OPINION
BRUGGINK, Judge.
This is a class action in which the plaintiff class members seek' compensation for the imposition of an easement for recreational trail use on their land. The court has previously ruled that the government did in fact take such an easement without compensation. Glosemeyer v. United States, 45 Fed.Cl. 771 (2000) (including the consolidated case of Moore v. United States, No. 93-134L). Because nearly 300 landowners opted into the class, the parties agreed to identify categories of property. The court could then determine compensation for a representative parcel from each category, in hopes that the parties could apply this ruling to the remaining parcels. Consequently, trial was conducted in St. Louis, Missouri from November 12 through 22, 2002 on thirteen representative parcels. At the conclusion of trial, the court issued a bench ruling. What follows is a heavily-edited version of that ruling.
DISCUSSION
At issue is the appropriate compensation for easements imposed on the respective parcels when the United States precluded plaintiffs from reclaiming control over the surface of their property upon what would otherwise have been the abandonment of a pre-existing railroad easement supporting the Missouri-Kansas-Texas (MKT) rail line. The circumstances of that abandonment and the subsequent imposition, through Congressional action, of the Katy Trail — now operated by the State of Missouri Department of Natural Resources (MDNR), is set out at Glosemeyer, 45 Fed.Cl. 771.
Before assessing the testimony and exhibits, we note at the outset that the parties and the court conducted a site visit of the representative parcels. That inspection was very helpful in giving context to the other evidence. The following facts were obvious from the inspection.
For a substantial majority of its length, the Katy Trail, like the old rail bed, follows the [749]*749course of the Missouri River. The course apparently attempts to take advantage of the pancake-flat terrain of the river’s flood plain, while at the same time minimizing the risk of flooding by occupying the plain’s extreme northern edge. On that edge, it typically bumps abruptly up against a bluff marking the beginning of hilly terrain. The bluff varies in height from negligible in a few places to a steep, sharp edge, 100 feet high. At most places it appears to be at least 40 to 50 feet high.
Because for most of its length the trail runs through flood plain, it is typically elevated on a substantial berm, composed of hard fill. In some places the elevation difference was substantial, perhaps ten or fifteen feet. Frequently it is paralleled by a ditch on one or both sides. Though the trail starts in the highly urbanized suburbs of St. Louis, it runs through agricultural land for most of its length.
In some cases, the trail bisects land owned by one of the plaintiffs. In many places, the Katy Trail follows a path very close to state Highway 94. On occasion, the road and the trail or a stream and the trail both bisect the same parcel. The trail, therefore, varies in the degree to which it creates a physical impediment to a vehicle wishing to go across from one adjoining parcel to another. In some places, crossing could be extremely difficult, but in most places it would be relatively easy. The evidence at trial was that the MDNR will permit adjoining agricultural users to establish regular crossing points. Many such agricultural crossings were apparent. Upon payment of a modest fee, MDNR would also permit new crossings in residential areas.
The parties offered both fact and expert witnesses to present their competing views as to the value of the easement taken. The court begins by discussing the experts and their reports. Edward Dinan testified on behalf of plaintiffs and Kevin Nunnink testified on behalf of the government. While both experts are plainly highly qualified to give expert evidence on valuation, for reasons set out below, the court concludes that, with one significant exception, it found Mr. Nunnink’s overall approach and the particulars of his appraisals more convincing. They were more usable to the court, more internally consistent and coherent, and fraught with fewer technical errors.
It is settled that a landowner claiming a physical, partial taking of property is entitled to the difference in value before and after the taking. In this case, each landowner suffered a partial taking in two senses. First, the new easement is less than the fee estate. Second, the new easement potentially negatively impacts a larger piece of land than the right of way itself. This is known as severance damage and constitutes a pedigreed element of compensation, assuming it can be proved. See Julius L. Sackman, Nichols on Eminent Domain § 16-3 — 16-7 (3d ed.). Both parties approached the calculation by considering each of these two elements. They thus made a “before and after” calculation — a determination of the fair market value of the entire affected parcel as if the easement did not exist and then another determination in light of the taking.
Though in agreement on the basics, the parties still disputed several particular points. First, Mr. Dinan’s appraisals valued each property as two separate parcels, the easement “corridor” and the remaining land, whereas Mr. Nunnink appraised the totality of the property as one parcel. Second, Mr. Dinan applied a “corridor factor” of 125% to the “before” value of the land underlying the right of way. Mr. Nunnink did not. Third, Mr. Dinan’s “after” calculation included no value for the land underlying the right of way. Mr. Nunnink, on the other hand, sometimes considered this land to have 50% of its value in the “after” calculation. Fourth, the experts disagree as to the actual .per acre valuation of the property. Finally, Mr. Dinan generally awarded severance damages and Mr. Nunnink generally did not. Each of these issues is dealt with in turn.
First, Mr. Dinan’s analysis treats the land underlying the right of way as if it is a distinct parcel with potentially a different highest and best use from that of the balance of the property. He thus calculated the per square foot or per acre value of this strip independently from the rest of the land. In his view, the highest and best use of the [750]*750parcel underlying the easement was as a corridor. Mr. Dinan has some experience in valuing corridors and he explained that the potential corridor could be used, for example, for light rail, fiber optic cable, or a gas pipeline. Because of this possibility, he first determined a value for the land in its natural condition, and then applied a 125% multiplier to account for the possibility of corridor use. Mr. Dinan on several occasions valued the right of way strip, prior to the adjustment factor, at a lower value than did Mr. Nunnink.
Mr. Nunnink, on the other hand, took a different approach and assumed that the right of way had the same highest and best use as the balance of the property. He thus calculated an overall “before” value as if the acreage underlying the right of way had the same value per square foot or per square acre as the rest of the property. He also contended that no upward adjustment was appropriate for the possibility that adjoining right of way parcels might be reassembled and used for a corridor. Ironically, many of his “before” values for this strip of land are higher than Dinan’s corridor-adjusted figures.
We agree with Mr. Nunnink that the record does not support a separate highest and best use for the corridor portion and we decline to adopt Mr. Dinan’s approach to the question of highest and best use. We recognize that the right of way frequently was physically distinct from the balance of a particular parcel. This was particularly true when the most desirable portion of the property was above the bluffs and the right of way was a physically inaccessible strip at the foot of the bluffs. We believe, however, that the record does not support a separate “corridor” highest and best use for the right of way parcel.
Mr. Dinan testified that appraisers have applied multipliers of six or seven times to certain corridors. These corridors, however, were either previously assembled or specifically designated for other uses. Examples of other uses include acquisition of a former railroad bed for a trail or expansion of light rail. The portions of the former railroad right of way in this case, however, remain legally unconnected. A purchaser of the right of way wishing to introduce light rail, or a gas pipeline, or a fiber optic cable, for example, would have nothing other than a hope of assembling sufficient adjoining parcels to put together a corridor. In effect, what is present here is just a recollection that there once was a corridor in this environment, evidence of which still exists in the form of extensive clearing and diking. While we appreciate that Mr. Dinan used a very conservative multiplier, there was no market data to support it under these circumstances. In the absence of better proof that there was specific demand for a corridor use in this vicinity, or market data showing that even the possibility of reassembling a corridor increases value, we accept Mr. Nunnink’s assessment that purchasers would not pay a premium to any single landowner because of the uncertainty of putting together a useable corridor. We thus conclude that the land underlying the right of way should be valued at the full value of and at the same highest and best use as the adjoining land.
Though we agree with Mr. Nunnink’s analysis that the land under the easement is not entitled to separate valuation as a corridor, Mr. Nunnink’s analysis contemplated another step. In his “after” valuations of the landowner property, he sometimes discounted that parcel by 50%. He made a determination about the right of way parcel which he characterized in the following ways: the extent to which it was “useable”; the extent to which the parcel had “impact” on the adjoining parcel; whether the “productivity” of the adjoining parcel was negatively affected; or whether it could have been “reclaimed” absent the trail. In short, Mr. Nunnink added a step which either substituted for his severance analysis, discussed below, or which revisited the question of whether the land underlying the right of way should be valued differently than the balance of the property. If the land was not useable, could not be reclaimed, did not impact the adjoining parcel, or did not negatively affect the adjoining parcel, he discounted direct damages by 50%.
We thus disagree with Mr. Nunnink’s downward adjustment to the corridor portion. It is either inconsistent with an under[751]*751lying assessment of the per acre value of the land or else it duplicates a severance determination. We agree with plaintiffs that no such adjustment is appropriate and that the right of way parcel should be diminished 100% in the “after” analysis because the landowners had no effective remaining use of the property.
When asked what utility remained to the landowners, the government suggested that the landowners could still use the land underlying the easement to determine the percentage of the building “footprint” tolerated on a parcel by zoning officials, because the total green space available would include the right of way. Given the large size of even the suburban parcels at issue here, we find that this retained “use” is too ephemeral to assign any value and the government offers no other explanation of the landowners’ remaining value. The landowners, in effect, retained no functional utility from the land underlying the easement. Mr. Dinan explained that an easement for trail use is tantamount to a taking of the entire fee. Even Mr. Nunnink conceded that if the right of way imposed had been for a road, the diminution would be 100% and we see no reason to treat the outcome differently here. Accordingly, the parties should value the land underlying the right of way in the “after” calculation at zero.
