OPINION

Myma Walker appeals from the trial court’s dismissal on summary judgment of her complaint against the City of Scottsdale (City) and the McCormick Ranch Property Owners’ Association (Association). The issue on appeal is whether the trial court properly construed A.R.S. § 33-1551, Arizona’s recreational use statute, to limit the defendants’ duty to Walker, thereby precluding her recovery.

FACTS AND PROCEDURAL HISTORY

Walker was injured in a fall which occurred while she was riding her bicycle along a bicycle path at 85th Street and Hayden Road in Scottsdale, Arizona. The property on which the bicycle path is located is a greenbelt area within McCormick Ranch, a planned, residential community in Scottsdale, Arizona. The property was owned by the Association. The City of Scottsdale owned an easement on the McCormick Ranch property for the bicycle path, which it constructed and maintained. Walker filed suit against the Association and the City for her injuries, alleging that her fall was caused by the negligence of the defendants in the design, construction, and maintenance of the bike path.

Both defendants moved for summary judgment, arguing that they owed no duty of care to Walker pursuant to A.R.S. § 33-1551. The trial court agreed and granted summary judgment for both defendants. After the trial court denied Walker’s motion for reconsideration or new trial, Walker appealed from the judgment and the order denying her motion. Walker argues that A.R.S. § 33-1551 cannot be read to preclude her suit under the facts presented in this case.

DISCUSSION

Arizona Revised Statutes § 33-1551, enacted in 1983, is a statute granting immunity from suit, with limited exceptions, to owners, lessees or occupants of certain types of property for injuries to persons who have made certain recreational uses of the property without paying an admission fee for such use. Subsection (A) of the statute provides that:

A. An owner, lessee or occupant of premises does not:

1. Owe any duty to a recreational user to keep the premises safe for such use.

2. Extend any assurance to a recreational user through the act of giving permission to enter the premises that the premises are safe for such use.

3. Incur liability for an injury to persons or property caused by any act of a recreational user.

Subsection (B) defines the statute’s key words, “premises” and “recreational user,” as follows:

B. As used in this section:

1. “Premises” means agricultural, range, mining or forest lands, and any other similar lands which by agreement are made available to a recreational user, and any building or structure on such lands.

2. “Recreational user” means a person to whom permission has been granted or implied without the payment of an admission fee or other consideration to enter upon premises to hunt, fish, trap, camp, hike, ride, swim or engage in similar recreational pursuits. The purchase of a state hunting, trapping or fishing license is not the payment of an admission fee or other consideration as provided in this section.

Subsection (C) lists the limited circumstances in which liability is not precluded:

C. This section does not limit the liability which otherwise exists for maintaining an attractive nuisance, or for wilful or malicious failure to guard or warn[*208] against a dangerous condition, use or activity.

Statutes of this general type are frequently referred to as “recreational use” statutes.

In this appeal, we must construe Arizona’s recreational use statute to determine whether it applies under the facts of this case to bar Walker’s suit. The narrow question presented is whether A.R.S. § 33-1551 limits the liability of those with an interest in a maintained bike path running through a greenbelt area of an urban, residential neighborhood in a suit brought by a bicyclist injured on the bike path. The trial court concluded that the limitations in the statute were applicable in this situation.

In reviewing the trial court’s decision, this court is not bound by the legal conclusions reached by the trial court. Gary Outdoor Advertising Co. v. Sun Lodge, Inc., 133 Ariz. 240, 650 P.2d 1222 (1982); Cecil Lawter Real Estate School, Inc. v. Town and Country Shopping Center Co., 143 Ariz. 527, 694 P.2d 815 (App.1984). An issue of statutory interpretation is one of law, and hence this court is free to draw its own conclusions regarding it. Arizona State Bd. of Accountancy v. Keebler, 115 Ariz. 239, 564 P.2d 928 (App.1977). Arizona’s appellate courts have not previously been called upon to construe and apply Arizona’s recreational use statute.

No contention is made in this case that Walker’s claim might fall within the exceptions to the immunity provided in subsection (C) of the statute. Walker does not charge the defendants with acting wilfully or maliciously or with maintaining an attractive nuisance. Walker’s contention is that the facts existing in this case cause it to fall outside the statute. Walker’s primary argument is that the property on which her injury occurred was not “premises” as that term is defined in subsection (B). Secondarily, she argues that she is not the type of “recreational user” to which the statute applies. She does not deny that she was using the property without paying a fee and that at least part of her purpose for riding on the bike path was recreational. Even so, she maintains that the riding of her bicycle was not the type of “riding” that the legislature meant to include within the statute.

