AI Help Detects Wildlife Health Issues in Real Time
During the spring, a disturbing pattern was played as seabirds off the coast of California died of domoic acid poisoning, caused by harmful algal blooms. An early indication shows when and where the spread of this problem began: surviving brown pelicans, red loon loon, and other species began arriving at rehabilitation centers in wildlife with signs of neurological disease. However, even if they painted the state map, these centers did not connect enough to whitewash the issue of the button. If staff at a center diagnose a sick bird, the other 40 miles of the road may not be aware of that information.
That’s why UC Davis researchers examined an early detection system that used artificial intelligence to classify admissions at rehabilitation centers, in hopes of sending wildlife agencies and researchers warning of growing problems among of sea birds and many other species of animals. Their system analyzes drinking reports generated at 30 California centers, listing information such as animal species, age, cause of acquisition, and diagnosis. Afterwards AI used natural language processing to categorize the reports, looking for patterns in the number of receptions associated with certain illnesses and injuries.
The researchers used five years of data and more than 200,000 records to create baselines on how often these conditions occur. If the system detects an anomaly – an uncommon number of cases in a given variant – it automatically generates an alert, which is carried by wildlife experts via the system dashboard, email, or text message. Because the system processes rehab center admission data in just a day or two, it can generate “prediagnostic” alerts, which is much quicker than waiting until diagnoses are confirmed.
In July, the group published a paper describing a test of their system in the journal Royal Society procedures. “We want to use the data in an integrated form to better help rehabbers see more photos, in addition to what they see in their individual centers,” said Devin Dombrowski, president of Wild Neighbors Database Project and one of the authors of the paper.
In a year of pilot study, the system identified several patterns that showed even more problems. The influx of seabirds with neurological symptoms such as shaking of the head and shaking of the whole body is alert. On postmortem examination, these birds, including the black and white water bird species of western grebes, were found to have domoic acid poisoning. A few months earlier, a high San Francisco Bay Area clinic admission rate for stone pigeons showing symptoms of neurological disease had raised another alert. Further examination established the parasite Sarcocystis calchasi as the cause.
Coauthor Terra Kelly, a veterinarian and epidemiologist at UC Davis, compared the system syndromic surveillance for people, who use electronic health records to monitor public health concerns, such as flu outbreaks, opioid overdoses, and the spread of the Zika virus and Covid-19. He pointed out that the animal alert system can also benefit human health. “Wild animals can serve as an early sign” of diseases like West Nile virus, he said. The disease, being killed more than 2,000 people since 1999, according to the Centers for Disease Control and Prevention, sick birds have always been detected before they are examined by domestic animals and humans.
In addition, Kelly said, “We’ll find the first animal invading species to appear in a central California.” For example, if the number of mourning pigeons admitted to wildlife centers suddenly changes, the system will generate an alert to signal to veterinarians that the Eurasian collared pigeon has arrived; they are an invading species that compete for food and can spread parasites on native pigeons.