The named plaintiffs in this class action own real property in Mallard Pointe, a residential neighborhood in Effingham County. Nearby, since 1986, Georgia-Pacific Consumer Products, LP has operated the Savannah River Mill, a facility that includes more than a hundred acres of sludge fields, into which Georgia-Pacific puts the solid waste generated at the Mill. As this solid waste decomposes, the plaintiffs say, hydrogen sulfide gas is released from the sludge fields. Alleging that their real property has been contaminated by this gas — and that, as a result, they have been exposed to noxious odors, their use and enj oyment of their property has been impaired, and the value of their property has diminished — the plaintiffs sued Georgia-Pacific for nuisance, trespass, and negligence. The plaintiffs sought not only to recover monetary damages for themselves, but they proposed to seek relief for a class of other nearby property owners.
The trial court permitted them to do so, certifying a class that consists of the owners of 67 parcels of real property in and around Mallard Pointe.1 Georgia-Pacific appealed the certification of the *525class,2 and the Court of Appeals affirmed, Georgia-Pacific Consumer Products, LP v. Ratner, 323 Ga. App. 203, 203-212 (746 SE2d 829) (2013), although three of its judges dissented. See id. at 213-221 (Branch, J., dissenting). Upon the petition of Georgia-Pacific, we issued a writ of certiorari to review the decision of the Court of Appeals. We conclude that the trial court abused its discretion when it certified the class, and we reverse the judgment of the Court of Appeals.
1. “The class action is an exception to the usual rule that litigation is conducted by and on behalf of the individual named parties only,” Comcast Corp. v. Behrend, _ U. S. _. _ (II) (133 SCt 1426, 185 LE2d 515) (2013) (citation and punctuation omitted), and consistent with its exceptional nature, a class action is permitted only in the limited circumstances described in OCGA § 9-11-23.3 The party seeking to represent a class “bear[s] the burden of proving that class certification is appropriate.” Carnett’s, Inc. v. Hammond, 279 Ga. 125, 127 (3) (610 SE2d 529) (2005) (citation omitted). See also McGarry v. Cingular Wireless, LLC, 267 Ga. App. 23, 25 (1) (599 SE2d *52634) (2004). In this case, to permit the certification of a class of plaintiffs, the named plaintiffs had to satisfy each of the four requirements described in OCGA § 9-11-23 (a) — numerosity,4 commonality,5 typicality,6 and adequacy of representation7 —• as well as the predominance requirement of OCGA § 9-11-23 (b) (3).8 See American Debt Foundation v. Hodzic, 312 Ga. App. 806, 808 (720 SE2d 283) (2011). To satisfy these requirements, it was not enough for the plaintiffs simply to have alleged that they were satisfied. Wal-Mart Stores, Inc. v. Dukes, _ U. S. _, _ (II) (A) (131 SCt 2541, 180 LE2d 374) (2011). See also Fortis Ins. Co. v. Kahn, 299 Ga. App. 319, 321-322 (1) (683 SE2d 4) (2009). Rather, the plaintiffs had to come forward with evidence to prove their satisfaction of the statutory requirements. See Dukes, _ U. S. at _ (II) (A). See also Jones v. Douglas County, 262 Ga. 317, 324 (2) (418 SE2d 19) (1992); Rite Aid of Ga. v. Peacock, 315 Ga. App. 573, 574-575 (1) (726 SE2d 577) (2012).
Whether to certify a class is a matter committed to the discretion of the trial court, State Farm Mut. Auto. Ins. Co. v. Mabry, 274 Ga. 498, 499-500 (1) (556 SE2d 114) (2001), but any exercise of that discretion must comport with the statutory requirements. Moreover, the certification of a class is appropriate only to the extent that “the trial court is satisfied, after a rigorous analysis, that [the statutory requirements] have been satisfied.” Gen. Tel. Co. of the Southwest v. Falcon, 457 U. S. 147, 161 (III) (102 SCt 2364, 72 LE2d 740) (1982). See also Rite Aid, 315 Ga. App. at 574-575 (1); Kahn, 299 Ga. App. at 321 (1). As a part of this rigorous analysis, “sometimes it may be necessary for the court to probe behind the pleadings before coming to rest on the certification question.” Falcon, 457 U. S. at 160 (II). Indeed, as the United States Supreme Court has explained:
Frequently that rigorous analysis will entail some overlap with the merits of the plaintiff’s underlying claim. That *527cannot be helped. The class determination generally involves considerations that are enmeshed in the factual and legal issues comprising the plaintiff’s cause of action.
