concurring in result.
[¶ 47.] I agree with the conclusion of the Court, but arrive at it by different means. On some factual and legal points, I disagree with both the Court and the dissent on how this case should be resolved. Trooper Swets may not have acted or relied on his dog’s “alert,” but his subjective state of mind is not relevant to our inquiry. If the dog alerted and if Swets knew it alerted, he had probable cause to search, even if he expressed at the time some other legal justification to search. State v. DeLaRosa, 2003 SD 18, 657 N.W.2d 683 (drug dog alert is sufficient probable cause). On the other hand, if he did not know or believe that his dog alerted, then, regardless of whether the dog in fact alerted, Swets did not have probable cause.
[¶ 48.] Whether a Fourth Amendment violation occurred “turns on an objective assessment of the officer’s actions in light of the facts and circumstances confronting [the officer] at the time,” and not on the officer’s actual state of mind at the time the challenged action was taken. Scott v. United States, 436 U.S. 128, 138, 98 S.Ct. 1717, 1723, 56 L.Ed.2d 168 (1978). We review two elements in probable cause determinations. First, we identify all rele*103vant facts known to the officer at the time of the search; and second, we decide, under a standard of objective reasonableness, whether those facts would give rise to a reasonable suspicion justifying probable cause to search. Ornelas v. United States, 517 U.S. 690, 696-697, 116 S.Ct. 1657, 1661-62,134 L.Ed.2d 911 (1996). A search cannot be retrospectively transformed into an unlawful seizure by virtue of the officer’s subjective intent. Maryland v. Macon, 472 U.S. 463, 471, 105 S.Ct. 2778, 2783, 86 L.Ed.2d 370 (1985).
[¶ 49.] Before we begin to examine this case, we must establish the legal reference point from which we view the trial court’s decision. As explained in State v. Lamont, 2001 SD 92, ¶ 21, 631 N.W.2d 603, 610, it is essential in these cases to keep in mind that fact findings in suppression hearings are reviewed for clear error, but ultimately, in reviewing decisions on motions to suppress evidence taken in warrantless searches, our standard of review is de novo. See State v. Morato, 2000 SD 149, ¶ 10, 619 N.W.2d 655, 659 (citing Ornelas, 517 U.S. at 699, 116 S.Ct. at 1663, 134 L.Ed.2d at 920). It is our duty to make our own legal assessment of the evidence to decide under the Fourth Amendment whether the officer’s actions were “objectively reasonable.” See Maryland v. Buie, 494 U.S. 325, 330, 110 S.Ct. 1093, 1096, 108 L.Ed.2d 276 (1990). Consequently, we are not constrained by the trial judge’s legal rationale for upholding the search. Equally important, we are not bound by a police officer’s subjective rationale. rkansas v. Sullivan, 532 U.S. 769, 771-72, 121 S.Ct. 1876, 1878, 149 L.Ed.2d 994 (2001); Whren v. United States, 517 U.S. 806, 813, 116 S.Ct. 1769, 1774, 135 L.Ed.2d 89 (1996).
1. Did Crockett alert?
[¶ 50.] This question is important, but not in the way that both the Court and the dissenters approach it. We need not decide whether the dog alerted based on how Trooper Swets reacted to the dog’s behavior. That is a backward analysis, made unnecessary by the expert testimony. We know the dog alerted because there is uncontradicted expert opinion that he did. The circuit court accepted that opinion. If we apply the clearly erroneous standard of review to that finding, as we must, then we are bound by it, for there is evidence in the record to support it.
[¶ 51.] I realize that the federal district court found to the contrary. But that court did not have the benefit of hearing the State’s expert. The federal court had to discern whether the dog alerted based on Trooper Swets’s response to the dog. With only that evidence, I concede the answer is equivocal. But the State’s expert on Crockett’s behavior settled the issue. Kyle Heyen is familiar with Crockett: he trained him and he re-certified him. Heyen described Crockett’s alert: “his tail intensifies in its wagging and he gets an increased — very intense sucking in through his nose that — it’s very audible, and you can hear it.” As the video tape was played for the circuit judge, Heyen pinpointed the exact time of Crockett’s alert as heard on the tape: “his nose started inhaling, sucking in, being a Hoover vacuum.” For the benefit of the circuit judge, Heyen himself indicated how this sounds. With Heyen’s assistance, the circuit court was evidently able to hear it because the court specifically found that the dog alerted.
