Ten best practices for effective phenological research

Abstract

The number and diversity of phenological studies has increased rapidly in recent years. Innovative experiments, field studies, citizen science projects, and analyses of newly available historical data are contributing insights that advance our understanding of ecological and evolutionary responses to the environment, particularly climate change. However, many phenological data sets have peculiarities that are not immediately obvious and can lead to mistakes in analyses and interpretation of results. This paper aims to help researchers, especially those new to the field of phenology, understand challenges and practices that are crucial for effective studies. For example, researchers may fail to account for sampling biases in phenological data, struggle to choose or design a volunteer data collection strategy that adequately fits their project's needs, or combine data sets in inappropriate ways. We describe ten best practices for designing studies of plant and animal phenology, evaluating data quality, and analyzing data. Practices include accounting for common biases in data, using effective citizen or community science methods, and employing appropriate data when investigating phenological mismatches. We present these best practices to help researchers entering the field take full advantage of the wealth of available data and approaches to advance our understanding of phenology and its implications for ecology.

Keywords: Citizen science; Community science; Historical data; Mismatch; Phenology network; Remote sensing.

Fig. 1

Interannual variation in the dates of the cherry blossom festival in Kyoto, Japan, shows earlier flowering times in the past 50 years; data were acquired from historical documents and recent records. Dates of full flowering of the mountain cherry (Prunus jamasakura) are shown for A the period from AD 801 to 1400 and B 1400 to 2008. Flowering dates are affected by both global climate change and urbanization. Data gaps exist for particularly old records, many of which were lost during natural disasters and fires. DOY refers to day-of-the-year (i.e., Jan. 1 = 1). Language and calendars changed during this record, requiring researchers to translate old records to modern Japanese language and the Gregorian calendar. C Historical record documenting cherry flowering. D People celebrating the cherry blossom festival in Tokyo. Figure from (Aono and Saito ; Aono and Kazui 2008) (images courtesy of Aono (A–C) and Hiroyoshi Higuchi (D))

Fig. 2

The theoretical effect of changes in population size on changes in first dates of phenological events, assuming constant sample effort. Dashed curve shows the distribution of phenological event dates for a population in a past year. The solid curves show two scenarios in which the mean date of the population occurs earlier now than it did in the past: the population size has either (a) increased or (b) declined. Arrows highlight changes in peak and first dates over time. Increases or decreases in sampling effort could similarly affect observed dates. Figure is not drawn to scale and is conceptual only. Miller-Rushing et al. (2008b) found empirical evidence for this phenomenon in bird populations in eastern Massachusetts, USA