Dukes, _ U. S. at _ (II) (A) (citation and punctuation omitted). See also Rite Aid, 315 Ga. App. at 575 (1); McGarry, 267 Ga. App. at 25 (1).
Upon our review of the record, we conclude that the plaintiffs failed to come forward with evidence sufficient to show the commonality of the particular class that was certified. Georgia-Pacific raises some fair questions about typicality and predominance as well,9 but we do not have to reach those questions today. Because commonality is lacking, the trial court abused its discretion when it certified the class, and the Court of Appeals should have reversed the certification.
2. To show commonality, the plaintiffs had to demonstrate that “[t]here are questions of law or fact common to the class,” OCGA §9-11-23 (a) (2), but as the United States Supreme Court recently explained in Dukes, “[t]hat language is easy to misread, since any competently crafted class complaint literally raises common questions.”_U. S. at_(II) (A) (citation and punctuation omitted). Every putative class action presents questions that are, in a sense, “common” to the class. But for commonality under OCGA § 9-11-23 (a) (2), not just any “common” questions will do. Commonality depends on the presence of a particular sort of “common” question, and simply reciting a list of questions that are “common” in another sense contributes nothing to the commonality inquiry.10 Dukes, _ U. S. at _ (II) (A).
*528To establish the sort of commonality that OCGA § 9-11-23 (a) (2) requires, the plaintiffs were required to show “that the class members have suffered the same injury.”11 Dukes, _ U. S. at _ (II) (A) (citation and punctuation omitted). See also Rite Aid, 315 Ga. App. at 575 (1) (a). To do so, the plaintiffs had to point to a “common contention” that each member of the class had suffered the same instance or course of wrongful conduct, see Dukes, _ U. S. at _ (II) (A), and the plaintiffs also had to show that this “common contention” “is capable of classwide resolution — which means that determination of its truth or falsity will resolve an issue that is central to the validity of each one of the claims in one stroke.”12 Id. “What matters to class certification is not the raising of common questions — even in droves — but, rather the capacity of a classwide proceeding to generate common answers apt to drive the resolution of the litigation.” Id. (citation and punctuation omitted; emphasis in original). Although Dukes makes these points more clearly than our own precedents, it is entirely consistent with those precedents. See Carnett’s, 279 Ga. at 129 (4) (“Acommon question is not enough when the answer may vary with each class member and is determinative of whether the member is properly part of the class.” (Citation omitted; emphasis in original.)).
The plaintiffs in this case pointed to a “common contention” that might properly form the basis for a finding of commonality, namely, *529that their properties were contaminated with hydrogen sulfide gas released from the sludge fields at the Mill. But pointing to a “common contention” is only the first step. The plaintiffs also had to show that this “common contention” is “capable of classwide resolution” with respect to the particular class that the trial court certified. And that is where they came up short.
We do not find in the record evidence by which the plaintiffs might be able to prove on a classwide basis that the entire area by which the class was defined, in fact, was contaminated with hydrogen sulfide gas from the sludge fields. There is, for instance, no scientific evidence of the amounts of gas released from the sludge fields, no evidence of the rate of release, no evidence of the extent to which the amounts released and rates of release varied over time, and no evidence of exactly how the gas would be expected to move through the air upon its release. As for the direction in which gas would be expected to move, the only evidence of local wind patterns is that the prevailing winds over the sludge fields blow away from the class area, and the winds blow from the sludge fields toward any specific location within the class area only about four percent of the time. There is no evidence of the rate at which the gas would be expected to dissipate following its release, and there is no evidence of air quality sampling across the class area at any time.13
The record does contain some anecdotal evidence of hydrogen sulfide gas in some areas around the Mill, but this anecdotal evidence is not enough to satisfy a rigorous analysis with respect to the commonality of the particular class that the trial court certified. First, there is evidence of multiple complaints to Georgia-Pacific about the corrosion of air-conditioning units in the vicinity of the Mill. But as best we can tell from the record, a number of these complaints did not even originate within the area by which the class was defined, and those that did appear to have originated in a geographically compact and relatively small portion of the class area, all in Mallard Pointe itself or otherwise along the far eastern edge of the class area, *530just across Fort Howard Road from the Mill.14 The class area stretches substantially to the west of the small and compact area from which these complaints came, and the complaints do not, therefore, tend to show that the class area as a whole was contaminated by hydrogen sulfide gas.
