(dissenting).
I join in Part III of the court’s opinion, which affirms the postconviction court’s summary denial of Rhodes’s third petition for postconviction relief. But I disagree with the court’s decision in Part TV, which affirms the summary denial of Rhodes’s fourth petition. The postconviction court abused its discretion by summarily denying the fourth petition because this petition raises rriaterial fact questions, in an entirely circumstantial case, that must be resolved at an evidentiary • hearing. Contrary to the court’s conclusion, Rhodes’s fourth petition is legally sufficient to establish the “innocence” prong to overcome the 2-year statute of limitations, Minn.Stat. § 590.01, subd. 4(b)(2) (2014). More specifically, Rhodes has alleged the existence of evidence that, if true and considered in the light most favorable to the petition, would establish by a clear and convincing standard that no reasonable jury would have convicted Rhodes -had the newly discovered evidence beeri presented at trial. Therefore, I would remand for an eviden-tiary hearing. I respectfully dissent.
’ I.
Under the postconviction statute; Minn. Stat. § 590.04, subd. 1 (2014), a postconviction court “shall” promptly grant an evi-dentiary hearing on a petition for postcon-viction relief unless “the petition and the files and records of the proceeding conclusively show that the petitioner is entitled to no relief.” (Emphasis added.) To determine whether an evidentiary hearing is required, the postconviction court must assume the facts alleged in the petition to be true, Bobo v. State, 820 N.W.2d 511, 516 (Minn.2012), and consider those facts in the light most favorable to, the petition, Riley v. State, 819 N.W.2d 162, 167 (Minn.2012). The postconviction court “must grant the evidentiary hearing whenever material facts are in dispute.” Wilson v. State, 726 N.W.2d 103, 107 (Minn.2007). “Any doubts about whether to conduct an evidentiary hearing should be -resolved in favor of [the petitioner].” Bobo, 820 N.W.2d at 516 (emphasis added). And an evidentiary hearing is “particularly important when the petition ‘attacks’ important evidence in a .circumstantial case.” Wilson, 726 N.W.2d at 107 (emphasis added) (citing Opsahl v. State, 677 N.W.2d 414, 423 (Minn.2004)). This is a wholly circumstantial case. State v. Rhodes (Rhodes II), 657 N.W.2d 823, 827 (Minn.2003) (stating that "“the evidence in this case is wholly circumstantial”).
The court holds that Rhodes’s fourth petition fails to establish the “innocence” prong under the newly-discovered-evidence exception to the postconviction statute of 'limitations. Minn.Stat. § 590.01, subd. 4(b)(2). The innocence prong requires ' Rhodes to allege the existence of evidence that, if true, would establish by' a clear and convincing standard that Rhodes is innocent. See id.; Miles v. State, 800 N.W.2d 778, 783 (Minn.2011). But the court does not discuss our recent precedent on the meaning and application of this innocence prong. Contrary to the approach suggested by the court, the word “innocent” in this statute requires neither a complete contradiction of incriminating evidence at trial nor a showing that the petitioner certainly or “unequivocally” did not commit the crime for which he was convicted. This approach is not supported by precedent and, moreover, it would be impossible to meet such a standard in á circumstantial case. Rather, our precedent indicates that this innocence prong refers to the “actual innocence” rule, which *793is established when no reasonable jury would have found proof of guilt beyond a reasonable doubt. See Brown v. State, 863 N.W.2d 781, 787-88 (Minn.2015); Riley,. 819 N.W.2d at 170.1 More specifically, to satisfy the innocence prong under Minn. Stat. § 590.01, subd. 4(b)(2), Rhodes must have alleged the existence of newly discovered evidence that, if-true and considered in the light most favorable to the petition, would establish by a clear and convincing standard that no reasonable jury would have found Rhodes, guilty of his wife’s murder beyond a reasonable doubt, had that newly discovered evidence been presented at trial.. In addition, because any newly-discovered-evidence claim necessarily involves new evidence the jury did not have before it, the posteonviction court should assess how a reasonable jury would react to the overall, newly supplemented record, which requires a judgment about the evidence as a whole and its likely effect on reasonable jurors in applying the reasonable-doubt standard.
