Donald H. Albrecht and the Rivermeadows Corporation (Albrecht) brought this action against the United States to quiet title to 265.01 acres of land along the Snake River in Teton County, Wyoming. Albrecht claimed title through mesne conveyances of five patents issued by the United States between 1913 and 1930 which described the land according to a 1902 survey. The government claimed that the survey's location of the Snake River was misplaced at certain points causing an underestimation of the surveyed acreage. Consequently, the government argued, title to lands omitted from the survey remained in the United States. The District Court for the District of Wyoming, finding the survey to be in gross error, entered judgment for the government. Albrecht v. United States, 529 F. Supp. 135 (D.Wyo.1981). We conclude the district court's determination of gross error was imprudently based; therefore, we reverse and remand for further determinations.

The Snake River is a braided channel that has higher flows in the summer caused by snowmelt and lower flows in the winter. The high flows erode material in the riverbed, causing shifts in the channels and gravel bars. In the past eighty years, due to manmade regulations of the Snake River,1 there has been a decrease in peak or flood flows and shorter periods of low flows, stabilizing the area along the Albrecht property.

The lands at issue were the subject of a survey conducted by Newell J. Burnham in 1902 and confirmed in 1903 by Special Agent William Lightfoot of the Surveyor General's Office. The survey included the establishment of "meander corners," monuments set at the intersection of section or township lines with the bank of the Snake River and "meander lines," lines which connect the meander corners and approximate the river's boundaries. The patents under which Albrecht claims title stated the conveyances were "according to the Official Plat of the Survey of the said Land." The survey indicated that the land covered by each of the patents bordered the Snake River.

In a public grant nothing passes by implication, and unless the grant is explicit with regard to the property conveyed, a construction will be adopted which favors the sovereign rather than the grantee. Walton v. United States, 415 F.2d 121 (10th Cir.1969). If the patents referred to an official plat, showing meander lines along or near a body of water, the plat is to be treated as part of the conveyance, and the patentee's title extends to the actual water line. Jeems Bayou Fishing & Hunting Club v. United States, 260 U.S. 561, 43 S. Ct. 205, 67 L. Ed. 402 (1923); Smith v. United States, 593 F.2d 982 (10th Cir.1979). However, this rule does not apply where there was a significant departure from the proper location of the meander line sufficient to show gross mistake or fraud; in that case, the meander line becomes a fixed boundary.2 Snake River Ranch v. United States, 542 F.2d 555 (10th Cir.1976); see, e.g., Walton v. United States, 415 F.2d 121 (10th Cir.1969). The government claims that in surveying the Snake River, Burnham mistakenly erected one of the meander corners along the bank of Fish Creek,3 which runs parallel to Snake River, thereby omitting a peninsula which separated the two waters. According to the government, the peninsula has since eroded into several islands. Albrecht's theory is, however, that Burnham correctly followed the Snake River, and that the lands alleged to be part of the former peninsula separating Fish Creek and Snake River have been created since 1902 and fall within the doctrine of accretion.4

The district court concluded that the survey was in gross error and that Albrecht owned only those areas west of the meander line. Albrecht argues that the district court's conclusion is not supported by clear and convincing evidence and, therefore, is clearly erroneous. The district court made the following findings of fact:

15. In his survey in this area, Burnham mistakenly meandered the bank of Fish Creek rather than the right, or west, bank of the Snake River and the disputed lands in question in this case were omitted from the survey. Burnham's field notes reflect his observation that he could not sight across the river because of "willows on sand bars."

16. The evidence established the existence of a peninsula in this area bordered on one side by Fish Creek and on the other side by the Snake River. The peninsula appears to have run the length of sections 3 and 4. The mouth of Fish Creek has changed location since the 1902 survey. The peninsula has been divided into several islands by the erosive force of the river.

17. The unsurveyed land between the meander line and the bank of the river is covered with vegetation, brush and trees. Many of the trees are cottonwoods, although other varieties are also found in the area. The ages of the trees vary considerably, with a large number of trees predating the Burnham survey.

Albrecht v. United States, 529 F. Supp. at 137.

These findings must not be set aside unless clearly erroneous. Fed.R.Civ.P. 52(a); Walton, 415 F.2d at 123-24; see also United States v. Ruby Co., 588 F.2d 697, 701 (9th Cir.1978), cert denied, 442 U.S. 917, 99 S. Ct. 2838, 61 L. Ed. 2d 284 (1979). Albrecht introduced testimony and other evidence at trial which demonstrated a dramatic difference between the soil and vegetation now existing east of the meander line and that found on the west of the meander line. Albrecht urges that this evidence, the survey itself, the circumstances under which the survey was completed, and its subsequent inspection and approval by Lightfoot compels a finding that Burnham constructed a true meander line along the Snake River. On the other hand, the government's case, which includes evidence of trees east of the meander line which predate the survey, a map which accompanied 1902 water permit applications, grazing leases, Burnham's field notes, and eyewitness descriptions of the area constitutes forceful evidence of the existence of a peninsula or islands at the time of the survey.

It is within the trial court's discretion to weigh the evidence and to determine the credibility of witnesses. Unless the findings are without support in the record, they may not be overturned and replaced by the judgment of this court. Lemaire v. United States, 826 F.2d 949 (10th Cir.1987). After a thorough examination of the record, we find it contains sufficient evidence to support the trial court's finding that Burnham's meander line was in error.