We turn first to the question of whether the greenbelt area of McCormick Ranch within which the accident occurred falls within the statutory definition of “premises.” The cardinal rule of statutory interpretation is to determine and effectuate the legislative intent behind the statute. Calvert v. Farmers Ins. Co., 144 Ariz. 291, 697 P.2d 684 (1985); Collins v. Stockwell, 137 Ariz. 416, 671 P.2d 394 (1983). This particular statute contains no statement of legislative purpose. Moreover, there is scant legislative history to shed light on whether the legislature intended to encompass property of this nature within the statute. The only information of any relevance which we glean from the legislative history is that the bill was intended to promote the use of “vast areas of land not now being used for recreational purposes” ^[1] and that much of the statutory language is taken from a model act which was proposed by the Council of State Governments in 1965. ^[2]

We recognize that the impetus for the Council of State Governments’ proposal of the model act was the growing awareness of the need for additional recreational areas to serve the general public. Suggested State Legislation, Vol. XXIV, Public Recreation on Private Lands: Limitation on Liability, pp. 150-52 (1965). It was felt that “in those instances where private owners are willing to make their land available to members of the general public without charge, it is possible to argue that every reasonable encouragement should be given[*209] to them.” Id. at 150. The stated purpose of the model act is “to encourage availability of private lands by limiting the liability of owners to situations in which they are compensated for the use of their property and to those in which injury results from malicious or willful acts of the owner.” Id.

Unfortunately, this information does not carry us far in determining whether the McCormick Ranch greenbelt area is encompassed within Arizona’s recreational use statute.

In order to determine the legislature’s intention as to what property is encompassed within the statute, we must look to the language used in the statute. The language of a statute is the most reliable evidence of its intent. State ex rel. Corbin v. Pickrell, 136 Ariz. 589, 667 P.2d 1304 (1983). It is clear from the language used in the statute that the legislature sought to place some limitations on the types of property falling within the immunity provided by the statute. Instead of granting immunity to an owner, lessee, or occupant of “any” premises, the legislature restricted the meaning of “premises” to “agricultural, range, mining or forest lands, and any other similar lands.” Where a statute expressly defines certain words and terms used in the statute, the court is bound by the legislative definition in all cases where rights of parties litigant are based upon statute. Serna v. Statewide Contractors, Inc., 6 Ariz.App. 12, 429 P.2d 504 (1967).

The defendants apparently concede that the property where plaintiff’s injury occurred is not “agricultural, range, mining or forest lands.” They argue, though, that the property falls within the phrase “any other similar lands.” They argue that a reading of the entire statute leads to the conclusion that in using the phrase “any other similar lands which by agreement áre made available to a recreational user,” the legislature meant to encompass any land made available by agreement to persons performing the activities listed in subsection (B)(2). They emphasize that the statute contains no express language limiting it to rural or semi-rural properties, and they urge us to find no implication of such limitation. They cite numerous cases in which courts have found recreational use statutes in other states to apply to properties located in urban areas such as city parks. They contend that these cases provide support for their argument that Arizona’s statute applies to the McCormick Ranch greenbelt area even though it is located in an urban area.

We find the cases cited by the defendants to be of little assistance to them. We note that approximately forty states have enacted some type of recreational use statute, many of them patterned after the model act. See Riksem v. City of Seattle, 47 Wash.App. 506, 513, 736 P.2d 275, 279 (1987). The statutes vary widely, though, concerning precisely to which lands the tort immunity is to be extended. See generally 3 S. Speiser, C. Krause & W. Ganns, The American Law of Torts, § 14:5 (1986); Annotation, Effect of Statute Limiting Landowner’s Liability for Personal Injury to Recreational User, 47 A.L.R. 4th 262 (1986). An examination of eases from our sister states shows that, in each instance, the language defining the property encompassed within the statute is much broader than that used in Arizona’s statute. For instance, in Riksem v. City of Seattle, 47 Wash.App. 506, 736 P.2d 275 (1987), the Washington appellate court found the City of Seattle immune from suit brought by a bicyclist injured on a trail in a city park. Washington’s recreational use statute which was quoted in the case, however, expressly extended immunity to “[a]ny public or private landowners or others in lawful possession and control of any lands whether rural or urban____” Id. at 509, 736 P.2d at 277.

