ON PETITION FOR REHEARING
Based on the appellant’s petition for rehearing, we vacate our prior opinion and replace it with this one, which revises only footnote 6.
Thomas Lynne Trejo appeals his convictions for conspiracy to possess with intent to distribute fifty grams or more of methamphetamine, in violation of 21 U.S.C. §§ 841(a)(1), 841(b)(1)(A), 846, and 851, and knowingly and intentionally possessing with intent to distribute fifty grams or more of methamphetamine, in violation of 21 U.S.C. §§ 841(a)(1), 841(b)(l)(A)(viii), and 18 U.S.C. § 2. On appeal, Trejo argues that the district court erred in denying his motion to suppress contraband discovered during a warrantless search of his vehicle and the fruits of that search. Specifically, Trejo challenges the district court’s finding that a positive alert from a law-enforcement canine trained in narcotics detection provided sufficient probable cause to justify the search. After careful review, we affirm.
I.
On January 6, 2012, a detective from the Polk County Sheriffs Office contacted Florida Highway Patrol Trooper Jason Lemery to request that Lemery be on the lookout (BOLO) for a white Chevy Tahoe with Florida license tag ADCT27 traveling southbound on Interstate 75. The Polk County detective informed Lemery that the vehicle was suspected of transporting narcotics. Based on this call, Lemery understood that he was to look for probable cause to stop the Tahoe, and to enlist the services of a law-enforcement canine trained in narcotics detection during any resulting stop to determine if the vehicle in fact contained narcotics.
Lemery situated his patrol car along Interstate 75 and waited for the Tahoe’s arrival, receiving updates on the vehicle’s location as it approached. That afternoon, the Tahoe passed in front of Lemery’s patrol car. Lemery initiated a traffic stop, maintaining that the Tahoe was traveling too closely to the vehicle in front of it and that the windows were unlawfully tinted.1 Trejo was driving the vehicle with his eventual co-defendant, Juaquin Ramirez, as his passenger.
Just a few minutes after Lemery initiated the traffic stop, Florida Highway Patrol Trooper Michael Jordan arrived with a drug-detection canine named Barny. Bar-ny is trained to detect the presence of several narcotics, including methamphetamine. When Barny detects narcotics, he normally “alerts” his handler, Jordan, to the presence of the narcotic odor by aggressively pawing and scratching at the location.
Jordan and Barny conducted an exterior sniff test of the Tahoe during the traffic stop. During the test, Barny alerted to the seams of both the driver’s side and passenger’s side front doors. Based on Barny’s alerts suggesting the presence of narcotics, Lemery and Jordan searched the vehicle’s interior without a warrant. The search revealed about four pounds of methamphetamine concealed inside and *567underneath the center console. Based upon this evidence, Trejo and Ramirez were charged with knowingly and intentionally possessing, and conspiracy to possess with the intent to distribute, fifty grams or more of methamphetamine.
Trejo filed a motion to suppress the contraband seized during the warrantless search of his vehicle, as well as any admissions made or evidence deriving from that search. Trejo argued that Barny’s positive alert, standing alone, did not provide the probable cause necessary to justify the search.
At a suppression hearing to address Trejo’s motion,2 the government presented evidence regarding Barny’s training and certifications to support the reliability of his field alerts. Trooper Jordan began working with Barny in 2009, although Bar-ny had some prior training as a drug detection dog. When Barny and Jordan began working together, the pair completed their first joint training and certification course, which required eighty hours of training over two weeks. The program was administered by Florida Highway Patrol Trooper Michael VanLeer3 and resulted in a certification issued by Florida’s Department of Highway Safety and Motor Vehicles Highway Patrol Academy. The standards for the initial certification course were governed by Florida regulations. See Fla. Admin. Code r. 11B-27.013.
After this initial joint certification, Jordan and Barny re-certified annually. The re-certifications that Jordan and Barny completed required the pair to pass annual exams, which were again administered by VanLeer in a manner consistent with the “standards and guidelines set forth by the Florida Highway Patrol.” Barny and Jordan had passed their most recent re-certification examination on September 21, 2011, just a few months before the January 6, 2012 traffic stop.
The pair also trained weekly. The weekly training format was similar to the initial and annual certification examinations — trainers would hide narcotics in various places and order the canines to find and alert their handlers if they detected a narcotics odor. In total, Jordan estimated that Barny participated in over eight hundred training sessions. And in those training sessions, Barny never falsely alerted to the presence of a narcotic odor when, in fact, no narcotics had ever been present. Only three times did Barny fail to identify a narcotic odor in training when an odor should have prompted an alert.
Trejo countered the government’s offer of proof of Barny’s reliability not by challenging the adequacy of Barny’s training, but by arguing that his real world search record was so poor that, even with the training, a positive alert in the field could not provide probable cause to conduct a warrantless search. Barny’s real world search records revealed that he had been called upon to conduct 311 sniff searches. On 303 of these calls, Barny alerted to the presence of narcotics. Of these 303 alerts, officers identified residue but no measurable contraband 107 times. In addition, on *56871 occasions officers could not even find any residue. Trejo argued that this amounted to at least a 28.5% false alert rate in real world situations, and at worst a false alert rate of around 58%. This high error rate in the field, Trejo argued, rendered the officers’ reliance on Barny’s positive alert to supply probable cause unconstitutional.
The district court denied Trejo’s motion to suppress, adopting the conclusions set forth in the magistrate judge’s Report and Recommendation. The magistrate judge concluded that the “preponderance of the evidence supports the conclusion that canine Barny is reliable.” The magistrate judge questioned Trejo’s characterizations of Barny’s error rate in the field, noting that Trejo faults Barny, who is trained to alert to narcotics odor, for alerting to residue (which gives off an odor). As a result, the magistrate judge concluded that the 23.5% field error rate — the rate at which Barny alerted but neither residue nor contraband was found — “is likely a better barometer of Barny’s reliability in. the field.” Even if the 58% field error rate were accepted, the magistrate judge noted that probable cause to search a vehicle requires only a “fair probability” that contraband will be found. Beyond that, the magistrate judge noted, “the training of a dog alone may be sufficient proof of reliability.” The magistrate judge thus concluded that “Barny’s alert gave officers probable cause to search the Chevy Tahoe pursuant to the automobile exception.”
Trejo was convicted at a bench trial based on stipulated facts, and now appeals his conviction and the denial of his motion to suppress.
II.
The propriety of a district court’s denial of a motion to suppress is a mixed question of law and fact. United States v. Virden, 488 F.3d 1317, 1321 (11th Cir.2007). We review the factual findings for clear error and the application of the law to the facts de novo. Id. We construe the facts in the light most favorable to the government here because it prevailed in the district court. See id.
The Fourth Amendment provides that “[t]he right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures, shall not be violated.” U.S. Const, amend. IV. A warrantless search is per se unreasonable unless it falls into one of the exceptions to the warrant requirement. United States v. Sokolow, 450 F.2d 324, 325 (5th Cir.1971).4 One such exception is the automobile exception, which permits warrantless searches of a vehicle when there is probable cause to believe it contains contraband or evidence of a crime. Virden, 488 F.3d at 1321-22. An alert from a narcotics detection canine establishes probable cause if “all the facts surrounding a dog’s alert, viewed through the lens of common sense, would make a reasonably prudent person think that a search would reveal contraband or evidence of a crime.” Harris, 133 S.Ct. at 1058. This analysis requires courts to examine the totality of the circumstances surrounding the canine’s alert, instead of applying “rigid rules, bright-line tests, and mechanistic inquiries.” Id. at 1055.
This Court has “long recognized that probable cause arises when a drug-trained canine alerts to drugs.” United States v. Tamari, 454 F.3d 1259, 1265 (11th Cir.2006) (quotation marks omitted). Before *569finding probable cause, we generally require that the dog’s alert be reliable, and have recognized that “training of a dog alone is sufficient proof of reliability.” United States v. Sentovich, 677 F.2d 834, 838 n. 8 (11th Cir.1982); see also United States v. Dunkley, 911 F.2d 522, 527 (11th Cir.1990) (citing Sentovich, 677 F.2d at 838 n. 8, with approval).
On appeal, Trejo argues that the Supreme Court’s recent decision in Florida v. Harris, — U.S. -, 133 S.Ct. 1050, 185 L.Ed.2d 61, raises the requirements for the government’s reliability showing. Tre-jo asserts that, in light of Harris, this Court must vacate and remand for the district court to consider in the first instance (1) whether the organizations that trained and certified Barny were “bona fide” and (2) whether Barny’s training program utilized double-blind precautions to prevent the certification of canines who respond to a handler’s cues as opposed to the odor of narcotics alone.5 Trejo also argues that the district court erred when it found Barny’s alert reliable despite his error rate in the field.
Trejo’s argument, however, reads too much into Harris. In Harris, the Supreme Court considered how a court should decide whether a drug detection dog’s alert during a traffic stop provides probable cause to search a vehicle. Id. at 1053. In Harris, the U.S. Supreme Court reviewed a Florida Supreme Court ruling that the Fourth Amendment requires the government to offer evidence from a checklist of items, emphasizing the need for evidence of not only the canine’s training and certification but also evidence of performance in the field, including false alerts. Id. at 1054-55. The Supreme Court rejected this approach as too formulaic for the totality-of-the-circumstances probable cause inquiry, noting that “a finding of a drug-detection dog’s reliability cannot depend on the State’s satisfaction of multiple, independent evidentiary requirements.” Id. at 1056.
