Effects of sampling methodology on phenology indices: Insights from sites across India and modelling.

Plant phenology is the study of the timing and extent of leaf, flower, and fruit production. Phenology data are used to study the drivers of cyclicity and seasonality of plant life-history stages, interactions with organisms such as pollinators, and effects of global change factors. Indices such as the timing of phenological events, and the proportion of individuals in a particular phenophase, seasonality, and synchrony have often been used to summarise plant phenology data. However, these indices have specific utilities and limitations and may be sensitive to sampling methodology, making cross-site comparisons challenging, particularly when data collection methods vary in terms of sample size, observation frequency, and the resolution at which phenophase intensity scores/values are recorded. We used fruiting phenology data from tropical trees across five sites in India to study the effects of sampling methodology on two indices: population-level synchrony (overlap) and seasonality. We supplemented these results with simulations of fast- and slow-changing phenologies to test for the effects of sampling methodology on these indices. We found that the overlap index is sensitive to the resolution of phenophase intensity measurement, with coarser intensity measures leading to overestimation of the overlap index. The seasonality index, on the other hand, was not affected by intensity resolution. Simulations indicated that finer intensity resolution is more important than frequency of observation to accurately estimate population synchrony and seasonality for fast- and slow-changing phenophases. Based on our findings, we provide recommendations for future study designs of tropical tree phenology research, particularly for long-term or cross-site studies.