(concurring in part and dissenting in part):
Although I concur in parts IV and V of the majority opinion, I must dissent from its affirmance, without modification, of the lower court’s award of forage loss damages from trespass by cattle and sheep.
The partition of the Gillmor family’s land was an obvious harbinger of this trespass action. The 1981 grazing season was already underway when the partition decree was entered. The parties in possession did not have time to fence miles of new boundary lines. Neither the parties nor their livestock knew the respective on-site boundary locations. During the 1981 grazing season, both Stephen and Edward undertook livestock operations on the partitioned properties. The livestock engaged in their traditional grazing and drifting patterns without regard to lines created by the decree from the courthouse. If the ranch resource was to be utilized, mutual trespassing and mixing of livestock were inevitable.
As could have been anticipated, both parties sued for relief. During closing arguments at the end of a four-day trial, Stephen’s counsel submitted to the lower court and counsel a document entitled “Plaintiff’s Trial Brief on the Issue of Damages,” containing the damage calculations set forth in the majority opinion. In its Memorandum Decision, the trial court unquestioningly adopted Stephen’s calculations, and the factors itemized in them became the only findings underlying its ultimate conclusions of certain dollar amounts of damage entitlement.
I. RESPONDENTS’ BURDEN AT TRIAL
I wholeheartedly endorse the general proposition that reasonable factual inferences may be logically deduced “from the establishment of other facts.” Wyatt v. Baughman, 121 Utah 98, 101, 239 P.2d 193, 198 (1951). See Ballow v. Monroe, 699 P.2d 719 (Utah 1985) (sufficient evidentiary basis is required before negligence can reasonably be inferred); Owen v. Burcham, 100 Idaho 441, 599 P.2d 1012, 1019 *467(1979) (inference is unreasonable if based on mere speculation and conjecture); Murray v. T.W. Dick Co., 398 A.2d 390, 392 (Me.1979) (inference must be based on probability and not on mere possibilities). There is, however, no evidence in the record to support many of the foundational facts that are indispensable to this Court’s affirmance, in total, of an award of damages for forage loss from: 169 cattle trespassing continuously for 75% of 10 months; 217 cattle trespassing continuously for 75% of 5.73 months; and 1,125 sheep trespassing continuously for 75% of 8 months.
In my opinion, the trial court’s ultimate findings regarding damages — to the extent they are unsupported by evidence in the record or by reasonable inferences drawn from evidence in the record — are clearly erroneous. The majority nonetheless wholly embraces the lower court’s damage findings and attempts to fill the evidentiary void with inferences that lack a factual basis and with unreasonable inferences built upon inferences.
[W]hen an inference of the probability of the ultimate fact must be drawn from facts whose existence is itself based only on an inference or a chain of inferences ... all prior links in the chain of inferences must be shown with the same certainty as is required in criminal cases, in order to support a final inference of the probability of the ultimate fact in issue. ... [Pjrior inferences must be established to the exclusion of any other reasonable theory rather than merely be a probability, in order that the last inference of the probability of the ultimate fact may be based thereon. This rule is not based on an application of the exact rules of logic, but upon the pragmatic principle that a certain quantum of proof is arbitrarily required when the courts are asked to take away life, liberty or property.
State v. Hall, 105 Utah 162, 145 P.2d 494, 497 (1944) (quoting New York Life Ins. Co. v. McNeely, 52 Ariz. 181, 79 P.2d 948, 954-55 (1938)). Stephen, by sheer force of ar-gumentive repetition has the majority believing that “Edward admitted he stocked his one-quarter share of the land with the same number of livestock for which he previously required the entire land ” (emphasis added). That idea was proposed in counsel’s question at trial, but it is not found in any testimony or admission of Edward. The record shows only that Edward stated he operated in 1981 with the same livestock numbers as in previous years. Thus, the majority’s initial or foundational fact — that Edward’s livestock numbers automatically equal full use of the forage resource — can only be reached by a far-fetched inference.
