Systematic data for phenological analyses and population trends
To analyze phenological change in adult flight periods of Lepidoptera, we use c. 20 years of systematically collected data from butterfly transects and moth traps. For butterflies, we combined data from two butterfly surveys conducted in agricultural landscapes in Finland during 1999-2017 (Fig. 2): 1) volunteer transects with varying length and sampling frequency, and 2) professional transects with standardized length and sampling frequency. For moths, we use data that have been collected in the Finnish national moth monitoring scheme (Nocturna) during 1993-2017 (Leinonen et al. 2016, 2017).
The raw data and how they were refined to fit the purposes of phenological analyses and population trend estimation in this study are described more in depth in Text S1. For example, to ensure enough phenological resolution, we removed all sampling events in the moth data with a sampling interval longer than ten days. All the observation dates for butterflies and moths were converted into Julian dates (day of year), to represent the timing of adult flight. To ensure sufficient data for each analyzed species per year, we included only species with observations from at least ten years during the 19- (butterflies) and 25- (moths) year study periods and with at least 30 observations from each of the included years. This procedure left us with a total of 289 species. As a consequence of this conservative inclusion of data, our set of study species does not contain rare species. Lastly, to exclude the effect of potentially differing phenological timing in recently colonized areas, we also excluded records of species from sites located further north of the estimated NRB for T1 (see section below and Text S2). The final data for phenological analyses and population trend estimations consisted of 725 106 and 355 970 abundance observations of the 289 species, respectively.