In Syrowik v. City of Detroit, 119 Mich. App. 343, 326 N.W.2d 507 (1982), the Michigan appellate court rejected the plaintiff's contention that Michigan’s recreational use statute should not apply to property located in urban areas because the activities listed in the statute, specifically, “fishing, hunting, trapping, camping, hiking, sightseeing or other similar outdoor recreational use,” are typically enjoyed in rural set[*210] tings. The court pointed out that the statute made no reference to a distinction between rural and urban areas. It found the City of Detroit immune under the statute for a tobogganing accident which occurred in a municipal playfield. However, the Michigan statute applied broadly to “the lands of another.” Id. at 346, 326 N.W.2d at 509. The statute contained no language limiting the types of lands to which it applied.

Similarly, the Idaho Supreme Court in McGhee v. City of Glenns Ferry, 111 Idaho 921, 729 P.2d 396 (1986), and the Ohio Supreme Court in Marrek v. Cleveland Metroparks Bd. of Comm’rs, 9 Ohio St.3d 194, 459 N.E.2d 873 (1984) found cities to be immune under the recreational use statutes of their states for accidents occurring in municipal parks. In both instances, the property encompassed within the act was broadly defined. The Idaho statute granted immunity to owners of the land defined as “private or public land, roads, trails, water, water courses, private or public ways and buildings, structures, and machinery or equipment when attached to or used on the realty.” McGhee, 111 Idaho at 921-22, 729 P.2d at 396-97. The Ohio statute applied to “premises” defined as “all privately-owned lands, ways, waters and any buildings and structures thereon, and all state-owned lands, ways, and waters leased to a private person, firm, organization, or corporation, including any buildings and structures thereon.” Marrek 9 Ohio St.3d at 197, 459 N.E.2d at 876.

A California appellate court also held that California’s recreational use statute is applicable to urban lands in Moore v. City of Torrance, 101 Cal.App.3d 66, 166 Cal. Rptr. 192 (1979), cert denied, 464 U.S. 915, 104 S.Ct. 277, 78 L.Ed.2d 257 (1983). In Delta Farms Reclamation Dist. v. Superior Court, 33 Cal.3d 699, 190 Cal.Rptr. 494, 660 P.2d 1168 (1983), cert. denied, 464 U.S. 915, 104 S.Ct. 277, 78 L.Ed.2d 257 (1983), the California Supreme Court limited Moore only to the extent that Moore held that California’s recreational use statute was intended to apply to public as well as private lands. As in all the other above-mentioned cases, the California statute grants immunity broadly to “[a]n owner of any estate or other interest in real property.”

In contrast to the statutes involved in these cases which place little or no limitation on the types of property encompassed within the statute, the Arizona legislature specifically limited immunity under its recreational use statute to an owner, lessee or occupant of “agricultural, range, mining or forest lands and any other similar lands.” Were we to accept the defendants' argument that the phrase “any other similar lands” expresses a legislative intent that the statute apply to any land upon which, by agreement, any of the stated recreational activities could be performed, the words, “agricultural, range, mining or forest lands” would become surplusage. We are forbidden to construe a statute in such a manner; instead each word, phrase, clause, and sentence must be given meaning so that no part of the statute will be void, inert, redundant, or trivial. City of Phoenix v. Yates, 69 Ariz. 68, 72, 208 P.2d 1147, 1149 (1949); United States v. Mehrmanesh, 689 F.2d 822 (9th Cir. 1982).

Clearly, the legislature chose the words “agricultural, range, mining or forest lands,” for a purpose, and we must look to the common elements existing among these four types of lands to determine whether the property on which Walker was injured can be considered “other similar lands.” The four types of lands listed in the statute normally are relatively large areas of land. Most frequently, these types of property are located outside urban areas in thinly populated rural or semi-rural locales. All four types of land listed in the statute have as their primary use economic activities which are compatible with incidental recreational use.

Also, property of this nature is often in a natural, undeveloped state, although the express language of the statute provides that the legislature would not exclude any buildings or structures which happen to exist on such lands. Moreover, when we look to the recreational activities[*211] enumerated in subsection (B)(2) of the statute, we see that, taken as a whole, they are pursuits that would normally be undertaken on large-sized tracts of land located in “the wilds” rather than in populated areas. Very few of the activities listed would be conducted on the property where plaintiff was injured. Finally, we are mindful that a statute in derogation of the common law, such as this one, must be strictly construed. Schilling v. Embree, 118 Ariz. 236, 575 P.2d 1262 (App.1977). We conclude from the language used in the statute that the legislature did not intend for property such as the McCormick Ranch greenbelt situated within the McCormick Ranch residential area within the City of Scottsdale to come within the purview of the statute.