There also is the testimony of two air-conditioning technicians, both of whom testified that they have observed corrosion in a number of air-conditioning units in the vicinity of the Mill, which was, they surmised, a result of excessive hydrogen sulfide exposure. But according to the record, to determine whether any particular corrosion actually was a result of hydrogen sulfide exposure, it would be necessary to test the corroded components, and neither of the technicians has submitted any parts for such testing. Consequently, their opinion that the corrosion is due to hydrogen sulfide exposure seems to amount to nothing more than conjecture. And in any event, one of the technicians testified that the corrosion he observed was near the Mill, but not within the class area. The other technician said that he saw corrosion in a few air-conditioning units within the class area and outside Mallard Pointe, but the locations at which he saw corrosion still were close in proximity — relative to the class area as a whole — to Mallard Pointe. Without more, evidence of corrosion at these locations does not tend to show any likelihood that the class area as a whole was contaminated.
There also was evidence from the named plaintiffs themselves, who noted that they had been exposed to noxious odors at their homes, which they attributed to hydrogen sulfide gas. But again, the named plaintiffs live in Mallard Pointe, which sits just across the road from the Mill itself. The idea that their testimony tends to show that the class area as a whole — which stretches for some distance to the west of Mallard Pointe — has been exposed to harmful hydrogen sulfide gas is a dubious one.
From the evidence of record, one fairly could conclude that a number of people in the vicinity of the Mill — especially those living in Mallard Pointe — have experienced problems that might be attributable to the alleged release of hydrogen sulfide gas from the sludge fields at the Mill. But the locations at which these people have experienced such problems (as evidenced by the record) do not seem to coincide with the full extent of the class area. Whether the *531“common contention” of harmful exposure can be proved on a class-wide basis is in doubt, at least as to the class as the trial court defined it. As Judge Branch aptly said in her dissent:
[T]he class appears to have been defined not by any logical determination of the actual effects of the Mill’s numerous and intermittent releases of hydrogen sulfide on persons and property over a period of years, but rather by means of arbitrarily drawn lines on a map.
Ratner, 323 Ga. App. at 218 (Branch, J., dissenting) (citation and punctuation omitted).15
No one should misunderstand us to say that commonality never can be shown in the context of environmental mass torts, that it cannot be shown in this case, or even that it cannot be shown in this case as to the class as the trial court defined it. It certainly is conceivable that the plaintiffs might show the requisite commonality of that class or another. Cf. Brenntag Mid South, Inc. v. Smart, 308 Ga. App. 899, 902 (2), n. 2, 903-904 (2) (a) (ii) (710 SE2d 569) (2011) (plaintiffs in class action asserting claims of environmental torts satisfied commonality requirement with proof that each class member actually was evacuated as a result of accidental release of acidic gas, and class was defined in terms of the locations covered by “evacuation orders as well as expert consultation as to ‘the actual geographical extent of the chemical plume as it spread’ ”). But if the plaintiffs are to satisfy the commonality requirement, they have some more work to do. The certification of the class was an abuse of discretion, and the Court of Appeals should have reversed it. Accordingly, we reverse the judgment of the Court of Appeals.
Judgment reversed.
