United States v. 269 Acres Located in Beaufort

PUBLISHED

UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT

No. 19-2212

UNITED STATES OF AMERICA,

Plaintiff – Appellant,

v.

269 ACRES, MORE OR LESS, LOCATED IN BEAUFORT COUNTY, STATE OF SOUTH CAROLINA; HAROLD E. TRASK, JR.; JOHN DONALD TRASK; JAMES HEIDE TRASK; MARGARET SCHEPER TRASK; WILLIAM D. TRASK, JR., as Trustee of the William D. Trask, Jr. Trust under Article Nine of the William D. Trask Revocable Trust dated February 17, 2000; SARAH T. BURRUS, as Trustee of the Sarah T. Burrus Trust under Article Nine of the William D. Trask Revocable Trust dated February 17, 2000; ROBERT EDWARD L. HOLT, III, as Trustee of the Kitty Trask Holt Family Trust and as Trustee of the Kitty Trask Holt Family Trust−Marital both under the Last Will and Testament of Kitty Trask Holt dated June 8, 2004; ROBERT EDWARD L. HOLT, III, as Trustee of the Kitty Trask Holt Family Trust and as Trustee of the Kitty Trask Holt Family Trust (EI#26−6501159) under the Last Will and Testament of Kitty Trask Holt dated June 8, 2004; SOUTH CAROLINA DEPARTMENT OF REVENUE,

Defendants – Appellees, and BEAUFORT COUNTY ASSESSOR; INTERNAL REVENUE SERVICE,

Defendants.

No. 20-1226

UNITED STATES OF AMERICA,

Plaintiff – Appellant,

v.

269 ACRES, MORE OR LESS, LOCATED IN BEAUFORT COUNTY, STATE OF SOUTH CAROLINA; HAROLD E. TRASK, JR.; JOHN DONALD TRASK; JAMES HEIDE TRASK; MARGARET SCHEPER TRASK; WILLIAM D. TRASK, JR., as Trustee of the William D. Trask, Jr. Trust under Article Nine of the William D. Trask Revocable Trust dated February 17, 2000; SARAH T. BURRUS, as Trustee of the Sarah T. Burrus Trust under Article Nine of the William D. Trask Revocable Trust dated February 17, 2000; ROBERT EDWARD L. HOLT, III, as Trustee of the Kitty Trask Holt Family Trust and as Trustee of the Kitty Trask Holt Family Trust−Marital both under the Last Will and Testament of Kitty Trask Holt dated June 8, 2004; ROBERT EDWARD L. HOLT, III, as Trustee of the Kitty Trask Holt Family Trust and as Trustee of the Kitty Trask Holt Family Trust (EI#26−6501159) under the Last Will and Testament of Kitty Trask Holt dated June 8, 2004; SOUTH CAROLINA DEPARTMENT OF REVENUE,

Defendants – Appellees, and BEAUFORT COUNTY ASSESSOR; INTERNAL REVENUE SERVICE,

Defendants.

No. 20-1281

UNITED STATES OF AMERICA,

Plaintiff – Appellee,

v.

269 ACRES, MORE OR LESS, LOCATED IN BEAUFORT COUNTY, STATE OF SOUTH CAROLINA; HAROLD E. TRASK, JR.; JOHN DONALD TRASK; JAMES HEIDE TRASK; MARGARET SCHEPER TRASK; WILLIAM D. TRASK, JR., as Trustee of the William D. Trask, Jr. Trust under Article Nine of the William D. Trask Revocable Trust dated February 17, 2000; SARAH T. BURRUS,

as Trustee of the Sarah T. Burrus Trust under Article Nine of the William D. Trask Revocable Trust dated February 17, 2000; ROBERT EDWARD L. HOLT, III, as Trustee of the Kitty Trask Holt Family Trust and as Trustee of the Kitty Trask Holt Family Trust−Marital both under the Last Will and Testament of Kitty Trask Holt dated June 8, 2004; ROBERT EDWARD L. HOLT, III, as Trustee of the Kitty Trask Holt Family Trust and as Trustee of the Kitty Trask Holt Family Trust (EI#26−6501159) under the Last Will and Testament of Kitty Trask Holt dated June 8, 2004,

Defendants – Appellants, and

BEAUFORT COUNTY ASSESSOR; SOUTH CAROLINA DEPARTMENT OF REVENUE; INTERNAL REVENUE SERVICE,

Defendants.

Appeals from the United States District Court for the District of South Carolina, at Beaufort. Richard M. Gergel, District Judge. (9:16-cv-02550-RMG)

Argued: December 10, 2020 Decided: April 16, 2021 Before KING, WYNN, and RICHARDSON, Circuit Judges.

Affirmed in part and reversed in part by published opinion. Judge Richardson wrote the opinion, in which Judge King and Judge Wynn joined.

ARGUED: Jeffrey Steven Beelaert, UNITED STATES DEPARTMENT OF JUSTICE, Washington, D.C., for Appellant/Cross-Appellee. Paul Allen Dominick, NEXSEN PRUET, LLC, Charleston, South Carolina, for Appellees/Cross-Appellants. ON BRIEF: Jeffrey Bossert Clark, Assistant Attorney General, Eric Grant, Deputy Assistant Attorney General, Daniel W. Kastner, Richard McMurtray, Environment and Natural Resources Division, UNITED STATES DEPARTMENT OF JUSTICE, Washington, D.C.; Peter M. McCoy, Jr., United States Attorney, Columbia, South Carolina, Lee E. Berlinsky, Assistant United States Attorney, OFFICE OF THE UNITED STATES ATTORNEY, Charleston, South Carolina, for Appellant/Cross-Appellee. Alexandra H. Austin, NEXSEN PRUET, LLC, Charleston, South Carolina, for Appellees/Cross-Appellants.