The recreational use statutes in many other jurisdictions have received a similarly narrow construction due to the language or legislative history of the particular statute. For example, in Tijerina v. Cornelius Christian Church, 273 Or. 58, 539 P.2d 634 (1975), the Oregon Supreme Court construed Oregon’s statute, containing language limiting the types of land within the purview of the statute similar to the language used in Arizona’s statute. The Oregon statute defined “land” within the statute as “agricultural land, range land, forest land, and lands adjacent or contiguous to the ocean shore____” Id. at 63 n. 1, 539 P.2d at 636 n. 1. The plaintiff in that case was injured on a softball field located on a three and one-half acre parcel, owned by the defendant, within the boundaries of a city. The defendant argued that its land should be considered “agricultural land” within the statute because the land produced a substantial growth of volunteer grain and various weeds. The court disagreed, noting that the legislative history of the statute left no doubt that the legislature intended by the restrictive definition of “land” to “limit its application to land holdings which tended to have recreational value but [were] not susceptible to adequate policing or correction of dangerous conditions.” Id. at 64, 539 P.2d at 637. It concluded that the phrase “agricultural land” must be interpreted narrowly to exclude the defendant’s land in light of this legislative purpose.

Washington’s statute was similarly worded when the Washington Court of Appeals decided Kucher v. Pierce County, 24 Wash.App. 281, 600 P.2d 683 (1979). The statute granted immunity for recreational use to “[a]ny public or private landowners or others in lawful possession and control of agricultural or forest lands or water areas or channels and rural lands adjacent to such areas or channels.” Id. at 283 n. 2, 600 P.2d at 685 n. 2. The issue in that case was whether a wooded area of a park located within the City of Tacoma, Washington, should be considered “forest lands” within the meaning of the statute. The court concluded from the legislative history of this statute that the words “agricultural and forest” were meant to limit the scope of the coverage of the act and that the legislature intended that there be room left for the application of the common law of premises liability as to other kinds of land, such as urban residential properties, not covered by the Act. Id. at 286, 600 P.2d at 686. The court observed that any statute in derogation of the common law must be strictly construed. The court went on to conclude that the word “forest” as used in the statute did not include the park land which was improved, routinely inspected, and inside the city of Tacoma. ^[3]

In McCarver v. Manson Park and Recreation Dist., 92 Wash.2d 370, 597 P.2d 1362 (1979), a wrongful death action arising from an accident in a public swimming area within a developed park, the issue was whether the park was within the class of protected land under Washington’s recreational user statute. The majority, in a 5 to 4 decision, found it was. The dissent urged an interpretation of Washington’s recreational user statute recognizing the recreational use as secondary, stating:

The arrangement was clear: While the primary use of the land would continue[*212] to be for agricultural or forestry, a secondary use, without fee, for recreation would be allowed by the landowner in exchange for a limitation of liability ... Nothing in the legislative background or history of the act nor in its language leads to any other conclusion than that ROW 4.24.200-.210 is intended to cover a situation where public recreational use is a secondary and not a primary use of the land or water.

Id. at 379-80, 597 P.2d at 1367. The majority declined to impose such a distinction, stating that “[Arguments to achieve such a result should appropriately be addressed to the legislature.” Id. at 379, 597 P.2d at 1366. ^[4]

The same distinction was made in Stephens v. United States, 472 F.Supp. 998 (C.D.Ill., 1979), a personal injury action brought by a swimmer who was injured when he dove into a lake located within a state park. The district court distinguished between areas used primarily for recreational purposes, to which the Illinois recreational user statute did not apply, and areas where the recreational uses were incidental only, to which the Illinois recreational user statute did apply.