All the Justices concur, except Benham and Hunstein, JJ., who dissent.Generally speaking, these 67 properties are in a contiguous area to the west of Fort Howard Road. The Mill is situated just to the east of Fort Howard Road. Of the properties in the class area, the lots in Mallard Pointe — a neighborhood that is located just off Fort Howard Road *525— and a few other properties on Fort Howard Road appear to he the closest in proximity to the sludge fields at the Mill. To be exact, the trial court defined the class as follows:
All citizens of the State of Georgia who, as of November 18, 2010, owned property lying, in whole or in part, within an area of land lying in Effingham County, Georgia, and bounded as follows:
On the east by a line running along the west side of the right of way of Fort Howard Road between the intersection of Fort Howard Road and Seckinger Ford Road and the south side of the right of way of the railroad line that serves the Georgia-Pacific plant;
On the north by a line running along the south side of said railroad right of way between For[t] Howard Road and Rincon-Stillwell Road;
On the west by a line running along the east side of the right of way of Rincon-Stillwell Road; [and]
On the south by a line running along the north side of the right of way of Bunyan Kessler Road beginning at Rincon-Stillwell Road and going in an easterly directly to a point at Latitude 32.30582649 and Longitude minus 81.21047668 and thence easterly along a straight line to the intersection of Fort Howard Road and Seckinger Ford Road.
Expressly excluded from membership in the class are [Georgia-Pacific], its related corporations and all directors, officers and employees of [Georgia-Pacific].
This class definition identifies the members of the class with enough precision.
See OCGA § 9-11-23 (g) (“A court’s order certifying a class or refusing to certify a class shall be appealable in the same manner as a final order to the appellate court which would otherwise have jurisdiction over the appeal from a final order in the action.”).
Many provisions of OCGA § 9-11-23 were borrowed from Federal Rule of Civil Procedure 23, and for this reason, when Georgia courts interpret and apply OCGA § 9-11-23, they commonly look to decisions of the federal courts interpreting and applying Rule 23. See, e.g., State Farm Mut. Auto. Ins. Co. v. Mabry, 274 Ga. 498, 499 (1) (556 SE2d 114) (2001); Rite Aid of Ga. v. Peacock, 315 Ga. App. 573, 574 (1) (726 SE2d 577) (2012); Brenntag Mid South, Inc. v. Smart, 308 Ga. App. 899, 903 (2) (710 SE2d 569) (2011); Fuller v. Heartwood 11, LLC, 301 Ga. App. 309, 312 (687 SE2d 287) (2009); Gay v. B. H. Transfer Co., 287 Ga. App. 610, 611, n. 2 (652 SE2d 200) (2007).
OCGA § 9-11-23 (a) (1) (“The class is so numerous that joinder of all members is impracticable!)]”).
OCGA § 9-11-23 (a) (2) (“There are questions of law or fact common to the class].]”).
OCGA § 9-11-23 (a) (3) (“The claims ... of the representative parties are typical of the claims ... of the class].]”).
OCGA § 9-11-23 (a) (4) (“The representative parties will fairly and adequately protect the interests of the class.”).
OCGA § 9-11-23 (b) (3) (“]T]he questions oflaw or fact common to the members ofthe class predominate over any questions affecting only individual members, and ... a class action is superior to other available methods for the fair and efficient adjudication of the controversy”). For a class action to proceed, all the requirements of OCGA § 9-11-23 (a) must be satisfied, as well as one ofthe three requirements described in OCGA § 9-11-23 (b). Here, the plaintiffs have relied exclusively upon OCGA § 9-11-23 (b) (3), and that is the only paragraph of subsection (b), therefore, with which we are concerned.
Georgia-Pacific does not dispute that the certified class is sufficiently numerous, see OCGA § 9-11-23 (a) (1), and aside from the question of typicality, they appear not to dispute that the named plaintiffs are adequate representatives of the class. See OCGA § 9-11-23 (a) (4). See also Horton v. Goose Creek Independent School Dist., 690 F2d 470, 485 (IV), n. 27 (5th Cir. 1982) (“The requirements [of typicality and adequacy] are closely related, for demanding typicality on the part of the representative helps ensure his adequacy as a representative”).