The majority’s other foundational fact is the reported shortage of “lamb feed” on Edward’s land, not on Stephen’s land. The undenied evidence was that Stephen’s sheep herd completely grazed Edward’s land before Edward’s sheep arrived. As a result, the “lamb feed” shortage is irrelevant and immaterial as either a foundational fact or as part of a logical sequence of thought. But the majority proceeds to “couple” the foregoing unsupported inferred fact and the irrelevant fact with Stephen’s spotty documentation of trespass through sheep and cattle counts. And presto — they have reached the “rational and logical” conclusion “that approximately three quarters of the time, Edward’s livestock would move onto surrounding lands of Stephen Gillmor in search of food.”1
*468The total damages awarded to Stephen for forage loss require speculative leaps over yawning gaps in the evidence, wide open spaces of time with phantom livestock viewed only by ghostriders. It was respondents’ burden to present at trial a sufficient evidentiary basis to establish the fact of damages with reasonable certainty and to provide the trier of fact with a reasonably certain basis for computing damages. Sawyers v. FMA Leasing Co., 722 P.2d 773, 774 (Utah 1986).
To prove damages, the plaintiff must prove two points. First, it must prove the fact of damages. The evidence must do more than merely give rise to speculation that damages in fact occurred; it must give rise to a reasonable probability that the plaintiff suffered damage.... Second, the plaintiff must prove the amount of damages. The level of persuasiveness required to establish the fact of loss is generally higher than that required to establish the amount of a loss.
Atkin Wright & Miles v. Mountain States Tel. & Tel. Co., 709 P.2d 330, 336 (Utah 1985) (emphasis in original).
The majority characterizes the main issue before us as whether the respondents met their second burden of putting forth sufficiently certain evidence of the amount of forage loss. The real problem, however, is that respondents have not met their initial burden of proving the fact of damages to the extent they calculated them. Because of this, they should not be allowed to recover for forage purportedly consumed by an unestablished number of livestock trespassing for an unsubstantiated number of days. I am compelled to hold that the trial court’s adopted ultimate findings on these two factors in the total forage loss computation are clearly erroneous. The damages awarded for forage loss should be reduced by subtracting the days and numbers of trespassing livestock for which there is no substantial supportive evidence in the record, as shown below.
The majority opinion ignores the issue of the quantum and nature of proof required to provide the legal basis necessary for a damage award for continuous animal trespass. It does so by glossing over the paucity of documented instances of trespass and the discrepancy between the number of trespassers observed and the numbers of trespassers for which damages were calculated. Although my colleagues prefer to obscure these deficiencies, I instead turn to a detailed consideration of the evidence of trespass as actually produced at trial, not as conjured up by respondents’ counsel.
II. TRESPASS AND UNIT VALUE OF FORAGE LOSS
The four factors used by the trial court in calculating the forage loss caused by Edward’s trespassing livestock were:
(1) presence of livestock on Stephen’s property;
(2) number of livestock thereon;
(3) number of days thereon; and
(4) unit value of forage utilized.
These four factors, if established by substantial, competent evidence received at trial, provide a reasonably certain basis for the computation of damages for forage loss due to trespassing livestock. See Bastian v. King, 661 P.2d 953, 957 (Utah 1983).
When viewed in the light most favorable to the court below, the evidence does provide substantial support for factors (1) and (4). However, I agree with appellant Edward Gillmor that the evidence presented at trial does not support the values assigned to the other two factors (number of livestock and number of days) in the forage loss formula adopted by the trial court and embraced by my colleagues.
III. EVIDENCE OF NUMBERS AND DURATION OF CATTLE TRESPASS
Edward testified that he owned and ran 386 mature cattle during 1981. By adopt*469ing Stephen’s forage loss computations, the lower court’s findings assumed either: (a) that 100% of Edward’s 386 cattle were on Stephen’s property 75% of the ten months from March 17, 1981 to February 17, 1982; or (b) that 75% of Edward’s cattle (289.5 head) were on Stephen’s lands 100% of that time.2 But neither route to the same result has support in the evidence.