RICHARDSON, Circuit Judge:

Deference to a district court’s factual findings is vital to our judicial system. Our deference respects the superior competence of local fact finders in reviewing the evidence while also promoting judicial economy. This appeal turns on that deference.

After a trial before a three-member land commission, a district court awarded compensation to Landowners after the government took an easement on their land, which is near a U.S. Marine Corps airbase in Beaufort, South Carolina. See U.S. Const. amend. V. The district court awarded the Landowners $4.4 million, apportioned attorney’s fees and litigation costs, and split the cost of the commission.

The government challenges the amount the district court awarded in compensation and the award of attorney’s fees. The Landowners cross-appealed to dispute the district court’s apportionment of the attorney’s fees and the commission costs.

Although we might have decided this case differently in the first instance, the deference we afford to the district court compels us to affirm its award of just compensation. See Anderson v. Bessemer City, 470 U.S. 564, 573–74 (1985) (a reviewing court may not “reverse the finding of the trier of fact simply because it is convinced that it would have decided the case differently”). Similarly, we cannot say the court abused its discretion in splitting the commission costs. But we do find that the district court legally erred in awarding attorney’s fees to the Landowners as the “prevailing party” in this litigation. See 28 U.S.C. § 2412(d). So we affirm in part and reverse in part. *5 I. Background

In 2016, the federal government filed an action to impose a permanent easement on 269.22 acres of land in Beaufort, South Carolina. See 10 U.S.C. § 2663; 40 U.S.C. §§ 3113–3114. The land is near a busy Marine Corps Air Station where the government trains fighter pilots. The easement restricts development in the military jets’ flight paths to ensure pilot and civilian safety, as takeoff and landing is the most dangerous part of training. Along with the action imposing the easement, the government deposited $1,091,000 with the district court as an estimate of the condemned property’s value. The court appointed a three-member land commission that conducted a bench trial to determine how much compensation the government owed the Landowners. See Fed. R. Civ. P. 71.1(h)(2). After the commission recommended just compensation of $5,311,313, the district court reviewed the evidence de novo and awarded the Landowners $4,441,410. See United States v. 269 Acres, No. 9:16-2550-RMG, 2019 WL 1450578, at *6 (D.S.C. Apr. 2, 2019). The district court also awarded the eligible Landowner attorney’s fees and costs under the Equal Access to Justice Act and split the cost of the commission equally between the government and the Landowners. United States v. 269 Acres, No. 9:16-2550-RMG, 2020 WL 219792, at *11 (D.S.C. Jan. 15, 2020). The government appealed the amount of compensation and the award of attorney’s fees, and the Landowners cross appealed based on the court’s apportionment of attorney’s fees and commission costs.

A. The Property and the easement

The Property, which the Landowners’ family has owned in fee simple since the midfifties, consists of two parcels. One parcel is 446.33 acres zoned for industrial use *6 (“Industrial Parcel”), of which the restrictive easement covers 179 acres. The second parcel is 90.22 acres zoned for residential use of up to one house per three acres (“Residential Parcel”), which is entirely covered by the restrictive easement. 1

The Property is undeveloped land that has been used for farming, recreation, and timber. It is bisected by power lines, and, while it has access to municipal utilities, those utilities are not connected to the Property. The Property also has significant road frontage

1 The Industrial Parcel is shown as Lots 1-4 on the map and the Residential Parcel is the portion labeled “Residential” on the bottom left. Unfortunately, the property lines are only visible in a color version. For context, a portion of the air base is visible on the right side of the map.

on Parker Drive, a newly paved two-lane road less than twenty miles from Interstate-95. It sits within the Marine Corps Air Station Overlay Zone, an area where activities that may interfere with flight patterns are limited by a Beaufort County ordinance: For example, that ordinance limits activities that produce smoke, glare, or electronic signals or attract waterfowl.

The easement is quite restrictive, prohibiting the property’s subdivision, human habitation, the installation of any structures or obstructions taller than 120 feet, activities that attract birds or produce visual hazards like smoke or glare, for-profit recreational activities, and activities that encourage more than 10 people to gather. The Landowners retain all responsibilities of land ownership and must bear all costs and liabilities “of any kind related to the ownership and maintenance of the [Encumbered Property].” J.A. 54. And even using the property for approved activities sometimes requires approval from the Navy.

B. The Industrial Parcel

At trial, the Landowners called an expert appraiser, Thomas Hartnett, and put forth their own lay appraisal of the fair value of the easement on the Industrial Parcel.

Hartnett concluded that the highest and best use of the Industrial Parcel was light industrial development. In his view, Beaufort was a “growing community” that he compared to “Charleston 25 years ago, or 20 years ago.” J.A. 234. Even so, he admitted that he did not know of any serious demand for the parcel or how far in the future such demand may manifest. Hartnett found no comparable arm’s-length sales in Beaufort County and so relied on four land sales that took place outside the county, three in Berkeley *8 County and one in Charleston County. For these sales, he applied downward adjustments of 15% to 43% for their size and 20% to 35% for their location. Based on those adjusted values, Hartnett opined that the encumbered portion of the Industrial Parcel’s value before the restrictive easement was $20,000 per acre, for a total value of $3,580,000 (179 acres × $20,000/acre). Then he concluded that the easement reduced the value of the encumbered land by 75 percent to $5,000 per acre for a total reduction in value of $895,000 (179 acres × $5,000/acre). 2 Thus, using his values, just compensation for the easement on the Industrial Parcel was $2,685,000 ($3,580,000 − $895,000). 3

As an alternative to Hartnett’s expert testimony, Landowner Harold Trask offered his lay opinion on the “before” value of the Industrial Parcel. Trask argued that the vacant encumbered land was worth $30,000 per acre before the easement for a total value of $5,370,000 (179 acres × $30,000/acre). Trask reached this valuation based on his experience “working with the family properties,” including sales of smaller tracts of land. J.A. 190. Trask acknowledged that his family had never received an offer to purchase the Industrial Parcel. But shortly before this litigation began, a solar company had offered to