The New Jersey courts have long held that New Jersey’s recreational use statute does not apply to land situated within residential, urban areas. In Boileau v. DeCecco, 125 N.J.Super. 263, 310 A.2d 497 (App.Div.1973), aff' d, 65 N.J. 234, 323 A.2d 449 (1974), the court refused to apply the statute to an injury occurring in a private swimming pool, noting that the activities statutorily described are those occurring in “the true outdoors.” 125 N.J.Super. at 267, 310 A.2d at 500. In Odar v. Chase Manhattan Bank, 138 N.J.Super. 464, 351 A.2d 389 (App.Div.1976), the court stated that the recreational use statute was intended to apply to nonresidential, rural or semi-rural land whereon the enumerated activity would be conducted. The court reached this conclusion even though the recreational use statute provided that immunity applied to “[a]n owner, lessee or occupant of premises,” representing a change from prior wording which had granted immunity to landowners of “woodlands and agricultural lands.” Id. at 467-68, 351 A.2d at 391. See also Magro v. City of Vineland, 148 N.J.Super. 34, 371 A.2d 815 (App.Div.1977).

In Harrison v. Middlesex Water Co., 158 N.J.Super. 368, 386 A.2d 405 (App.Div.1978), the appellate division of the New Jersey Superior Court attempted to break from its long tradition of restrictive application of the statute. In that case the court concluded that the New Jersey statute provided immunity to a water company which owned 136 acres of land, including a 94-acre lake sitting in the midst of a largely residential area. The court concluded that:

[T]he application of the statute depends more upon the nature of the property, in terms of its use for the type of public recreational activity contemplated by the Legislature, than it depends upon the community or neighborhood in which the property is located. There may be a greater coincidence in the approved activity, the type of property and rural or semi-rural areas. We would not expect to find much hunting or trapping in suburbia, though stretches of land there may still support horseback riding and hiking. But the key factors must be the activity and the kind of property and use to which it is put.

Id. at 381-82, 386 A.2d at 412 (citations omitted). The appellate division’s opinion, however, was quickly overruled by the New Jersey Supreme Court in Harrison v. Middlesex Water Co., 80 N J. 391, 403 A.2d 910 (1979). It found that the meaning of the New Jersey recreational use statute was not clear and that interpretation of the word “premises” must be made. In giving a narrow interpretation to the word “premises” the court observed:

The use in the current statute of the word “premises” is in conjunction with provisions that failure to “post” such land in accordance with the posting statute, N.J.S.A. 23:7-1 to -8 would not af-[*213] feet a landowner’s immunity. N.J.S.A. 2A:42A-3. The mention of the posting statutes is a strong indication that the kind of premises which the legislature contemplated when it enacted the Landowner’s Liability Act was primarily undeveloped, open and expansive rural and semi-rural properties where hunting, fishing and trapping might be expected to take place. This is supplemented by the specific inclusion of horseback riding, skiing, and toboganning in addition to other “outdoor sport, game and recreational activity.” N.J.S.A. 2A:42A-2. These are endeavors which can be accommodated, under normal conditions, only upon large sized tracts of rural or semi-rural lands, or other lands having similar characteristics.

Id. at 399-400, 403 A.2d at 914. The court noted that owners of such large rural or semi-rural lands would have difficulty in defending their lands from trespassers or, indeed, even in taking precautions to render them safe for invited persons engaging in these kinds of energetic outdoor activities. Id. at 400, 403 A.2d at 914. The court felt that the act would clearly go beyond the legislature’s goals were it construed to grant a blanket immunity to all property owners, particularly to those owning lands in densely populated urban or suburban areas, without regard to the characteristics of their properties. Id. at 400-01, 403 A.2d at 914. It observed that statutes granting immunity from tort liabilities should be given a narrow range.

Like the courts in the cases cited above, we construe our statute strictly, attentive to the restrictive language chosen by the legislature, in determining the extent of our legislature’s abrogation of traditional common law liability. We conclude from the language chosen by our legislature that it did not intend to grant a blanket immunity to all landowners without regard to the characteristics of their property. In determining what properties are to receive the protection of the statute, characteristics such as size, naturalness, primary and secondary uses of the land, remoteness or isolation from populated areas would all be considered. While the line might be hard to draw in some instances, we are satisfied that the statutory immunity does not extend to the property in question.

Having reached the conclusion that the property where the accident occurred does not fall within the definition of “premises”, it is unnecessary for us to consider Walker’s other contention that she was not a “recreational user” as defined in the statute.

CONCLUSION

For the reasons previously stated, we find that the trial court erred in holding that A.R.S. § 33-1551 applies to the property where Walker’s injuries occurred. The case is reversed and remanded for further proceedings in accordance with this opinion.

NOTE: The Honorable MICHAEL C. NELSON of the Apache County Superior Court, State of Arizona, has been authorized to participate in this matter by the Chief Justice of the Arizona Supreme Court, pursuant to Ariz. Const. Art. VI, § 3.