[¶ 1] The plaintiffs, John and Mary Mark-ley, appeal from the judgment entered in the Superior Court (York County, Fritzsche J.) in favor of the defendants, Scott and Debra Semle. The Markleys contend, inter alia, that the court erred in concluding that they had failed to prove by a preponderance of the evidence the location of the common boundary between their property and the Semles’ property. Finding no error, we affirm.
I.
[¶2] The Markleys and the Semles own adjoining parcels of land in Hollis. The Markleys initiated this action pursuant to the Declaratory Judgments Act, 14 M.R.S.A §§ 5951-5963 (1980), seeking, inter alia, a judgment declaring the location of the common boundary between their property and the Semles’ property. The Semles counterclaimed, requesting, inter alia, that the court “[djetermine the true boundary line between the property of the Defendants and Plaintiffs.”
[¶ 3] During the three-day bench trial, the parties introduced over 120 deeds into evidence. The Markleys presented the expert testimony of Walter Dunlap, a professional land surveyor, whose testimony required almost two days. Dunlap opined that he had identified the rock wall that is the common boundary between the Markleys’ and the Semles’ properties. The Semles did not present a competing expert, choosing instead to cross-examine Dunlap extensively. At the conclusion of the trial, the court took the matter under advisement. The court subsequently concluded that it “is unable to adequately loeate the common boundary between the land of the plaintiffs and the land of the defendants,” and it entered a judgment in favor of the Semles on the Markleys’ complaint. Noting that the Semles did not pursue their counterclaim at trial, the court entered a judgment in favor of the Markleys on that counterclaim. The Markleys and the Semles both appealed from the court’s judgment. The Semles, however, have not pursued their cross appeal.
II.
[¶ 4] In a boundary dispute, “what the boundaries are is a question of law, but the location of the boundaries on the face of the earth is a question of fact.” White v. Zela, 1997 ME 8, ¶ 3, 687 A.2d 645, 646. The Markleys and the Semles do not dispute what constitutes their common boundary; rather, they dispute the location of that boundary on the face of the earth.1 The trial court concluded, without setting forth its factual findings, that the evidence failed to establish the location of the common boundary. Pursuant to M.R.Civ.P. 52(a), “[i]n all actions tried upon the facts without a jury ... the Superior Court justice ... shall, upon the request of a party made as a motion within 5 days after notice of the decision, ... find the facts specially and state separately its conclusions of law thereon.” If a party does not move for specific findings of fact, we assume “that the trial court found all of the facts necessary to support its decision.” Mariello v. Giguere, 667 A.2d 588, 591 (Me.1995). Generally, we review the trial court’s implicit *947factual findings for clear error. See Board of Overseers of the Bar v. Sylvester, 650 A.2d 702, 704 (Me.1994). However, when factual findings, albeit implied, are adverse to the party with the burden of proof, we will reverse them only when the record compels a contrary conclusion. See Falvo v. Pejepscot Indus. Park, Inc., 1997 ME 66, ¶ 10, 691 A.2d 1240, 1243.
[¶5] We have stated that “[a]n action for declaratory judgment is an appropriate vehicle for establishing rights in real property.” Hodgdon v. Campbell, 411 A.2d 667, 669 (Me.1980). In a declaratory judgment action, “the allocation of the burden of proof ... must be determined by reference to the substantive gravamen of the complaint. The party who asserts the affirmative of the controlling issues in the case, whether or not he is the nominal plaintiff in the action, bears the risk of non-persuasion.” Id. at 670-71. “In an action to establish a boundary line, the party asserting a specific location of such a line bears the burden of presenting credible evidence to establish that location.” Ollison v. Village of Climax Springs, 916 S.W.2d 198, 203 (Mo.1996) (en banc) (citations and quotations omitted); cf. Hodgdon, 411 A.2d at 671 (stating that in a quiet title action the party seeking a declaratory judgment bears the burden of proving “better title than that of the defendant”); Chappell v. Donnelly, 113 N.C.App. 626, 439 S.E.2d 802, 805 (1994) (stating that in a quiet title action the plaintiff bears the burden of establishing the on-the-ground location of the boundary lines that he or she asserts).
[¶ 6] In this case, the Markleys bore the burden of proving the location of the boundary line that they urged the court to adopt. The court concluded that they had failed to satisfy their evidentiary burden because “there are too many ambiguities and uncertainties in the various deeds to allow me to have adequate confidence in the correctness of any common boundary I might find.” The court explained generally that
[t]he difficulty in establishing a precise boundary comes from many factors. The land is deep in the woods and there has been no compelling need for precision in the past. The deed descriptions have changed markedly over the years. Former monuments of trees, stakes and the like disappear over time. The Hollis-Wa-terboro town line, which formed the earliest westerly boundary of the plaintiffs’ parcel, is not currently known. Former deeds and surveys may have contained several errors.