Albrecht brings to our attention a recent decision, United States v. Albrecht, No. C79-113-K (D.Wyo. April 25, 1985), which concerned lands adjacent to appellants' land, including part of the same peninsula. In that case, it was found that the government's peninsula theory was "speculative at best." We are limited to reviewing the record before us and examining the trial court's findings of fact under the clearly erroneous standard. In the present case, the trial court's finding of the existence of a peninsula survives that scrutiny quite well.

Because some discrepancies in surveys are to be expected,5 courts require that the error be sufficiently obvious, or "gross" to have charged the subsequent purchasers with notice of fraud or mistake and thereby make them aware that the meander line, not the water line, was intended as the boundary.6 Snake River Ranch v. United States, 395 F. Supp. 886 (D.Wyo.1975), aff'd, 542 F.2d 555 (10th Cir.1976). In deciding a question of gross error, a court must determine the amount and proportion of acreage between the meander line and the shore existing at the time of the survey. This determination of the proportion, or ratio, of the patented acreage and the omitted acreage is referred to as "the substantial omission test." United States v. Walton, 266 F. Supp. 257 (D.Wyo.1967), aff'd, 415 F.2d 121 (1969); Smith v. United States, 593 F.2d 982 (10th Cir.1979). This comparison is balanced with the circumstances surrounding the original survey and the type and comparative value of the land at that time. United States v. Lane, 260 U.S. 662, 43 S. Ct. 236, 67 L. Ed. 448 (1923); see also Zager, 338 F. Supp. at 989-90.

Although gross error must necessarily be determined on a case-by-case basis, several decisions provide guidance. In Lane, the Court held that the value of the omitted land and the difficulties surrounding the survey mitigated against any greater accuracy than that resulting in a 2.8 to 1 ratio of patented to omitted land. The Zager court found that where the land around a Wisconsin lake was remote, wild, and wet, and the ratio of surveyed land to omitted land was only 2 to 1, there was no preponderance of gross error. In fact, the Zager court noted that as of 1960, a 42% understatement (roughly 1.4 to 1) was the smallest omission to have prompted a finding of gross error.7 Similarly, the court in Wackerli v. Morton, 390 F. Supp. 962 (D.Idaho 1975), held that based on the character, terrain, and minimal value of the land at the time of the survey, a 2.5 to 1 ratio of surveyed lands to omitted lands did not warrant a conclusion of gross error. Conversely, a 1 to 3 ratio of surveyed lands to omitted lands has resulted in a finding of gross error. Walton, 266 F. Supp. at 264-65.

In the case at hand, the trial court found the ratio of patented land to omitted land to be 615.99 to 265.01, or approximately 2.3 to 1. The trial court then looked to the patented and omitted land pertaining to Albrecht's northern property and produced a ratio of 261.728 to 251. On the basis of this calculation, the court held the patents did not convey title to the water's edge. Albrecht, 529 F. Supp. at 138.

Albrecht points out that the trial court's approach is inconsistent with another recent decision of the Wyoming district court:Fraud or gross error must be shown with respect to each segment of a meander line that is asserted to be a boundary of an upland lot; fraud or gross error cannot be imputed to particular segments of a meander line survey from fraud or gross error in other places along the meander line of the same survey or other surveys.

Snake River Ranch, 395 F. Supp. at 900. We agree that the two results conflict. The district court apparently found gross error in the survey near the northern property and then imputed that error to the remainder of the meander line, following exactly the approach it condemned six years earlier in Snake River Ranch.

While we find the district court's analysis unacceptable, the traditional approach of comparing the aggregate patented and omitted acreage involved in the particular action9 is equally unacceptable. The traditional approach invites gamesmanship in the selection of areas to be joined for adjudication. For instance, by joining a parcel which includes a section of a meander line which suffers from fraud or gross error with other parcels which benefit from an accurate meander line, so as to make the aggregate omitted land comparatively small, a landowner could "cure" the "erroneous" parcel. The more logical approach is to examine and compare each patent separately. This will ensure that error (or accuracy) is not imputed to other sections along the meander line.

Albrecht further argues that some of the parcels presently lie within 900 feet of the Burnham meander line, a limit which has been held not to show gross error "in all events." Snake River Ranch, 395 F. Supp. at 901. The width of the omitted lands is relevant but only insofar as it relates to the area of that parcel and the character of the land. Surveyors establish meander lines only to calculate acreage, not to establish boundaries. United States v. Pappas, 814 F.2d 1342 (9th Cir.1987). Land comes in varying shapes; the area, not just one dimension, should be dispositive of the gross error issue.

Albrecht also objects to the district court's use of current measurements in his comparison of patented to omitted lands. We agree with Albrecht that it is the area of the respective parcels as of the time of the survey which controls the gross error issue.10 Snake River Ranch, 542 F.2d at 558. Accordingly, findings as to the size of each patent and the associated omitted land as of the date of the Burnham survey are necessary for proper application of the substantial omission test.11

Moreover, each ratio of patented land to omitted land must be viewed in light of the circumstances existing at the time of the survey. One must consider the date of the survey, the remoteness and value of the land, and the difficulty surrounding the work to determine whether the accuracy obtained was proportionate to the effort and cost required to obtain it. Lane, 260 U.S. at 665, 43 S.Ct. at 237; Snake River Ranch, 395 F. Supp. at 900-01; see also Zager, 338 F. Supp. 989-90.

The judgment of the district court is REVERSED and the case is REMANDED. On remand, the district court shall make individual comparisons of the patented and omitted land pertaining to each of the five patents as of the date of the survey. The court shall then decide whether the degree of accuracy obtained was reasonable in light of the circumstances surrounding the survey.