2 In coming to this conclusion, Hartnett did not consider comparable sales because he could not find properties with similar encumbrances. He concluded that the Landowners may continue to use the encumbered part of the parcel “for such [purposes] as planting and gardening, hunting, and other uses in keeping with the restrictions of the easement.” J.A. 1046 3 Even though he was their own expert, the Landowners rejected Hartnett’s appraisal of the Property’s value after the easement and argued based on their belief that the Industrial Parcel had no post-easement value, and thus they were entitled to the full $3,580,000 in compensation for the Industrial Parcel.

lease portions of the parcel for $1,000 an acre to operate a solar farm for up to 30 years. And Landowner James Trask acknowledged that the solar lease was “the best financial option available” for the use of the parcel at that time. J.A. 327. The Landowners, despite this offer, still put forth a valuation of $5,370,000 before the easement. The Landowners then argued that the encumbered portion of the Industrial Parcel had “no value” after the easement. J.A. 196, 464–65. So they sought the entire $5,370,000 as just compensation for the encumbered part of the Industrial Parcel. The Landowners also claimed that the easement on the encumbered portion of the Industrial Parcel (179 acres) reduced the value of the unencumbered portion (267.33 acres) by 20% because the easement reduced road access and required a 50-foot setback from the encumbered part of the parcel. So they argued they were also entitled to $1,603,980 (267.33 acres × (0.2 × $30,000/acre)) for that portion.

The government disputed the Landowners’ valuations. The government’s expert, Haywood Newkirk, concluded that the encumbered portion of the Industrial Parcel was worth only $7,500 per acre before the easement, for a total value of $1,342,500 (179 acres × $7,500/acre). Newkirk believed that the parcel’s highest and best use was its current one (farming, recreation, and timber). He projected that the demand for industrial development in the area was “very small.” J.A. 396; see also J.A. 900. Newkirk noted that over the last 50 years, roughly 225 acres of industrial-zoned land had been sold near the Industrial Parcel, an average of only four acres per year, and extensive industrial space remained unoccupied nearby.

The government bolstered Newkirk’s projection with the testimony of Kimberly Statler, a regional workforce adviser at the South Carolina Department of Commerce. Statler testified that Beaufort’s market for industrial property was “very, very small” and “pretty much nonexistent.” J.A. 92. She explained that Beaufort was not “a logical location” for industrial development because it is not near major highways or ports. J.A. 96. She also opined that incoming industrial developers do not want to purchase raw land like the Industrial Parcel; instead, they prefer property with existing infrastructure. Some evidence suggested that infrastructure improvements would cost at least $16 million, and Statler did not think a developer would purchase the parcel without those improvements.

To conduct his appraisal, Newkirk considered four comparable land sales, two in Beaufort County, one near Charleston, and one near Savannah, Georgia. Some of those sales were unusual, however: one involved land with soon-to-expire wetlands permits and another was a bank’s sale of a park bought at a foreclosure auction to the City of Beaufort. Newkirk then adjusted those sales to reflect the characteristics of the Industrial Parcel and concluded that the parcel was worth $7,500 per acre before the easement. He then determined that the easement reduced the value of the encumbered property to $3,000 per acre, leading to a lost value of $805,500 (179 acres × ($7,500/acre − $3,000/acre)).

C. The Residential Parcel

Hartnett, who once again was the Landowners’ appraiser, concluded that the highest and best use of the Residential Parcel was residential development consistent with its zoning status. At first, he misunderstood the zoning acreage requirement but later claimed *11 that his error made no difference. 4 Hartnett testified that although he did not observe any demand for residential development near the Residential Parcel, the vacant land was still “ideal for” such development: “it’s close to Beaufort, it’s 90 acres of land, it is in a residential area, [and] it does have public utilities available to it.” J.A. 269. But Hartnett noted that he did not know when development could happen—it could manifest five to fifteen years in the future.

To value the parcel, Hartnett considered three sales of undeveloped land in Beaufort County between 2014 and 2015. Those parcels had sold for $14,000 to $55,000 per acre. Hartnett then applied a downward adjustment of 10% to 28% for size and up to 25% for location to each parcel. J.A. 225, 1043. He thus valued the Residential Parcel at $18,500 per acre before the easement, for a total value of $1,669,070 (90.22 acres × $18,500/acre). J.A. 225, 1044. As he had done for the Industrial Parcel, Hartnett concluded that the easement reduced the value by 75% to $4,625 per acre, for a post-easement value of $417,267 (90.22 acres × $4,625/acre). J.A. 1063. Thus, in his view, just compensation for the Residential Parcel was $1,251,803 ($1,669,070 − $417,267). J.A. 221.

Again, the Landowners disagreed with their expert’s post-easement valuation. As they had done with the Industrial Parcel, the Landowners claimed that the Residential Parcel was worth $30,000 per acre before the easement and had no value after the easement.

4 Hartnett originally understood the applicable zoning standard for the Residential Parcel as permitting “two houses per acre,” J.A. 1030, when it allowed only one house “every three acres,” J.A. 263.

So they argued just compensation for the easement was $2,706,600 (90.22 acres × 30,000/acre).

By contrast, the government’s expert, Newkirk, opined that the highest and best use of the Residential Parcel was residential development of one to three lots along the road and agricultural or recreational use otherwise. And he concluded that only 83.95 acres of the 90.22-acre tract was usable and thus available for development at all. But Newkirk identified limited demand for single family homes in the area and believed the market showed “little signs of growth.” J.A. 858.

To arrive at his valuation, Newkirk considered three land sales in Beaufort County. After applying downward adjustments, he concluded that the Residential Parcel was worth $4,800 per acre, or $402,960 in total (83.95 acres × $4,800/acre), before the taking. Based on comparable recreational land sales, Newkirk concluded that the Residential Parcel was worth $3,000 per acre, or $270,660 in total (90.22 acres × $3,000/acre), after the easement. J.A. 431, 927. 5 So he found the compensation owed for the easement on the Residential Parcel was $132,300 ($402,960 − $270,660).