The Markleys did not request findings of fact pursuant to Rule 52(a) before initiating this appeal. We therefore must assume that the court found those facts necessary to support its conclusion that the location of the boundary line was not determinable from the evidence presented. See Mariello, 667 A.2d at 591. We may vacate the court’s judgment only if we find that the evidence compelled a contrary conclusion. See Falvo, 1997 ME 66, f 10, 691 A.2d at 1243.
III.
[¶ 7] We conclude that the evidence presented in this case, although voluminous,2 did not compel the trial court to conclude that the common boundary between the Markleys’ property and the Semles’ property may be located on the face of the earth.
A Inconsistent Deed Descriptions
[¶ 8] The Markleys claim title to their property pursuant to a deed from Richard Hobson. That deed describes the Markleys’ property as a three-sided parcel:
[bjeginning at two crotehed white oak trees near land formerly owned by Nahum *948Thompson, thence. running westerly 40 rods to land of Heirs of Benjamin Day, thence Southerly by said Day land about 70 rods to land of Heirs of Joseph L. Benson, thence by said Benson heirs land to the place of beginning.
Dunlap testified that the Hobson-Markley deed is “deficient ... in terms of its specificity. It does give us clues as to where to place it ... and given ancillary information and background information for placing it, it is possible to place it, but on its face, it’s very difficult to tell the dimensions or the location ... of the parcel.”
[¶ 9] The description set forth in the Hobson-Markley deed is substantially similar to the descriptions set forth in several other deeds in the property’s chain of title: the 1913 deed from William Hobson to Ralph Hobson; the 1897 deed from Violet Benson to William Hobson; and the 1893 deed from John Linscott to Violet Benson. However, the 1827 deed from James Gillpatriek to Joseph Linscott contains a different description, one that describes a six-sided parcel:
begining [sic] at an oak at Daniel Johnsons [sic] corner, thence riming [sic] South forty five and one half degrees West twenty four rods to a yellow oak, thence south eight degrees west, fifty two rods to a stake, thence Northeast forty two roads to an oak, thence southeast five rods to a white oak, thence Northeast fifteen rods to an oak stake, thence to the first mentioned bounds_■
In contrast, the 1806 deed to James Gillpa-trick from the Town Proprietors, the original deed in the Markleys’ chain of title, describes a four-sided parcel:
[b]eginning at the W. corner of lands of Joseph Linscott’s land on Waterborough line running N. lijé W. 50 rods to lands of James Johnson’s [sic], thence N.E. about 30 rods to lands, of Carrol Tarbox, thence S.E. about 40 rods to the N. corner of said Linscott’s land, thence S.W. 45 rods to Waterborough line....
Dunlap testified that he did not believe that the Proprietors-Gillpatrick deed described the same tract of land as did the Gillpatrick-Linscott deed.
[¶ 10] In his “Surveyor’s Report,” Dunlap described the deeds in the Markleys’ chain of title as containing “substantial and unexplained variation in the language.” According to Dunlap’s report:
[i]t has been said that the early settlers marked off their claims and received subsequent deeds from the Proprietors, thus lending a great amount of weight to the position of stone walls which may well have pre-dated the deed language and by which the settlers actually expected to claim. It is helpful in this case to note the pattern of walls which conform in many respects to the deed from Gilpatrick [sic], upon which this survey is based. The deed from the Proprietors seems to indicate a rectangle of larger acreage and the loss of this pattern cannot be explained.
Dunlap noted that he had relied upon stone walls in his analysis as “corroborative evidence of where the boundaries lie.” He testified, however, that stone walls had purposes other than demarcating property boundaries; for instance, they sometimes were used by a single owner to delimit his or her holdings for certain uses. None of the deeds in the Markleys’ chain of title contain any references to stone walls.
