OPINION

ROBB, Chief Judge.

Case Summary and Issues

Kristine Bunch was convicted by a jury in 1996 of felony murder for the death of her young son, Anthony ("Tony"), in a fire at their mobile home and sentenced to sixty years.1 In 2006, Bunch began pursuing post-conviction relief, which was ultimately denied by the post-conviction court in 2010. In this appeal from the denial of post-conviction relief, Bunch raises three issues that we expand and restate as four: 1) whether the post-conviction court erred in concluding fire victim toxicology evidence offered at the post-conviction hearing was not newly-discovered evidence; 2) whether the post-conviction court erred in concluding fire investigation technique evidence offered at the post-conviction hearing was not newly-discovered evidence; 3) whether the post-conviction court erred in denying her relief on the basis of a failure by the State to turn over exculpatory evidence in contravention of the dictates of Brady v. Maryland; and 4) whether the post-conviction court erred in denying her relief because of ineffective assistance of trial counsel. We conclude the fire victim toxicology evidence does constitute newly-discovered evidence and the post-conviction court clearly erred in denying Bunch relief on this claim. We also conclude the State's failure to turn over a report from the ATF testing of floor samples violates Brady and the post-conviction court also clearly erred in denying Bunch relief on this claim. Because either of these two errors warrants a new trial, we need not address the remaining issues. We reverse and remand for a new trial.

Facts and Procedural History

As stated in the direct appeal:

Bunch, 697 N.E.2d at 1256 (footnote omitted).

At the jury trial, no witness testified to seeing Bunch set the fire or hearing her talk about doing so; there was no evidence Bunch had purchased a liquid accelerant and no evidence of flammable liquid on the clothes she was wearing; and there was no testimony regarding a motive for her setting the fire. The State's case relied largely on expert testimony describing two points of origin for the fire from visual inspection and testing of floor samples showing evidence of a liquid accelerant. Brian Frank, assistant chief investigator for the Indiana State Fire Marshall's office, testified to the existence of "V" burn patterns on horizontal surfaces that indicated to him areas of combustion. See Record of (Trial) Proceedings ("Trial Record") at 819 (Frank explaining that "[a] V pattern ... points you back down to the area where the fire started."). In addition, a hole was burned through the floor near the bed in the south bedroom. "[I]t's unusual for a fire to burn in a downward manner.... Something has to draw the fire down. Liquid accelerant would do that." Id. at 829. Based upon his training, experience, and observations at the scene, Frank opined:

Id. at 831. Frank also testified that a hydrocarbon sniffer and a canine both indicated the presence of hydrocarbons2 at the fire scene and samples were taken of those areas. The samples were primarily of the flooring throughout the mobile home, but also included mattress ticking, tack strips and paneling, and the nightgown Bunch was wearing at the time of the fire.

William Kinard of the federal Bureau of Alcohol, Tobacco, and Firearms testified that he received ten samples taken from the fire scene and was asked to determine whether there was evidence of an accelerating material present in those samples. Kinard testified that his analysis by gas chromatography found evidence of a heavy petroleum distillate, such as diesel fuel or kerosene, in five of the flooring samples, although two of the samples tested with different carbon number readings putting them "on the ... downhill side or the tailend of a heavy petroleum distillate...." Id. at 907. Four of the samples testing positive came from the living room and one from the bedroom. A written report summarizing these findings was also entered into evidence.

The forensic pathologist who conducted the autopsy on Tony testified that "within reasonable medical certainty," id. at 779, Tony died from smoke inhalation. He further testified that Tony's blood had a carbon monoxide saturation reading of eighty percent, which indicated "we're probably looking at an event that happened over a period of minutes, in terms of the absorption of carbon monoxide." Id. at 782.

In addition to the expert testimony, the State presented evidence that firefighter Ron Clark entered the mobile home to attempt to rescue Tony and encountered an obstacle in the path he took between the living room and south bedroom. After the fire had been put out and the investigation began, a recliner or swiveling chair was identified as partially obstructing the doorway between the living room and the bedroom. Bunch's mother testified that the chair had always been in that location so as not to obstruct a nearby heating vent and it "stuck out a little bit, but not really." See id. at 1077-78. Clark testified that he had climbed over both ceiling tiles and furniture in entering the bedroom, id. at 669, but also gave a statement after the fire about his entry into the bedroom in which he stated "that wall was already destroyed by fire.... There was some structural members that were weak that I know that I had knocked down going through at that time[,]" id. at 679. A neighbor who arrived to help testified that Bunch told him Tony was behind a locked door, although the door between the living room and the south bedroom had been removed some time prior to the fire. Connie Land, manager of the mobile home park, testified that when Bunch and her mother returned to the mobile home a couple of days after the fire to thank people who had helped, Bunch gave an account of how the fire started in which she recounted that someone else had been in the trailer with her and Tony and had "sprinkled her with, I'm not sure if she said it was either kerosene or gasoline." Id. at 789. Bunch's mother testified, however, that she was with Bunch the entire time they were at the mobile home and Bunch did not relate that version of the fire to Land. Id. at 1058. Bunch's mother also testified that following the fire, she and Land were engaged in litigation over the disposition of the mobile home, and there were hard feelings between the families. Id. at 1059-60.

The State also presented evidence that Bunch's injuries were minor and consistent with a brief but direct exposure to flame. Finally, the State presented evidence of the various accounts Bunch gave of the time leading up to the fire and her actions upon discovering it. She variously recounted that she and Tony had gone to sleep on the couch in the living room or in the bedroom; that she did or did not see smoke when she woke up; and that she saw flames in the bedroom doorway or on the floor by the bed. She also recounted Tony speaking to her as she tried to put the fire out.

Bunch offered evidence of past electrical problems in the mobile home, and Tom Hulse, Bunch's expert witness and an arson investigator, testified that in his opinion, the origin of the fire "should be classified as undetermined, with an explanation that there is a probability that it is an accidental fire and not an intentional fire." Id. at 1130. Hulse testified the fire did not reach a stage in fire development known as "flashover." Id. at 1151. The jury found Bunch guilty of both felony murder and arson. The trial court entered judgment of conviction on both counts, and despite stating the arson conviction was merged into the felony murder conviction, ordered a fifty-year sentence for the arson conviction to be served concurrently with a sixty-year sentence for the felony murder conviction.

On direct appeal, our supreme court remanded the case to the trial court with directions to vacate the arson conviction because although the trial court "may have intended to vacate the arson conviction,... the record remains susceptible to the conclusion that Bunch stands convicted of arson and was given a concurrent fifty year sentence[,]" which would violate double jeopardy. Bunch, 697 N.E.2d at 1257. The court also considered Bunch's challenges to a jury instruction informing the jury that the State was not required to prove a motive, to the sufficiency of the evidence, and to the reasonableness of her sentence. The court held the jury instruction issue was waived and that her sentence was not manifestly unreasonable. Id. at 1257, 1258. In holding there was sufficient evidence to support the felony murder conviction, the court specifically pointed to the following, among other pieces of evidence:

Id. at 1257. Accordingly, Bunch's conviction of felony murder and her sixty-year sentence were affirmed. Id. at 1258.

In 2006, Bunch filed a petition for post-conviction relief, which was amended in 2008, claiming newly-discovered evidence in the form of advances in the field of fire science, a violation of due process by the State in failing to disclose certain evidence, and ineffective assistance of trial counsel. At the post-conviction hearing, held in October of 2009, Bunch presented testimony from four experts who testified to advances in fire science that she contends undermine the State's theory and evidence at her trial and disprove each point of evidence cited by the Indiana Supreme Court in affirming her conviction. Bunch also presented testimony from her trial counsel. The post-conviction court issued extensive findings of fact and conclusions of law and denied Bunch's petition. Bunch now appeals.3 Additional facts will be provided where appropriate.

Discussion and Decision4

I. Post-Conviction Standard of Review

In post-conviction proceedings, the petitioner bears the burden of proof by a preponderance of the evidence. Henley v. State, 881 N.E.2d 639, 643 (Ind.2008). "When appealing from the denial of post-conviction relief, the petitioner stands in the position of one appealing from a negative judgment." Fisher v. State, 810 N.E.2d 674, 679 (Ind.2004). "To prevail on appeal from the denial of post-conviction relief, a petitioner must show that the evidence as a whole leads unerringly and unmistakably to a conclusion opposite that reached by the post-conviction court." Kubsch v. State, 934 N.E.2d 1138, 1144 (Ind.2010), reh'g denied.

