CHRISTOPHER R. COOPER, United States District Judge
While millions of bison once roamed the entire North American plains, most of the few thousand left in the wild today live in or near Yellowstone National Park. Hoping to protect these remaining bison, Buffalo Field Campaign and other environmental groups petitioned the Fish and Wildlife Service to add the Yellowstone bison population to the federal endangered species list. After the Service made a threshold "90-day" determination that their petition failed to present sufficient scientific evidence that listing the bison may be warranted, they brought suit under the Administrative Procedure Act, alleging that the Service's determination was arbitrary and capricious. Because the Court agrees that the Service applied an improper standard when evaluating Buffalo Field's petition, it will grant Buffalo Field's motion for summary judgment, deny the Service's cross-motion, and remand the case for the agency to conduct a new 90-day finding using the proper standard.
I. Background
A. The Endangered Species Act and Citizen Petitions
The Endangered Species Act serves to protect threatened or endangered species. Under the Act, an "endangered" species is one that is "in danger of extinction throughout all or a significant portion of its range" and a "threatened" species is one that is "likely to become an endangered species within the foreseeable future throughout all or a significant portion of its range." 16 U.S.C. § 1532(6), (20). A *106species can be endangered or threatened by any one of five factors: (1) "the present or threatened destruction, modification, or curtailment of its habitat or range;" (2) "overutilization for commercial, recreational, scientific, or educational purposes;" (3) "disease or predation;" (4) "the inadequacy of existing regulatory mechanisms;" or (5) "other natural or manmade factors affecting its continued existence." Id. § 1533(a)(1). The Secretary of the Interior is responsible for determining whether a species is endangered or threatened, and must base such determinations "solely on ... the best scientific and commercial data available." Id. § 1533(b)(1)(A).
Individuals may petition the Secretary "to add a species to, or to remove a species from" the list of endangered and threatened species. Id. § 1533(b)(3)(A). When the Secretary receives such a petition, he is directed "[t]o the maximum extent practicable, within 90 days after receiving the petition" to "make a finding as to whether the petition presents substantial scientific or commercial information indicating that the petitioned action may be warranted." Id. Service regulations at the relevant time defined "substantial information" as "that amount of information that would lead a reasonable person to believe that the measure proposed in the petition may be warranted." 50 C.F.R. § 424.14(b)(1) (2016).1 If the Secretary concludes that there is substantial evidence, then the petition advances to further review and, within 12 months and following public comment, the Secretary must determine whether the petitioned action is warranted or not and, if it is warranted, publish a proposed implementing regulation. 16 U.S.C. § 1533(b)(3)(B). This additional consideration is known as a "12 month review."
The "substantial evidence" standard applied at the 90-day finding period is not a rigorous one. A petitioner need not present "conclusive evidence regarding" threats to a species. Humane Soc'y of U.S. v. Pritzker, 75 F.Supp.3d 1, 14 (D.D.C. 2014) ; see also Ctr. for Biological Diversity v. Morgenweck, 351 F.Supp.2d 1137, 1140 (D. Colo. 2004) ("[T]he ESA does not require such conclusive evidence that listing is warranted to go to the next step."); Moden v. U.S. Fish & Wildlife Service, 281 F.Supp.2d 1193, 1203 (D. Or. 2003) ("[T]he standard in reviewing a petition to delist does not require conclusive evidence that delisting is warranted."). And in making its 90-day determination, the Service is confined to the information contained in the petition or the Service's files. See, e.g., McCrary v. Gutierrez, 2010 WL 520762, at *2 (N.D. Cal. Feb. 8, 2010) ; Colorado River Cutthroat Trout v. Kempthorne, 448 F.Supp.2d 170, 176 (D.D.C. 2006).
B. The Yellowstone Bison 90-Day Finding
Before European settlers arrived, bison occupied millions of square kilometers across North America. A.R. 376. By 1889, however, they had been driven near extinction, with less than 1,000 still alive in the wild. A.R. 371. Many of the few hundred remaining bison were captured and sent to zoos or private ranches. Id. Today, bison are nearly extinct from their historic range. A.R. 376. The largest extant herd of bison in the wild now occupies roughly 20,000 square kilometers in the area around and inside Yellowstone National Park. A.R. 376. Moreover, the Yellowstone bison are the only significant herd of bison with no evidence of hybridization with cattle, thus representing a genetically important *107population. A.R. 379. As of June 2014, there were an estimated 4,900 bison in the Yellowstone population. A.R. 548.
