In 2015, we remanded this case to the Minnesota Tax Court after concluding that the court's valuation of relator KCP Hastings, LLC's shopping mall was not reasonably supported by the record. KCP Hastings, LLC v. County of Dakota , 868 N.W.2d 268, 276 (Minn. 2015). Following the remand, the tax court received further evidence from the parties and ultimately reached a new valuation for KCP's property. KCP again appeals, arguing the tax court committed several errors that resulted in a valuation of its property that is not supported by the record. We conclude that the tax court did not err, with two exceptions: the assignment of value to an outlot on the property, and the use of a building area other than the one stipulated to by the parties. Accordingly, we affirm in part, reverse in part, and remand to the tax court for further proceedings.

FACTS

KCP owns real property in Hastings, which consists of a shopping mall and a surrounding parking lot. After the County assessed property taxes on the property for 2010, 2011, and 2012 based on the estimated market value of the property, KCP appealed each assessment to the tax court, which consolidated the appeals. Following a 2-day trial in 2014, the tax court adopted market valuations that exceeded the estimates in the County's original assessments.

*777To reach these market valuations, the tax court considered five valuation approaches: the County's sales-comparison approach (based on the sales value of other, similar properties), the County's cost approach (based on the cost to build an equivalent property), the County's direct-capitalization approach (based on capitalizing one year's income), KCP's income approach, which more specifically was a discounted-cash-flow, or DCF, approach (based on forecasting income for several years), and KCP's sales-comparison approach. The tax court gave no weight to the County's cost approach, "little weight" to the County's direct-capitalization approach, and "little weight" to KCP's sales-comparison approach. Although the tax court concluded that a DCF analysis was an appropriate valuation method, it gave KCP's DCF analysis no weight because the court was "unable to replicate" many of the calculations included in KCP's expert's report. Notably, the data necessary to replicate those calculations was contained in a spreadsheet that KCP did not produce during discovery "because it was not 'asked for,' " and when KCP attempted to introduce it at trial, "[t]he tax court sustained the County's objection and excluded the evidence on the basis of 'unfair surprise,' stating ... that it would 'muddle along without' the spreadsheet." KCP Hastings, LLC , 868 N.W.2d at 271. Consequently, of the five approaches considered in the 2014 trial, the tax court was left with the County's sales-comparison approach, upon which it based its valuations.

KCP appealed, arguing that the tax court clearly erred by using the mall's gross building area rather than its gross leasable area, and in rejecting KCP's sales-comparison and DCF approaches. Id. at 270. KCP also asserted that the tax court abused its discretion by using only the County's sales-comparison approach to value KCP's property. Id. We affirmed in part and reversed in part, concluding that the tax court did not clearly err by using the gross building area or by rejecting KCP's sales-comparison approach. Id. at 269-70, 273. But we concluded that the tax court clearly erred when it rejected KCP's DCF analysis and consequently the tax court abused its discretion by relying solely on the County's sales-comparison approach. Id. at 270. We therefore remanded the case to the tax court, giving it the option to reopen the record and conduct further evidentiary hearings. Id. at 276.

Following the remand, the tax court invited the parties to brief whether it should reopen the record or permit supplemental briefing. The County moved to reopen the record so that it could submit its own DCF analysis. KCP argued that the record should be reopened only to permit it to admit its underlying DCF calculations and that the County should not be permitted to introduce evidence that the County had previously argued was speculative.1 The tax court granted the County's motion and permitted both parties to supplement their appraisals with DCF analyses.

On December 29, 2016, the tax court filed an order that calculated the valuations for KCP's property. KCP moved for amended findings, arguing that the tax court's capitalization rate was "illogical," that the court should not have assigned value to an unmarketable outlot, and that the gross building area used by the court was unsupported by the record. The County also moved for amended findings, asking the tax court to provide greater mathematical detail on how it arrived at its DCF value conclusions. Following a hearing, the tax court granted KCP's motion to correct *778the capitalization rate, but denied KCP's other motions. At the hearing, the tax court and parties agreed that instead of having the tax court provide greater mathematical detail on its DCF calculations, the court would instead "provid[e] the parties with specific parameters from which the parties could make their own DCF calculations, agree on the result if possible, and if not, submit their separate calculations to the court for resolution."