In addition, the post-conviction court entered findings of fact and conclusions of law in accordance with Indiana Post-Conviction Rule 1(6). "The post-conviction court is the sole judge of the weight of the evidence and the credibility of witnesses." Woods v. State, 701 N.E.2d 1208, 1210 (Ind.1998), cert. denied, 528 U.S. 861, 120 S.Ct. 150, 145 L.Ed.2d 128 (1999). Although we do not defer to the post-conviction court's legal conclusions, Wilson v. State, 799 N.E.2d 51, 53 (Ind.Ct.App. 2003), "[a] post-conviction court's findings and judgment will be reversed only upon a showing of clear error—that which leaves us with a definite and firm conviction that a mistake has been made[,]" Ben-Yisrayl v. State, 729 N.E.2d 102, 106 (Ind.2000) (citation and quotation marks omitted), cert. denied, 534 U.S. 830, 122 S.Ct. 73, 151 L.Ed.2d 38 (2001). "In short, the question before us is whether there is any way the [post-conviction] court could have reached its decision." Id. (quotation omitted).

II. Newly-Discovered Evidence

Post-Conviction Rule 1(1)(a)(4) provides:

Newly-discovered evidence mandates a new trial only when the defendant demonstrates each of the following nine requirements:

Taylor v. State, 840 N.E.2d 324, 329-30 (Ind.2006) (quoting Carter v. State, 738 N.E.2d 665, 671 (Ind.2000)). The reviewing court "analyzes these nine factors with care, as the basis for newly discovered evidence should be received with great caution and the alleged new evidence carefully scrutinized." Id. at 330 (internal quotations omitted). The burden of showing all nine requirements rests with the post-conviction petitioner. Webster v. State, 699 N.E.2d 266, 269 (Ind.1998).

In her petition for post-conviction relief, Bunch made two claims of newly-discovered evidence she asserted entitled her to a new trial: advances in science regarding fire victim toxicology and advances in science relating to fire investigation techniques. The post-conviction court concluded Bunch had failed to prove she has newly-discovered evidence, in part because:

Appellant's Appendix at 20-27. Bunch contends on appeal the post-conviction court committed clear error in so concluding.

A. Victim Toxicology Evidence on Post-Conviction

Bunch first contends that advances in the field of victim toxicology analysis constitute newly-discovered evidence entitling her to a new trial. In support of her claim, Bunch presented to the post-conviction court the testimony of Jamie McAllister, who has bachelor's and master's degrees in fire protection engineering and at the time of the hearing was pursuing a doctorate in toxicology. In addition, she has fourteen years' experience as a firefighter responding to approximately 1,000 fires and nine years' experience as a fire investigator investigating between 200 and 300 fires. The "main focus" of her work as a fire investigator has been "the effects of products of combustion on victims, why they don't get out and how their autopsy reports and pathology information can help ... understand more about how the fire may have started." Post-Conviction Relief Hearing Transcript ("PCR Hrg. Tr.") at 275. In her opinion, "the fire originated in the concealed space above the south bedroom in between the ceiling and the roof[,]" id. at 276-77, and was not caused by an accelerant, id. at 278.

McAllister relied on her review of eye-witness testimony about the fire, photographs of the damage to the mobile home, and Tony's toxicology results in concluding the fire originated as an under-ventilated confined fire in the space above the ceiling in the bedroom where Tony was found before burning through the ceiling tiles and dropping to the floor, thereafter spreading into the living room and through the mobile home. Specifically, she looked to the testimony of neighbors Tom Claxton and Robert Parkinson, both of whom testified at trial that they first saw smoke and flames in the south bedroom. See Trial Record at 600 (Claxton testifying that when he looked into the south bedroom through a window, the room was obscured by smoke and he saw flames on the floor in the back of the room), id. at 616-19 (Parkinson testifying that he looked in the door of the mobile home toward the south bedroom and saw "a lot of thick, black smoke" but "a very small amount" of fire, if any, in the living room); see also id. at 713, 719 (Bunch stating in two police interviews that she first saw fire on the floor of the south bedroom). In addition, McAllister reviewed photos of the mobile home after the fire had been put out and noted there was less damage in the living room than in the south bedroom and the damage had not spread equilaterally to other areas adjoining the living room but away from the bedroom. Based on this evidence in combination, McAllister concluded the fire did not originate in the living room, but instead in the south bedroom.

McAllister also reviewed toxicology and autopsy reports which showed Tony's carbon monoxide saturation (also referred to as carboxyhemoglobin or COHb level) was eighty percent and he had inhaled smoke and soot, yet there was no thermal damage to his respiratory system. McAllister drew several conclusions from this evidence. First, McAllister noted that different fires produce different amounts of carbon monoxide based on ventilation conditions with under-ventilated fires producing greater amounts. See PCR Hrg. Tr. at 279 ("[W]hen you have a fire that doesn't have enough air to burn completely... you'll get a higher production of [carbon monoxide]...."). Using a chart modeled on the Coburn-Forster-Kane equation correlating COHb levels to carbon monoxide exposure for a person of Tony's size and weight, McAllister showed that a typical under-ventilated fire yielding 5,000 to 10,000 parts per million of carbon monoxide could cause an eighty percent COHb level in five to ten minutes once the fire is no longer contained. A typical well-ventilated fire, such as one burning in an open room, would yield approximately 1,000 parts per million of carbon monoxide, which would not cause an eighty percent COHb level for well over sixty minutes.5 Under-ventilated fires also yield higher concentrations of soot. McAllister opined that Tony would have had thermal injuries to his respiratory system and a low COHb level if he had been in proximity to an incendiary fire set in an open area of the bedroom.

McAllister also noted that Bunch's medical examination and toxicology results were consistent with this conclusion. She noted that Bunch was likely also exposed to carbon monoxide, impairing her senses and judgment. Consistent with the conditions caused by an under-ventilated fire, Bunch's hospital discharge summary showed there was soot in her respiratory system, indicating she had inhaled a great deal of smoke. Moreover, that her blood gases were normal when checked at the hospital was consistent with the conclusion that she was not in the room where the fire started, and with the facts that she had been given oxygen at the hospital and was not tested until some time after the fire. Finally, McAllister opined that Bunch's injuries—first degree burns on her face and forearms, a second degree burn on the tip of her nose, and some singeing of the hair around her face—were not consistent with intentionally setting a fire with a liquid accelerant.

Conceding there might be other explanations for each individual fact which factored into her opinion, McAllister emphasized that each fact "has to be considered with, within the whole entire collection of data. And all of the data ... and the hypotheses have to be consistent with the facts of the case." Id. at 341. Thus, considering all the data together, McAllister concluded it was most consistent with an under-ventilated fire started accidentally in a confined space in the south bedroom.

McAllister also testified the consideration of a fire victim's physiological condition did not become a recognized component of fire origin analysis until after 2001, the first time a chapter on fire-related deaths appeared in the National Fire Protection Association 921 Guide for Fire and Explosion Investigations ("NFPA 921"), which is "a peer reviewed and generally accepted standard in the fire investigation community." Travelers Prop. & Cas. Corp. v. General Elec. Co., 150 F.Supp.2d 360, 366 (D.Conn.2001).6 NFPA 921 is revised approximately every three years due to new information and advances in science.

Bunch contends that contrary to the post-conviction court's conclusion, McAllister's testimony meets each of the requirements of newly-discovered evidence.

