Waterfowl occurrence and residence time as indicators of H5 and H7 avian influenza in North American Poultry

Abstract

Avian influenza (AI) affects wild aquatic birds and poses hazards to human health, food security, and wildlife conservation globally. Accordingly, there is a recognized need for new methods and tools to help quantify the dynamic interaction between wild bird hosts and commercial poultry. Using satellite-marked waterfowl, we applied Bayesian joint hierarchical modeling to concurrently model species distributions, residency times, migration timing, and disease occurrence probability under an integrated animal movement and disease distribution modeling framework. Our results indicate that migratory waterfowl are positively related to AI occurrence over North America such that as waterfowl occurrence probability or residence time increase at a given location, so too does the chance of a commercial poultry AI outbreak. Analyses also suggest that AI occurrence probability is greatest during our observed waterfowl northward migration, and less during the southward migration. Methodologically, we found that when modeling disparate facets of disease systems at the wildlife-agriculture interface, it is essential that multiscale spatial patterns be addressed to avoid mistakenly inferring a disease process or disease-environment relationship from a pattern evaluated at the improper spatial scale. The study offers important insights into migratory waterfowl ecology and AI disease dynamics that aid in better preparing for future outbreaks.

Figure 1

Track Map. Telemetry tracks for 42 blue-winged teal over grid depicting poultry abundance. Black lines show individual bird tracks, legend describes estimated number of poultry (chickens and ducks) per km². To better display poultry abundance in this figure, zero was defined as less than 100 poultry/km².

Figure 2

Locations of Avian Influenza Events (AIP) by virus subtype. AIP cover the period 2004–2019 and included 440 events used for model training and 49 (20%) randomly selected for model validation. Locations shown in the color red signify those retained for validation.

Figure 3

Weekly comparison of historic AI Events (AIPs) with waterfowl latitudinal displacement. Left vertical axis corresponds to smooth black line and represents weekly latitudinal displacement of waterfowl measured in degrees latitude. Horizontal axis provides the week of year. Dotted line intersecting zero signifies no net waterfowl displacement, with portions above zero indicating relative northward movement and portions below showing net southward change. Vertical axis at right corresponds to gray histogram for the weekly counts of all AIP summed over the period 2004–2019. Note apparent temporal correlation between waterfowl spring migration (Weeks 5–18) and increasing AIP as well as that between fall migration (weeks 35–42) and decreased AIP counts.

Figure 4

Random field density comparison. Spatial field resulting from the (A) base model (Model1), (B) base model with clustering effect and other covariates, (C) joint-model for residence time (Model5), and (D) joint-model for occurrence (Model6). Red areas highlight where models under predict AIP, cooler blue regions indicate where models over predict. Maps are plotted on the same scale using the same palette and can be interpreted as “mapped residuals”. Note that random field densities for models incorporating telemetry (C,D) are reduced relative to models without telemetry (A,B), but, even the telemetry models underestimate AIPs over several areas in the upper Midwestern and Northwestern U.S., as well as in central Mexico. Residual error indicates that additional factors or variables are needed to fully explain the AIPs at these locations.

Figure 5

Model estimated clustering covariate. Non-linear, smooth line describes AIP probability as a function of proximity to other, adjacently located AIP occurrences. Vertical axis provides probability (logit scale) and the horizontal axis gives distance in kilometers. Dashed gray lines display 95% Credible Interval. Note that the credible interval includes zero at distances greater than approximately 170 km.

Figure 6

Model estimated temporal effect. Non-linear response of AIP probability as a function of time-structured bird latitudinal displacement. Vertical axis provides probability (logit scale) and the horizontal axis gives time in weeks. Dashed gray lines display 95% Credible Interval. Note that the credible interval includes zero between Weeks 20–35, a time period corresponding with the waterfowl breeding season.

Figure 7

Mapped AIP probabilities predicted by joint-model (Model5). Weeks 7, 15, 18, 23, 33, and 51 of the year are labeled as A through F respectively. Color coding and legend reflect relative probability (0.00–1.00) of H5 or H7 occurrence in North American Poultry with darker colors signifying increased likelihood. Note that C representing week 18 of the year (late April – early May) was predicted to exhibit the greatest AIP probability.