The court also must choose between the expert’s competing calculations of the price per square foot or price per acre of the subject property. In determining the “before” value of the entire parcel, both appraisers, with the exception of one parcel, exclusively relied on a comparable sales approach. Frequently, the values were very similar and the two appraisers often utilized the same comparable sales. As indicated above, however, the court was more impressed with Mr. Nunnink’s determinations of value and therefore we used his figures. We judged his analysis to be more comprehensive, rigorous and transparent. In part, this was due to the fact that on parcels with residences, Mr. Dinan’s analysis conflates land value into the value of the residence, making his calculations less transparent.
Lastly, the court turns to the issue of severance damage. As indicated above, severance damage is more than just theoretically available. Indeed, Mr. Nunnink concluded that the Heck Industrial Parcel had suffered such impact. As to the other parcels, however, we find that severance damage was not proved.
Plaintiffs offered two kinds of severance damage evidence. The first was the testimony of landowners or persons familiar with' particular parcels. This evidence was completely credible. The landowners and others were sincere and persuasive in explaining to the court that their proximity to the trail on occasion resulted in trespass, litter, loss of privacy, .fear of liability, and even danger. The most compelling evidence concerned the largest of the representative parcels, that belonging to Ms. Dorothy Moore, the lead plaintiff. Other landowners testified to incursions and inconveniences similar in kind but of substantially less degree than those affecting Ms. Moore’s land. For that reason, the court begins the severance damage analysis in the context of her property.
Ms. Moore and her husband, along with associated family or corporate interests developed several hundred acres of upland along the Missouri River over the past fifty years. The parcel at issue consists of 168 acres of woods and pastureland. Her property is relatively near the confluence of the Missouri and Mississippi rivers in what is now a highly-developed area within greater St. Louis. There is no flood plain at this location. Hence the property has a commanding view directly onto the river, although access from any particular location is severely restricted by high bluffs and dense vegetation. Ms. Moore lives on the property, operates a riding stable and breeds basset hounds. In addition, she has peacocks, llamas and other animals. The Katy Trail runs at the lowest point of her property along the river, with a small portion of land between the trail and the river.
This little Eden is so attractive that people are frequently drawn off the trail to approach the animals. People have also trespassed through the main entrance to her property, searching for a way to get to the trail. Ms. Moore formerly kept the slope below her house cleared in order to have a [752]*752better view of the river. She now has allowed it to grow up to give her some privacy from the people on the trail. Of even greater concern is the fact that Ms. Moore’s land is one of the few remaining large, undeveloped river tracts in the St. Louis area. It thus attracts hunters who enter onto her land without permission. Some have been extremely assertive and unresponsive to threats of arrest. One of Ms. Moore’s Mends, Paul Gragg, has been fired upon in his attempts to police her property to keep off hunters.
As to Ms. Moore, Mr. Dinan testified that the highest and best use of her land was as a “residential development.” That is, the marketplace would assign a value to the land based on its potential development into a high-end, large-lot subdivision and not on its current use. In this respect, Mr. Nunnink was in full agreement.
Mr. Dinan, however, reduced the value of the undeveloped portion of Ms. Moore’s land by two percent because it is “negatively impacted by the trail.” Mr. Dinan readily conceded that his two percent adjustment was a judgment call, but he based this reduction on “discussions with the market” (other brokers) and discussions with the landowners. In the narrative portion of his report, he explains that the downward adjustment was due to the “traffic on the trail, additional risks associated with public access that is not monitored, the incompatibility of farm operations with recreational uses, as well as the general nuisance and lack of control by the property owner.” (emphasis added).
Mr. Dinan also supported his severance figures with a paired sales analysis, Plaintiffs’ Exhibit R. Mr. Dinan created a matrix of five paired sales of agricultural land. One series of the agricultural sales abutted the Katy Trail and one series did not. In both series, he adjusted the actual sales prices in an attempt to approximate a hypothetical model sale of 100 acres. He used adjustment increments of five percent to adjust for a number of subjective factors. The factors included size, location, topography, zoning and utility. In light of his adjusted figures, he concluded that abutting parcels, on average, sold for 7.9% less per acre than comparable non-abutting land.