B. Newly-Discovered Evidence Requirements

1) The evidence has been discovered since trial. As McAllister testified, fire victim toxicology analysis first appeared in the leading fire investigation guide in 2001, five years after Bunch's trial. Bunch contends the post-conviction court's finding that McAllister's testimony was just an interpretation of evidence that existed at the time of trial "misses the point" because although the toxicology results existed at that time, this particular interpretation and application of the results did not. Brief of the Appellant at 25. Bunch does not dispute that the factual evidence existed at the time of her trial; rather, she claims that the current scientific analysis of the factual evidence was not then recognized. The amicus note that three state legislatures have recently passed resolutions supporting judicial review of cases in which faulty science is alleged to have contributed to an arson conviction, implicitly acknowledging the "transformative advancements" in the science of fire investigation since Bunch's trial. Brief of the Amicus Curiae at 12 (citing 2010 resolutions from Oklahoma, Nebraska, and Arizona).7

Bunch analogizes this evidence to DNA analysis, which has been considered newly discovered evidence even though the DNA evidence itself existed at the time of trial. As Chief Justice Roberts has explained, "DNA testing has an unparalleled ability both to exonerate the wrongly convicted and to identify the guilty. It has the potential to significantly improve both the criminal justice system and police investigative practices. The Federal Government and the States have recognized this, and have developed special approaches to ensure that this evidentiary tool can be effectively incorporated into established criminal procedure—usually but not always through legislation." District Attorney's Office for Third Judicial Dist. v. Osborne, 557 U.S. 52, 129 S.Ct. 2308, 2312, 174 L.Ed.2d 38 (2009); see also Sewell v. State, 592 N.E.2d 705, 708 (Ind.Ct.App. 1992) (noting, in deciding petitioner's right to discovery of rape kit for purpose of obtaining DNA analysis, that "[a]dvances in technology may yield potential for exculpation where none previously existed."), trans. denied; State v. Behn, 375 N.J.Super. 409, 868 A.2d 329, 343 (2005) ("[I]t is well-known by now that the use of DNA testing has upset many convictions which took place before the technique was developed."), cert. denied, 183 N.J. 591, 874 A.2d 1108 (N.J.2005). In Indiana, the potential for post-conviction DNA testing and analysis to affect the outcome of a criminal case is recognized and addressed by Indiana Code chapter 35-38-7. But see Pinkins v. State, 799 N.E.2d 1079, 1092 (Ind.Ct.App.2003) (noting that in order to be granted a new trial based on subsequent DNA testing, the statute "dictates that the results must be favorable—not the fact that the science of the testing has become more advanced.... If the statute provided otherwise and permitted the scientific procedure alone to satisfy the standard, a DNA scientist could simply testify that a method performed in the original testing had been superseded, and the petitioner could then be afforded a new trial without any additional showing."), trans. denied.

We do not believe fire victim toxicology analysis can be precisely equated with DNA analysis, because DNA analysis can now definitively prove whether or not a DNA sample is from a given person, whereas fire victim toxicology analysis only tends to prove one fire origin scenario is more likely than another. However, we do agree with Bunch that, just as the evolving science of DNA analysis became accepted as the scientifically reliable method for accurately interpreting even previously-existing DNA evidence, fire victim toxicology analysis has become recognized as a scientifically reliable method to better interpret existing evidence, and that it has done so since the time of Bunch's trial. The post-conviction court's finding that Bunch's post-conviction evidence is just "different packaging" for the same conclusion does not give appropriate due to the science which has emerged since Bunch's trial to support that conclusion. Thus, we agree with Bunch that the fire victim toxicology analysis offered by McAllister has been discovered since her trial.

2) The evidence is material and relevant. "In the criminal context, evidence is relevant if it tends to prove or disprove a material fact or sheds any light on the guilt or innocence of the accused." State v. Lovett, 943 N.E.2d 409, 411 (Ind. Ct.App.2011). Evidence is material "if there is a reasonable likelihood that it might have affected the outcome of the trial." Samaniego v. State, 679 N.E.2d 944, 948 (Ind.Ct.App.1997), trans. denied. Bunch contends McAllister's fire victim toxicology analysis showing that Tony's injuries are not consistent with an open fire directly rebuts the State's theory at trial and therefore disproves a material fact and is relevant to the issue of her guilt.

The information charging Bunch alleged:

Trial Record at 8 (emphases added). All parties and witnesses agree that the conclusion of Bunch's experts at trial and on post-conviction is the same; namely, that the origin of the fire is undetermined. At trial, the State proceeded on the theory that the undetermined origin coupled with the circumstantial evidence indicated arson. However, the victim toxicology analysis specifically addresses the question of whether Bunch intentionally set the fire in the living room of the mobile home using a liquid accelerant as charged by the State. The post-conviction toxicology evidence would tend to disprove the State's theory of the case and is reasonably likely to affect the outcome of a trial on these charges, and it is therefore material and relevant.

3) The evidence is not cumulative. Cumulative evidence is "additional evidence that supports a fact established by the existing evidence.... [T]o be considered cumulative, evidence should be of the same kind or character. That is, evidence will not be considered cumulative if it tends to prove the same facts, but in a materially different way." In re Paternity of H.R.M., 864 N.E.2d 442, 451 (Ind.Ct. App.2007) (internal quotations, alterations, and citations omitted). Cf. Pinkins, 799 N.E.2d at 1092-93 (holding that subsequent DNA testing results did not constitute newly-discovered evidence because "[e]ven if it could be assumed that some of the scientific procedures proffered at the original trial were erroneous," both the evidence offered at trial and the subsequent tests were consistent and therefore the subsequent test results were merely cumulative). The State points out that, consistent with McAllister's estimate, the forensic pathologist who conducted Tony's autopsy testified Tony's carbon monoxide saturation level could have reached eighty percent over a period of minutes, but the pathologist was not concerned with the manner in which such level could have been obtained. That is, the pathologist was not making, nor had he been asked to make, any fire origin determination in so testifying. The State also points out that McAllister, like Bunch's trial expert, testified the fire began in the ceiling. McAllister, however, additionally testified to the relation between Tony's physical condition and her origin determination. Finally, although McAllister's ultimate opinion—that the cause of the fire is undetermined—was also the opinion given by Bunch's trial expert, the method by which she arrived at that conclusion was considerably different and no testimony akin to hers that Tony's physical condition was inconsistent with a fire set in an open room was offered at trial. Although Hulse, Bunch's trial expert, did state that a concealed fire in the ceiling could not be discounted as a possible cause, he mentioned a ceiling fire in going through the various accidental causes—including Tony playing with lighters or matches—that should have been eliminated before concluding the fire was arson, rather than as an actual opinion as to the origin of the fire. His opinion of the origin of the fire was that it started in the south bedroom and "ascended to ceiling level and came out of this room at the upper elevations." Trial Record at 1126. This is in direct opposition to McAllister's opinion. Therefore, the evidence is not cumulative.

4) The evidence is not merely impeaching. "`Impeachment' is defined as `[t]he act of discrediting a witness, as by catching the witness in a lie or by demonstrating that the witness has been convicted of a criminal offense.'" Taylor v. State, 840 N.E.2d 324, 330 n. 1 (Ind.2006) (emphasis omitted) (quoting Black's Law Dictionary 768 (8th ed. 2004)). It is important to distinguish "merely impeaching" evidence from impeaching evidence. "Evidence which destroys or obliterates the testimony upon which a conviction was obtained is not appropriately considered as merely impeaching evidence." Wilson v. State, 677 N.E.2d 586, 588 (Ind.Ct.App. 1997) (citing Dennis v. State, 103 Ind. 142, 151-52, 2 N.E. 349, 355 (1885)). In this regard, this factor has to be considered in conjunction with the final factor: is the evidence of such significance that it would have an effect on the result of the trial? See State v. McCraney, 719 N.E.2d 1187, 1190 (Ind.1999) (holding proffered newly-discovered evidence is not merely impeaching because, although inconsistent with prior testimony, it "serves as freestanding evidence of [defendant's] innocence and does not merely call into question [the witness's] testimony"); Francis v. State, 544 N.E.2d 1385, 1387 (Ind. 1989) (holding proffered newly discovered evidence in the form of eyewitness testimony from two disinterested witnesses placing defendant away from the scene of crime did not merely impeach the victim's testimony, but had "independent, probative merit of its own"). The post-conviction court found McAllister's testimony was merely impeaching of the testimony of the State's trial experts. However, unlike cases in which a witness testifies one way at trial and then later recants or testifies differently on post-conviction, see, e.g., McVey v. State, 863 N.E.2d 434, 446 (Ind.Ct.App. 2007) (noting that molestation victim's affidavit contradicting her trial testimony was merely impeaching because it would "merely serve to cast doubt on [her] trial testimony [and] place her credibility at issue"), trans. denied, McAllister's testimony was free-standing evidence that Bunch did not set multiple incendiary fires in the mobile home and offers a new, exculpatory explanation for Tony's death. It is not, therefore, merely impeaching evidence.