[¶ 11] Dunlap testified that he was unable to locate physically any of the monuments referenced by the Gillpatrick-Linseott deed. Each reference in that deed is to a monument that no longer exists; the deed makes no reference to abutters. He nevertheless believed that he had discerned the location of the former oak at Daniel Johnston’s corner. During cross-examination, however, Dunlap acknowledged that as of June 30, 1827, the date of the transfer from Gillpatriek to Lin-seott, Daniel Johnston did not own the property at which Dunlap had pinpointed the location of the former oak at Daniel Johnston’s comer. That property was transferred to Daniel Johnston by James Johnston pursuant to a December 25, 1827, deed. Dunlap admitted that he had not researched whether Daniel Johnston had owned any other property in Hollis as of June 30, 1827. The Semles then introduced into evidence a May 9, 1816, deed from Nathaniel Smith to Samuel Hodsdon and to Daniel Johnston, *949demonstrating that as of June 30,1827, Daniel Johnston did own other property in Hollis. In response to this testimony, the court observed that “it seems that we need to know where the Nathaniel Smith to Hodsdon and Johnston piece is, whether that’s in this general area or whether it’s in a — in a whole different part of town, to see whether it creates an ambiguity or not.”
[¶ 12] On redirect, Dunlap opined that the Smith-Hodsdon/Johnston deed did not appear to describe any land in the area of the Markleys’ and the Semles’ properties because it mentioned a mill privilege and he was not aware of any mills in that area. He also explained that “[i]t may very well be that Dan Johnson [sic] was a tenant of that land, and since the last two names are both Johnson [sic], there may be in a [sic] familial relationship. Dan Johnson [sic] may have occupied that land and everyone in the neighborhood may have known that that was where his comer was.” Dunlap indicated that he had seen similar phenomena in his experience as a surveyor, “often not, but it happens....”
[¶13] The Markleys contend that these inconsistencies in the deed descriptions are not fatal to Dunlap’s analysis. They note', for example, that Dunlap attempted to reconcile the Hobson-Markley deed’s description of a three-sided parcel with the Gillpa-trick-Linscott deed’s description of a six-sided parcel. Dunlap testified that the call, “thence Southerly by said Day land about 70 rods to land of Heirs of Joseph L. Benson,” in the Hobson-Markley deed equates to the calls, “thence runing [sic] South forty five and one half degrees West twenty four rods to a yellow oak, thence south eight degrees west, fifty two rods to a stake” [for a total of 76 rods], in the Gillpatrick-Linscott deed. Dunlap’s hypotheses may have been sufficiently credible to permit a finding by the trial court that despite the deeds’ inconsistencies, Dunlap’s analysis successfully identified the location of the common boundary. In the absence of a request for findings of fact, however, we must assume that the court made the findings necessary to support the conclusion that Dunlap’s analysis was not persuasive. We cannot conclude that the court was compelled to accept Dunlap s explanations of the variations.
B. The Unknown Location of the Town Line
[¶14] Dunlap testified that he did not locate the town line between Hollis and Wa-terboro. He acknowledged that the 1806 Town Proprietors’ plan contained “one predominant feature” from a surveying standpoint: the properties shared a common westerly boundary, the town line between Waterboro and what was then Phillipsburg. Dunlap stated that the town line is thus an important monument. Dunlap further testified that the Proprietors’ 1806 plan was “very helpful” in determining the locations of the various land transfers in the Mark-leys’ and in the Semles’ chains of title because it provided “some sort of a geometric pattern, to try to match the deeds that were granted to the various landowners.” The original deed in the Markleys’ chain of title, the 1806 deed from the Proprietors to Gill-patrick, uses the Hollis-Waterboro town line, denoted “N. 11% W.,” as a controlling monument. Dunlap agreed with the Semles’ assertions that “the limits of the Markley holdings are determined based on that original deed” from the Proprietors, and that “[i]n order to determine where the original parcel was, you needed to know where the town line is.”
[¶ 15] In his Surveyor’s Report, Dunlap addressed the location of the town line as follows:
[t]he location of the town line between Hollis and Waterborough continues to be a source of conflict. Clearly the Proprietors intended all the grantees by the Knight plan of 1806 to run along the town line, where it was specifically granted, on a bearing of N 11% W.
Pursuant to an act of the Legislature in 1865, the line common to Hollis and Water-borough was run and marked by stones at crossings of roads. The bearing at that time was specified as being ‘South about seven degrees East’. There is no way to relate the new and old locations without an on ground survey of large proportions....
*950At trial, Dunlap examined a Hollis tax map and noted that although it depicted the town line, it did not depict Day’s Lane, an existing road in the area of the Markleys’ property. Dunlap attributed the absence of any reference to Day’s Lane on the tax map to “a misapprehension on someone’s part” as to whether Day’s Lane constituted the town line. Dunlap then examined Plaintiffs’ Exhibit 3, a map of Hollis, and testified that on that map Day’s Lane was depicted distinctly from the town line.
