Figure legends
Figure 1 Paired day of year observations and change in presence for select species over the study period (1992-2018). For each species, the top plot graphs observations by the year and numeric day of year of every sighting throughout the study period, slightly jittered for visibility; The upper line represents the end of flight (estimated as the 0.9 quantile day of year observation) and the bottom line represents the onset of flight (estimated as the 0.1 quantile day of year observation). For each species, the bottom plot describes the change in abundance over the study period; each point represents the total proportion of trips on which the species was present for each year, the slopes of these plots reflect trends in relative abundance incorporated into the list length model. Axis ticks for all plots represent the years 2000 and 2010.
Figure 2. Correlations between annual changes in relative abundance and 0.1 quantile (a), 0.9 quantile (b), flight period (c). Each point (a, b & c) represents a single species, shaped by voltinism and colored by range type. Y axis values represent the annual change in relative abundance estimated from list length analysis. (a) change in start of flight is measured as the slope of the 0.1 quantile (advances or delays); (b) change in the end of flight is measured as the slope of the 0.9 quantile (advances or delays); (c) change in total flight period is measured as the difference between the 0.9 and 0.1 quantiles (elongations or contractions).
Figure 3. Structural equations models. Lines indicate the direction of causality. Listed adjacent to lines are the scaled regression coefficients, the relative magnitude of the effects with respect to one changing standard deviation of the predictor. The a priori model is represented by solid lines, and post hocmodifications are represented by dashed lines.