OPINION

SINGLETON, Judge.

At the conclusion of an investigatory stop Marlene Moran was charged with driving while intoxicated, AS 28.35.030. She moved to suppress all the evidence discovered contending that the stop was illegal. After listening to the testimony of Trooper Duncan, the arresting officer, the trial court dismissed the case for "insufficient probable cause for traffic stop." The state appeals. We have jurisdiction over this appeal. State v. Michel, 634 P.2d 383, 385 (Alaska App. 1981). We reverse and remand for reinstatement of prosecution.

Trooper Troy Duncan testified at the evidentiary hearing held on March 3 and 8, 1983. Moran did not testify, nor dispute Trooper Duncan's testimony at the hearing, and Judge Cline apparently accepted the accuracy of Trooper Duncan's statements.

On December 4, 1982, in Fairbanks, Trooper Duncan was working the graveyard shift. The temperature was about -20°. Visibility was clear; there was no ice fog; and streetlights were illuminated. At approximately 11:45 p.m., Trooper Duncan observed Moran make a left turn onto a four-lane road. She turned into the outside lane, which constitutes an illegal turn in violation of 13 AAC 02.200(b)(1). Trooper Duncan felt that this violation warranted further observation of Moran. He followed her for about one-half mile, during which time he observed her tires touch the center line and then drift over to the fog lines on the side of the road three times. While there was fairly heavy traffic on this Saturday night Trooper Duncan did not observe anyone take evasive action due to Moran's driving.

When Moran weaved the third time, Trooper Duncan signaled for her to stop. Trooper Duncan testified that in light of his observation, he suspected that Moran might be driving while intoxicated. Moran unsatisfactorily performed field sobriety tests and was arrested. She was taken to the police station where she submitted to a breathalyzer examination which yielded a .14% blood alcohol level.

The trial court found that Trooper Duncan's observations did not constitute "probable cause" to believe that Moran had committed a crime. The judge rejected the state's contention that the situation warranted an investigatory stop by pointing out that nothing Trooper Duncan observed indicated that Moran had in fact endangered anyone.

Judge Cline utilized an incorrect legal test. As prior cases establish, an investigatory stop is differentiated from an arrest on the basis of (1) its purpose, see Coleman v. State, 553 P.2d 40, 44 (Alaska 1976), (2) the magnitude of the intrusion, see Howard v. State, 664 P.2d 603, 608 (Alaska App., 1983), and (3) the quantum of information necessary to justify the intrusion, i.e., reasonable suspicion rather than probable cause, see Coleman v. State, 553 P.2d at 45-46. Moran does not contend that the magnitude of the intrusion — stopping her vehicle and subjecting her to field sobriety tests — exceeded the scope of an investigatory stop. See id. at 46 n. 19 (the court refused to distinguish between investigatory stops of pedestrians and motorists in assessing the magnitude of the intrusion).

Turning to the purpose for which a stop may be made, prior Alaska decisions require "reasonable suspicion that imminent public danger exists or serious harm to persons or property has recently occurred." See id. at 46. In Ebona v. State, 577 P.2d 698 (Alaska 1978), the court concluded that suspicion of drunk driving met the Coleman test. The court said, "significant dangers to persons or property that can possibly result when the operator's capacity to control a motor vehicle is impaired are apparent." Id. at 701.

We are thus left with the third criterion, the quantum of information an officer must have in order to justify a stop. Generally this has been described as reasonable suspicion and differentiated from the probable cause required for an arrest. Necessarily these concepts are somewhat vague, but as Professor LaFave points out:

By contrast [to probable cause], as is suggested by the reference in Terry to a reasonable belief "that criminal activity may be afoot," it would seem clear that the more-probable-than-not standard is never applicable to a brief stopping for investigation. Rather, it will suffice that there exists a substantial possibility that criminal conduct has occurred, is occurring, or is about to occur. As stated in United States v. Holland [ 510 F.2d 453 (9th Cir. 1975)]:

Clearly, the officers were not required to rule out all possibility of innocent behavior before initiating a brief stop and request for identification. The test is founded suspicion * * *. Even if it was equally probable that the vehicle or its occupants were innocent of any wrongdoing, police officers must be permitted to act before their reasonable belief is verified by escape or fruition of the harm it was their duty to prevent.