II.
At Rhodes’s trial, Dr. Michael McGee, a medical expert for the State, testified in relevant part that Rhodes’s wife, Jane, “received some type of trauma to the outer surface of the skin in the neck area ... with enough force to cause breakage of blood vessels.” When asked if that external neck trauma could “have been done with a hand,, in particular a hand used .,,. in the V position?” Dr. McGee replied, “I believe that is possible, yes.” Based on this testimony, the State argued at trial that “[Rhodes] knock[ed] his wife out of the boat with a blow to the neck.” In my view of the State’s entirely circumstantial case, this testimony by Dr. McGee — that Jane’s neck hemorrhaging was caused by external force — was the most critically incriminating evidence of the entire trial.2
In his fourth . posteonviction petition, Rhodes presents scientific evidence that allegedly refutes Dr. McGee’s conclusions regarding the cause of Jane’s neck hemorrhaging. This evidence includes recent scientific literature on the causes of neck hemorrhaging, and reports from seven forensic experts hired by the Minnesota Innocence Project in 2012 and 2013, who examined Jane’s autopsy and the evidence from Rhodes’s trial. Rhodes’s experts include several chief medical examiners, professors of pathology at prominent universities, and reputable forensic pathologists. In the context of jane’s neck hemorrhaging, Rhodes’s petition and his expert reports rely most heavily on two scientific articles, referred to as Alexander & Jent-zen (2011) and Pollanen (2009).
Alexander .& Jentzen (2011) advances new scientific knowledge regarding hemorrhaging in certain neck muscles caused by drowning, which allegedly refutes previously accepted scientific knowledge that hemorrhaging in such muscles “do[es] not occur in drowning and should always raise the suspicion of foul play.” According to Dr. Jentzen, this article established that *794this previous knowledge was “erroneous” because such neck hemorrhaging can be explained by elevated venous pressure and rupture of congested blood vessels, caused by drowning reactions such as coughing, gagging, vomiting, and abdominal contractions.
The Pollanen (2009) article proposed that, when a dead body is angled downward (a “head down position;” which commonly occurs for drowned bodies), hemorrhagic lividity of the soft tissue of the neck (extravascular rupture and leakage of blood vessels due to gravitational pressure after death), causes “pseudo bruises” in the neck that may lead to “misidentification of violent neck injury.”
After analyzing the evidence in Rhodes’s trial and applying the relevant scientific literature, Rhodes’s forensic experts concluded that “[Dr. McGee] misinterpreted postmortem artifacts as antemortem injuries”; “I do not believe there is evidence of [premortem] trauma — consistent with the story [that Jane] accidently fell overboard and drowned”; “the hemorrhages in [Jane’s] neck were postmortem”; “I would not consider this death a homicide”; and the death “should have been classified as an accident.” More specifically, based on the application of science from Alexander & Jentzen (2011) and Pollanen (2009), Dr. Hyma opined that “[t]he hemorrhages in the neck do not appear to have been caused by external pressure.” (Emphasis added.) Dr. Wigren concluded that Jane’s neck hemorrhages are “likely explained” by the natural, internal processes proposed by Alexander & Jentzen (2011) and Pollanen (2009); that the type of hemorrhaging in Jane’s neck is “not associated with blunt force injury ”; and that the hemorrhages were “not consistent with blunt force injury to the neck.” (Emphasis added.) The opinions of Rhodes’s experts as a whole, and in particular the conclusions by Dr. Hyma and Dr. Wigren, completely contradict Dr. McGee’s critical trial testimony that Jane’s neck hethor-rhages were caused by external trauma “to the outer surface of the skin in the neck area ... with enough force to cause breakage of blood vessels.”3
The court selectively emphasizes that two of Rhodes’s experts, Drs. Jentzen and Rao, used equivocal language in parts of their reports. They stated that “the hemorrhage in [Jane’s] neck could have occurred during' the drowning process or postmortem, as opposed to pre-mortem external pressure”; and the “hemorrhages in [Jane’s] neck could be attributable to something other than external pressure.” (Emphasis added.) But any concern about equivocation, from two of seven of the forensic experts, goes to the weight of the evidence alleged, which must be evaluated at an evidentiary hearing where these experts and the state’s experts would be allowed to testify, to be cross-examined, and to further explain their conclusions. At this stage, the only question is whether Rhodes alleged the existence of evidence that is legally sufficient to warrant relief, which requires the postconviction court to consider the evidence as true, in the light most favorable to Rhodes, with any doubts resolved in his favor. Under this “light *795most favorable” standard, the only fair assessment of Rhodes’s petition as a whole, including all seven of the expert reports and the scientific, literature, is that Dr. McGee’s testimony about external neck trauma was erroneous, and that Jane’s neck hemorrhages were caused by natural drowning processes. Moreover, even if I assumed that there was not a “complete” contradiction of Dr. McGee’s testimony,, such a complete contradiction is not required to warrant an evidentiary hearing. At the very least, Rhodes’s fourth petition presents a dispute of material fact, which must be.resolved at an evidentiary hearing. Wilson v. State, 726 N.W.2d 103, 107-08 (Minn.2007).