5) The evidence is not privileged or incompetent. Incompetent evidence is "evidence irrelevant to the issue or otherwise inadmissible." Powell v. Alabama, 287 U.S. 45, 69, 53 S.Ct. 55, 77 L.Ed. 158 (1932). We have already determined the evidence is relevant to the issue of Bunch's guilt or innocence and there seems to be no question it is not privileged. The post-conviction court did not find that the proffered evidence was privileged or incompetent, and the State does not argue on appeal that it was.

6) Due diligence was used to discover the evidence in time for trial. A motion for a new trial based on newly-discovered evidence is subject to a hostile inference of want of due diligence in the absence of a clear showing to the contrary. Denney v. State, 695 N.E.2d 90, 93 (Ind. 1998). A party may not rest upon its abstract conclusion about or assertion of its exercise of due diligence but must give a particularized showing that all methods of discovery reasonably available to counsel were used and could not uncover the newly found information. Hawkins v. Cannon, 826 N.E.2d 658, 663 (Ind.Ct.App. 2005), trans. denied. As noted above, discussion of fire victim toxicology analysis became a recognized component of fire investigation after it first appeared in NFPA 921 in 2001, five years after Bunch's trial. McAllister testified NFPA 921 is updated and reissued approximately every three years. Therefore, even if toxicology analysis began to be recognized in the three years leading up to the 2001 revision, it would have been recognized at the earliest in 1998, two years after Bunch's trial. Even with due diligence, there would have been no way for Bunch to have discovered this analysis prior to her trial. See Behn, 868 A.2d at 343 (ordering new trial on basis of studies on composition bullet lead analysis that had not yet been developed prior to defendant's trial).

7) The evidence is worthy of credit. In general, "[w]hether a witness's testimony at a post-conviction hearing is worthy of credit is a factual determination to be made by the trial judge who has the opportunity to see and hear the witness testify." Whedon v. State, 900 N.E.2d 498, 504 (Ind.Ct.App.2009), aff'd, 905 N.E.2d 1108 (Ind.2009). It is not within the province of the appellate court to replace a trial judge's assessment of witness credibility with its own. McCraney, 719 N.E.2d at 1191. Thus, if the post-conviction court expressly finds that the testimony of a fact witness is or is not worthy of credit, we must accept that determination. See, e.g., Carter, 738 N.E.2d at 672 (affirming post-conviction court's "distrust" of newly-found witness who claimed to be the shooter in an incident for which defendant was convicted of attempted murder because the post-conviction court had the "benefit of viewing [the witness's] face and his reactions during testimony"; the witness did not come forward until the post-conviction stage; and the witness's factual testimony was wholly inconsistent with the factual testimony of others who witnessed the shooting); McCraney, 719 N.E.2d at 1191 (affirming grant of new trial based in part on post-conviction court's express finding that witness's testimony on post-conviction that his trial testimony against defendant was false was worthy of credit); Reed v. State, 508 N.E.2d 4, 6 (Ind.1987) (affirming post-conviction court's conclusion that witness's post-conviction testimony recanting his trial testimony placing defendant at the scene of the crime was not worthy of credit because witness's testimony was "of course an attempt to impeach his prior testimony" and post-conviction court was well within its discretion to doubt the recanting testimony).

However, here, the post-conviction court did not find McAllister's testimony was not worthy of credit because it doubted her credibility or veracity based upon a firsthand evaluation of her demeanor; the post-conviction court found her testimony not worthy of credit because it was in conflict with trial evidence, without considering that new science—rather than new or different facts—formed the basis for her opinion. In other words, the post-conviction court did not find McAllister unworthy of credit on the basis of her demeanor; it found her expert opinion unworthy of credit on the basis of its foundation. Under these circumstances, we do not think it necessary or appropriate to impute a personal credibility determination to which we must defer to the post-conviction court. To do so would virtually eviscerate appellate review of post-conviction denials because we would have to speculate in every instance that the post-conviction court could have concluded the witness was not credible based on his or her demeanor.

Thus, although we would defer to the post-conviction court's assessment of fact witnesses—for instance, a trial witness now recanting trial testimony or a new witness offering never-before-heard exculpatory testimony—we will not defer in this case to the post-conviction court's assessment of an expert's scientific evidence. We have the ability to assess McAllister's expert testimony ourselves because her credentials and the basis for her opinion are part of the record. The post-conviction court found McAllister's testimony was "not reliable" because she did not establish the scientific principles for her conclusion and because McAllister's conclusions contradicted undisputed evidence and eyewitness testimony from the trial. In making such a determination, the post-conviction court did not rely on her demeanor, but rather on the stated bases for her opinion and review of the trial record. We have the same information before us, and therefore are able to independently assess whether McAllister's testimony is worthy of credit without invading the province of the post-conviction court. Cf. GKN Co. v. Magness, 744 N.E.2d 397, 401 (Ind.2001) (noting situations in which appellate courts have held that de novo review is appropriate where a reviewing court is in as good a position as the trial court to make a decision on a particular issue based upon the evidence presented).

Indiana Evidence Rule 702 provides:

The post-conviction court allowed McAllister to testify as an expert.8 McAllister described her training and experience and the evidence and method upon which she relied in forming her opinion. She described her familiarity with NFPA 921, which, as previously discussed, is now the leading guide for fire investigation. She did, as the post-conviction court found, refer during her testimony generally to "numerous studies" measuring carbon monoxide production in comparison with ventilation conditions. See, e.g., PCR Hrg. Tr. at 280. However, in explaining her conclusions, she specifically referred to the Coburn-Forster-Kane equation as the basis for the chart she produced to demonstrate the COHb levels a person of Tony's size was likely to have registered based upon an open versus a confined fire. She then used that chart, in conjunction with the trial testimony and exhibits, in forming her opinion as to the origin of the fire.

The post-conviction court also found that McAllister's opinion contradicted undisputed facts from the trial record, as the dissent notes. See Op. at 305-07 (Crone, J., dissenting). McAllister's opinion that the fire likely started in the confined space above the ceiling in the south bedroom and smoldered for some time until it finally burned through the ceiling tiles and dropped to the floor of the bedroom is not inconsistent with the undisputed facts the post-conviction court cited. The post-conviction court noted that Bunch "always said she first saw the fire on the floor." Appellant's App. at 25. That Bunch stated she first saw fire on the floor of the south bedroom is not equivalent to stating, let alone proving, the fire started on the floor,9 and therefore is not inconsistent with McAllister's opinion regarding the origin of the fire. Likewise, McAllister's opinion is not inconsistent with the trial testimony of other witnesses who testified they saw fire on the floor of the bedroom. McAllister's testimony that the fire started in the south bedroom and spread into the living room and beyond is also not inconsistent with the trial testimony of witnesses who testified they saw fire in the living room. Parkinson in particular, when asked if he saw fire on the ceiling, testified that "in one corner of the living room, there were flames coming out above that," trial record at 618; that he saw smoke "coming out of the top of the room," id. at 620; and that he saw smoke "high on the wall," id. at 1099. Although he testified he saw fire on the living room floor, he never testified that he saw fire only on the floor. McAllister's testimony that an under-ventilated fire in a confined space would produce a great deal of smoke upon becoming exposed to an open area is consistent with witness testimony regarding the smoky conditions in the mobile home,10 and is more consistent with the medical reports that both Tony and Bunch had soot in their respiratory systems than the State's position at trial that a fire was set in the open.

The post-conviction court also found McAllister's testimony was not worthy of credit because it contradicted the trial testimony of investigator Frank—who was able to see the condition of the mobile home following the fire firsthand—that the fire did not start in the ceiling. His testimony, however, does not indicate he made any serious investigation into the ceiling as a possible source of the fire:

Trial Record at 877-78. Frank's testimony indicates because he believed the fire started on the floor, it therefore did not start in the ceiling. He did not testify to any other reasons for coming to the conclusion the fire did not start in the ceiling.