The following is a summary of all the evidence of trespasses by Edward’s cattle:
[[Image here]]
There was no evidence of any trespassing cattle counted on Stephen’s lands until April 3, 1981. The seventeen days from March 17 to April 2 should not have been used by the trial court in its computation of forage losses. From April 3-30, 1981, Edward’s cattle were sighted six times, in counts basically ranging from 100 to 150 pairs (cow and calf). Viewing each count in the light most favorable to the lower court, that is, using the maximum number counted on any one day and believing that those counted were in fact on Stephen’s lands, even between counts, there were 150 cows trespassing for twenty-five days in that month.
Edward’s son, Edward L. Gillmor, Jr., was called as a witness on behalf of Ste*470phen. He testified that 224 of Edward’s cows were completely on Edward’s separately controlled, private, and leased lands until June 9, 1981. This left 162 head available as possible trespassers on Stephen’s lands during the period from May 1 through June 9, 1981. This evidence coincides with the fact that Stephen’s trespassing cattle counts did not exceed 150 head during the spring grazing season.
There was evidence that 100 trespassing cattle were observed and counted on May 15, 1981. This one sighting, however, is not a sufficient basis for the lower court’s award of damages for 289.5 head trespassing for the entire thirty days of May. Likewise, testimony that a “large bunch” was spotted on June 5 cannot alone provide sufficient support for any forage loss award for cattle trespass in June.
Between July 8 and 13, there were three counts of trespassing cattle, in decreasing numbers from 75 head to 47 to 24. At best, this evidence would support compensation for trespass by 75 head for only six days in July. On August 5th, 42 pairs were observed and driven back to Edward’s land. A similar number was counted on August 27. Other counts of 10 to 11 pairs took place on August 3, 6, 9 and 31. The best this evidence can muster is support for a damage award that includes 11 head trespassing for twenty-seven days and 42 head trespassing for twenty-three days in August.
There was no evidence of sightings or counts of trespassing cattle in September, 1981. On October 14, according to the testimony at trial, there was a “bunch” in the steer pasture, land partly owned by Edward. On another, unspecified date in October, a “large number” was seen. This evidence standing alone is insufficient to support a damage calculation for cattle trespass using any days in September or October.
On November 2 and 3, 204 pairs were counted on trucks loaded out from the summer range. Since there was no evidence of grazing, this count and this month cannot be used in computing damages for forage loss. On December 12, 84 head were counted in the area of the disputed Swaner lease, north of the Salt Lake International Airport. This count is allowed for one day of trespass by 84 head because Stephen was, as the trial court found, entitled to possession of those leaselands. In January, 1982, 50 head were counted in the 17th North area, near the Great Salt Lake, for periods of eight days and six days. The testimony, however, was that Edward owned part of this area.
The above review of the evidence presented at trial yields a total of 5,647 days of cattle trespass, or 188 cattle AUMs. Each cattle AUM has a value of $7.96. The evidence thus supports an award of damages for forage loss due to trespass by Edward’s cattle only in the amount of $1,496.48.
IV. EVIDENCE OF NUMBERS AND DURATION OF TRESPASS BY SHEEP
I conclude that the trial court’s computation of damages for trespass by Edward’s sheep is likewise deficient and lacking in evidentiary support. Edward testified that he owned about 2,700 mature sheep during 1981. He operated them in two main herds, consisting of approximately 1,125 head and 1,575 head. Only the former herd was involved in the trespass action before us. The lower court’s findings assumed that all 1,125 were trespassing on Stephen’s property 75% of the eight months from March 24, 1981 to November 24, 1981. But that result is not supported by the evidence.
The following is a summary of all the evidence of trespass by Edward’s sheep:
*471[[Image here]]
There is no evidence of any sheep counted on Stephen’s lands until March 31. The seven days from March 24 to March 30 should not have been used by the trial court in computing damages from trespassing sheep. Beginning on March 31 and continuing to April 19, five counts of 1,000 to 1,200 ewes were reported. On April 6 and 7, counts of only 400, 75 and a “large concentration” were made. Stephen testified that 1,000 to 1,200 head were driven off his land on April 20. Viewed in the light most favorable to the lower court, that is, using the maximum number counted and believing that those counted were in fact on Stephen’s lands, even between counts, there were 1,200 sheep trespassing for twenty days. On April 24, 26 and 27, 300 to 350 ewes were counted. That evidence supports, at most, damages for trespass by 350 sheep for four days.
