Work

Deriving exploratory temperature-based phenological indicators under data-limited conditions: integrating ERA5 and citizen science

Year: 2025

Type: article

Abstract: Understanding phenological responses to temperature variability in tropical and subtropical systems remains a major research gap, hindered by the scarcity of long-term observational datasets. We present an exploratory workflow that integrates temperature indicators derived from ERA5-Land reanalysis with opportunistic flowering records from citizen science to predict the onset of flowering in Crateva religiosa, a geographically widespread yet understudied tropical-subtropical deciduous tree species. Using partial least squares (PLS) regression on 93 opportunistic citizen science records, we identified potential thermal indicators, including low-percentile temperatures and long-window variability metrics, that together explained 71.18% of the variation in flowering date (RMSE = 12.13 days). Our findings highlight the potential thermal sensitivity signals of C. religiosa and demonstrate that biologically informed indicators can extract potential predictive value from coarse-resolution and noisy datasets. This approach illustrates a pragmatic workflow for deriving exploratory phenological variables in data-limited regions. It may serve as a valuable entry point for biodiversity monitoring, climate impact assessments, and early-stage evaluation of phenological sensitivity in understudied non-model species.

Source: Ecological Indicators

DOI: 10.1016/j.ecolind.2025.114484