3 W. LaFave, Search and Seizure § 9.3, at 65-66 (1978) (footnotes omitted; emphasis supplied).

Recently the United States Supreme Court elaborated on this standard:

Courts have used a variety of terms to capture the elusive concept of what cause is sufficient to authorize police to stop a person. Terms like "articulable reasons" and "founded suspicion" are not self-defining; they fall short of providing clear guidance dispositive of the myriad factual situations that arise. But the essence of all that has been written is that the totality of the circumstances — the whole picture — must be taken into account. Based upon that whole picture the detaining officers must have a particularized and objective basis for suspecting the particular person stopped of criminal activity.

United States v. Cortez, 449 U.S. 411, 417-18, 101 S.Ct. 690, 695, 66 L.Ed.2d 621, 628-29 (1981). Professor LaFave notes:

At another point the court in Cortez said the question was whether the agents "could reasonably surmise that the particular vehicle they stopped was engaged in criminal activity." This is better than the "reasonable suspicion" formula, see note 15 supra, as "surmise" means to "form a notion on slight proof." Webster's Third New International Dictionary 2301 (1961).

3 W. LaFave, supra, at 25 n. 15.1 (Supp. 1983) (emphasis in original). In Delaware v. Prouse, 440 U.S. 648, 661, 99 S.Ct. 1391, 1400, 59 L.Ed.2d 660, 672 (1979). The court made it clear that the stop may not be truly random. There must be some circumstances that differentiate the driver stopped from the motoring public in general.

With these standards in mind we find that Trooper Duncan's observations of Moran more than satisfied the reasonable suspicion test and justified an investigatory stop to determine whether she was driving while intoxicated. Judge Cline apparently believed that a policeman may not make an investigatory stop unless he sees the person he ultimately stops do something dangerous. This is an incorrect reading of Coleman. It is sufficient if the officer observes facts which lead him to reasonably believe that the person to be stopped is dangerous.

The facts of this case are virtually indistinguishable from those discussed by our supreme court in Ebona v. State, 577 P.2d 698 (Alaska 1978). In Ebona, Juneau police officers first observed Ebona leaving a cafe in a somewhat intoxicated condition. About one and one-half hours later, officers aware of this prior observation saw a vehicle which they knew to belong to Ebona "swerving — going from the center line to the righthand edge," but not going into any other lane. Id. at 699. After following the vehicle about one and one-half blocks, the officers stopped it and arrested Ebona for OMVI. At Ebona's trial, one of the officers testified that he had no reason to believe that Ebona's vehicle had recently caused harm to persons or property, but he did testify that although he was not sure that Ebona was endangering anyone else, the officer was concerned for Ebona's safety. Id. at 770-71.

Although there are a few distinctions between the cases, we are satisfied that Ebona is controlling. While Trooper Duncan did not observe Moran prior to her driving exhibiting indicia of intoxication as the officers who observed Ebona testified, he did observe her make an illegal left turn. Ebona's observed driving, in contrast, had satisfied all legal requirements. Trooper Duncan followed Moran for about one-half mile, while the Juneau officers followed Ebona only one and one-half blocks. We do not know how many times Ebona swerved, but the description of his driving is identical to Trooper Duncan's description of Moran's driving. Trooper Duncan did observe Moran swerve three separate times before stopping her. While Moran's actions may have been consistent with innocence, it was also consistent with drunk driving. Trooper Duncan did not engage in a purely random stop in signalling Moran to stop. The trial court erred in invalidating the stop and suppressing evidence that Moran failed the field sobriety tests and the breathalyzer examination.

The judgment of the district court is REVERSED.