[¶ 52.] Although my personal viewing of the videotape is not dispositive on this point, in listening to it myself, at the point where Heyen stated that the dog alerted, I can hear increased sniffing, and some additional sounds, not inconsistent with the way Heyen described them. In declaring Heyen “not credible,” the dissent wrests *104from the trial judge a question that every appellate court in this land has said is ultimately a factual determination to be made by trial judges.
2. Did Trooper Swets “know” Crockett alerted?
[¶ 53.] This is a more difficult question. The answer is crucial, however, because probable cause must be based on the facts “known” to the officer at the time of the search. Ornelas, 517 U.S. at 696-97, 116 S.Ct. at 1661-62. We have both direct and circumstantial evidence to examine on this question. First, Trooper Swets testified that, although he never told his fellow officers, he nonetheless believed that Crockett had alerted to the trunk area of defendant’s vehicle. He referred to the alert in his affidavit in support of a search warrant, signed the same day as the incident, and in his report of October -9, 2000. On the question whether Swets had knowledge that his dog alerted, the court simply found that the dog did alert and that thereafter Swets had probable cause to search.
[¶ 54.] The remainder of the evidence is circumstantial. And it is mixed, i.e., some of it supports his belief that the dog alerted, and some does not. Let us examine the supporting evidence. Swets gave Crockett a toy immediately after the alert on the trunk. The toy is Crockett’s reward as Heyen explained:
The toy is the reward for the dog, which is delivered or the dog retrieves upon conclusion of a search. The only purpose for that is to show the search is done. They’ve hunted; they’ve alerted; they’ve found. They’ve pinpointed; the exercise is done.
Does Crockett get a reward if he does not alert in a search? In the federal hearing, Swets confirmed that if he gives the dog the toy that means that he has seen an alert or an indication. The videotape shows the dog returning to the patrol car with a toy in his mouth. Immediately after putting the dog back in the patrol car, and making a radio call, Swets went over to defendant and asked him if he had any marijuana in the car -or had used marijuana. Swets had asked defendant shortly after the stop if he had any marijuana or cocaine in the car and now asking him again after the sniff search could suggest that Swets felt there was reason to re-inquire. Also corroborating Swets’s belief that the dog alerted is the fact that shortly afterwards he asked one of defendant’s passengers “if there is any reason why my drug dog would alert to your vehicle today.” Of course, that is essentially the same question he asked defendant before he took the dog to sniff around the car.
[¶ 55.] The evidence against Swets’s knowledge that the dog alerted is significant. The dissent outlines this evidence and most of it need not be recited again here. Most troubling is the conversation between Swets and Deputy Shafer. Swets took Shafer aside, presumably to be out of the hearing of the defendant and his passengers, and suggested that the false identification gave them probable cause to search. Swets said, “Something is up.” But he said not a word about Crockett alerting or even that he suspected that drugs were in the car. ■
[¶56.] These remarks are perplexing. But as for the comment Swets made to Shafer — “Crockett shows some interest, but” — I think there is an explanation. ' In listening to the tape,' this remark was made in connection with the fact that Swets noticed that the passenger had a large amount of cash in his pocket. It was then that Swets remarked about Crockett showing interest. Significantly, Swets’s report of October 9, 2000, written long *105before the issue of whether the dog alerted, states:
Deputy Shafer stated he felt the card was fraudulent. I informed Shafer the passenger had quite a bit of cash on [sic] his pocket and my dog showed some interest on the passengers [sic] side door seam.
This comment in the report is separate from the comment on the dog’s alert: “I took my dog around the vehicle. He alerted to the odor of an illegal drug coming from the vehicle at the trunk seam.”
[¶ 57.] The question is, from this contradictory evidence, was there sufficient evidence in the record to support the trial court’s implicit finding that Swets knew his dog alerted and therefore had probable cause to search the vehicle based on that alert? Aside from the conflicting evidence from Trooper Swets, we have what the trial court considered as corroborative testimony. Here is where I part company •with the Court and the trial judge most decidedly. The trial court placed reliance on the other two officers to corroborate Trooper Swets. A careful reading of the record fails to support the trial court’s findings. First, there is the testimony of Trooper Thomas. He testified that the “the dog sniffed a little bit more, stiffened up, and then the officer gave him a toy.” This corroborates Swets’s testimony to the extent that it confirms that the dog’s behavior changed at the trunk of the car. It does not establish that Swets knew the dog alerted, or, for that matter, that Thomas knew that the dog alerted.