D. The land commission

The land commission accepted Hartnett’s pre-easement valuations for both parcels. In doing so, it rejected the Landowners’ higher valuation and the government’s lower valuation. The commission then rejected all of the proposed post-easement estimates. Instead, it concluded that the easement reduced the value of the encumbered portion of the

5 Newkirk used the full 90.22 acres to calculate the parcel’s value after the easement because the excluded wetlands had recreational value.

Industrial Parcel by 91% and the value of the unencumbered portion of the Industrial Parcel by 10%. So the commission recommended a total compensatory award of $5,311,313 (($3,580,000 × .91) + ($1,669,070 × .91) + (267.33 acres × ($20,000/acre × 0.1))). The commission also concluded that the government’s litigation positions were not substantially justified and recommended awarding attorney’s fees and costs to the Landowners.

E. The district court

The government objected to the award recommended by the commission. It argued that Hartnett relied on incomparable land sales, ignored the high cost of the infrastructure required to make development feasible, and did not rebut the presumption that the Property’s current use was its highest and best one. The court rejected those arguments, finding that the land was unique and the easement was extremely restrictive, which required both experts to rely on their expertise and judgment to identify reasonable comparisons. 269 Acres, 2019 WL 1450578, at *4. The court also found that the expert testimony showed a reasonable probability that the land could be used for future industrial development. Id. And because it found that Hartnett’s comparisons were more apt, it adopted his (and therefore the commission’s) pre-taking values. Id. at *5.

But the court rejected both the commission’s and Hartnett’s after-taking valuation, questioning Hartnett’s failure to identify comparable sales and the commission’s unsupported 91% devaluation. Id. at *6. Instead, it adopted Newkirk’s after-taking valuation of $3,000 per acre. Id. at *5–6. The area was quickly developing like other parts *14 of the surrounding area, the court explained, and the government, aware of this development, had tried to take the land before it appreciated in value. Id. at *6.

The district court thus found that the encumbered portion of the Industrial Parcel was worth $20,000 per acre before the taking and $3,000 per acre after, requiring compensation of $3,043,000 (179 acres × ($20,000/acre − $3,000/acre)). Id. And it found that the Residential Parcel was worth $18,500 per acre before the easement and $3,000 per acre after, requiring compensation of $1,398,410 (90.22 acres × (18,500/acre − $3,000/acre)). Id. So in total, it ordered the government to pay $4,441,410 in compensation. Id. 6

In a separate opinion, the district court awarded the Landowners attorney’s fees and litigation expenses under the Equal Access to Justice Act. 269 Acres, 2020 WL 219792, at *11. It determined that the Landowners were entitled to those fees because they were the “prevailing party” in the litigation and the government’s position was not “substantially justified.” Id. at 7–8; see 28 U.S.C. § 2412(d)(1)(A). The court decided that the Landowners were the “prevailing party” because their highest valuation of the Property was closer to the final compensation awarded than the government’s valuation. 269 Acres, 2020 WL 219792, at *7. In making this determination the court only took Hartnett’s valuation into account, ignoring Harold and John Trask’s estimates because they valued the total property, not just the encumbered property. Id. at *6 n.5. It then found that the government’s position was not “substantially justified” because the government had

The district court found the unencumbered property’s value was unaffected. See 6

269 Acres, 2019 WL 1450578, at *6. *15 changed its valuation three times and sought to discover “irrelevant” personal information of the Landowners. Id.

Considering cost-of-living adjustments and the complexity of the case, the court awarded $593,301 in attorney’s fees and $37,998.60 in litigation costs to the Landowners. Id. at *9. It then apportioned these amounts among the seven Landowners, finding that only William Trask, as the only party eligible to receive reimbursement based on his limited net worth, could receive his share of the fees, $84,757.29 ($593,301 / 7) and costs, $5,428.37 ($37,998.60 / 7). Id. at 9–10. Finally, the court equally divided the cost of the land commission between the government and the Landowners at a cost of $35,575 per side. Id. at *11.

Both the government and the Landowners timely appealed. We have jurisdiction under 28 U.S.C. § 1291. II. Analysis

We affirm the district court’s award of just compensation and the splitting of the commission costs because of the deference we must afford to the district court. But we reverse the award of attorney’s fees and costs because the Landowners were not the “prevailing party” in the litigation. 28 U.S.C. § 2412(d)(1)(A).

A. Just compensation

When the government uses its power of eminent domain to take private property rights, the Constitution requires it to pay “just compensation.” U.S. Const. amend. V. “[J]ust compensation” is the “amount of money necessary to put a landowner in as good a pecuniary position, but no better, as if his property had not been taken.” United States v. *16 69.1 Acres, 942 F.2d 290, 292 (4th Cir. 1991). For an easement, that amount is “the difference between the value of the property before and after the Government’s easement was imposed.” United States v. Va. Elec. & Power Co., 365 U.S. 624, 632 (1961). The relevant “value” is the “fair market value” of the property, which is “‘what a willing buyer would pay in cash to a willing seller’ at the time of the taking.” United States v. 564.54 Acres, 441 U.S. 506, 511 (1979) (quoting United States v. Miller, 317 U.S. 369, 374 (1943)). Landowners bear the burden of establishing the fair market value of their property. United States ex rel. Tenn. Valley Auth. v. Powelson, 319 U.S. 266, 273 (1943).