[¶ 16] The court engaged in its own questioning of Dunlap regarding the location of the town line. Through this questioning, the court clarified that Dunlap thought that it lay somewhere to the west of Day’s Lane (although he did not know where), and that moving the hypothetical location of the town line could have the effect of shifting Dunlap’s approximations of the parcels’ locations.
[¶ 17] The Markleys contend that Dunlap did not have to locate the town line definitively “because the town line was not a common boundary between land of Plaintiffs and land of Defendants.” Our inquiry, however, is not whether the trial court could have accepted Dunlap’s analysis notwithstanding his failure to locate the town line. Rather, we must determine whether the court was compelled to accept Dunlap’s testimony about the location of the common boundary. Again, we conclude that it was not.
IV.
[¶ 18] Lastly, we reject the Markleys’ contention that the court was obligated to declare a location of the common boundary. The Markleys posit that “[i]f the court did not agree with the Plaintiffs’ and Defendants’ assertions as to the location of the boundaries, it had more than ample evidence before it with which to determine the boundaries.” We disagree. In Boynton v. Adams, 331 A.2d 370 (Me.1975), the plaintiff sought to establish title to, and to recover possession of, certain real property. The defendant was an abutting property owner. See id. at 372. The parties disputed the location of their common boundary on the face of the earth. See id. at 374. ■ The trial court entered a judgment for the defendant, finding “without extended opinion” that the plaintiff had failed to satisfy his burden of proof. Id. at 372. We denied the plaintiffs appeal, concluding that our review of the evidence did not demonstrate that “the presiding Justice was clearly wrong in concluding that the plaintiff ‘failed to sustain his burden of proof....’” Id. at 375.
[¶ 19] The Súpreme Court of Connecticut reached a similar result in Steinman v. Maier, 179 Conn. 574, 427 A.2d 828 (1980) (per curiam). In Steinman, the plaintiff brought an action to establish the location of the common boundary between his property and that of an adjoining landowner. See id. at 829. The trial court entered a judgment for the defendants, finding that deficiencies in the plaintiff’s evidence prevented the plaintiff from satisfying his burden of proof. See id. at 829-30. The Supreme Court affirmed, reasoning that “[a]lthough the plaintiff’s evidence was not directly contradicted by other evidence, the trial court was-not bound-to accept it at face value. The sifting and weighing of-evidence is peculiarly the function of the trier.” Id. at 830 (citation omitted).3
[¶ 20] In this ease, the trial court listened attentively to the parties’ presentation of evidence. The court sought clarification from Dunlap at several different points during his testimony by actively questioning him. During the parties’ closing arguments, the court asked each party to respond to specific theories and interpretations of the evidence that had been advanced. After taking the matter under advisement, the court concluded that the Markleys had failed to present sufficient evidence to support their assertion that a rock wall identified by Dunlap constituted the common boundary line. This ruling does *951not reflect a failure on the part of the court. It only reflects a failure of the plaintiffs’ proof.4
The entry is:
Judgment affirmed.
. The Markleys acknowledged in their appellate brief that “[t]he nature of the dispute between Plaintiffs and Defendants is such that there is no dispute as to what the boundaries of their respective properties are, rather the dispute is based upon where those boundaries are located on the face of the earth.”
. For instance, Plaintiffs' Exhibit 20 consisted of more than 120 deeds from the Markleys’ chain of title, the Semles’ chain of title, and their abutting landowners' chains of title. This multiplicity of deeds was necessitated by the fact that die Mark-leys’ deed, the Semles’ deed, and many of their adjoining landowners’ deeds are "abutter’s deeds” — deeds that do not set forth precise distances or bearings but rather describe the limits of property in terms of the adjoining parcels of land. To locate property described in an abut-ter’s deed, one must use extrinsic evidence (e.g., adjoining landowners’ deeds) to determine the locations of the adjoining parcels. Dunlap acknowledged that to formulate his opinion as to the location of the boundaiy line, he had to review "[a] substantial number of deeds for the subject parcel and for all the adjoining parcels within a very large radius.”
. Cf. Cutts v. Casey, 278 N.C. 390, 180 S.E.2d 297, 307-08 (1971) (observing that in a trespass action in which both parties claim title to the land involved, one parly's failure to carry his or her burden of proof does not relieve the other party of the burden to show title in himself or in herself, and noting that "[t]here are cases involving a disputed title to land in which neither party can carry the burden of proof'); Beebe v. Reichert, 172 Neb. 172, 108 N.W.2d 804 (1961) (affirming trial court's finding that plaintiff's proof failed to establish location of boundary line).
. Because we hold that the court was not compelled to find that the evidence presented at trial established the location of the common hotmd-ary, we need not reach the Markleys' other assignments of error,