Finally, the post-conviction court is correct that it is undisputed that windows in the bedroom were broken soon after the fire became evident. In a generic sense, it is true that this created ventilation in the room, but in a scientific sense, it is not necessarily true that this "ventilated" the fire. McAllister's opinion was that this was an "under-ventilated" fire because while it burned in the concealed space above the ceiling, it did not have adequate oxygen to burn completely. That windows were broken after the fire burned through the ceiling and dropped down into the room, at which point the fire became ventilated simply by the volume of air in the larger room regardless of whether windows were broken, does not undercut her conclusion.

Indiana Evidence Rule 702 is intended "to liberalize, rather than to constrict, the admission of reliable scientific evidence." Turner v. State, 953 N.E.2d 1039, 1050 (Ind.2011). In Turner, our supreme court held on direct appeal that the trial court did not err in allowing firearms tool mark identification testimony even where the expert's conclusion was equivocal, he did not formally describe his testing method, and he did not pinpoint other research supporting his conclusion because although these facts "inform the fact finder's judgment on weighing this evidence, [they do] not render the evidence inadmissible." Id. at 1051. McAllister described the method by which she reached an unequivocal conclusion that the fire, though still of undetermined origin, did not start as the State alleged, noting specifically the necessity for a valid hypothesis to be consistent with the facts of a case.11 See PCR Hrg. Tr. at 341. Her methodology is sufficiently supported by the generally-accepted scientific principles set forth in NFPA 921, see Farmland Mut. Ins. Cos. v. Chief Indus., Inc., 170 P.3d 832, 836 (Colo.App. 2007) (noting that "a number of courts have held that [NFPA 921] ... is an accepted reference for fire investigators"), and her conclusion is sufficiently consistent with the evidence of record to be worthy of credit.

8) The evidence can be produced upon a retrial. Although the post-conviction court concluded that McAllister's testimony could not be produced upon a retrial of the case, it did so based on the alleged lack of scientific reliability. Having determined above that her testimony was sufficiently supported by scientific methods to be worthy of credit, we conclude the evidence could be produced upon retrial by either McAllister's own testimony, that of her associate, or another similar expert. Cf. Thompson v. State, 796 N.E.2d 834, 839 (Ind.Ct.App.2003) (statement proffered as newly-discovered evidence did not qualify as a dying declaration and therefore the evidence could not be produced at a new trial), trans. denied.

9) The evidence will probably produce a different result at retrial. "In ruling on whether the evidence would produce a different result, the trial court may properly consider the weight that a reasonable trier of fact would give it and while so doing may also evaluate its probable impact on a new trial in light of all the facts and circumstances shown at the original trial of the case." Reed v. State, 702 N.E.2d 685, 691 (Ind.1998). "[T]he defendant must raise a strong presumption that the result at any subsequent trial in all probability would be different." Id.; cf. Taylor, 840 N.E.2d at 330 (there was overwhelming evidence of defendant's guilt so although proffered new evidence "might have weakened the State's case," it was not enough to make it probable that a different result would be obtained). "A sufficient probability of a different result upon retrial is present when the omitted evidence creates a reasonable doubt that did not otherwise exist." Fox v. State, 568 N.E.2d 1006, 1008 (Ind.1991).

Bunch contends the fire victim toxicology evidence would likely produce a different outcome on retrial because it "does not merely call into question some of the State's evidence," it establishes her innocence. Br. of the Appellant at 29. There was no direct evidence at Bunch's trial that she set multiple fires in the mobile home through use of an accelerant. No one saw her set the fire or heard her refer to setting the fire, there was no evidence she had purchased a liquid accelerant, and there was no motive offered for her to intentionally set a fire. McAllister's testimony shows the State's theory of fire purposefully set in an open room could not have caused the injuries from which Tony was shown to have died. McAllister's conclusion is, if not entirely consistent with the eyewitness testimony, at least not inconsistent with it: that several witnesses testified they saw fire on the floor does not mean the fire started there; testimony regarding large amounts of thick smoke and evidence Tony inhaled soot is consistent with an under-ventilated fire; and testimony that there was no fire on the living room floor in the early stages of the fire is consistent with an accidental fire beginning in the ceiling in the south bedroom.

The largely circumstantial evidence at trial did not overwhelming prove Bunch's guilt. Bunch was alone in the trailer with Tony when the fire started. She gave inconsistent statements regarding where Tony was immediately prior to the fire and what she heard him say and saw him do once she discovered the fire, but never deviated from her observation that the fire started in the bedroom. She told a neighbor that Tony was behind a locked door when there was no door at all to the bedroom. She allegedly told another neighbor that "somebody" had sprinkled kerosene or gasoline12 on her before the fire, but that neighbor's credibility was suspect given testimony that no such conversation occurred and that after the fire but before trial, the two families were engaged in litigation against each other. A chair partially obstructed the doorway from the living room into the bedroom, but had been in that same position for a considerable length of time. However, she had no apparent motive to set the fire, no one saw her do it or heard her talk about it, she had not recently purchased an accelerant, and her clothing did not indicate the presence of an accelerant when tested. McAllister's fire victim toxicology analysis evidence, although not definitively able to disprove that Bunch committed a crime, at least creates a reasonable doubt that she committed the crime with which she was charged: knowingly using an accelerant to start a fire in the living room of her mobile home. See Record of Proceedings at 8. In reaching that conclusion, we acknowledge the inconsistencies in Bunch's statements, but note also that she was asked to give two statements on the day of the fire, including one while she was still at the hospital, and that only excerpts of those statements were introduced into evidence. The State's evidence was largely circumstantial, none of Bunch's statements are actually inculpatory, and the toxicology evidence supports her defense of innocence. We conclude it is probable this evidence would produce a different result if offered on retrial.

Summary. Bunch has met her burden of proving the fire victim toxicology analysis evidence meets all nine requirements of newly-discovered evidence. We therefore hold the post-conviction court clearly erred in denying her petition for post-conviction relief on this claim.

III. Brady Violation

In Brady v. Maryland, the United States Supreme Court held that "the suppression by the prosecution of evidence favorable to the accused upon request violates due process where the evidence is material either to guilt or to punishment, irrespective of the good faith or bad faith of the prosecution." 373 U.S. 83, 87, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963). "To prevail on a Brady claim, a defendant must establish: (1) that the prosecution suppressed evidence; (2) that the evidence was favorable to the defense; and (3) that the evidence was material to an issue at trial." Minnick v. State, 698 N.E.2d 745, 755 (Ind.1998) (citing Brady, 373 U.S. at 87, 83 S.Ct. 1194), cert. denied, 528 U.S. 1006, 120 S.Ct. 501, 145 L.Ed.2d 387 (1999). Evidence is material under Brady "only if there is a reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different. A `reasonable probability' is a probability sufficient to undermine confidence in the outcome." United States v. Bagley, 473 U.S. 667, 682, 105 S.Ct. 3375, 87 L.Ed.2d 481 (1985). However, the State will not be found to have suppressed material evidence if it was available to a defendant through the exercise of reasonable diligence. Conner v. State, 711 N.E.2d 1238, 1246 (Ind.1999), cert. denied, 531 U.S. 829, 121 S.Ct. 81, 148 L.Ed.2d 43 (2000). "Favorable evidence" includes both exculpatory evidence and impeachment evidence. See Prewitt v. State, 819 N.E.2d 393, 401 (Ind.Ct.App.2004), trans. denied. Suppression of Brady evidence is constitutional error warranting a new trial. Turney v. State, 759 N.E.2d 671, 675 (Ind.Ct.App.2001), trans. denied.

