testified at the hearing that the flume would have to be 300 feet long or less to qualify as an inlet controlled structure. Estimates of the flume’s length around the time it was rebuilt in 1931 were 200-300 feet. Further, even using Manning’s formula, with the more realistic slope of .00157 documented in 1921, Expert Anderson concluded 20 cfs was a reasonable flume capacity. The Chief Water Judge’s quantification of the flume’s capacity was supported by the evidence in the record.
¶47 The Chief Water Judge correctly determined that the Master’s findings as to flume capacity are clearly erroneous. The record demonstrates that the Master erred because the evidence did not support his determinations as to flume capacities. The Chief Water Judge depended on reliable expert testimony to determine the capacities of the two flumes and associated water rights. His findings were supported by substantial evidence, the effect of which he did not misapprehend. We are not left with the impression that he made a mistake.
¶48 3. Did the Chief Water Judge correctly determine that portions of the claimants’ water rights had been abandoned or never perfected?
¶49 The Master determined that the claimants’ predecessors in interest abandoned additional water claimed under the 1895,1902, and 1913 NOAs that they were unable to use because of the flume’s limitations. The Master also determined that if the 1904 and 1906 appropriations ever had been perfected, any interest Skelton may have had in those appropriations had been abandoned. The Chief Water Judge determined that Duncan and Skelton did not have rights to the 1895 claim and effectively adopted the Master’s determination that [342]*342water was abandoned for the claims based on the 1902 and 1913 NOAs. The claimants appeal these determinations on the grounds that the 1895 right was perfected in the ditch that washed out; and that the claimants’ predecessors lacked the requisite intent to abandon their rights, demonstrated by a continuous struggle to repair and expand the flume.
a. The 1895 right.
¶50 “In order to make good his claim to the right as of the date at which it was initiated, the possessor must show some contractual relation between himself and the original appropriator.” See Osnes Livestock Co., 103 Mont. at 290, 62 P.2d at 209 (quotation omitted). When land is sold, water used to irrigate the land and appurtenant thereto passes as part of the transaction unless specifically excluded. Section 85-2-403, MCA; Lensing v. Day & Hansen Sec. Co., 67 Mont. 382, 215 P. 999 (1923). Where a particular quantity of water is specified in a land sale deed, the presumption is that the seller retains any water it owns in excess of that quantity. See Kofoed v. Bray, 69 Mont. 78, 84, 220 P. 532, 534 (1923); Castillo v. Kunnemann, 197 Mont. 190, 197, 642 P.2d 1019, 1024 (1982).
¶51 The Chief Water Judge’s determination regarding the 1895 water right responded to evidence set forth by the claimants that purported to refute Pondera’s contention that the right had not been perfected on lands owned by Skelton. The judge observed that only three of the exhibits the claimants relied upon contained any information regarding the B.P. Clark right, and that two of those contained only a weak connection. Our review of the record confirms this determination. The primary piece of evidence on which the Chief Water Judge relied in addressing Skelton’s claim was the original B.P. Clark NOA. As the Chief Water Judge explained, none of the lands referenced in that NOA are currently owned by Skelton. Without some other evidence to support their claim, Skelton cannot claim a water right based on this NOA. The lands owned by Duncan and subject to the B.P. Clark NOA were conveyed in a chain of title that stemmed to Armedia Clark. The deed conveying those lands conveyed 4.5 cfs from the 1902 appropriation, reserved certain other ditch capacity related to that right and did not reference the 1895 B.P. Clark appropriation. Pursuant to the rule set forth in Kofoed, the Water Court concluded Armedia Clark had reserved any portion of the B.P. Clark right used to irrigate the land in that transfer. The fact that water was used in a ditch is not enough to cast doubt upon the Water Court’s determination as to Skelton’s claim; and Duncan presents no argument to refute the [343]*343Water Court’s reasoning, instead claiming it obtained any portion of the right it abandoned through prescriptive use. Because neither claimant has any right to water based on the 1895 NOA, we need not address their arguments related to abandonment of this right. We conclude that the Water Court correctly addressed this issue.
b. The 1902 and 1913 rights.
¶52 Abandonment of a water right requires both non-use and intent to abandon. 79 Ranch v. Pitsch, 204 Mont. 426, 432, 666 P.2d 215, 218 (1983). Whether a water right has been abandoned is a question of fact that depends on “the evidence of the conduct, acts, and intent of the parties claiming the usufruct of the water.” Heavirland, ¶ 31 (quotation omitted). Intent to abandon water “need not be proved directly, but may be inferred from all the circumstances of the case.” Heavirland, ¶ 31 (quotation omitted).
¶53 The objector bears the initial burden of showing a long period of continuous non-use of the claimed water right. This showing creates a rebuttable presumption of abandonment. 79 Ranch, 204 Mont. at 432-33, 666 P.2d at 218. The burden then shifts to the claimants to produce “[s]pecific evidence explaining or excusing the long period of non-use of the particular water rights on the specific property....” In re Musselshell River Drainage Area, 255 Mont. 43, 51, 840 P.2d 577, 582 (1992). Partial use of a water right does not necessarily show intent to use the entire right or prevent a finding of partial abandonment. Holmstrom Land Co. v. Meagher Co. Newlan Creek Water Dist., 185 Mont. 409, 424, 605 P.2d 1060, 1069 (1979).
