To help address the underrepresentation of arthropods and Asian biodiversity from climate-change assessments, we carried out year-long, weekly sampling campaigns with Malaise traps at different elevations and latitudes in Gaoligongshan National Park in southwestern China. From these 623 samples, we barcoded 10,524 beetles and compared scenarios of climate-change-induced biodiversity loss, by designating seasonal, elevational, and latitudinal subsets of beetles as communities that plausibly could go extinct as a group, which we call “loss sets”. The availability of a published mitochondrial-genome-based phylogeny of the Coleoptera allowed us to compare the loss of species diversity with and without accounting for phylogenetic relatedness. We hypothesised that phylogenetic relatedness would mitigate extinction, since the extinction of any loss set would result in the disappearance of all its species but only part of its evolutionary history, which is still extant in the remaining loss sets. We found different patterns of community clustering by season and latitude, depending on whether phylogenetic information was incorporated. However, accounting for phylogeny only slightly mitigated the amount of biodiversity loss under climate change scenarios, against our expectations: there is no phylogenetic “escape clause” for biodiversity conservation. We achieve the same results whether phylogenetic information was derived from the mitogenome phylogeny or from a de novo barcode-gene tree. We encourage interested researchers to use this data set to study lineage-specific community assembly patterns in conjunction with life-history traits and environmental covariates.