After the tax court provided the parties with its parameters, the parties exchanged memoranda disputing the DCF calculations. On January 11, 2018, the tax court filed its final order, setting out its final valuation of the subject property:

County KCP County 2014 Tax Remand Remand 2018 Tax Assessor Court Appraisal Appraisal Court 2010 $4,791,600 $5,535,000 $6,167,500 $3,250,000 $5,307,100 2011 $4,821,700 $5,258,200 $5,514,100 $2,850,000 $4,856,000 2012 $4,821,700 $4,995,300 $5,033,600 $2,800,000 $5,221,800

KCP sought review by writ of certiorari. After briefing was completed in the spring of 2018, we stayed the appeal pending our decision in County of Hennepin v. Bhakta , 922 N.W.2d 194 (Minn. 2019). After Bhakta was decided, we dissolved the stay.2

ANALYSIS

KCP presents multiple issues for our review. On appeal from the tax court's valuation of property, we generally consider whether the court lacked jurisdiction, whether its decision is supported by the evidence and in conformity with the law, and whether the court committed any error of law. See Eden Prairie Mall, LLC v. County of Hennepin , 797 N.W.2d 186, 189 (Minn. 2011). We review the tax court's legal conclusions de novo, but will not disturb factual findings in support of a valuation determination unless they are not reasonably supported by the record as a whole. See Cont'l Retail, LLC. v. County of Hennepin , 801 N.W.2d 395, 398-99 (Minn. 2011) (explaining that we defer to the tax court's valuation determinations unless the court has clearly overvalued or undervalued the property).

I.

Following remand, the tax court permitted the County to introduce its own DCF analysis. KCP argues that this decision was both unfair and exceeded the scope of remand. "The decision to admit or exclude evidence rests with the tax court, and its rulings will not be disturbed absent an error of law or abuse of discretion." Cont'l Retail, LLC , 801 N.W.2d at 399. A court abuses its discretion on remand when it exceeds the scope of the remand. See Eden Prairie Mall, LLC v. County of Hennepin , 830 N.W.2d 16, 20 (Minn. 2013) ("[T]he tax court must execute our remand order according to its instructions and has no power to modify those instructions.").

KCP argues that the admission of the County's DCF analysis "depart[ed] from notions of basic fairness" because the County had previously argued that such analyses are inherently unreliable. Permitting the County to admit that evidence in *779the face of that objection, KCP contends, was an abuse of discretion. We disagree. No such departure occurred here. Before the tax court's 2014 order, the County argued DCF analyses were inherently speculative. Ultimately, we disagreed, and we permitted the tax court to accept KCP's DCF analysis. It is not unfair to permit the County to take a different position on remand after its original position was rejected by an appellate court. On the contrary, it would arguably be unfair to the County to conclude that the tax court should rely on DCF analyses, but then prohibit the County from providing that evidence to the court.3 We therefore conclude that it was not unfair for the tax court to admit the County's DCF analysis.

KCP also argues that admitting the County's DCF analysis violated the scope of remand. In the conclusion of our opinion remanding this case to the tax court, we stated:

We conclude that the tax court clearly erred by rejecting KCP's DCF analysis and abused its discretion when it failed to consider the income approach in its final valuation. The court's valuation of Westview Mall therefore was not supported by the record as a whole. Accordingly, we remand to the tax court for further proceedings consistent with this opinion. The tax court may, if necessary, reopen the record and conduct a further evidentiary hearing.