¶54 Non-use, “while not conclusive, is evidence of an intention to abandon.” 79 Ranch, 204 Mont. at 432, 666 P.2d at 218. This Court has stated that nine years of non-use is “certainly very potent evidence, if it stood alone, of an intention to abandon.” Smith v. Hope Mining Co., 18 Mont. 432, 438, 45 P. 632, 634 (1896). We have upheld the Chief Water Judge’s finding that an approximately 23-year period of non-use raised a rebuttable presumption of abandonment and that the water rights had been abandoned. In re Adjudication of Water Rights of Clark Fork River, 254 Mont. 11, 17, 833 P.2d 1120, 1124 (1992).
¶55 If an appropriator’s needs exceed the capacity of his means of diversion, moreover, then the capacity of the diversion measures the extent of his water right. Bailey v. Tintinger, 45 Mont. 154, 178, 122 P. 575, 583 (1912).
“[The holder of a water right] is entitled to only the beneficial use of the amount of water called for by his appropriation or decree [344]*344when he has need therefor, and providing his distributing system has a sufficient capacity to carry such an amount of water. When his ditches are incapable of carrying the amount of water decreed to him, his right is measured by the capacity of a system of distribution, regardless of his needs.”
McDonald v. State, 220 Mont. 519, 526, 722 P.2d 598, 602 (1986) (quoting Tucker v. Missoula Light & Water Co., 77 Mont. 91, 101-02, 250 P. 11, 15 (1926) (internal citations omitted)).
¶56 We agree with the Chief Water Judge that the Master did not clearly err in determining that the claimants abandoned any water they claimed to have used that exceeded the flumes’ capacities. Neither claimant had a right to water claimed from the 1895 appropriation. The flume’s capacity limited how much water could be diverted and put to beneficial use from the other appropriations. Thus, the period of non-use for these claims ranges from 18 to 29 years — a period sufficient to raise a presumption of abandonment. Duncan and Skelton failed to present evidence sufficient to rebut this presumption. The flume’s expansion in 1931 does not conclusively demonstrate an intent to “incrementally develop” the water right as the claimants suggest. The efforts involved in maintaining the pre-1931 flume do not demonstrate an intention to maintain rights in excess of the flume’s capacity. All this evidence signals is an attempt to continue use of the amount of water carried by the 1912 flume; and to appropriate an additional quantity of water in 1931, to meet the parties’ predecessors’ needs. The Master and Water Court did not err in finding claimed water in excess of the flumes’ capacities had been abandoned.
¶57 4. Did the Chief Water Judge correctly adopt the Master’s conclusion that the claimants did not acquire any water rights by adverse possession?
¶58 The claimants contend that, if they did lose their interest in the 1895 appropriation, they or their predecessors reacquired ownership of that right through adverse possession. The Chief Water Judge approved and adopted the Master’s finding that the evidence does not support Duncan’s or Skelton’s claims to a prescriptive right to this appropriation.
¶59 In order to acquire a water right by adverse use, the claimant must show that the use has been continuous for the statutory period, exclusive, open, under a claim of right, hostile, and an invasion of another’s rights, which the latter had a chance to prevent. Havre Irrigation Co. v. Majerus, 132 Mont. 410, 415, 318 P.2d 1076, 1078 [345]*345(1957); Smith v. Krutar, 153 Mont. 325, 330, 457 P.2d 459, 462 (1969). The burden of proving an adverse use rests upon the party alleging it. Krutar, 153 Mont. at 329, 457 P.2d at 461.
¶60 Applying these rules, in order to prove adverse use, Duncan or Skelton must show that they or their predecessors used water at a time when the owner of the right to use the water had need of it, used it in such a substantial manner as to notify the owner that it was being deprived of water to which it was entitled; and that during all of that period, the owner could have maintained an action against him for so using the water. Krutar, 153 Mont. at 330, 457 P.2d at 462.
¶61 The Master found that “[a]dverse possession does not apply to Skelton.” The Water Court’s determination was consistent with this conclusion: Duncan’s predecessor presumably used all or close to all of the 4.5 cfs water right carried in the 4.6 cfs capacity 1912 flume before Skelton could attempt to use it, as Duncan’s property was located upstream of Skelton’s. The Master concluded, and the Water Court ultimately agreed, that Skelton is not entitled to claim any water from the 1895 appropriation based on adverse possession.
¶62 The Master similarly concluded that Duncan was not entitled to any portion of the 1895 appropriation based on adverse use. The Master found that, for some of Duncan’s claims, “[t]here is simply insufficient evidence to support adverse possession of any of these shares of the 1895 and 1902 appropriations by [Duncan’s predecessor].” For other shares, the Master found that the testimony presented at the hearing was not sufficient to show that Duncan’s use was exclusive or hostile. The Water Court did not address these findings and apparently adopted them.
¶63 A reviewing Court cannot substitute its own judgment for the trier of fact. Desaye, 250 Mont. at 324, 820 P.2d at 1288. We must give due regard, as did the Chief Water Judge, to the Master’s ability to judge the credibility of the witnesses. Desaye, 250 Mont. at 324, 820 P.2d at 1287-88 (citing M. R. Civ. P 52(a)). Based on a thorough review of the record, we hold that the Master’s findings regarding adverse possession, which the Chief Water Judge adopted, are supported by substantial evidence, the court has not misapprehended the effect of the evidence, and the court did not commit a mistake in interpreting the evidence on this issue. The Water Court correctly concluded that the claimants did not adversely possess rights to the 1895 appropriation.
[346]*346CONCLUSION
¶64 For the reasons set forth above, the Water Court’s opinion is affirmed.
CHIEF JUSTICE McGRATH, JUSTICE McKINNON, and DISTRICT JUDGE TUCKER, sitting for former JUSTICE MORRIS concur.