A particular property may be “adaptable to several uses” so “just compensation is measured by the use that would bring the highest price—the ‘highest and best’ use.” 69.1 Acres, 942 F.2d at 292. The highest and best use of a parcel “is presumed to be its current use.” Id. But compensation is not limited to the current use of the property, as it “also includes any additional market value it may command because of the prospects for developing it to the ‘highest and best use’ for which it is suitable.” United States v. 480.00 Acres of Land, 557 F.3d 1297, 1307 (11th Cir. 2009); accord Mitchell v. United States, 267 U.S. 341, 344–45 (1925) (value includes adaptability for potential uses). Indeed, much of the value of a piece of land may be its future potential. But to base compensation on a use other than the current one, the landowner must produce evidence that the proffered use is “‘reasonably probable’ and that the probability has a real market value.” 69.1 Acres, 942 F.2d at 292. A showing of physical adaptability to a proposed use is not enough; there must be demand for the use in the “reasonably near future.” Olson v. United States, 292 U.S. 246, 255 (1934).

The best evidence of property value comes from comparable land sales, where the more similar the land is the more probative the sale price is. United States v. Whitehurst, 337 F.2d 765, 775 (4th Cir. 1964); 480.00 Acres, 557 F.3d at 1307. “Courts have generally recognized that comparability is a function of three variables: the respective characteristics of the properties, their geographic proximity to each other, and the closeness in time of the sales.” United States v. 68.94 Acres of Land, 918 F.2d 389, 399 (3d Cir. 1990) (citing United States v. 320.0 Acres of Land, 605 F.2d 762, 798 (5th Cir. 1979)). Very often, the most comparable sales will differ in these aspects, even significantly. 320.0 Acres, 605 F.2d at 798. Because identifying the relevant comparisons and making value adjustments often turns into a battle of experts, we defer to the district court in weighing the varying opinions those experts offer, reversing only when the district court has made a clear error. See United States v. 1,061.14 Acres of Land, 491 F.2d 700, 701 (8th Cir. 1974); United States v. 124.84 Acres of Land, 387 F.2d 912, 915 (7th Cir. 1968).

Our review of a district court’s determinations of a property’s highest and best use, the probability of future demand, and the ultimate award of just compensation is equally deferential under the clear error standard. See 69.1 Acres, 942 F.2d at 293–94. This is especially true when the district court adopts the findings of a land commission. See, e.g., United States v. Certain Parcels of Land Located in Fairfax and Loudoun Cntys., 384 F.2d 677, 679 (4th Cir. 1967). To reverse under the clear error standard, we would need to be “left with the definite and firm conviction that a mistake has been committed.” United States v. U.S. Gypsum Co., 333 U.S. 364, 395 (1948). Even if we are “convinced *18 that had [we] been sitting as the trier of fact, [we] would have weighed evidence differently,” as long as “there are two permissible view of the evidence,” we cannot find clear error. Anderson, 470 U.S. at 574.

The government argues that the district court clearly erred in accepting the Landowners’ pre-taking valuations for the Industrial and Residential Parcels. The overarching theme of their argument is that the valuations hinged on speculative future demand, and that speculative future demand could not rebut the presumption of its current use being the highest and best use. In making that argument, the government contends that (1) many of the district courts’ factual findings were clearly erroneous; (2) Hartnett’s land comparisons were inapplicable and should not have been credited; and, (3) the Landowners presented insufficient evidence of demand for the Property in the reasonably near future. The first two arguments are intensely factual and ask us to second guess both the land commission’s and the district court’s credibility determinations and weighing of *19 the evidence. On this record, we cannot find that the district court clearly erred in making its factual determinations, 7 including its decision to credit Hartnett’s comparable sales. 8

7 The government marshals a plethora of arguments against the valuations, but most turn on factual determinations that we cannot find to be clearly erroneous. We only address a few of their factual arguments. For example, the government points out that Hartnett used the wrong zoning requirements in evaluating the potential uses of the Residential Parcel. But Hartnett acknowledged his mistake and said it made little difference to his valuation, which the district court seemed to accept. See 269 Acres, 2019 WL 1450578, at *5. The government also offered Statler’s deposition into evidence at trial to show that industrial demand in this area was low. But the commission and the court found that Statler’s testimony was not very probative since her relevant experience took place near the recession. Those credibility and weight determinations are within the purview of the district court. The government also argues that it would cost around $16 million to ready the land for development, and investment developers would not purchase the land without it. According to the government, neither the court nor Hartnett accounted for that cost. But many of the sale comparisons the district court credited involved undeveloped land that also lacked infrastructure. At their base, this is a factual determination that turns on the weight given to conflicting evidence, which is in the province of the district court. So we cannot conclude that the district court clearly erred in making any of these or the other factual determinations about which the government complains. 8 The government argues that the land sales Hartnett relied on are not sufficiently similar to the Property because of their sizes, locations, infrastructure, and ready development. But comparisons are rarely identical. See 320.0 Acres, 605 F.2d at 798 (“In most cases, of course, there are no open market sales ‘ideally’ comparable.”). That is why courts rely on experts to select good comparisons and make adjustments for the condemned property’s differences. See id. at 798, 802; 68.94 Acres, 918 F.2d at 393. Even the government’s appraiser admitted the Property was unique and looked outside Beaufort for comparable sales. And his land-sale comparisons were also significantly different from the Property. Identifying comparable land sales is a highly factual endeavor and often turns on a battle of expert opinions. The district court is at the apex of its discretion in weighing such testimony. See United States v. Wateree Power Co., 220 F.2d 226, 230 (4th Cir. 1955); 124.84 Acres, 387 F.2d at 915; United States v. Banisadr Bldg. Joint Venture, 65 F.3d 374, 377–78 (4th Cir. 1995); United States v. 7,216.50 Acres of Land, 507 F. Supp. 228, 237 (D.S.C. 1980). So on this record we cannot conclude that the district court clearly erred in choosing to credit Hartnett’s comparisons.

The government’s third argument has more merit. The government posits an array of facts that suggest demand for industrial and residential development on the Property was speculative: The Landowners had received no offers to purchase the Property; the best available use of the Property at that time was a solar-farm lease; the potential demand for the Property was unknown and possibly five to fifteen years in the future; most people moving into the area want to purchase condos or apartments, not large pieces of undeveloped land; there had been no major industrial developments nearby; and the closest industrial spaces had not been leased. And merely speculative demand cannot suffice.

