Changes in suitability distribution
To assess the impact of climate change on the suitability of Bangladeshi butterflies, I followed four steps. First, I calculated the centroid (centre point (latitude and longitude)) and the area (summed the pixel value (1/0) and multiplied it with the pixel resolution (21.625 km2)) for each suitability map using the ‘raster’ R package (Hijmans et al., 2015). Second, I calculated the distance and angle between the centroids using the ‘geosphere’ R package (Hijmans, 2017). Third, for each species, I calculated the regions where the suitability range expanded, contracted, or remained the same (overlapped) using the ‘raster’ R package (Hijmans et al., 2015). I stacked all the rasters (species-wise) to determine the overall expanded, contracted, and overlapped regions. Finally, I calculated the potential change in realised niche distribution between the current and the future suitability distributions using the ‘dismo’ R package (Hijmans et al., 2017). For all these combinations, I compared the current suitability with the four future climatic scenarios (ssp126, ssp245, ssp370, and ssp585); however, for the clarity purpose, I only used the results of the ssp126 in the main manuscript, and the rest of the results are in the supplementary section. It should be noted that I only measured the climate exposure rather than a full measure of climatic vulnerability, given I did not account for species sensitivity and adaptability to climatic change.