KCP Hastings, LLC , 868 N.W.2d at 276. Nothing in this language prevented the tax court from accepting new evidence. To the contrary, we expressly allowed the tax court to "reopen the record and conduct a further evidentiary hearing," so long as those proceedings were "consistent with this opinion." Allowing the County to offer its DCF analysis was consistent with our 2015 opinion. Accordingly, we conclude that the tax court did not violate the scope of remand by doing so.

II.

An appraisal may determine the value of a fee-simple interest in property, or it may determine the value of some inferior interest. The tax court found that KCP's expert's appraisal was a leased-fee appraisal rather than a fee-simple appraisal, and therefore rejected it as not properly valuing the property. KCP argues that this finding was clearly erroneous.

An interest in real property consists of one or more rights to the property; the entire "bundle of rights" to the property, which consists of all such interests, is a fee-simple interest. See Cont'l Retail, LLC , 801 N.W.2d at 401 ; see also Contos v. Herbst , 278 N.W.2d 732, 737 (Minn. 1979) ("[T]he entire 'bundle of rights' ... comprise[s] the fee simple interest in the parcel ...."). An appraisal may value the entire bundle of rights or only some subset of them. See Appraisal Inst., The Appraisal of Real Estate 69-70 (14th ed. 2013). Property taxes are determined based on the fee-simple interest in the property. Minn. Stat. § 273.11, subd. 1 (2018) ("[A]ll property shall be valued at its market value" and also noting that "a leasehold estate" or "some lesser value" is not market value); see Cont'l Retail, LLC , 801 N.W.2d at 401 ("The statute contemplates valuation of the entire, unencumbered interest in the real property, and not a lesser estate.").

The parties disagree whether KCP's expert assessed the value of a fee-simple interest in the property or whether he assessed the value of some lesser estate. To perform a DCF analysis, an assessor *780needs to determine the projected cash flow of the property, which is based in part on the income the property is projected to produce. The Appraisal of Real Estate , supra at 529-30. Because KCP's property is a shopping mall, KCP's income primarily comes from the rent it collects renting out storefronts to retail-store tenants.

To determine rental income, KCP's expert had to predict the terms (specifically the amount of rent) of the leases KCP's tenants would enter into during the 10-year DCF period.4 These predictions form the basis of the parties' dispute. In predicting lease terms for the 10 years, KCP's expert predicted that tenants would enter into 5-year leases with rent values that remained constant for the 5-year term. The County argues, and the tax court found, that by keeping the rent constant for these projected leases, KCP's expert assessed a property interest inferior to that of the fee-simple interest, specifically a leased-fee interest.

A leased-fee interest is the interest retained by a landlord after leasing a piece of real property to others. Cont'l Retail, LLC , 801 N.W.2d at 401. Although the specific interests vary from lease to lease, they typically include "rent to be paid by the lessee under stipulated terms," "the right of repossession at the termination of the lease," and "default provisions." The Appraisal of Real Estate , supra at 72. A leased-fee interest is inferior to a fee-simple interest because it does not typically include, inter alia, the right of use and occupancy of the property. Id. at 73.

Reviewing the record, we conclude that the tax court did not clearly err in finding that KCP's expert used a leased-fee interest appraisal. While tenants may have entered into multi-year leases with flat rental rates, those flat rental rates do not account for other existing interests in the property, which must be included to determine the fee-simple value. More specifically, although the value of the leased-fee interest in the property may have remained constant over the 5-year leases that formed the basis for the expert's opinion, that does not mean that the value of the fee-simple interest likewise remained constant. On the contrary, KCP's expert assumed that market rents would increase by 3 percent starting in 2014.

A hypothetical tenant best illustrates this point. Suppose that in 2014, Tenant leased a space from KCP for $1000 per month. Under the leases as assumed by KCP's expert, Tenant would pay $12,000 a year each year for a total of $60,000. But the expert's analysis assumed that the market supports increasing the rent by 3 percent every year. Thus, but for the multi-year, flat-rate nature of the leases, in 2014, the space could be rented for $12,000; in 2015, $12,360; in 2016, $12,730.80; in 2017, $13,112.72; and in 2018, $13,506.11, for a total of $63,709.63. The value of KCP's leased-fee interest was therefore $60,000, the amount it would collect over the course of the 5 years. But each year, Tenant could sell its interest (the right to use and possess the space) to a third-party tenant at the increasing market rate, thus collecting profits of $360, $730.80, $1,112.72, and $1,506.11, respectively. By failing to account for that value, KCP's expert failed to include the value of the tenant's right of use and possession, and consequently failed to value the fee-simple interest.

