Morphological disparity and taxonomic diversity are distinct measures of biodiversity, typically expected to evolve synergistically. However, evidence from mass extinctions indicates they can be decoupled. Mass extinctions lead to a drastic loss of diversity, but their impact on disparity remains unclear. