Environmental factors found to be key predictors of avian flu outbreaks in Europe

Several local factors—including the minimum temperature reached in autumn, the water level in lakes and ponds in winter, and the presence of mute swans (Cygnus olor)—could be key to predicting the potential of an outbreak of highly pathogenic avian flu (HPAI) occurring in Europe.

The findings, published in Scientific Reports, are derived from a machine learning model trained on the characteristics of 21st-century European HPAI outbreaks and could help improve future monitoring programs.

HPAI outbreaks are a serious concern for both animal and public health. A wave of HPAI outbreaks across the Northern Hemisphere during 2022 was associated with an increase in the number of avian influenza virus infections in mammals, which subsequently increased the likelihood of a spillover event in humans. To reduce the chances of such an event occurring, it is critical for scientists to understand the underlying factors which can increase the likelihood of an HPAI outbreak.

Researcher Joacim Rocklöv and colleagues trained a machine learning model on the characteristics of every HPAI outbreak reported in Europe between 2006 and 2021. The characteristics assessed were all identified as potential outbreak predictors and included: the seasonal temperature and precipitation conditions in the region; the local wild bird population; the local farmed poultry density; and the seasonal vegetation density and water level in the region. The authors then tested the accuracy of their model using the outbreak data for 2022 and 2023.

The authors found that the coldest recorded temperature in autumn had the greatest effect on the likelihood of an outbreak occurring. However, the actual effect varied considerably by region. In some areas, warmer minimum temperatures were associated with a higher likelihood of an outbreak; while in others, they were associated with a lower likelihood. Cold winter and spring temperatures were also both associated with an increase in the likelihood of outbreak.

However, a low vegetation density between October and December, and a lower-than-expected water level in lakes and ponds between January and March, were both associated with a decrease in the outbreak likelihood. The authors also noted that the presence of a local population of mute swans was associated with an increase in the likelihood of an outbreak.

The authors suggest that the results could be used to help tailor regional HPAI surveillance programs across Europe, increasing the likelihood of identifying an outbreak in its earliest stages.

Several local factors—including the minimum temperature reached in autumn, the water level in lakes and ponds in winter, and the presence of mute swans (Cygnus olor)—could be key to predicting the potential of an outbreak of highly pathogenic avian flu (HPAI) occurring in Europe.

The findings, published in Scientific Reports, are derived from a machine learning model trained on the characteristics of 21st-century European HPAI outbreaks and could help improve future monitoring programs.

HPAI outbreaks are a serious concern for both animal and public health. A wave of HPAI outbreaks across the Northern Hemisphere during 2022 was associated with an increase in the number of avian influenza virus infections in mammals, which subsequently increased the likelihood of a spillover event in humans. To reduce the chances of such an event occurring, it is critical for scientists to understand the underlying factors which can increase the likelihood of an HPAI outbreak.

Researcher Joacim Rocklöv and colleagues trained a machine learning model on the characteristics of every HPAI outbreak reported in Europe between 2006 and 2021. The characteristics assessed were all identified as potential outbreak predictors and included: the seasonal temperature and precipitation conditions in the region; the local wild bird population; the local farmed poultry density; and the seasonal vegetation density and water level in the region. The authors then tested the accuracy of their model using the outbreak data for 2022 and 2023.

The authors found that the coldest recorded temperature in autumn had the greatest effect on the likelihood of an outbreak occurring. However, the actual effect varied considerably by region. In some areas, warmer minimum temperatures were associated with a higher likelihood of an outbreak; while in others, they were associated with a lower likelihood. Cold winter and spring temperatures were also both associated with an increase in the likelihood of outbreak.

However, a low vegetation density between October and December, and a lower-than-expected water level in lakes and ponds between January and March, were both associated with a decrease in the outbreak likelihood. The authors also noted that the presence of a local population of mute swans was associated with an increase in the likelihood of an outbreak.

The authors suggest that the results could be used to help tailor regional HPAI surveillance programs across Europe, increasing the likelihood of identifying an outbreak in its earliest stages.