Classification: Documented

Roberto Wagner Fernandes: three women, two decades, and an identification after death

Florida investigators linked Roberto Wagner Fernandes by DNA to the killings of Kimberly Dietz-Livesey, Sia Demas and Jessica Good in 2000-2001. The trail runs through the Brazilian trial over Danyelle Bouças Fernandes’s death, an exhumation, and suspicions about other crimes that remain unproven.

Miami, Flórida, Estados Unidos · See stories from this place and nearby

Begin with the women

On 22 June 2000, Kimberly Dietz-Livesey was found in a suitcase beside a road in Cooper City, Florida. On 9 August, Sia Demas was found in a travel bag near Dania Beach. More than a year later, on 30 August 2001, Jessica Good was recovered from Biscayne Bay in Miami. These dates and places appear in a statement by the Broward Sheriff’s Office. Kimberly was 35, Sia 21 and Jessica 24, according to reporting around the police announcement.

These were separate killings in different places and years, investigated by different agencies. The first two women suffered fatal beatings; Jessica died from stab wounds. The difference in the third case initially made the connection harder to see. Police applied the term serial killer after physical evidence linked the murders. There was no American criminal conviction: the man identified by investigators died before he could be tried.

Before Florida: Danyelle’s death

On 18 November 1996, Danyelle Amaral Bouças Fernandes, a 27-year-old teacher, died after being shot twice in the Londrina, Paraná apartment she shared with Roberto. The Folha de Londrina reported the case the next day and said her husband had left the building with their daughter and become the principal suspect. In a statement reported days later, Fernandes acknowledged firing the gun but said the shots occurred during a struggle and claimed self-defence.

The case went before a Brazilian jury, which acquitted him. That result matters: Danyelle’s death and his account of firing the gun are documented, but he was not convicted of murdering her. Retrospective articles sometimes call her his “first victim” after the discoveries in the United States. This dossier does not turn that later interpretation into a verdict the Brazilian jury never reached. Its map marker is in central Miami, an approximate centre for the American investigation, not at Danyelle’s home.

Another voice in the 1996 record

Reporting on 22 November 1996 preserved conflicting accounts of a prior assault. Solange Nunes Ramalho approached police and said Fernandes had attacked her during an encounter and that she had sought medical attention. He gave a different version, saying a dispute over money led her to attack him. The newspaper reported both her statement and the defence’s response. Her account matters because a woman involved in sex work made a violence complaint that entered the record; the available account does not show a judgment proving that allegation.

The Folha de S.Paulo found in 2022 that the assault proceeding ended after the suspect’s death. It also reported a later accusation, in a sealed case, of attempted sexual abuse of a minor; that case became time-barred. Neither is a conviction. Nor does the jury’s acquittal in Danyelle’s death erase Solange’s account. The distinctions between a complaint, a proceeding, an acquittal and a forensic conclusion are essential to reading this file.

How Jessica’s case produced a name

After Jessica’s death, Miami police had Fernandes as a person of interest, but he returned to Brazil before they could interview him. Decades later, Broward detective Zachary Scott explained on Cold Case Files: Miami that Jessica’s boyfriend had relayed information about the person she planned to meet. Scott described a lead involving a tour company and a vehicle. It helps explain why Miami had a name while the two Broward murders still lacked an identified suspect.

That interview was recorded decades after the events, and its automatically generated transcript contains transcription errors. The account is therefore attributed investigative recollection, not an original witness statement or laboratory report. The sheriff’s official statement makes the narrower point that suspicion arose in the Good case and that Fernandes went back to Brazil. No public evidence establishes that all three women met him in the same way.

The missing link between three investigations

Two crime scenes contained compatible fingerprints, and biological profiles were preserved. Broward and Miami investigators cooperated, but the identity remained unknown for years. In 2011, renewed analysis of evidence in Jessica’s case yielded a DNA profile matching the unidentified profile in the Broward killings. In his 2025 interview, Scott said extra resources for old Miami homicides enabled the processing. The immediate finding was that the same man was linked to all three killings.