In the postconviction court’s summary denial of Rhodes’s fourth petition, it erroneously reached the following conclusion: “In light of the contradicting testimony presented by equally qualified forensic pathologists, the Court cannot conclude that the scientific literature and expert opinions offered by [Rhodes] (even if true) make- it highly probable that [Rhodes] is innocent.” (Emphasis added.) This conclusion by the district court, made “in light of contradicting testimony” by “equally qualified forensic pathologists” (referring to Rhodes’s experts and the State’s expert Dr. McGee), is an abuse of discretion. This is an erroneous application of the standard for granting an evidentiary hearing, which requires the facts alleged in the petition to be considered as true and in the light most favorable to the petition. By weighing the “contradicting” and “equally qualified” testimony by Dr. McGee, the postconviction court failed to properly' apply this standard. And it did what can be done only after an evidentiary hearing. Whether Dr. McGee’s opinions are “equally qualified,” compared with the forensic experts offered by Rhodes, is a question of credibility and evidentiary weight, which can be considered and resolved only by live testimony at an evidentiary hearing — not on a summary denial. Even more importantly, the fact that the postconviction 'court apparently determined that Rhodes’s experts and literature were “contradicted] ” by Dr. McGee’s testimony is actually a basis for requiring an evidentiary hearing, not denying one, because -the contradiction presents a dispute ‘ of material fact. See Riley, 819 N.W.2d at 167; Wilson, 726 N.W.2d at 107-08.
The scientific evidence presented by Rhodes, when considered as true and in the light most favorable to the petition, alleges important issues of material fact regarding Rhodes’s newly-discovered-evidence claim, which must' be resolved at an evidentiary hearing. Even if we determine that Rhodes is not entitled ‘to a new trial on the record before us, an evidentia-ry hearing must be held if issues of material fact remain. See Wilson, 726 N.W.2d at 106-08. In Wilson, a postconviction petitioner presented an affidavit of a forensic expert, who questioned the scientific methods and testimony of a State witness related to ballistics tests. Id. at 106-06. Although we held that, on the record bé-fore us, the petitioner was “not entitled to a new trial,” we remanded for an eviden-tiary hearing ‘because the petitioner “raised important issues of material fact by submitting [the forénsic expert affidavit, which] raise[d] serious questions about the scientific methods used ... and the opinion testimony by the [police officer].” Id. at 106-07 (emphasis added). We concluded that an evidentiary hearing was required because the forensic expert’s affidavit, • “if true, indicates that the officer’s testimony may have been inaccurate or even unfounded.” Id. at 108.
Similarly here, according to the literature and expert reports presented by Rhodes’s . fourth petition, regarding changes in scientific knowledge on drown*796ing-related internal neck hemorrhaging, Dr. McGee’s testimony at trial that Jane’s neck hemorrhages were caused by external force may be “inaccurate or even unfounded” under today’s science. Indeed, under the “light most .favorable”, standard, we must assume that it is inaccurate based on- the contradictious .presented by Rhodes’s - scientific evidence. Because Rhodes’s fourth petition, considered in the light most favorable to Rhodes, raises “important issues of material fact” and “serious questions” about the opinion testimony of Dr. McGee, an evidentiary hearing is required, Wilson, 726 N.W.2d at 106-08, which is “particularly important” here because Rhodes’s fourth petition “attacks” the most critical evidence in the State’s wholly circumstantial case, id. at 107 (citing Opsahl, 677 N.W.2d at 428); see Rhodes II, 657 N.W.2d at 827 (stating that “the evidence in this case is wholly circumstantial”).4
In the alternative, the court concludes that, “even if we assume that the alleged scientific evidence fully refutes Dr. McGee’s medical testimony, that assumption is legally insufficient to establish the innocence prong because, as we held in Rhodes II, [657 N.W.2d at 846,] Rhodes’s murder conviction is independently supported by nonmedical (i.e., nonscientific) evidence,” including “physical and motive evidence, testimony as to Rhodes’ conduct, and inconsistencies in Rhodes’ statements.” The court holds that its conclusions 14 years ago from Rhodes II “apply with equal force here.”
The problem is that Rhodes II’s conclusions do not apply equally here for at least two' reasons. First and most obviously, at the time that Rhodes II was considered and decided (2002-2008), the presently asserted, new scientific knowledge on the causes of drowning-related hemorrhaging (Alexander & Jentzen 2011, Pollanen 2009, and the seven expert reports) did not yet exist. Despite Rhodes IPs holding on the sufficiency of the evidence, 657 N.W.2d at 846, the science presented in Rhodes’s fouith petition'1 is new, distinct, and much more critically damaging to Dr. McGee’s medical testimony compared with the evidence asserted in Rhodes II. The evidence in this petition changes the record that would have been heard by the jury, and this new supplemented record must be considered as a whole, and not by melding it with our conclusion from 14 years ago based on different, much less advanced scientific evidence related to the causes of Jane’s injuries. Our conclusion 14 years ago that the medical evidence asserted in Rhodes’s first postconviction petition was insufficient for relief in Rhodes II does not necessitate a determination now that the distinct scientific 'evidence presented here is legally insufficient to warrant an eviden-tiary hearing because nonscientific evidence supported the conviction. The court presents a false dichotomy between scientific evidence and nonscientific evidence, *797under which no scientific evidence of any type would.ever be sufficient to exonerate Rhodes because in Rhodes II we concluded that the nonscientific evidence was. sufficient. This cannot be correct. Imagine, hypothetically, that a defendant’s.-conviction was affirmed on direct appeal because nonscientific evidence was sufficient to support the verdict. Then, in a postcon-viction proceeding, the defendant presented newly discovered DNA evidence that showed, to a 99.99% probability, that the defendant did not kill the victim. , Would this court hold that this DNA evidence is legally insufficient for an evidentiary hearing because it already concluded, under an earlier direct appeal, that certain nonscientific evidence was sufficient to support the conviction? Of course not. The point here is that in a newly-discovered-evidence case, the newly supplemented record must be considered anew, as a whole, because the newly alleged evidence -may push or pull on the remainder of the evidence at trial (or from previous postconviction proceedings) in new or different ways, such that the newly balanced record would require a reasonable jury to reach a different conclusion. In other words, merely because scientific evidence in a previous proceeding could not outweigh the nonscientific evidence then does not preclude new and distinct scientific evidence from tipping that balance now. As applied here, the nonscientific evidence from Rhodes’s trial may no longer be independently sufficient to support his conviction, as held in Rhodes II, if it is considered fully within a newly supplemented record and outweighed by distinctly' compelling scientific evidence that -refutes the most critical evidence offered by the State at trial.
Second, the list of “nohmedical” evidence sufficient for Rhodes’s conviction, as described by the Rhodes II decision, now has holes that must be addressed and cannot simply be pasted into the current, newly supplemented record. One of the items in the list, “physical evidence,” refers entirely to the State’s expert'testimony from Captain Chandler regarding the expected resurfacing time Of a drowned body based on the water temperature and depth at the drowning location. See Rhodes II, 657 N.W.2d at 829-30, 840, 843 (“Chandler found it improbable that a body could sink in Green Lake [at the 40-foot-depth location identified by Rhodes], resurface, and then float nine-tenths of a mile in 13 hours, given the lake temperature..’.. [Physical evidence proves that, under the , circumstances, a body under 40 feet of water would not refloat for three to four weeks.... [P]hysical evidence suggests that Jane fell overboard in a different part of the lake.”).
. But we now know that Captain Chandler testified incorrectly at trial regarding the water temperature at a 40-foot depth and the resulting 3-to-4 week resurfacing time. Although Captain Chandler testified at trial that Green Lake was 39 degrees, we now know that the actual temperature, according to a survey by the Minnesota DNR, was 69 degrees.5 Thus, the list of *798“sufficient” nonmedical evidence from Rhodes II no longer holds up. Now, all we are left with is weak evidence of motive;6 conflicting witness testimony regarding relatively innocuous conduct by Rhodes;7 observations of a boat on Green Lake;8 and relatively minor inconsistencies in Rhodes’s statements, given three separate times over a 3-month period.9 See Rhodes II, 657 N.W.2d at 846. Thus, the decision here becomes even more troubling. In essence, the court is now relying.on only part of the conclusion from Rhodes II on the sufficiency of the nonmedical evidence that was reached after an evidentiary hearing; adding that conclusion to a newly supplemented record with distinct scientific evidence that refutes critical state testimony; and concluding that Rhodes’s fourth petition is- legally insufficient, without allowing an evidentiary hearing to actually weigh this newly supplemented record.
The provisions of Chapter 590, and our case law interpreting those provisions, are designed to reduce the burden of meritless postconviction proceedings. Those provisions are not intended, however, to bar claims that may have merit. Based on the material factual issues presented by Rhodes’s fourth petition,' I cannot conclusively say that Rhodes is entitled to no *799relief. Minn.Stat. § 590.04, subd. 1; Wilson, 726 N.W.2d at 106-08; see Bobo, 820 N.W.2d at 516-20. Here, the newly discovered evidence, if proven to be true, and considered in the light most favorable to Rhodes, would establish by a clear-and-convincing standard that no reasonable jury would have convicted Rhodes of first-degree premeditated murder beyond a reasonable doubt had the newly discovered evidence been presented at trial. See Minn.Stat. § 590.01, subd. 4(b)(2); Brown, 863 N.W.2d at 787-88; Riley, 819 N.W.2d at 170. Under our well-established precedent, “any doubt’* must be resolved , in favor of holding an evidentiary hearing, and holding such a hearing is “particularly important” where newly discovered evidence attacks important evidence in a wholly circumstantial case like this one. Bobo, 820 N.W.2d at 516; Wilson, 726 N.W.2d at 107; Opsahl, 677 N.W.2d at 423. For these reasons, I would reverse and remand for an evidentiary hearing. I respectfully dissent.
. At Rhodes’s posteonviction hearing in 2001, one expert opined that Dr. McGee’s medical testimony was "the most paramount of the entire trial” and the question on the hand in the "V” position was "the murder question. That’s the question: Did he Mil her?”