[¶ 58.] Yet the trial court found that “Trooper Thomas witnessed Crockett alert to the presence of the odor of illegal drugs in the trunk area of the defendant’s vehicle.” How the court reached this finding is questionable. In response to, the prosecutor’s question, “what did the dog do?” Thomas volunteered the opinion that “he appeared to alert on the vehicle.” Defense counsel immediately objected to that remark because there was no foundation to establish that Thomas knew how to determine a drug dog alert. The trial court sustained the objection and later sustained another defense objection for the same reason. To use that answer, then, as the trial court did, was erroneous. After all, the judge had agreed that Thomas had not been qualified to give this opinion but it nonetheless used the opinion in its findings. Certainly, we should not rely on that opinion. But even if Thomas’s opinion was properly considered, Thomas did not communicate his opinion to Swets. Thomas was there in a backup capacity. He did not prepare a report because he saw his role as purely secondary. It was Swets who decided to search the car; he was the trained drug dog officer; his knowledge was essential to create the basis for probable cause.
[¶ 59.] Even more doubtful, however, is the trial court’s reliance on Deputy Shafer’s testimony that Swets told him that the dog had alerted. First, Swets denied or did not recall that he told either Thomas or Shafer that his dog had alerted. Second, in explaining the way Swets had informed him, Shafer testified that he interpreted Swets’s remark, “Crockett shows some interest ....” to mean that the dog alerted. But that was obviously a misinterpretation. Indeed, Swets confirmed for the trial court that “showing some interest” is not an alert. Thus, Shafer’s testimony that Swets told him the dog alerted was a misinterpretation of what Swets said. The trial court was clearly erroneous in this finding.
[¶ 60.] The Court takes the view that because the circuit court believed Swets, we should defer to the trial judge’s firsthand perception of credibility. It is true that we leave credibility decisions to the *106trial court, unless we can say that its finding was clearly erroneous. However, the trial court’s reliance on the “corroboration” from the other two officers is problematic' Now, after reviewing the videotape, I cannot ignore the witness of my own eyes. What I see causes me to doubt. I see no verbal indication in the videotape that Trooper Swets thought he was searching based on his dog’s alert. To the contrary, his remarks seem to indicate that he was seeking to search on some other basis.
[¶ 61.] On the other hand, how much weight should we give to the indirect and passing remarks made by the officers as they went about their business? This is not a case like State v. Hanson, 243 Ga.App. 532, 532 S.E.2d 715 (Ga.App.2000), where a videotape showed that a Deputy Sheriff was not credible in contending that he smelled marijuana during a traffic stop. In that case, the Deputy commented outside the hearing of the driver, solely for purposes of the videotape, that the driver and his passenger appeared to be very nervous. He was using the videotape, to make verbal notes of his observations during the traffic stop. But, tellingly, he made no remark, as an aside or otherwise, that he smelled the odor of marijuana. Id. at 533, 532 S.E.2d at 717. The trial court found: “The inclusion of an assertion in the officer’s report that he smelled marijuana is not persuasive in light of the officer’s obvious effort to inject relevant comments sotto voce, which convinces the Court the officer’s testimony on that issue is not credible.” Id. The appellate court agreed, concluding that credibility decisions are for the trial judge. Here, on the other hand, it is obvious from viewing the tape that the officers were not using the video recording to make “verbal notes” or to intentionally record their internal thought processes for the benefit of some future decision maker.
[¶62.] In the end, we have an expert opinion that the dog alerted, although this opinion does not retrospectively create probable cause at the time of the search. Probable cause must be formed on the information confronting the officers at the time, not as later revealed. State v. Meyer, 1998 SD 122, ¶ 23, 587 N.W.2d at 724 (citing State v. Heumiller, 317 N.W.2d 126, 129 (S.D.1982)). The State bears the burden of justifying a warrantless entry into a constitutionally protected area. Lamont, 2001 SD 92 at ¶22, 631 N.W.2d at 610. The trial court found that the State met its burden. I conclude, based on my own legal assessment of the evidence, that the videotape is inconclusive. All we have, then, is whether Trooper Swets is to be believed. Questions of witness credibility are reserved to the trial court, and we have no authority to substitute our judgment for that of the trial judge, unless we can say that the record clearly refutes it. Regardless of whether Trooper Swets communicated his knowledge of his dog’s alert, if the dog alerted and if Swets knew it alerted, he had probable cause to search, even if he expressed some other legal justification to search. The trial judge believed him when Swets testified that he saw his dog alert. With only an inconclusive videotape to rebut Swets’s testimony, the trial judge’s credibility finding must stand. Therefore, I concur in result.