Investigators still needed to show that he was Fernandes. Fingerprints taken in Brazil after Danyelle’s death were compared with traces from the Florida scenes and pointed to him. DNA and fingerprints played different roles: the first joined the murders; the second helped name the unknown profile. A circumstantial lead from 2001 thus gained the support of physical evidence years later. The sequence matters because it avoids the false impression that police possessed one conclusive test from the beginning and ignored it.

The dead man who had to be identified

When American investigators sought a sample from Fernandes, they learned he had reportedly died in a 2005 plane crash in Paraguay. News of a crash did not establish the identity of the buried remains. In his 2025 interview, Scott described doubts investigators had about the grave and transfer records. They were historical doubts during the inquiry, not evidence that Fernandes survived.

After cooperation between Brazilian and American authorities, the grave was opened in late 2020 or early 2021. Broward’s statement says remains were found and their DNA profile was consistent with evidence from all three scenes. Scott later added that a family reference sample helped establish the remains’ identity before comparison with the American profiles. The exhumation answered two distinct questions: who was in the grave, and whether that person’s DNA matched the murders.

What the 2021 announcement established

On 31 August 2021, the Broward Sheriff’s Office and Miami police publicly announced that the evidence identified Roberto Wagner Fernandes as responsible for Kimberly’s, Sia’s and Jessica’s deaths. CBS Miami, reporting on the news conference, recorded the explanation of DNA, fingerprints and exhumation, and the response of Michael Livesey, Kimberly’s former husband, after two decades of waiting for an answer. The police conclusion was supported by the forensic match authorities described.

It was not a court judgment. Fernandes had died and could not be charged, challenge the evidence or be convicted of the Florida crimes. “Identified by police” is more precise than “convicted.” His death also foreclosed questioning about how the encounters occurred, why these women were targeted and whether he committed other crimes. The finding answered a crucial forensic question for relatives without resolving every historical question.

The number beyond three

Broward officials said in 2021 that Fernandes might be connected to additional killings. In the 2025 podcast, Scott repeated his personal suspicion, partly because more than a year passed between Sia’s and Jessica’s deaths. He also said that while examining other cases, investigators found at least one linked to a different perpetrator. A pattern or an interval cannot, on its own, add victims to a suspect’s count.

In the sources reviewed for this dossier, three Florida killings are the cases linked by DNA and named in the police announcement. Danyelle’s death belongs to a separate legal and evidential category: Fernandes was tried and acquitted in Brazil. The violence complaints and time-barred case are not “additional murders.” Any higher total requires identified cases, their own evidence and their investigation status. “Possible serial killer” remains cautious about further victims; the police identification in the three American cases was explicit.

What the label can hide

Police and some news accounts noted that the three women had histories of drug use and sex work. Such details may explain an investigative geographic lead, but they do not define their lives or diminish their deaths. Local reporting on Danyelle recalled her work as a teacher in Londrina and the street later named for her. Speaking to Folha de Londrina, researcher Silvana Mariano discussed how the testimony of women who do sex work may be devalued. That is her analysis of the social context, not proof that a specific bias determined the jury’s decision.

The investigation crossed borders and decades because physical traces were retained and retested. It also exposes fragmentation: Broward held two linked cases; Miami had a person of interest; Londrina held a different set of records. The question connecting the files is not how often Fernandes can be called a “monster,” but how evidence crossed jurisdictions and what answer could finally be given to the families.

Evidence boundary and further reading

Documented: three women were murdered in South Florida in 2000-2001; police linked the crimes by DNA and identified Fernandes after exhumation. Danyelle was shot in Londrina in 1996, and the jury acquitted her husband. Reported without a conviction: Solange’s assault allegation and another later accusation. Open: whether Fernandes committed further homicides, and which cases they might be. The public sources used here provide no American judgment because he died before the announcement.

Follow the trail through the sheriff’s statement for names, dates and tests; then read the 1996 Londrina reporting to distinguish what was known then from later conclusions; finally hear Scott describe the international obstacles in 2025. The podcast adds firsthand investigative context, though its machine transcript should be checked against the audio. An open question does not permit anonymous victims to be assigned to Fernandes: respecting the three named women and the families who waited for an answer is part of this case’s accuracy.

Open in Google Maps · Approximate location