We agree that the Landowners’ evidence of demand in the reasonably near future for the particular property at issue is tenuous and a reasonable court could find it insufficient. But in such fact-intensive areas, we defer to resident district courts, who have both knowledge of the local community and fact-finding expertise. Cf. Ornelas v. United States, 517 U.S. 690, 700 (1996) (noting the importance of deference to resident judges). And the very point of deference is to accept a reasonable decision even though we might have decided it differently. See Anderson, 470 U.S. at 573–74.

We applied this deference in 69.1 Acres. There, we held that a district court did not clearly err in finding “that the highest and best use” of a condemned property was commercial sand mining, even though the land was not currently being used for that purpose. 942 F.2d at 291–92. The landowners had to show that a “market exists for the mineral that would justify its extraction in the reasonably foreseeable future” to rebut the presumption of current use. 69.1 Acres, 942 F.2d at 292. So they presented an appraiser *21 who considered comparable sales to value the property and a businessman in the sand-mining industry who testified that “[w]hile the current demands for sand in the [] area are being met by existing sources . . . his company was interested in purchasing reserves for development within the next twenty years, and possibly within the next five years.” Id. at 293.

The government there, just as in this case, relied on United States v. Whitehurst, 337 F.2d 765 (4th Cir. 1964), for the proposition that the land “may not be valued on the basis of conjectural future demand for it. There must be some objective support for the future demand, including volume and duration.” 69.1 Acres, 942 F.2d at 293 (quoting Whitehurst, 337 F.2d at 771). The government reasoned from Whitehurst that the landowner had not shown the “volume and duration” of future demand for sand from the property so sand mining could not be the property’s highest and best use. Id. But we rejected that argument and confined Whitehurst to its “context,” the disfavored use of the “income capitalization” method of calculating just compensation. Id. We then held that “the landowner [did] not have to show an imminent demand for the sand from his property. He just ha[d] to show that there [was] a reasonable probability that the sand w[ould] be needed and wanted at a near enough point in the future to affect the current value of the property.” Id. at 294 (footnote omitted). It was enough that a person in the industry was interested in purchasing the land at that time to secure sand reserves that might be extracted within twenty years and that similar sales in the area had been made. Id.

The “person in the industry” in 69.1 Acres tied the future demand for sand to current demand for obtaining the particular property for sand reserves. And we found that *22 evidence sufficient. But in United States v. Wateree Power Co., 220 F.2d 226 (4th Cir. 1955), we upheld an award where the connection of future and current demand was even less apparent. There, we found no clear error in the conclusion that a property’s highest and best use was not its current use of growing timber but its potential industrial use. Id. at 230. The government argued that future demand for the property lacked any support as shown by recent sales and the lack of any industry in the area. Id. at 228–30. The landowners countered with testimony that the particular property was suitable for future industrial use and that industry was growing in areas of the state with similar attributes. Id. at 228–29. The district court and the commission, we held, did not clearly err in finding evidence of non-speculative future demand because the landowners “show[ed] that a market for these properties existed or was reasonably likely to exist in the near future.” Id. at 230–31. As industrial development was “not remote or speculative,” we held that it could be considered. Id. at 232 (quoting United States v. 1,532.63 Acres of Land, 86 F. Supp. 467, 470 (W.D.S.C. 1949)).

The government tries to distinguish Wateree by arguing that the industry experts there showed actual demand for the land and the proposed future use. But those experts had not shown specific demand for that particular property or even property in that particular area. They had merely opined that the land had physical attributes favorable for industrial use and that similar locations across the state had shown such demand. Id. at 228–29. We found that the evidence permitted the district court to conclude that the taken land might have been used for industrial purposes “within the reasonable foreseeable future.” Id. at 231. And that is all that is needed here.

Hartnett’s testimony satisfies that threshold showing. He opined that the Property was suitable for industrial and residential development based on the growth of the area and similar sales and growth in other parts of the surrounding area. And general demand in similar areas can affect a property’s market value, even if the owners cannot produce direct evidence of current demand for their specific tract. Id. The commission and the district court agreed with Hartnett’s assessment, and we cannot find their reliance clearly erroneous.

Instances when deference to a district court’s factual finding was overcome because the future demand was too speculative provide useful foils. In Olson v. United States, 292 U.S. 246, 260 (1934), the Supreme Court found that the evidence of demand for a future use of a particular property was speculative because any potential future demand turned on a purely theoretical use of the land as a reservoir, which would require acquiring many other lots from private individuals, Native American tribes, and sovereign proprietors. In the face of those complexities, the physical adaptability of that land to use as a reservoir was not enough. Id. at 256–57. Similarly, in United States ex rel. & for Use of Tennessee Valley Authority v. Powelson, 319 U.S. 266, 274–76, 281 (1943), the Supreme Court found the potential future use of the condemned land as a dam was too speculative because building a dam would require uniting hundreds of individually owned pieces of land, including by the use of eminent domain.

But contrary to Olson and Powelson, the proposed uses of this Property do not require massive preparatory undertakings, the acquisition of other pieces of land, or the *24 additional use of eminent domain. The Property is physically and legally adaptable to the proposed uses on its own.

The deference we afford district courts in this area requires us to affirm. A resident district judge is in a better position to weigh comparable sales, dueling experts, prospective uses, the local economy, and other factors bearing on its decision. The government’s argument that the Property’s current agricultural and recreational use has not been rebutted has real force. The Landowners produced little evidence tying the generalized demand to potential demand for their particular property. So a district court could well have rejected the Landowner’s position.