The bison population resides principally, but not wholly, within the boundaries of Yellowstone National Park. A.R. 5. During the winter, bison can range outside the Park onto public and private lands. A.R. 376. Within its territory, the bison population is organized into two separate herds, known as the "Central" and "Northern" herds. A.R. 385. Some scientific literature has suggested that these herds are in fact genetically distinct populations of bison, each of which should be preserved. See Natalie D. Halbert et al., "Genetic Population Substructure in Bison at Yellowstone National Park," 103 J. Heredity 360, 367 (2012). Other experts have contended that the two herds were artificially created and thus no distinction should be maintained. See Patrick J. White & Rick L. Wallen, "Yellowstone Bison-Should We Preserve Artificial Population Substructure or Rely on Ecological Processes?", 103 J. Heredity 751, 752 (2012).
At present, the bison population in Yellowstone is listed neither as endangered nor threatened. Instead, it is managed by the Interagency Bison Management Plan for the State of Montana and Yellowstone National Park (the "IBMP" or "Management Plan"). The IBMP was adopted in 2000 after a decade of negotiation and planning among the National Park Service, the U.S. Department of Agriculture, and the State of Montana. A.R. 4064. The Plan sought "to continue research and take conservative but protective steps toward cooperative management of the bison while protecting Montana's brucellosis class-free status." A.R. 4066. (Brucellosis is a disease that can be transmitted from bison to cattle and that causes reproductive failure in infected animals. A.R. 404.)
The IBMP sets a target of 3,000 for the entire population of bison in Yellowstone, covering both the Central and Northern herds. A.R. 4093. It also establishes boundaries for the population's territory. A.R. 4072. During the spring, pursuant to the Plan, bison are hazed back into the boundaries of Yellowstone National Park and off of public and private lands outside the Park. A.R. 4072. In addition, the Plan provides for the capture and testing of bison for brucellosis when they leave the Park during the winter migration. Id. at 4072-73. It also details the methods used to monitor pregnant bison, since one of the chief means of transmission of brucellosis is through the remains of miscarriages or live births. Id. at 4072. Finally, the Plan provides that if the bison population in late-winter/early-spring exceeds 3,000, steps may be taken to reduce the population by capturing bison exiting the Park and removing them to quarantine or slaughter. A.R. 4093.
On November 13, 2014, Western Watersheds Project and Buffalo Field Campaign filed a citizen petition (the "First Petition") to list the population of bison in and around Yellowstone National Park as an endangered or threatened species. A.R. 370. A second petition was filed on March 2, 2015 by James Horsley (the "Second Petition"). A.R. 39. Two previous petitions, one submitted in 1999 and the other in 2011, had been denied by the U.S. Fish and Wildlife Service. A.R. 372.
Both petitions contended that the Yellowstone bison population should be listed as threatened or endangered. A.R. 390 (First Petition); A.R. 85 (Second Petition). The First Petition argued that the bison population was threatened or endangered because of restrictions in its range due to historical loss, livestock grazing, infrastructure and development, and invasive species. A.R. 398-400. The First Petition cited overutilization from hunting, disease, and climate change as threats to the bison's *108survival. A.R. 402-06, 413. Finally, the First Petition contended that the existing regulatory mechanism-the IBMP-was inadequate, in part because it (1) was primarily designed to protect against brucellosis-which, the Petition contends, is not nearly as significant a threat from bison-to-cattle transmission as the IBMP claims-rather than to ensure the survival of the bison, and (2) fails to account for the two distinct genetic herds when setting a population target and thus sets too low a population target to ensure the genetic survival of both herds (and thereby the entire population). A.R. 406-10. The Second Petition took similar aim at the IBMP and the population target it set, arguing that they are inadequate to maintain the genetic diversity of the bison, that the Plan focused on brucellosis prevention rather than bison preservation, and that over-utilization and loss of habitat threatened the species. A.R. 70, 84, 85-86.