Mr. Nunnink, however, disagreed with Mr. Dinan and did his own paired sales analysis, Defendant’s Exhibit 1. It consists of five case studies of home or residential lot sales. Studies One and Two involved homes or lots within Quinlan Acres, a small subdivision in Booneville, Missouri, a small town toward the western terminus of the trail. As it happens, Quinlan Acres in its undeveloped condition is one of the representative parcels. After the date of the taking, the Quinlans developed a portion of their land into the subdivision used by Mr. Nunnink for his comparable sales. Some of the lots abut the trail; others do not. Mr. Nunnink concluded that there was no meaningful difference in the sales prices of the abutting and non-abutting houses or lots.
Mr. Jerome Quinlan, owner of the land in its undeveloped state, testified concerning trespassers, loss of privacy and litter. He also testified, in connection with the comparability of the paired sales analysis, that lots away from the trail were fundamentally different from lots abutting the trail because the former sloped considerably and required a more expensive foundation. He also indicated that lots away from the trail were worth more than lots next to the trail.
Mr. Nunnink explained his disagreement with Mr. Quinlan and testified that the proximity of the trail to Quinlan Acres would not detract from its highest and best use; indeed it might make the land more attractive to wealthier purchasers interested in ready access to a recreational trail. In his report, he recites his conclusion that the “Katy Trail will affect the subject’s neighborhood by providing a significant recreational amenity to residents.”
We find it unnecessary to resolve this difference in opinion about Mr. Nunnink’s Case Studies One and Two because the remaining case studies seem more useful to contrast the presence of the trail as the only meaningful distinction. Case Studies Three, Four and Five all involved parcels in St. Charles County, the eastern terminus of trail. They consist of pairings of virtually identical houses in a subdivision abutting the trail. The only [753]*753significant difference in the pairings is whether the house is actually on the trail. Mr. Nunnink could find no meaningful difference in price.
Mr. Nunnink also expressed serious misgivings with Plaintiffs’ Exhibit R and cited two reasons why it was not a true matched pairs analysis. First, instead of adjusting the properties to a corresponding paired sale, the properties were adjusted to conform to a hypothetical ideal model. Second, the sales were not truly paired, because too many adjustments were made to the properties. In addition, after examining the individual parcels, he concluded that the parcels in Plaintiffs’ Exhibit R had too many unique features to ever be meaningfully matched. For example, one parcel was under a power line, one had a ditch through the middle, one included a borrow pit site, and another was crossed by a highway.
We accept Mr. Nunnink’s criticisms of Plaintiffs’ Exhibit R. We note also that Mr. Dinan’s adjustments were done in gross increments which admittedly were highly subjective. This means that the resulting 7.9% difference easily could have been inadvertently skewed. Finally, we note that there appeared to be virtually no repeatable pattern to the assignment of particular percentages of severance impact to any particular parcel. As we noted above, the Moore parcel, at least insofar as its present use is concerned, appeared to be the most physically affected by the trail. Yet, Mr. Dinan assigned only a 2-3% severance impact to that parcel. As Mr. Nunnink explained, paired sales studies are inherently problematic, and should only be used when the evidence is reliable and the results consistently show differences above the level of background static and error.
We also disagree with Mr. Dinan’s narrative explanation of the reasons for severance damage for the Moore parcel. Mr. Dinan’s narrative explanations in his reports are not individualized for each valued property. For example, Ms. Moore does not conduct a farm operation. Mr. Dinan’s narrative obviously was crafted in connection with another parcel and applied here without further consideration. The same language appears verbatim in the Hollrah appraisal. Moreover, in the Jensen appraisal, Mr. Dinan uses virtually identical language that the “site will be negatively impacted by the trail conversion” for virtually the same reasons. Yet in the Jensen appraisal, Mr. Dinan did not assign severance damages.
Mr. Nunnink acknowledged that severance damages might be appropriate for properties bisected by the trail. While some of the parcels were bisected, in no instance was the court persuaded that it materially affected the accessibility or use of the remaining parcel. Frequently, the trail was only one of several barriers. Highway 94, for example, often paralleled the trail in bisecting property-
Additionally, there was no convincing evidence that isolated trespassing or litter would materially affect the value of the land. Though several witnesses testified regarding these problems, many of these parcels consisted primarily of high quality bottom land. Such land, according to Mr. Nunnink, is sold exclusively on the basis of productivity per useable acre. He would expect that intermittent litter and trespassing would be irrelevant to potential yield.