But there was some evidence of industrial demand in the area—including for a nearby industrial park—and even the government’s appraiser acknowledged the rapid growth Beaufort was experiencing. J.A. 190–93, 232–34, 256, 270, 280–81, 845. And the Landowners’ appraiser testified that development of the Residential Parcel could occur within five to fifteen years. See 69.1 Acres, 942 F.2d at 293–94; Wateree Power, 220 F.2d 230–32. This potential industrial and residential demand in the near future was supported by demand for other land with similar physical attributes. Wateree, 220 F.2d at 230–31. We cannot say that this evidence, taken as a whole, can only be interpreted to show speculative demand. So even if we would have found differently in the first instance, the district court acted within its bounds in finding that this evidence was enough to rebut a presumption of current usage and establish demand in the reasonably near future that affected the Property’s value. Accordingly, we affirm the district court’s award of compensation.

B. Attorney’s fees and litigation expenses

Generally, parties pay their own litigation costs. But the Equal Access to Justice Act deviates from this rule by rendering “the United States liable for attorney’s fees for which it would not otherwise be liable.” Ardestani v. I.N.S., 502 U.S. 129, 137 (1991). Under the Act, the court must award attorney’s fees and other litigation costs to the private party if: (1) the claimant is an eligible individual with a net worth of less than “$2,000,000 at the time the civil action was filed” and (2) the claimant is the “prevailing party” in the litigation unless (3) the district court finds that the government’s position was “substantially justified” or “special circumstances make an award unjust.” 28 U.S.C. § 2412(d)(1)(A), (d)(2)(B). The parties here agree that only William Trask is eligible to receive attorney’s fees and costs given the net worth of the other Landowners. Their dispute instead centers on whether the Landowners were the “prevailing party.”

In an eminent-domain proceeding, the “prevailing party” is “the party whose highest trial valuation of the property is closest to the final judgment.” United States v. 515 Granby, LLC, 736 F.3d 309, 314 (4th Cir. 2013) (citing 28 U.S.C. § 2412(d)(2)(H)). 9 Put another way, to receive attorney’s fees, the Landowners’ highest valuation of the Property at trial must be as close to the final compensation awarded as the government’s highest

9 “Prevailing party” is statutorily defined for eminent-domain proceedings as “a party who obtains a final judgment (other than by settlement), exclusive of interest, the amount of which is at least as close to the highest valuation of the property involved that is attested to at trial on behalf of the property owner as it is to the highest valuation of the property involved that is attested to at trial on behalf of the Government.” 28 U.S.C. § 2412(d)(2)(H).

valuation. See id. We review this determination de novo. Goldstein v. Moatz, 445 F.3d 747, 751 (4th Cir. 2006).

The government argues that it was the “prevailing party” because the Landowners’ requested compensation of $9,680,580 is farther from the $4,441,410 award than the government’s proposed compensation of $937,800. But the district court rejected this simple math because the Landowners’ request of $9,680,580 was for the “entire property” and not just the “property involved.” 269 Acres, 2020 WL 219792, at *6. As a result, it regarded the requested $9,680,580 as irrelevant. Id. at *6 & n.5. The district court instead relied on the Landowners’ expert Hartnett’s valuation of $3,936,802, which was the closest valuation to the actual award. Id. at *6. That reliance was misplaced.

“[T]he highest valuation of the property involved that is attested to at trial” is the highest amount that the party requests for just compensation at trial. 28 U.S.C. § 2412(d)(2)(H); see United States v. 50.50 Acres of Land, 931 F.2d 1349, 1358–59 (9th Cir. 1991); United States v. 5,507.38 Acres of Land, 832 F.2d 882, 883 (5th Cir. 1987). It is not limited to valuations of the property interests that the district court ultimately concludes require compensation. Rather, any property interest a party seeks just compensation for at trial is “the property involved.” See Wateree, 220 F.2d at 231 (noting that the devaluation of attached land is relevant).

It is clear from the record that the Landowners asked for $9,680,580 in just compensation for the easement on the entire property, including the unencumbered portion. They reached this figure by considering the encumbered property to be worthless and the unencumbered property devalued by 20%. That estimate was used by the Landowners in *27 their closing argument to show their total damages. And indeed, in asking the commission for attorney’s fees, the Landowners themselves argued that their highest valuation at trial was $9,680,580. So $9,680,580 is the highest valuation of the Property the Landowners put forth at trial, which is farther from the district court’s award of $4.4 million than the government’s valuation of $937,800.

In arguing $9,680,580 is not the appropriate valuation to consider, the Landowners now ask us to consider this issue at each stage separately. They contend that before the commission they had the closest valuation to that adopted by the commission. Then, before the district court, they adopted the commission’s value ($5,311,313), which was the closest to the district court’s award. But we do not consider the commission and the district court to be separate stages for determining who prevailed. The statute requires us to compare the “final judgment” to the highest valuation “attested to at trial.” § 2412(d)(2)(H). The final judgment was the district court’s award of $4.4 million in compensation, so the relevant trial must be the proceedings that led to that final judgment. And in eminent-domain cases involving commissions, it is common for proceedings before the commission to be considered part of the trial, just as the district court did here. 269 Acres, 2020 WL 219792, at *6 n.3; see United States v. Harrell, 642 F.3d 907, 911, 916 (10th Cir. 2011).

The Landowners insist Harrell, 642 F.3d 907, supports their novel contention. In that case, the property owners argued that they should not be held to the $33 million valuation used before the land commission for awarding attorney’s fees because they had since adopted the commission’s $6.1 million valuation in briefing before the district court. Id. at 910. The Tenth Circuit explicitly rejected the owners’ argument based on “[a] strict *28 construction of § 2412(d)(2)(H).” Id. at 915–16. The court noted, however, that even if it discounted the property owners’ testimony before the commission, the property owners had still testified to a valuation as high as $33 million before the district court. Id. at 916. That alternative argument does not persuade us to consider the two stages separately. We instead find the Tenth Circuit’s explicit rejection of the Landowners’ argument to be more persuasive.