On December 15, 2015-over six months after the Second Petition was filed and a year following the First-the U.S. Fish and Wildlife Service denied the two petitions. A.R. 16. The Service ultimately held that the petitions had failed to present substantial scientific or commercial evidence indicating that listing the Yellowstone bison may be warranted. A.R. 2. The Service noted that, in light of the current stable-to-increasing bison population, the petitions failed to present substantial evidence of a threat from loss of historic range or disease. A.R. 3, 11. With respect to livestock grazing, development and infrastructure, and invasive species-threats identified in the First Petition-the Service stated that that petition did not provide evidence of how these issues affected the Yellowstone bison. A.R. 5-6. The Service also dismissed concerns over genomic extinction or climate change as threats. A.R. 14-15. Finally, with respect to the contentions regarding the IBMP and its population target, the Service stated that the IBMP already set separate population targets for the two herds and that maintenance of the two distinct populations might not be crucial for the survival of the species. A.R. 9.
Buffalo Field Campaign and Western Watershed Project, both of whom filed the First Petition, as well as Friends of Animals (collectively "Buffalo Field"), subsequently filed suit against Secretary of the Interior Ryan Zinke as well as the U.S. Fish and Wildlife Service and its Acting Director Jim Kurth (collectively "the Service") challenging the Service's 90-day finding. Compl. ¶ 1. They alleged that the Service's conclusion was arbitrary and capricious under the Administrative Procedure Act ("APA") because the Service ignored the plain language of the Endangered Species Act, failed to follow the requirement to make decisions based on the best scientific and commercial data available, and applied an improper evidentiary standard. Id. ¶ 6. Buffalo Field and the Service filed cross-motions for summary judgment, and the Court held a hearing on January 18, 2018. After considering the parties' arguments in their briefs and at the hearing, the Court will grant Buffalo Field's motion for summary judgment and deny the Service's.
II. Standard of Review
"Summary judgment is the proper mechanism for deciding, as a matter of law, whether an agency action is supported by the administrative record and consistent with the APA standard of review." Blue Ocean Institute v. Gutierrez, 585 F.Supp.2d 36, 41 (D.D.C. 2008). In evaluating the case on summary judgment, the Court applies the standard of review from the APA.
Under the APA, a court shall hold unlawful and set aside agency action that is arbitrary and capricious, an abuse *109of discretion, or otherwise not in accordance with the law. 5 U.S.C. § 706(2)(A). Agency action is arbitrary and capricious when "the agency has relied on factors which Congress has not intended it to consider, entirely failed to consider an important aspect of the problem, offered an explanation for its decision that runs counter to the evidence before the agency, or is so implausible that it could not be ascribed to a difference in view or the product of agency expertise." Motor Vehicle Mfrs. Ass'n of U.S., Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43, 103 S.Ct. 2856, 77 L.Ed.2d 443 (1983). An agency's decision is an abuse of discretion if the agency has applied an incorrect legal standard in making its decision. See Price v. District of Columbia, 792 F.3d 112, 114 (D.C. Cir. 2015) ("An abuse of discretion occurs by definition when the court does not apply the correct legal standard ...").
The scope of review under the arbitrary and capricious standard is "narrow," and the Court "is not to substitute its judgment for that of the agency." Motor Vehicle Mfrs., 463 U.S. at 43, 103 S.Ct. 2856. Additionally, the Court is confined to the administrative record in conducting its review. See, e.g., Marcum v. Salazar, 751 F.Supp.2d 74, 78 (D.D.C. 2010).
III. Analysis
Buffalo Field contends that the Service's 90-day finding is arbitrary and capricious because the Service (1) applied the incorrect evidentiary standard for a 90-day finding, (2) did not apply the best available science standard, and (3) failed to consider or otherwise ignored evidence indicating that the bison are threatened or endangered. See generally Mem. P. & A. Supp. Pls.' Mot. Summ. J. ("Buffalo Field MSJ"). Because the Court agrees that the Service failed to apply the proper standard for a 90-day finding, summary judgment is appropriate for Buffalo Field.
A. Whether the Service applied the proper standard for a 90-day finding.
Buffalo Field argues that the Service's determination is arbitrary and capricious because the Service applied an improperly heightened evidentiary standard to its petition. As evidence of this, Buffalo Field points to the 90-day finding's reliance on the study on the two distinct subpopulations of bison by Patrick J. White and Richard L. Wallen-which argued that the distinct populations were artificially created and thus should not be maintained separately-over that by Natalie D. Halbert et al.-which argued for a change in management based on the existence of two separate genetic herds. Buffalo Field MSJ at 1. Additionally, it points to internal Service documents from the reviewing and drafting process that it argues further illustrate an improper standard was applied. Id. at 21-23.