In sum, we find there is no reliable, objective evidence supporting severance damages. While Mr. Dinan’s intuitive judgment that the trail has an impact is entitled to some weight, we find it no more reliable than Mr. Nunnink’s contrary assessment, which has some support in a priced sales analysis. Moreover, Mr. Dinan’s adjustments are particularly problematic in view of the inconsistency in the narrative portions of the reports and the lack of explanation for substantial differences in the precise reduction in value.
None of this minimizes the testimony of the individual fact witnesses. Indeed, it indirectly confirmed the court’s prior determination that the current use is different in kind from a railroad use. In the final analysis, however, the court can only award severance damages if there is some reliable proof that these physical intrusions, concerns, and annoyances have actually translated into loss in market value. This analysis does not center on the perceptions of particular landowners. [754]*754Instead, the question is whether the buying public, in light of the taking, would view the property as less attractive. Any other basis for an award would amount to tort damages. Application to particular parcels
In light of the above rulings, we make the following more particularized findings as to the individual parcels.
Jensen:
The parties agree on the size of the right of way parcel: 0.17 acres. We accept Mr. Nunnink’s assessment as to valuation, albeit not his 50% downward adjustment. On this appraisal in particular, Mr. Dinan’s valuation of the land itself was not internally consistent. Accordingly, we assign a decrease in value to the property of $6,000.
Moore:
In this instance, Mr. Nunnink’s determined the right of way parcel to be 6.06 acres and assigned it a value of approximately $15,000 per acre. Mr. Dinan’s figure for size is 5.88 acres, while his per acre value is substantially lower, at approximately $1,000 per acre. As to size, we have no reason to treat one of these figures as more accurate in this instance. Accordingly we use an averaged figure of 5.97 acres. We hold the government to its higher assessment of value because it is consistent with the court’s acceptance of why no severance damages are awarded. The trail adds to the value of the upland; not the contrary.
Rehmeier Farms:
As to size of parcel, we accept Mr. Nunnink’s larger figure of 1.9 acres. Mr. Dinan assumed a uniform 50 foot wide right of way, whereas it was in fact wider in places. We also accept Mr. Nunnink’s higher value, $2,500 per acre, of land, as explained above. Harrison:
The parties agree on the size of the right of way: 54,100 square feet. We accept Mr. Nunnink’s per acre price of $1,850.
Hollrah:
Lacking any better way to determine which expert is correct on the size of the right of way, the court averages the two: 0.91 acres. The court accepts Mr. Nunnink’s valuation of $2,600 per acre.
Heck Industrial:
We rely exclusively on Mr. Nunnink’s valuation here. This is not prejudicial to plaintiffs, as Mr. Dinan’s appraisal would yield a smaller recovery. Mr. Dinan’s appraisal, which included a cost and income approach, as well as comparable sales, was very difficult to follow and required several material post-appraisal adjustments. We thus assume a right of way size of 9,856 square feet and a value of $1.75 per square foot. In addition, unique to this parcel, we accept Mr. Nunnink’s assessment that the balance of this small industrial parcel suffered impact damages measured by the cost of maintaining a fence to keep out trespassers, namely, $2,000. Heck Farm:
Mr. Nunnink declined to quibble with Mr. Dinan’s figure on the size of the right of way, 8.91 acres. We use Mr. Nunnink’s valuation for the land, $1,425 per acre.
Lensing:
The parties agree on the size of the right of way: 1.6 acres. We use Mr. Nunnink’s figure of $1,600 per acre.
Richard:
The parties agree on the size of the right of way: 6.2 acres. We adopt Mr. Nunnink’s higher figure on value of $775 per acre. Holland:
The parties disagree on the precise size of the parcel, and neither figure is plainly incorrect. Accordingly, we average the size of the two: 60,000 square feet. We assign Mr. Nunnink’s value of $.02 per square foot. Quinlan:
The parties’ figures on the size of the right of way differed here as well. There is no clearly correct result, so we take the average of the two: 1.39 acres. We use Mr. Nunnink’s value of $2,500 per acre.
Heaton:
The parties are in agreement on the size of the right of way: 2.25 acres. We use Mr. Nunnink’s value of $650 per acre.
Farmers Livestock:
The parties agree on the size of the easement: 1.47 acres. We use Mr. Nunnink’s figure for value of $7,800 per acre.
[755]*755CONCLUSION
The parties are directed to consult to attempt to determine the precise recovery for each representative plaintiff, including statutory interest from the date of taking to a proposed date of judgment. On or before January 24, 2003, the parties shall file a joint status report indicating their preferences with respect to resolving the claims of remaining plaintiffs, and whether a Rule 54(b) judgment should be entered as to representative plaintiffs.