KCP attempts to avoid this conclusion by arguing that Minn. Stat. § 273.11, subd. 1, requires property to be valued at its market value and that the market during the assessment period generally supported *781multi-year leases without annually increasing rent. But this is effectively an argument that market values should be based on the value of the leased-fee interest in the property, not the fee-simple interest. We cannot square this argument with the plain language of the statute. See Cont'l Retail, LLC , 801 N.W.2d at 401 (explaining that "[t]he statute contemplates valuation of the entire , unencumbered interest in the real property, and not a lesser estate" (emphasis added)). As we stated in Continental Retail , the "fee simple interest of the property is the absolute ownership unencumbered by any other interest or estate , subject only to the limitations imposed by the governmental powers of taxation, eminent domain, police power, and escheat." Id. (emphasis added) (citation omitted) (internal quotation marks omitted). By looking only at terms of leases KCP could enter into, and not factoring in growth that could occur if the property was unencumbered by those leases, KCP's expert failed to determine the market value of the fee-simple interest in the property. Accordingly, the tax court did not clearly err in rejecting that analysis.

III.

To determine revenue lost due to vacant storefronts, the tax court had to determine vacancy rates. The tax court rejected the vacancy rates used by KCP's expert because they "assume[d] a constant 15% vacancy rate across the entire analysis period," even though the record indicated that vacancy rates were increasing during the years at issue. KCP claims this was erroneous because the tax court "misunderstood" its appraiser's vacancy rate.

KCP argues that its expert used two types of vacancies to determine the total vacancy rate. While KCP concedes that the expert "forecasted [a] stable market vacancy," he also forecasted increasing "lag vacancies," which accounted for the increasing vacancy rates noted by the tax court. In support of this argument, KCP directs us to the portion of its expert's report where the expert discusses vacancy rates.

Having reviewed the report, we are unpersuaded. Although the expert does explain that vacancy rates may be split into "general vacancy" and "absorption and turnover vacancy," the report appears to assume a total vacancy rate of 15 percent per year, which is deducted from the calculated absorption and turnover vacancy to determine the general vacancy for each year. Consequently, although the absorption and turnover vacancy may have varied from year to year, the total vacancy rate did not. This conclusion confirms the tax court's finding that the vacancy rates used by KCP's expert were not supported by the record (which showed varying vacancy rates), and therefore the tax court did not clearly err in rejecting KCP's expert's rates.

IV.

To determine the market value of the property, the tax court had to estimate how much the property would sell for at the end of the DCF period. The Appraisal of Real Estate , supra at 517. To do so, the income for the property for the year following the end of the DCF forecast is estimated, and then a terminal capitalization rate and a discount rate are applied to the estimate. Id. KCP argues the rates used by the tax court were clearly erroneous.

Three types of capitalization rates are relevant to this discussion; one can be calculated for a particular property based on known market conditions, and the other two are estimated to predict future value. First, a "going-in capitalization rate" is a calculated rate, determined by dividing a property's net operating income for the first year after purchase by the present *782value of the property. Id. Second, a "terminal capitalization rate" is an estimated rate, applied to the expected net income of a property for the year immediately following the end of a DCF projection period, to derive the resale price or value of a property. Id. Such rates are "generally, though not necessarily, higher than the going-in capitalization rate" to "reflect the reduction in the remaining economic life of the property and the greater risk associated with estimating net operating income at the end of the projection period." Id. Finally, a "discount rate" is an estimated rate, applied to convert the future capitalization into a present value. Id. at 510.

Reviewing the record, we conclude that the terminal capitalization rate and discount rate selected by the tax court are not clearly erroneous. To determine the discount rate, the tax court relied primarily upon surveys of capitalization rates. Those surveys showed discount rates between 9.25 percent and 11.8 percent, with rates generally (but not uniformly) increasing between 2010 and 2011 and decreasing between 2011 and 2012. The rates used by the tax court were both within the range of the rates found by the surveys and followed the trend over time from 2010 to 2012 (by remaining steady between 2010 and 2011 and decreasing between 2011 and 2012).

Having determined the discount rate, the tax court then reduced the rate, because both experts agreed that the terminal capitalization rate should be less than the discount rate. Again, relying primarily on capitalization rate surveys, the tax court determined the deduction to be made each year. Those deductions are supported by the evidence and are not clearly erroneous.

KCP disagrees for two reasons. First, it argues that, before determining the discount and terminal capitalization rates, the tax court noted that the most comparable property had a going-in rate of 9.9 percent. Based on evidence presented by the County, KCP argues that the terminal capitalization rate should have been at least 0.50 percent higher than that going-in rate for all 3 years, and the discount rate should have been a further 0.50-0.75 percent higher than that. Thus, according to KCP, the discount rate for all 3 years should have been at least 11.00 percent, as opposed to the 10.75 percent, 10.75 percent, and 10.50 percent values used by the tax court.

KCP misstates the County's evidence. Although the County's expert did use a terminal capitalization rate that was 0.50 percent higher than the 9.9 percent going-in rate, his expert report noted that increases between 0.25 percent and 1.00 percent are typical; not that a 0.50 percent increase was required. The tax court's terminal capitalization rate for 2010 (10.25 percent) was at least 0.25 percent higher than the 9.9 percent going-in rate. Therefore, KCP's argument fails regarding that year. Moreover, for 2011 and 2012, it is not clear from the record that going-in rates would have remained at 9.9 percent. On the contrary, the County's survey evidence showed that terminal capitalization rates declined over that time period, suggesting that going-in rates likely also would have declined. Accordingly, we reject KCP's argument that the rates used by the tax court were mathematically incompatible with a going-in rate of 9.9 percent.

Second, KCP argues that the tax court used an incorrect procedure to determine the discount and terminal capitalization rates because, rather than determining a discount rate and then subtracting basis points to determine the terminal capitalization rate, the proper procedure (according to KCP) is to determine the terminal capitalization rate and then add basis points to determine the discount rate. We disagree.

*783The tax court had survey evidence of the predominant terminal capitalization and discount rates in the market at the relevant times. Although The Appraisal of Real Estate states that discount rates can be calculated from the terminal capitalization rate, that possibility does not render rates determined by survey evidence invalid. Because evidence in the record supports the rates used by the tax court, we conclude that those rates are not clearly erroneous.

V.

In addition to determining the cash flow of the shopping mall, the tax court assigned value to a portion of the mall's parking lot (the "outlot") on the basis that the outlot could be sold and developed. KCP argues that such a sale is not possible, due to zoning restrictions. We review the factual question of whether the outlot could be developed for clear error, but to the extent that question turns on issues of law, our review is de novo. Antonello v. Comm'r of Revenue , 884 N.W.2d 640, 643-44 (Minn. 2016).

Before the tax court, KCP argued that the outlot could not be sold and developed because the zoning code for the City of Hastings requires one parking space per 200 square feet of retail space, or 647 parking spaces for the 129,475 square feet of net rentable area of KCP's mall. KCP further argued that the mall had only 700 parking spaces, including those occupying the outlot, and thus developing the outlot would have reduced the number of mall parking spaces below the requirements of the zoning code. See Hastings City Code tit. XV, app. B.

The tax court recognized this limitation, but concluded that it could nevertheless assign sale-and-development value to the outlot because "there remains the possibility that a buyer of the outlot could seek a variance," or in the alternative "a buyer of the outlot could enter into a reciprocal easement with KCP ... that would allow customers of a business constructed on the outlot to share parking with customers of the main shopping center, and vice versa." We conclude that the tax court clearly erred in making these speculative findings.

We have previously held that "[e]vidence of value for uses prohibited by an ordinance may be introduced and considered only where there is evidence showing a reasonable probability that the ordinance will be changed in the near future." Hedberg & Sons Co. v. County of Hennepin , 305 Minn. 80, 232 N.W.2d 743, 750 (1975) (emphasis omitted) (citation omitted) (internal quotation marks omitted). Here, there was no evidence that the zoning code would be changed-or that a variance would be granted-by the city. Further, any development of the outlot would come with its own parking requirements. See Hastings City Code tit. XV, app. B (listing parking-space requirements for commercial property developments). Without any evidence that the reduced parking availability would satisfy the parking requirements for both KCP's shopping mall and the commercial development on the outlot, the tax court nonetheless assumed a reciprocal easement would satisfy the parking requirements of the zoning code. But the mere possibility that KCP and the outlot purchaser could enter into a reciprocal easement is not "a reasonable probability" that such an easement would allow KCP to meet the zoning code's requirements if it sold the outlot. Accordingly, we conclude that the tax court clearly erred when it concluded that the outlot should be valued based on the possibility of its sale and development.

We recognize, however, that the outlot may add some value to the mall. For example, a mall with 700 parking spaces might be worth more than a mall with 675 *784parking spaces. The tax court, however, is in the best position to make this factual determination. Therefore, we remand for the tax court to determine if the outlot adds any value to KCP's property when the use of the outlot is limited to parking, as opposed to commercial development.

VI.

As part of its value determinations, the tax court determined the area of the mall. The tax court used a gross building area of 153,749 square feet.5 KCP argues that this value was clearly erroneous in light of a stipulation by the parties to a different gross building area.

"[P]arties may stipulate as to what evidence shall be considered by the trier of fact." Lappinen v. Union Ore Co. , 224 Minn. 395, 29 N.W.2d 8, 17 (1947). "Where the parties stipulate as to the facts, the effect of the stipulation is to take the place of evidence." Anderson v. Anderson , 303 Minn. 26, 225 N.W.2d 837, 840 (1975). "As long as a stipulation remains in effect it is binding not only on the parties, but on both the trial and appellate court." Lappinen , 29 N.W.2d at 17. "Because such a stipulation excludes consideration of other evidence, other evidence may not be considered, even though it may find its way into the record, except where it clearly appears that the parties have abandoned the stipulation." Id.

In its findings, the tax court identified two reasons for departing from the stipulated building area. First, it concluded that KCP had abandoned the stipulation because KCP's expert's report used 153,749 square feet. We disagree.

We have previously considered whether a stipulation was abandoned in Gethsemane Lutheran Church v. Zacho , 253 Minn. 469, 92 N.W.2d 905 (1958). In that case, during a pretrial hearing the parties stipulated that the intervenors were members of the Lutheran High School Association of Greater St. Paul. Id. at 912-13. At the start of the trial, however, the first argument raised by the association was that the intervenors "never were members" of the association. Id. at 913. Indeed, a review of the transcript of that hearing shows that a substantial portion of the testimony offered went to the issue of whether the intervenors had ever been members of the Association. Accordingly, we concluded that "the stipulation entered into at the pretrial conference was abandoned" because "from the very beginning of the trial [the] issue was litigated by both parties." Id.

Here, in contrast, the parties stipulated to the gross building area on February 12, 2014. But the expert report that allegedly abandoned the stipulation was written (and had been filed with the tax court) before that stipulation was reached. It is obvious that a party cannot abandon a stipulation that has not yet been made. Moreover, even if the date of trial were used as the purported date of abandonment,6 we conclude KCP did not abandon its stipulation by offering its expert's report as evidence. We favor stipulations "as a means of simplifying and expediting litigation."