As other district courts have explained, "[a]t the 90-day stage, the question is not whether the designation is warranted, only whether it may be." Center for Biological Diversity v. Kempthorne, 2007 WL 163244, at *7 (N.D. Cal. Jan. 19, 2007). It follows that "the standard requiring consideration of whether a 'reasonable person' would conclude that action 'may be warranted' contemplates that where there is disagreement among reasonable scientists, the Service should make the 'may be warranted' finding." Id.; see also McCrary, 2010 WL 520762, at *6 ; Center for Biological Diversity v. Kempthorne, 2008 WL 659822, at *12 (D. Ariz. March 6, 2008). After all, if reasonable scientists disagree-and one of those positions would indicate listing is warranted-a reasonable person could choose to agree with the scientist who supports the petition and, as a result, that listing may be warranted. At the hearing, the Service did not contest *110that this is the proper standard for a 90-day hearing.
It naturally follows that the 90-day standard does not allow the Service to simply discount scientific studies that support the petition or to resolve reasonable extant scientific disputes against the petition. Unless the Service explains why the scientific studies that the petition cites are unreliable, irrelevant, or otherwise unreasonable to credit, the Service must credit the evidence presented. See Pritzker, 75 F.Supp.3d at 14 ; cf. Morgenweck, 351 F.Supp.2d at 1142 (the Service need not "blindly accept statements in petitions that constitute unscientific data or conclusions, [or] information FWS knows to be obsolete"). In other words, if two pieces of scientific evidence conflict, the Service must credit the supporting evidence unless that evidence is unreliable, irrelevant, or otherwise unreasonable to credit.
This case presents a relatively straightforward example of such a scientific dispute. In their paper, Halbert et al. argue that there are two genetically distinct subpopulations of bison in Yellowstone and that this conclusion undermines the current management practices (and population levels) for the bison. See A.R. 721. Put another way, Halbert's study calls into question the 3,000 bison population target set by the IBMP for the entire herd by indicating that the two herds-Central and Northern-each need a population that is large enough to ensure that herd's individual survival. Since other studies have suggested that around 3,000 bison are needed to ensure a herd's survival, see A.R. 688, 1269, this suggests that the 3,000 bison population target for both herds is too low to ensure that each herd will survive. In contrast, White and Wallen argue that the two subpopulations are artificially created and management practices should not attempt to maintain them-thereby indicating that the current target is sufficient and that any hazing or culling to reduce population levels to that number is not a threat to the species' survival. See A.R. 674.
In its 90-day finding, the Service had two responses to the Halbert study. First, it stated that the IBMP already sets population management targets for the two herds separately. A.R. 9. While that may be the case, the Service's position ignores the inference from the Halbert study that the overall population target is either too low in light of other studies indicating 3,000 bison are needed for each herd to ensure survival, or inaccurate because it was set before recognizing that the two herds need to be maintained individually. Second, the Service simply adopted White and Wallen's conclusion that "maintenance of subpopulation genetic differentiation and overall genetic diversity may not be crucial for preserving genes" and that the Service therefore need not maintain the two subpopulations. A.R. 9-10. The Service offered no explanation for why Halbert et al.'s conclusion was irrelevant, unreliable, or unsubstantiated, nor any discussion why a reasonable person could not rely on the Halbert study over the White and Wallen study. In so doing, the Service appears to have taken it upon itself to resolve a disagreement among reasonable scientists, an observation echoed by arguments made in its brief defending the side it picked. See Mem. Supp. Fed. Defs.' Cross-Mot. Summ. J., at 27-28; Fed. Defs.' Reply Supp. Cross-Mot. Summ. J., at 14-19. The Service thereby applied an inappropriately heightened standard to the evaluation of Buffalo Field's petition. See McCrary, 2010 WL 520762, at *7 ; Kempthorne, 2007 WL 163244, at *7.
The Service responds that the Court should defer to the Service's expertise in evaluating the evidence presented in citizen petitions. That is certainly true with respect to the Service's conclusion that a *111specific study is unscientific, obsolete, or unreliable. But the Service made no such conclusions here: it offered no explanation whatsoever for its rejection of the Halbert study (other than its reliance on the White and Wallen study). Similarly, the Court might give deference to the Service's conclusion that a reasonable scientist would not rely on Halbert et al. But again, the Service provides no such justification for its conclusion here. Ultimately, the Service simply picked a side in an ongoing debate in the scientific community, which is improper at the 90-day finding stage. The Court need not defer to an agency's application of the improper legal standard.