Bunch contends when she subpoenaed the ATF file from her case for post-conviction purposes, she "discovered a wealth of exculpatory material never disclosed to her during her trial." Br. of the Appellant at 42. Specifically, she claims William Kinard's testimony and report offered at her trial are directly contradicted by an earlier report.13 Kinard testified that he conducted gas chromatography tests on ten samples taken from the fire scene to determine whether an accelerating material was present. Trial Record at 891. He further testified that a heavy petroleum distillate was present in five of the samples (samples 4, 5, 6, 8, and 10). Id. at 907-08. Kinard conceded on cross-examination that the carbon numbers found in samples 4, 5, and 10, although consistent with each other, were different from the carbon numbers found in samples 6 and 8. He nonetheless believed the carbon numbers for all five samples were within the range indicating the presence of a heavy petroleum distillate. Kinard's written report disclosed to Bunch prior to trial and offered into evidence at trial, consistent with Kinard's testimony, states:

Id. at 899 (State's Exhibit ATF 1) (referred to hereafter as "Report # 2"). The report that Bunch did not have until the post-conviction proceedings states:

Petitioner's PCR Exhibit ATF 3 (referred to hereafter as "Report # 1").18 In addition, until the post-conviction proceedings, Bunch did not have the actual gas chromatographs for the samples. John DeHaan, her expert on post-conviction, testified his evaluation of the gas chromatographs shows samples 6 and 8 tested negative for a heavy petroleum distillate and he therefore agrees with the conclusions in Report # 1 that was not made available to Bunch prior to her trial. Bunch contends the determination that sample 8 did not exhibit the presence of a heavy petroleum distillate is especially critical because it is the sole sample from the bedroom and the State relied heavily at trial upon the presence of a heavy petroleum distillate in that sample to prove there were multiple fires and support its theory the fire was intentionally set.

With respect to the undisclosed evidence, the post-conviction court concluded:

Appellant's App. at 34-41.

A. The State Suppressed Evidence

Bunch contends the post-conviction court erred in determining the complete ATF documents were work product exempt from Brady, citing several out-of-jurisdiction cases determining that because Brady is of constitutional dimension, it trumps rules prohibiting discovery of work product. See, e.g., Castleberry v. Crisp, 414 F.Supp. 945, 953 (N.D.Okla.1976) (noting that the work product discovery rule "cannot, of course, be applied in a manner which derogates a defendant's" Brady rights). The report at issue herein is not akin to a police report or other investigative notes which are protected as work product, and the State never alleged that they were. See Goudy v. State, 689 N.E.2d 686, 695 (Ind.1997) (noting "investigative police reports are not discoverable and are considered protected as the work product of the prosecutor"). Even if they were, however, the "defendant's right to fundamental due process outweighs the State's interest in nondisclosure." Sewell v. State, 592 N.E.2d 705, 707 n. 4 (Ind.Ct. App.1992), trans. denied. Thus, the Brady rule can require disclosure of evidence not otherwise discoverable if the evidence is shown to be exculpatory. Id. at 707.

Bunch also contends the post-conviction court erred in relying on the fact the State did not know of the contents of the ATF file because prosecutors have a duty "to learn of any favorable evidence known to the others acting on the government's behalf in the case...." Kyles v. Whitley, 514 U.S. 419, 437, 115 S.Ct. 1555, 131 L.Ed.2d 490 (1995). In Farris v. State, 732 N.E.2d 230 (Ind.Ct.App.2000), the State deposed a witness who testified the defendant was involved in a robbery. Five days prior to the defendant's trial, the witness recanted his testimony as to the defendant's involvement on the errata sheet to his deposition. The State called the witness at trial and the witness testified consistent with his recantation that the defendant had not been involved in the robbery. Several weeks after he was found guilty of robbery, the defendant became aware of the witness's errata sheet and filed a motion to set aside the verdict on the basis of a Brady violation. Without further elaboration, this court held the State had suppressed evidence from the defense because even though the prosecutor was unaware the errata sheet existed, the State "had the errata sheet in its possession before trial and did not disclose it to the defendant." Id. at 233. It would appear, therefore, that since the State took the witness's deposition, even if the errata sheet did not actually make its way from the reporter transcribing the deposition to the prosecutor, it was considered to be in the State's possession. See Martinez v. Wainwright, 621 F.2d 184, 186-87 (5th Cir.1980) ("The duty to produce requested evidence falls on the state; there is no suggestion in Brady that different `arms' of the government are severable entities[,]" citing United States v. Deutsch, 475 F.2d 55, 57 (5th Cir.1973), holding the United States Attorney suppressed evidence in the possession of the Post Office Department, and Yanetta v. State, 320 So.2d 23, 24 (Fla.Ct.App. 1975), holding defendant entitled to discovery of information not just in the physical possession of the State but also obtainable from the FBI). In this case, the ATF was acting at the behest of the State in testing the samples and providing a report. That the ATF kept the complete file on its premises does not mitigate the State's obligation to disclose exculpatory evidence in that file.

Finally, Bunch contends the post-conviction court erred in finding that she did not demonstrate reasonable diligence when she made no specific request for the ATF file despite having Report # 2. See Kyles, 514 U.S. at 433, 115 S.Ct. 1555 ("[A] defendant's failure to request favorable evidence [does] not leave the Government free of all obligation."); see also United States v. Agurs, 427 U.S. 97, 108, 96 S.Ct. 2392, 49 L.Ed.2d 342 (1976) (holding the Government has a duty to volunteer exculpatory evidence never requested, or requested only in a general way, though only when suppression of the evidence would be of such significance as to deny the defendant a fair trial). A defendant is not required to request specific pieces of exculpatory evidence in order to be entitled to them, and is in fact entitled to rely on the State's representation that it has fulfilled its duty to provide any such evidence. See Strickler v. Greene, 527 U.S. 263, 289, 119 S.Ct. 1936, 144 L.Ed.2d 286 (1999). Although it is true Bunch was able to receive the full ATF file upon requesting it in these post-conviction proceedings, "the availability of information is not measured in terms of whether the information is easy or difficult to obtain but by whether the information is in the possession of some arm of the state." Crivens v. Roth, 172 F.3d 991, 997-98 (7th Cir.1999). This case is unlike Denney, 695 N.E.2d 90, in which our supreme court held the defendant failed to exercise due diligence to discover results from a courtordered blood test. That a blood test was done was known to the defendant himself and was evident on the face of the record, therefore the defendant had the opportunity to pursue the results. Here, however, Bunch had no reason to know from the fact of Report # 2 that there was a Report #1, let alone that it had not just shown equivocal results for two of the samples, as in Kinard's trial testimony regarding samples 6 and 8, but actually showed negative results for those two samples.

The State asserted at oral argument that it is the State's right and obligation to decide what evidence in its possession is favorable to the accused and therefore required to be turned over to the defense in discovery. We do not disagree that it is initially the State's call, but we note that the State then has to live with the consequences of its decision. The State has the affirmative duty to turn over exculpatory evidence even in the absence of a specific request by the defendant. However, if the State determines not all evidence available to it is required to be turned over and it does not know for a fact that the defendant already has the evidence, the State runs the risk that a court reviewing a subsequent Brady claim may disagree with its assessment. The State concedes it has the obligation to turn over Brady evidence and concedes it did not turn over the entire ATF file to Bunch prior to her trial; it therefore failed to satisfy its obligation in this case. We conclude the State suppressed the complete ATF file.

B. The Evidence was Favorable to Bunch

The second showing required to establish a Brady violation is that the suppressed evidence was favorable to the defense. The post-conviction court concluded the evidence was not favorable to Bunch because her own expert agreed with Kinard's trial testimony that three of the five samples show the presence of kerosene. The post-conviction court specifically noted that DeHaan confirmed the accuracy of Kinard's conclusions as to 80% of the samples tested. The post-conviction court did not, however, note the significance of the disagreement with regard to the other two samples. Kinard testified that samples 6 and 8 also showed the existence of a heavy petroleum distillate. Samples 4, 5, 6, and 10 were taken from the living room; sample 8 was the only sample taken from the south bedroom.

That Kinard identified a heavy petroleum distillate in samples taken from both the living room and the bedroom was a lynchpin of the State's theory that there were multiple, and therefore incendiary, fires. Kinard acknowledged that the carbon numbers identified in samples 6 and 8 were different from those identified in samples 4, 5, and 10, and samples 6 and 8 were therefore on the "tailend of a heavy petroleum distillate." Trial Record at 907. Nonetheless, he adhered to the conclusion expressed in Report # 2 that samples 6 and 8 tested positive for a heavy petroleum distillate. Report # 1, however, concluded samples 6 and 8 tested negative, a conclusion with which DeHaan agreed at the post-conviction hearing after reviewing the gas chromatographs. If sample 8—the only sample taken from the bedroom— tested negative, then there is no evidence supporting an incendiary fire in the bedroom, and no evidence supporting multiple incendiary fires. If sample 6—taken from what was identified as a pour pattern in the living room—tested negative, then the significance of the alleged burn pattern is significantly undercut. The undisclosed evidence is therefore exculpatory, as it directly contradicts the State's theory of the case.