The U.S. Supreme Court criticized the Florida Supreme Court’s overreliance on field data in particular, noting that those records “may markedly overstate a dog’s real false positives.” Id. at 1056-57. “The better measure of a dog’s reliability,” the Supreme Court concluded, “comes away from the field, in controlled testing environments.” Id. at 1057. For that reason, the Supreme Court concluded that
evidence of a dog’s satisfactory performance in a certification or training program can itself provide sufficient reason to trust his alert. If a bona fide organization has certified a dog after testing his reliability in a controlled setting, a court can presume (subject to any conflicting evidence offered) that the dog’s alert provides probable cause to search. The same is true, even in the absence of formal certification, if the dog has recently and successfully completed a training program that evaluated his proficiency in locating drugs.
Id. at 1057. The Supreme Court went on to note that “[i]f the State has produced proof from controlled settings that a dog performs reliably in detecting drugs, and the defendant has not contested that showing, then the court should find probable cause. If, in contrast, the defendant has challenged the State’s case (by disputing the reliability of the dog overall or of a *570particular alert), then the court should weigh the competing evidence.” Id. at 1058.
Harris does not unsettle this Court’s jurisprudence that an alert from a canine trained in narcotics detection can establish probable cause. See Sentovich, 677 F.2d at 838 n. 8. Rather, it recognizes, consistent with Sentovich, that a court may “presume” that a dog’s alert provides probable cause to search under certain circumstances, such as when the canine is certified by a “bona fide” organization or “has recently and successfully completed a training program.” Harris, 133 S.Ct. at 1057. By no means does Harris impose a categorical requirement, as Trejo suggests, that the district court make a factual finding that the canine is certified by a “bona fide” organization before finding that the dog’s record in controlled settings supports a finding of reliability. See id. at 1057-58. Indeed, imposing a categorical evidentiary requirement like this would contradict the flexible and commonsense totality-of-the-circumstances inquiry Harris demands. See id.
Here, the training that Trooper Jordan and his canine Barny completed is sufficient to support a presumption that Bar-ny’s alert was reliable absent credible evidence to the contrary. Indeed, Jordan and Barny’s training is quite similar to the training programs the Supreme Court considered in Harris. In Harris, the certification and training program for the initial certification was offered by state police departments and the refresher course was offered by “a private company that specializes in testing and certifying” narcotics detection dogs. Id. at 1054. The Supreme Court discussed both of these programs with approval. Id. at 1058. Here, the original certification course was offered by the Florida Highway Patrol and run by one of the State Troopers. And although Trejo questions the legitimacy of the organizations under whose standards Jordan and Barny were annually recerti-fied, the recertification course was taught by the very same instructor who offered the initial course. That the district court did not declare the certifying organizations — the Florida Highway Patrol, the International Forensic Research Institute, and the National Forensic Science Technology Center — to be “bona fide” in so many words makes no difference. And even if the certifying organizations were not “bona fide” within the meaning of Harris, the evidence clearly established that Barny and Jordan had “recently and successfully completed a training program,” and so the presumption of reliability was permissible absent a timely challenge to the adequacy of the training program.
On that point, Trejo, like the defendant in Harris, “declined to challenge in the trial court any aspect of [the canine’s] training.” Id. at 1058. To be sure, he now “raise[s] questions about that training’s adequacy — for example, whether the programs ... used enough blind testing,” and asks us to remand for a determination of whether the program used double-blind testing protocols. Id. But Trejo “never voiced those doubts in the trial court, and cannot do so for the first time here.” Id.; see also United States v. Garcia-Sandobal, 703 F.3d 1278, 1283 (11th Cir.2013) (noting that a party’s failure to file a specific written objection to a proposed finding in a magistrate judge’s Report and Recommendation waives the party’s right to review of that finding). As Trejo’s case came to the trial court, Barny had successfully completed drug-detection courses each year for the three years leading up to. the search of Trejo’s vehicle and maintained his proficiency through weekly training courses. Such a training record, viewed alone and unchallenged before the trial court, is sufficient to establish a drug de*571tection dog’s reliability. See Harris, 133 S.Ct. at 1058.6
Trejo’s cross-examination focusing on Barny’s field performance failed to rebut the government’s case for Barny’s reliability. Cf. id. As the Supreme Court noted in Harris, records of a dog’s field performance in most cases have “relatively limited import.” Id. at 1056. Because of the risk that field data “may markedly overstate a dog’s real false positives,” a dog’s performance in controlled testing environments— where Barny’s performance was exceptional and nearly flawless — is a better measure of a canine’s reliability than field performance. Id. at 1056-57. As a result, even considering the proffered field error rate, the positive alert from a drug detection dog with an excellent training record, coupled with the BOLO indicating that Trejo’s Tahoe was suspected of transporting narcotics, established a “fair probability” that drugs would be found in the vehicle. This is all that probable cause requires. See Virden, 488 F.3d at 1322.
III.
For these reasons, we do not agree with Trejo’s argument that we must remand to the district court to make a factual finding that the organizations issuing Barny’s certifications were “bona fide” or that the training programs utilized double-blind training protocols. Neither do we agree that Barny’s field error rate undermines the conclusion that his alerts are reliable indicia of the presence of narcotics sufficient to support probable cause. We therefore AFFIRM.