Based on Harold and John Trask’s testimony, the Landowners attested before the commission that the just compensation owed for the easement was around $9.6 million. Because the government’s $937,800 value is closer to the district court’s final award of $4.4 million, the government, not the Landowners, is the “prevailing party” in this litigation. So we reverse the district court’s award of attorney’s fees and costs to the Landowners. As a result, we need not decide whether the district court erred in apportioning the attorney’s fees based on the eligibility of the Landowners.

C. Commission fee

In their cross appeal, the Landowners argue that the district court abused its discretion by dividing the cost of the commission between them and the government. Although there is very little law in this area, the district court retains great discretion in splitting commission costs. The district court did not abuse that discretion. 10

10 The practice of sharing the costs, however, appears to be exceedingly rare, as the government can only marshal a handful of district court cases doing so and the Landowners ably distinguish them. See United States v. 400 Acres of Land, No. 2:15-cv-01743-MMD- NJK, Dkt. 523 (D. Nev. Sept. 26, 2019) (ordering the government to pay 75% and the landowners to pay 25% of the commission costs and rejecting an argument based on Federal Rule of Civil Procedure 71.1(l) like that asserted by the Landowners in our case);

The governing rule is Federal Rule of Civil Procedure 71.1(l), which merely states that “[c]osts are not subject to Rule 54(d).” Rule 54(d) explains that costs are generally awarded to the prevailing party but may only be imposed against the federal government “to the extent allowed by law.” But while those rules tell us how commission costs should not be treated, they tell us little about how commission costs are to be treated.

The advisory note for Rule 71.1(l) arguably suggests that commission fees should be paid by the government, especially when the government is the prevailing party (as it is here). 11 But the advisory note says that it seeks to summarize existing caselaw pertaining to the rule’s topic. It does not state the rule. See United States v. Carey, 120 F.3d 509, 512

United States v. 275.81 Acres of Land, No. 3:09-cv-00233-DWA, Dkt. 144, at 2 (W.D. Pa. July 9, 2013) (splitting costs equally but per a consent agreement); United States v. 99,223.7238 Acres, No. 2:06-933 RB/RHS, Dkt. 128, at 1–2 (D.N.M. June 19, 2008) (splitting costs equally but by joint stipulation). 11 The advisory committee’s note states:

Since the condemnor will normally be the prevailing party and since he should not recover his costs against the property owner, Rule 54(d), which provides generally that costs shall go to the prevailing party, is made inapplicable. Without attempting to state what the rule on costs is, the effect of subdivision (l) is that costs shall be awarded in accordance with the law that has developed in condemnation cases. This has been summarized as follows: “Costs of condemnation proceedings are not assessable against the condemnee, unless by stipulation he agrees to assume some or all of them. Such normal expenses of the proceeding as bills for publication of notice, commissioners’ fees, the cost of transporting commissioners and jurors to take a view, fees for attorneys to represent defendants who have failed to answer, and witness’ fees, are properly charged to the Government, though not taxed as costs.

Fed. R. Civ. P. 71.1 advisory committee’s note to subdivision (l) (emphasis added). *30 (4th Cir. 1997) (“[T]he Advisory Committee Note is not the law; the rule is.”). So because the rule sets outer bounds as to what may not be done, the district court is left with discretion over what should be done.

Using that discretion, the district court turned to Rule 53(g), which addresses compensation for masters. Under that rule, masters may be compensated by either or both parties. See Fed R. Civ. P. 53(g) (“The compensation must be paid either . . . by a party or parties[] or from a fund or subject matter of the action within the court’s control.”). The Landowners argue that the canon of expressio unius ext exclusio makes reliance on that rule misplaced because Rule 71.1 explicitly references specific parts of Rule 53, but does not reference subpart (g). See Fed. R. Civ. P. 71.1. We reject that argument. Rule 71.1(l) excludes reliance on Rule 54; its silence about reliance on Rule 53(g) therefore implies that it may be considered. Nothing in Rule 71.1 displaces reliance on relevant rules other than Rule 54. So the district court was within its authority to look to Rule 53(g) for guidance on how to split the commission costs. See, e.g., Guardian Pipeline, L.L.C. v. 295.49 Acres of Land, No. 08-C-0028, 2008 WL 2482005, at *4 (E.D. Wis. June 18, 2008).

The Landowners argue that even if the district court may rely on Rule 53(g), it abused its discretion by not considering the various equitable factors listed in Rule 53(g)(3). See Fed. R. Civ. P. 53(g)(3) (“The court must allocate payment among the parties after considering the nature and amount of the controversy, the parties’ means, and the extent to which any party is more responsible than other parties for the reference to a master.”). But the court’s use of Rule 53(g) as a guide to resolving this issue does not bind it to all of the rule’s requirements. It has broad discretion to determine which portions of *31 Rule 53(g) it finds persuasive in answering this question. So we hold that the district court acted within its discretion in dividing the commission costs.

• • * *

Determining “just compensation” often turns on highly factual determinations, such as the identification of comparable land sales, the state and growth potential of the local economy, and the credibility of dueling experts and various reports. Comparing and weighing this mountain of evidence can be difficult. Those closest to the ground, who live near the relevant areas and can hear the witnesses’ testimony firsthand, are in the best position to evaluate the evidence. Commissions and district courts are there. We are not.

Although we may not have resolved the question of the Property’s future demand the same way, the district court was within its discretion to weigh the evidence and conclude that the Landowners had shown a non-speculative demand for industrial and residential development in the reasonably near future. So we cannot say that the district court clearly erred in calculating its award of just compensation. Similarly, the district court has broad discretion in apportioning commission costs, and we uphold its decision to do so. But identifying the “prevailing party” for awarding attorney’s fees is a legal question that we review de novo. And we find that the district court erred in making that determination. Accordingly, the district court’s judgment is

AFFIRMED IN PART AND REVERSED IN PART.