The trial court — and the majority of the Court of Appeals — appear to have fallen into the very analytical trap against which the United States Supreme Court warned in Dukes. In its order certifying the class, the trial court recited a list of “common” questions that, although “common” in a sense, do not show the sort of commonality required for OCGA § 9-11-23 (a) (2). For instance, the trial court noted that the case involves “common” questions concerning:
• The materials used by Georgia-Pacific at the Mill, including the ways in which those materials are stored and used;
• The manner in which Georgia-Pacific operates the Mill, including its sludge fields;
• The reasons for releases of hydrogen sulfide gas from the sludge fields;
• The potential effects of a release of hydrogen sulfide gas, including the toxic and corrosive properties of the gas;
*528• The nature and adequacy of precautions taken by Georgia-Pacific to prevent the release of hydrogen sulfide gas;
• The “liability of Georgia-Pacific”;
• The “common affirmative defenses raised by Georgia-Pacific”; and
• The remedies available to members of the class, including the appropriate measure of damages, exemplary damages, and expenses of litigation.
The majority in the Court of Appeals recited a similar list of “common” questions. See Ratner, 323 Ga. App. at 208 (1) (b). No doubt, these questions may be “common” in a sense, but they do not show the commonality that the statute requires, and “[rjeciting these questions is not sufficient to [warrant] class certification.” Dukes, _ U. S. at _ (II) (A).
To be clear, when it referred in Dukes to the need to demonstrate “that the class members have suffered the same injury,” we do not understand the Supreme Court to have meant that all the class members must have sustained precisely the same damages. See In re Deepwater Horizon, 739 F3d 790, 810-811 (IV) (A) (5th Cir. 2014) (“[T]he legal requirement that class members have all ‘suffered the same injur/ can be satisfied by an instance of the defendant’s injurious conduct, even when the resulting injurious effects — the damages — are diverse.”).
At class certification, plaintiffs need not prove the merit of their “common contention” • — • that is, they need not definitively prove that each and every member of the class did, in fact, suffer the “same injury.” But they have to show that their “common contention” is capable of being proved at once for the whole class. To do that, they have to point to some evidence by which they could prove the merit of their “common contention.”
The dissent relies on Smith v. ConocoPhillips Pipe Line Co., 2014 U.S. Dist. LEXIS 43172 (E.D. Mo. Mar. 31, 2014), and Jackson v. Unocal Corp., 262 P3d 874 (Colo. 2011). In each of these cases, the courts were presented with scientific evidence that tended to show the contamination of the proposed class area. Smith, 2014 U.S. Dist. LEXIS 43172, at *11-14 (IV) (A); Jackson, 262 P3d at 887-888 (IV) (A). As Jackson notes, “[u]sually, scientific or objective evidence closely ties the spread of the alleged pollution or contamination to the proposed class boundaries, as many mass environmental tort cases demonstrate.” 262 P3d at 887 (IV) (A) (citation and punctuation omitted). But “where plaintiffs fail to produce evidence linking the proposed boundary to the environmental hazard, courts have denied certification.” Smith, 2014 U.S. Dist. LEXIS 43172, at *10 (IV) (A) (citations omitted).
There is evidence of complaints that do not involve the corrosion of air-conditioning units, but it appears that most of those complaints did not originate within the class area and that those that did — like the complaints about air-conditioning units — originated in or near Mallard Pointe.
The record bears out the idea that the class was defined ‘by means of arbitrarily drawn lines on a map.” According to the testimony of record, the area by which the class was defined was identified in significant part by a real estate appraiser, who simply drove around in the vicinity of the Mill, looking for “major roads” and other “specific geographical boundaries” that “seemed reasonable” to bound the class area. When “it seemed like we were getting too far away from [the Mill],” the appraiser apparently would turn back and look for a closer boundary. The appraiser testified that he went in search of the boundaries of the class area without any scientific analysis of air quality or expected gas movements, without knowledge of local wind patterns, and without knowledge of other potential nearby sources of gases that might cause noxious odors. The appraiser also did not rely upon any appraisal methodology in helping to select boundaries for the class area, admitting explicitly that “it’s a seat of the pants kind of thing.”