The evidence is also impeaching, as Report # 1's finding that samples 6 and 8 tested negative for heavy petroleum distillates could have been used to impeach Kinard's assertion in Report # 2 and at trial that despite the differing carbon numbers and initial assessment, samples 6 and 8 showed the presence of a heavy petroleum distillate. Because the State did not turn over the entire ATF file, Bunch did not have the benefit of knowing that the finding in Report # 1 more strongly supported her defense than the State's theory. She also was not able to point out the change and pursue the reason the findings were altered and Kinard testified differently at trial. As either exculpatory or impeaching evidence, Report # 1 was favorable to the defense.

C. The Evidence was Material

To show the evidence was material, Bunch must show there is a reasonable probability that the result of the trial would have been different had it been disclosed to her. See Hayden v. State, 830 N.E.2d 923, 931 (Ind.Ct.App.2005), trans. denied. Bunch contends Kinard's testimony identifying heavy petroleum distillates in floor samples from both the living room and the south bedroom formed part of the basis for the State's theory that two incendiary fires were set in the mobile home and was likely critical to the jury's decision. Bunch contends the undisclosed evidence defeated the State's two-fire theory and would have bolstered the attempted impeachment of Kinard's testimony that although the composition of samples 6 and 8 was different from the others, it was still within the range of a heavy petroleum distillate. In short, Bunch contends Kinard's testimony would have been so undermined by this additional evidence that the entirety of the evidence would have been cast in a different light more favorable to Bunch.

We note that Bunch cross-examined Kinard about the differences in the results obtained for samples 4, 5, and 10 versus samples 6 and 8. We also note the non-scientific evidence the State alleges supports the conviction; namely, that Bunch was the only adult in the trailer at the time of the fire, she gave inconsistent statements to neighbors and police, and there was testimony that the doorway into the bedroom was at least partially obstructed. Nonetheless, the State proceeded primarily on the theory that there were multiple fires in the mobile home, and because multiple, non-communicating fires are more likely than not to be incendiary, Bunch was responsible for setting the fires. Kinard's testimony that floor samples from both the living room and the bedroom contained evidence of a flammable or combustible liquid is the strongest, if not the only, evidence that there were two separate fires in the mobile home.

The undisclosed evidence directly contradicts Kinard's trial testimony supporting fires originating in two places. Thus, there is a reasonable probability that but for the prosecutorial failure to disclose this evidence favorable to Bunch, the result of the trial would have been different. Bagley, 473 U.S. at 682, 105 S.Ct. 3375. And since the suppression of this evidence undermines confidence in the outcome, see id., we conclude that this federal constitutional error was not harmless beyond a reasonable doubt. See Chapman v. California, 386 U.S. 18, 24, 87 S.Ct. 824, 17 L.Ed.2d 705 (1967).

The post-conviction court clearly erred in concluding there was no Brady violation in the State's failure to disclose material exculpatory or impeaching evidence to Bunch prior to trial. Bunch is also entitled to a new trial on the basis of this violation.

Conclusion

The post-conviction court clearly erred in determining Bunch was not entitled to a new trial on the basis of the fire victim toxicology analysis evidence, as the evidence meets each of the nine requirements to be newly-discovered evidence. The post-conviction court also clearly erred in determining Bunch was not entitled to a new trial on the basis of a Brady violation by the State. Because our resolution of these issues is dispositive, we do not address Bunch's remaining newly-discovered evidence claim or her claim of ineffective assistance of counsel. We reverse the post-conviction court's denial of Bunch's petition for post-conviction relief, and remand for a new trial.

Reversed and remanded.

NAJAM, J., concurs.

CRONE, J., dissents with separate opinion.

CRONE, Judge, dissenting.

"To prevail on appeal from the denial of post-conviction relief, a petitioner must show that the evidence as a whole leads unerringly and unmistakably to a conclusion opposite that reached by the post-conviction court." Kubsch, 934 N.E.2d at 1144. I believe that Bunch has failed to meet this burden as to any of her claims, and therefore I respectfully dissent.

1. Newly Discovered Evidence: Victim Toxicology

Newly discovered evidence mandates a new trial only when the post-conviction petitioner demonstrates each of the following nine requirements:

Taylor, 840 N.E.2d at 329-30 (quoting Carter, 738 N.E.2d at 671). The burden of establishing all nine prerequisites for a new trial rests with the petitioner. Webster, 699 N.E.2d at 269.

I have no quarrel with the majority's determination that "the fire victim toxicology analysis offered by [Jamie] McAllister has been discovered since [Bunch's] trial" and that, "[e]ven with due diligence, there would have been no way for Bunch to have discovered this analysis prior to her trial." Op. at 290-91, 292-93. That said, I respectfully disagree with the majority's determination that the fire victim toxicology evidence is material and relevant and that it will probably produce a different result at trial, all of which hinges on its determination that the evidence is worthy of credit.

"Whether a witness's testimony at a post-conviction hearing is worthy of credit is a factual determination to be made by the trial judge who has the opportunity to see and hear the witness testify." Whedon, 900 N.E.2d at 504. "It is not within an appellate court's province to replace a trial judge's assessment of credibility with its own." Id. One could argue that the judge's credibility assessment is entitled to even greater deference in this case, given that he also presided at Bunch's trial and thus was able to evaluate the credibility of those witnesses (and Bunch's demeanor) as well. See Fox v. State, 568 N.E.2d 1006, 1007 (Ind.1991) (in determining whether evidence would probably produce a different result at retrial, "the judge may properly consider the weight that a reasonable trier of fact would give it and, while so doing, may also evaluate its probable impact on a new trial in light of all the facts and circumstances shown at the original trial of the case") (emphasis added), opinion on reh'g.

The majority says,

Op. at 293.

The majority's sentiments regarding demeanor are well taken, but they are simply beside the point here. To the extent the majority contends that the post-conviction court failed to consider that "new science" formed the basis for McAllister's opinion, I disagree. Although the post-conviction court was indeed skeptical that the principles underlying McAllister's opinions were "generally accepted in the relevant scientific community," perhaps because of her vague reference to "numerous studies," its primary concern was that those opinions were based on factual premises that conflicted with the evidence presented at trial. I agree with the majority that we are in as good a position as the post-conviction court to assess the validity of the foundation for McAllister's opinions because we have the same information before us. That said, I do not believe that we are in as good a position as the post-conviction court—especially in this case, where the same judge presided at trial—to make the substantive determination of whether a witness's opinions are sufficiently credible to merit a new trial. The two analyses are not the same.19

In this case, the post-conviction court specifically found that McAllister's scientific opinions were not worthy of credit based on the following considerations: (1) trial evidence (including statements from Bunch herself) that the fire started on the floor, not in the ceiling, as McAllister opined;20 (2) Bunch's statements that "there was not much smoke when she first observed the fire—clearly contradicting what would have had to have happened if this was a ceiling compartment fire as McAllister concludes," Appellant's App. at 25; (3) Rob Parkinson's testimony that he saw fire only on the living room floor, not in the ceiling; (4) Brian Frank's testimony that he found no evidence that the fire started in the ceiling;21 (5) trial evidence that the fire was ventilated "very early ... by breaking the windows to the south bedroom where the child was found dead and by an open entry door to the mobile home," id. at 21; and (6) McAllister's failure to present "evidence to support her estimate of the temperature of the fire in general or at any specific time," id.22 All of the foregoing findings are supported by evidence in the record, and all of them support the post-conviction court's conclusion that McAllister's scientific opinions are not worthy of credit and thus are neither material nor relevant and would probably not produce a different result on retrial. In concluding otherwise, the majority has improperly reweighed evidence and reassessed witness credibility.

The majority has also minimized Bunch's self-incriminating words and deeds, which the post-conviction court described as follows:

Id. at 25-26. Tellingly, the majority does not (and indeed cannot) contradict the post-conviction court's findings and conclusions on this point.

Additionally, the majority asserts that "McAllister opined that Bunch's injuries— first degree burns on her face and forearms, a second degree burn on the tip of her nose, and some singeing of the hair around her face—were not consistent with intentionally setting a fire with a liquid accelerant." Op. at 287. In fact, McAllister testified as follows:

PCR Tr. at 313.