. Rhodes’s experts are not offered merely to "impeach” Dr. McGee, as the court states. Rhodes’s experts are not merely taking a second look at the same trial record, disagreeing with Dr. McGee, and reaching different conclusions, whiph would be impeachment prohibited by Minn.Stat. § 590.01, subd. 4(b)(2). Rather, here Rhodes’s forensic experts are offered to explain and apply newly discovered scientific evidence, from Alexander & Jentzen (2011) and Pollanen (2009), to a newly supplemented record. It is the application of that scientific evidence, and not merely hindsight disagreements by new experts, which shows that Dr, McGee's conclusions are scientifically erroneous,
. The court declines to address the "shifted science” rule proposed by Rhodes for overcoming the 2-year statute of limitations under •Minn.Stat. § 590.01, subd. .4(c). Rhodes argues that, in the context of "shifted science,” the publication date of a scientific article does not determine the date on which science changed because scientific consensus evolves “gradually” through multiple experts and articles in the community. With this argument, Rhodes impliedly invites us to adopt a novel “shifted science” rule, in which the accrual date under subdivision 4(c) depends not on publication dates, but on the date that new scientific knowledge becomes generally accepted. I would adopt this rule, or some version of it, and remand for an evidentiary hearing to include expert testimony on whether-and when the shifted science alleged in Rhodes's fourth petition became "generally accepted" in the relevant scientific community. See Roby v. State, 787 N.W.2d 186, 191 & n. 3 (Minn.2010); see also State v. Tanksley, 809 N.W.2d 706, 708 n. 1 (Minn.2012),
. The postconviction court concluded that the DNR report was untimely under the "due diligence” requirement of the newly-discovered-evidence exception, Minn.Stat. § 590.01, subd. 4(b)(2), because the DNR report could have been discovered by due diligence before the 2-year statute of limitations expired in 2007. I conclude that Rhodes's submission of this DNR report is not untimely because it is not reasonable to expect a petitioner in Rhodes's circumstances, by due diligence, to discover this DNR data and file a petition for postconviction relief on that basis before 2007. And the evidence here is not a disputed opinion, expert or otherwise, or some other kind of ephemeral evidence where the passage of time or fading memories makes admission problematic; it is temperature data maintained by an official State agency *798and data that is foundational to expert testimony critical to the State’s case.
. The State argued that -.Rhodes had a motive to kill Jane because of (1) a nonsexual relationship with another woman that ended 13 months before Jane’s death; (2) a one-time consultation with a divorce attorney 15 months before Jane’s death; (3) life insurance; and (4) household debt from a home mortgage and loans on a car and a boat.
. For example, on the night of the drowning, one witness saw a man, matching Rhodes’s description, drive a boat to shore. The man “took off running” at "top speed” across the street to the Northern Inn. But another wit,ness testified that a man, matching Rhodes’s description, returned his boat to shore, and "walked” across the street to the Northern Inn. During the search for Jane, witnesses observed Rhodes's conduct to be consistent with "someone who just lost a loved one” — he was described by various witnesses as "crying,” "very upset,” "frantic,” “very emotional,” "wanting to save his wife,” "deeply agitated," and "devastated" with "huge sobs.” But one witness testified that although "once in a while [Rhodes would] be emotional and cry,” the witness did not "notice any tears.” Another witness said that Rhodes was “not [crying] in my opinion,” although that witness "couldn’t see [Rhodes’s] face” and wasn't "paying much attention.”
. For example, seven witnesses testified to seeing or hearing sounds from a boat on Green Lake on the night of Jane’s drowning. Some of these witnesses testified to seeing a boat driving in "nine and eight” patterns with rapid speed and sharp circles. Five of these witnesses testified to hearing sounds that were not distressful, such as "laughter,” "hooting and hollering,” "partying it up,” and “having fun out there.” One of these seven witnesses heard “moaning sounds” and a woman’s voice scream, "Stop. No. It Hurts.” Rhodes told police that, while he and his wife were taking an intimate boat ride; they saw and heard another boat without any lights that was "partying it up” and "tearing around.” .
.These inconsistencies, collected from statements made by Rhodes on August 3, 1996, August 15, 1996, and October 10, 1996, include the following: (1) that the boat was going "slow” or near "top speed,” although police confirmed that the boat’s speedometer was not working properly; (2) that Jane fell overboard near the "back,” the "side,” or the “back side” of the boat; (3) that the boat was headed north or headed toward shore at the time Jane fell overboard; (4) that when Jane fell overboard, Rhodes missed the throttle on the first grab and then turned around the boat, or he “immediately” turned around (without mentioning whether a throttle grab was missed); and (5) that Jane did not scream when she fell overboard, or there “might have been like a muffled scream” or a "cut off scream ... [but it] would be pure speculation on my part.” '