On May 25, an uncounted number of sheep were observed being “driven.” This vague evidence of one sighting cannot support the damages awarded for 1,125 head trespassing continually from April 28 to May 31 and on through the end of June, a month when no counts or observations were made.
Forty bucks counted in the buck pasture (apparently co-owned property) on July 3, and a sighting without numbers on July 22 are insufficient to support the finding that 1,125 head trespassed for thirty-one days in July. The fact that Stephen’s evidence did not show any counts of trespassing sheep between April 27 and August 22 coincides with Edward’s evidence that the 1,100 head herd was on his separate, individually owned land and leases from April to mid-July. Edward’s evidence on this point corroborated similar testimony by his son, Edward Jr., elicited as evidence in Stephen’s behalf.
In August there were two sightings, but no counts. On August 21, a mix of Edward’s and Stephen’s sheep occurred. Stephen testified that on August 21 and 22 the parties separated their sheep. He said 1,000 of Edward’s were separated and over 300 remained in the timber. At best, this *472evidence will support a damage computation based on 1,300 sheep trespassing for two days. Stephen’s son, James, counted 450 head on August 24, providing sufficient evidence to include this number for one additional day of trespass.
There was one sheep sighting in September, but no count or estimate of trespassers appears in the record. This evidence, standing alone, is insufficient to support the award of damages for trespass by 1,125 sheep for the entire month of September. James counted 1,200 to 1,300 trespassing sheep on October 6, when another mix and separation took place. On November 20, 1,100 pairs of ewes and lambs were loaded out from the summer range, sufficient to support the award of damages for that number of trespassing sheep for that one day. Edward’s uncontradicted evidence was that the herd was once again on his individually-controlled property from mid-October until loaded out at the end of the summer grazing season. This coincides with Stephen’s lack of counts of trespassing sheep during the same time frame.
My review of the evidence of trespasses by Edward’s sheep yields a total of 31,850 animal days, or 1,062 sheep AUMs. Each sheep AUM has a value of $1.20. The record evidence thus supports an award of damages for forage loss due to trespass by sheep only in the amount of $1,274.40.
V. CONCLUSION
In sum, I would affirm the judgment below, but only after modifying it. “The court may reverse, affirm, or modify any order or judgment appealed from.” R.Utah Ct.App.Proc. 30(a). The damages awarded for forage loss due to cattle trespass should be reduced from $17,504.04 to $1,496.48; damages awarded for forage loss due to sheep trespass should be reduced from $8,100.00 to $1,274.40.
. Although this conclusion may be consistent with the propensities of men, it does not take into account the propensities of livestock or the effects of the vagaries of weather on a range livestock operation. We can safely infer that range forage grows where Mother Nature plants it and bestows moisture. Forage does not grow uniformly on the range in terms of its dispersion or stage of maturity, and cattle and sheep do not order from the same menu. Nor does moisture come uniformly from season to season or year to year. We can also reasonably infer that livestock will go where the "grass is greener” if they know where it is; it could be anywhere on the partitioned properties. Even this conclusion must be qualified by the availability and nearness of water to drink and by the nature of certain livestock to habitually "locate" in the same place, year after year. Based on the feed, water and habit factors, the livestock will congregate in different areas which, in this case, could be on lands apportioned to either of the parties. One further propensity of livestock is *468that, for the first year or two after being placed on a new range with room to roam, they will wander about until a comfortable location is found. One would therefore expect Stephen’s new livestock to wander more than Edward’s. Accordingly, if the inferences in this case are to be drawn from the known passions, propensities and prejudices of livestock and Mother Nature, rather than from those of men, the ultimate conclusions reached by the majority simply do not make good sense.
. If this is a valid assumption, the trial court should have similarly computed percentages for Stephen’s livestock trespassing on Edward's 25% of the property.