Although the State's primary theory at trial was that Bunch used kerosene as an accelerant, the jury heard evidence that investigators found an empty can of gasoline just outside the door of Bunch's trailer and later recovered from inside the trailer "the remains of a melted gas nozzle that went on that can." Trial Tr. at 744.24 Curiously, in her statements to investigators, Bunch mentioned gasoline at least twice. See id. at 706 ("I tried to smother [the fire]. I just threw [a blanket] right over it, and the flames just went right through it. And I tried taking a pillow and beating it; and it was like, I don't know, gasoline or something; because it just went shooooooo, right all over the doorway by the time I got back."); id. at 713-14 ("I ran over here; grabbed the blanket; and brought it back and laid it across that flame. And I went and got the pillow and I started hitting it with the pillow. And then it was just, I don't know—the bed, I mean, the dresser, everything just went fooooooooo. And it was just, I just thought gas was in there all over the place."). During his rebuttal closing argument, the prosecutor conceded that "[w]e don't know that gasoline wasn't used" to start the fire, id. at 1426, and Bunch's expert witness, John DeHaan, acknowledged at the post-conviction hearing that gasoline "is considerably more volatile than kerosene is" and thus a fire investigator could "lose enough of the ... most volatile components ... such that [the investigator] couldn't identify it if it was actually present in significant quantities." PCR Tr. at 166. Expert witness John Malooly made a similar acknowledgement. See id. at 451 ("Q Then after the flashover any kerosene or any petroleum distillate that would have been on top of that linoleum would have evaporated, correct? A If it was on the free surface of the linoleum yes, but we wouldn't generally take a sample there.").25 All of this goes to say that Bunch has failed to convince me that she is entitled to a new trial based on the victim toxicology evidence.

2. Brady Violation

"To prevail on a Brady claim, a defendant must establish: (1) that the prosecution suppressed evidence; (2) that the evidence was favorable to the defense; and (3) that the evidence was material to an issue at trial." Minnick, 698 N.E.2d at 755. "[T]he State will not be found to have suppressed material information if that information was available to a defendant through the exercise of reasonable diligence." Conner, 711 N.E.2d at 1246. "Evidence is `material' only if there is a `reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different.'" Minnick, 698 N.E.2d at 755 (quoting Bagley, 473 U.S. at 685, 105 S.Ct. 3375). Bunch's Brady claim is based largely on insinuation and innuendo, and, unlike the majority, I believe that it fails in all respects.

First, Bunch knew that the ATF had performed testing on the flooring samples taken from her trailer, and she received the ATF's "final" report from the State before trial. As such, Bunch would have known that the ATF possessed raw test data (and perhaps other information related to the testing), but she failed to request it. "Where, as here, alleged Brady material was available to [the defendant] through the exercise of reasonable diligence, [she] cannot obtain a new trial by insisting that the government should have conducted [her] investigation for [her]." U.S. v. Morris, 80 F.3d 1151, 1170 (7th Cir.1996), cert. denied.

More importantly, the raw test data was not suppressed because Kinard read the gas chromatograph results to the jury at trial. Kinard's interpretation of the results provided fodder for Bunch's cross-examination, specifically regarding samples C6 and C8, which are the focus of Bunch's Brady allegations.26 The State points out that the ATF's "draft" report "does not `reveal' hidden proof that samples C6 and C8 did not contain HPD; it proves what [Kinard] testified to at Bunch's trial, namely, that he considered samples testing at or above the minimum limit for kerosene in ASTM 1387[27] to contain HPD." Appellee's Br. at 25. In other words, the evidence at issue is not as favorable as Bunch claims.

Nor is it material. The gas chromatograph results were disclosed at trial, and the jury considered that evidence along with Bunch's vigorous cross-examination of Kinard regarding those results and ultimately found her guilty beyond a reasonable doubt. In my view, there is no reasonable probability that disclosure of the ATF's "draft" report would have affected the outcome of the trial. Consequently, I find no merit in Bunch's Brady claim.

* * *

Because the majority has reversed based on the two foregoing issues and does not address the two remaining issues raised in Bunch's petition for post-conviction relief, I briefly address those issues below to explain my decision to affirm the post-conviction court's ruling in all respects.

3. Newly Discovered Evidence: Fire Investigation Techniques

As to fire investigation techniques, Bunch presented the testimony of John DeHaan, who has approximately forty years of experience in fire and explosion investigations. DeHaan testified that at the time of Bunch's trial, although the fire investigation community was aware of the process called "flashover,"28 its effects were not well known, and that many of the fire artifacts previously believed to indicate an incendiary fire (such as burn patterns, floor charring, holes in the floor, and V-patterns) are now recognized as products of flashover. DeHaan testified that he did not see any evidence of an intentional cause behind this fire. PCR Tr. at 111. He testified that the origin of this fire should be classified as undetermined, id. at 184, although he believed it most likely started somewhere in the wall structure between the south bedroom and living room. Id. at 132.

Bunch also presented the testimony of John Malooly, a fire investigator of over thirty years, who was a certified fire investigator and certified explosives specialist with ATF and who participated in, for example, the investigation of the bombing of the Murrah Building in Oklahoma City. Based on his review of the evidence at Bunch's trial, Malooly opined that "there is not a basis for the conclusion that this was an incendiary or an arson fire" and that what the State pointed to as evidence of an accelerant was "evidence of a post-flashover burning." Id. at 381. Malooly agreed with DeHaan that the effects of flashover were not widely known or understood by fire investigators in the mid-1990s. He also testified that the experts in Bunch's trial did not accurately describe flashover, either the scientific basis for flashover or whether flashover occurred in this fire. He further rebutted the State's trial evidence regarding burn patterns, the melting of an aluminum carpet tack strip, holes burned in the floor, and V-patterns as proving the use of a liquid accelerant.

Bunch contends that this evidence meets all nine criteria of newly-discovered evidence and entitles her to a new trial. The post-conviction court concluded that this evidence has not been discovered since trial, was cumulative, and was merely impeaching of the State's witnesses. I agree with the post-conviction court's assessment as to all three criteria, but because the failure to establish even one will defeat a newly-discovered evidence claim, I limit my analysis to impeachment.

Bunch contends that the evidence is not merely impeaching because it "obliterates the testimony upon which [her] conviction was obtained," see Wilson, 677 N.E.2d at 588, by undercutting each of the bases on which the State's fire investigation witnesses premised their opinions that this was an incendiary fire. But Hulse's trial testimony undercut the bases of the State's investigation as well. State's witness Frank testified that V-patterns and other burn patterns at the scene indicated intentionally set fires; Hulse testified that V-patterns are not persuasive of an incendiary fire. Frank testified that there were two separate fires caused by liquid accelerant, and Hulse testified that there was no liquid accelerant present and that there was but one fire originating in the south bedroom. DeHaan's and Malooly's testimony may impeach the State's witnesses, but Bunch's trial expert's testimony did so as well. As such, I cannot conclude that the post-conviction court clearly erred in determining that the post-conviction evidence was merely impeaching and in denying Bunch post-conviction relief on this claim.

4. Ineffective Assistance of Trial Counsel

Finally, Bunch raises numerous claims of ineffective assistance of trial counsel, including that counsel failed to call or consult with various experts, failed to present allegedly exculpatory evidence, and either elicited or failed to object to allegedly damaging evidence at trial. Our standard of review is well settled:

Harley v. State, 952 N.E.2d 301, 303 (Ind. Ct.App.2011) (citations omitted).

Randolph v. State, 802 N.E.2d 1008, 1013 (Ind.Ct.App.2004) (citations omitted), trans. denied. "The judicial scrutiny of counsel's performance is highly deferential and should not be exercised through the distortions of hindsight." Talley v. State, 736 N.E.2d 766, 769 (Ind.Ct.App.2000).

Here, Bunch's experienced trial counsel called experts, aggressively cross-examined witnesses, and made numerous objections. The evidence of Bunch's guilt, albeit circumstantial, was strong, especially her own incriminating words and actions both during and after the fire. None of the claims of deficient performance alleged by Bunch, even if true, would rise to the level sufficient to establish a reasonable probability that, but for those alleged errors, the result of her trial would have been different. Consequently, I would affirm the post-conviction court.