Finally, the Service contended at the hearing that by not allowing it to resolve scientific discrepancies or disputes, the Court will collapse any distinction between the 90-day finding and the 12-month review. This concern is overstated. The Court is not requiring the Service to accept any and all positions that a petition advances. Nor is it requiring the Service to credit scientific studies it knows to be obsolete, wholly unreliable, or irrelevant. See Morgenweck, 351 F.Supp.2d at 1142 ; cf. Pritzker, 75 F.Supp.3d at 14 (indicating that the Service can disregard obsolete studies or unsupported allegations). Even Buffalo Field agreed at the hearing that the Service can use its expertise in this fashion. Thus, the Service will not be required to grant every petition that appears before it-it simply cannot resolve outstanding disputes between reasonable scientists over relevant matters in the course of a 90-day review.
In sum, by resolving an outstanding scientific dispute, the Service applied an improperly heightened standard in making its 90-day determination. This renders the Service's decision arbitrary and capricious, and entitles Buffalo Field to summary judgment in its favor.2
B. The appropriate remedy.
Buffalo Field argues that if the Service's decision is arbitrary and capricious, the Court should direct the Service to begin a 12-month review rather than simply remand the case to the Service. The Court, however, agrees with the Service that remand is the appropriate remedy.
As the D.C. Circuit has repeatedly explained, a "district court reviewing a final agency action 'does not perform its normal role but instead sits as an appellate tribunal.' " Palisades General Hosp. Inc. v. Leavitt, 426 F.3d 400, 403 (D.C. Cir. 2005) (citation omitted). Thus, "when a court reviewing agency action determines that an agency made an error of law, the court's inquiry is at an end: the case must be remanded to the agency for further action consistent with the corrected legal standards." County of Los Angeles v. Shalala, 192 F.3d 1005, 1011 (D.C. Cir. 1999) ; see also North Carolina Fisheries Ass'n, Inc. v. Gutierrez, 550 F.3d 16, 20 (D.C. Cir. 2008). Indeed, the D.C. Circuit has held that district courts commit error when they impose specific relief-such as that Buffalo Field requests-on an agency. See, e.g., Palisades General, 426 F.3d at 403 ("[T]he district court had jurisdiction only to vacate the Secretary's decision rejecting the hospital's revised wage data and to remand for further action consistent with its opinion."); County of Los Angeles, 192 F.3d at 1011 ("Not only was it unnecessary for the [district] court to retain jurisdiction to devise a specific remedy for the Secretary to follow, but it was error to do so."). In light of this precedent, the Court concludes *112that it would be improper to direct the Service to begin a 12-month review.
The cases that Buffalo Field cites where a court did direct the Service to begin a 12 month review rather than simply remand-including one from this District-do not persuade the Court otherwise. In most of them, including the case from this District, the Service's error resulted from its reliance on third-party data or information. See Western Watersheds Project v. Kempthorne, 2009 WL 10678130, at *9 (D. Idaho March 31, 2009) (Service erred by soliciting information from third party scientist when making 90-day determination); Colorado River, 448 F.Supp.2d at 176-177 (Service erred by soliciting information from states and other agencies when making 90-day review); Morgenweck, 351 F.Supp.2d at 1143 (D. Colo. 2004) (same). Thus, in those cases, the Service effectively began a 12-month review during the 90-day review stage. As such, the courts were simply directing the Service to continue what it had in essence already begun. But the Service has not committed that error here-it has looked solely at the information in the two petitions and its files. In other words, the Service has not begun a de facto 12-month review disguised as a 90-day one. Given this distinction, the cases that Buffalo Field cites are inapposite. The appropriate remedy is to remand to the Service to allow it to conduct a 90-day finding using the appropriate standard.
IV. Conclusion
The Service argues that ultimately, Buffalo Field's concern over the two subpopulations is irrelevant because the bison population is growing. The Service may have a point, and on remand the Service may well be able to reach the same outcome after applying the proper standard. But to do so, the Service must explain why the evidence supporting the petition is unreliable, irrelevant, or otherwise unreasonable to credit rather than simply pick and choose between contradictory scientific studies.
For the foregoing reasons, the Court grants Buffalo Field's motion for summary judgment and denies the Service's cross-motion for summary judgment. A separate Order will accompany this Memorandum Opinion.