*785Maranda v. Maranda , 449 N.W.2d 158, 164 (Minn. 1989). As Lappinen acknowledged, evidence contradicting a stipulation "may find its way into the record," even after the parties have agreed to it. 29 N.W.2d at 17. Thus, the mere inclusion of a gross building area different than the stipulated area in KCP's expert's report is not enough to show abandonment. At a minimum, there must be some indication that the fact subject to stipulation was subsequently litigated in some way. A review of the transcripts in this case shows that the gross building area was not one of the litigated issues.

Second, the tax court also concluded that it was required to depart from the stipulated gross building area because we "affirmed [the tax court's] findings as to the size of the building" in our review of the 2014 order. This conclusion reads too much into the first appeal.

In KCP's first appeal, the parties disputed whether the gross building area or the gross leasable area was the appropriate area to use for appraisal purposes. We resolved that dispute in favor of the County, concluding that the tax court did not clearly err by using the gross building area. KCP Hastings, LLC , 868 N.W.2d at 273. We explicitly noted, however, that the parties did not ask us to resolve the issue of whether the tax court erred in deviating from the stipulated measurement. Id. at 270 n.1. Consequently, this specific issue was neither presented to us nor decided in our former opinion. See John Wright & Assocs., Inc. v. City of Red Wing , 256 Minn. 101, 97 N.W.2d 432, 434 (1959) (concluding that the district court's order following an appeal was not inconsistent with our opinion in the appeal because the issue was not "presented to the court or decided under our former opinion"). We therefore conclude that our prior opinion did not preclude the tax court from using the stipulated gross building area. And because KCP did not abandon that stipulation, both we and the tax court are bound by it. Accordingly, the tax court clearly erred by using a different gross building area.

VII.

After determining the value of KCP's mall using the sales and income approaches, the tax court had to decide the weight to be given to each approach to make a final valuation determination. KCP argues the tax court erred in the weight it assigned to each approach. We review the weight given by the tax court to differing valuation approaches for an abuse of discretion. See KCP Hastings, LLC , 868 N.W.2d at 276 (concluding the tax court abused its discretion by failing to place any weight on the income approach).

After reviewing the evidence, the tax court found that the income approach should be given a weight of approximately 70 percent and the sales-comparison approach should be given a weight of approximately 30 percent. KCP argues that, in our prior opinion, we noted that the income approach is often afforded "over-riding weight," and that the 70 percent/30 percent division used by the tax court failed to do so. In support of this position, KCP relies on Montgomery Ward & Co., Inc. v. County of Hennepin , 482 N.W.2d 785 (Minn. 1992). In that case, we concluded that giving 75 percent weight to the income approach, 15 percent weight to the sales-comparison approach, and 10 percent weight to the cost approach gave overriding weight to the income approach. Id. at 791. KCP argues that Montgomery Ward set the floor for determining overriding weight.

KCP overreads Montgomery Ward . Nowhere in that case did we state or otherwise suggest that 75 percent was the minimum to constitute overriding weight. On the contrary, we noted that "real estate appraisal is at best an imprecise art." Id. And we have held in other cases that the *786weight to be given to each value approach will vary from case to case based on the quantity and quality of the available data. See Nw. Racquet Swim & Health Clubs, Inc. v. County of Dakota , 557 N.W.2d 582, 587 (Minn. 1997). We likewise decline to establish a hard limit on what constitutes overriding weight in this case. Instead, because KCP has not presented any argument leading us to the firm conviction that the weights used by the tax court were incorrect, we conclude that the tax court did not abuse its discretion when it gave 70 percent weight to the income approach and 30 percent weight to the sales-comparison approach.

CONCLUSION

For the foregoing reasons, we affirm in part, reverse in part, and remand to the tax court, with directions to correct the valuation for KCP's property by using the stipulated gross building area and limiting the value added by the outlot to use as additional parking.

Affirmed in part, reversed in part, and remanded.