Disentangling Seasonality from Co-Occurrence: Anomaly-Driven Networks of Migratory Waterbirds

Understanding how migratory waterbird species co-vary through time can reveal guild structure and guide monitoring in dynamic coastal wetlands, yet seasonal phenology can inflate simple co-occurrence signals. Here, we used standardized monthly bird counts from Yongan Wetland, Taiwan (36 survey months across two survey blocks: November 2014 and January–August 2015, and October 2016–December 2018) to infer de-seasonalized interspecific associations. We analyzed 50 regularly recorded species and a focal subset of 13 shorebirds and ducks. For each species, we transformed raw counts to monthly anomalies that remove recurrent seasonal patterns, then quantified pairwise Spearman correlations and controlled multiple testing using Benjamini–Hochberg FDR (q ≤ 0.05) to construct association networks. The anomaly-based network revealed strong guild structure: positive links were concentrated within dabbling ducks and within shorebirds, consistent with shared habitat use and foraging regimes, whereas negative links were fewer and suggested potential niche partitioning or spatiotemporal segregation. Robustness analyses (moving-block bootstrap stability, a circular-shift null comparison, and log-transformed anomaly sensitivity) supported that the main network patterns were unlikely to arise from chance alignment. Our framework provides a transparent, time-series–based approach for disentangling phenology from association inference, offering a practical framework for wetland monitoring and hypothesis generation about waterbird community dynamics.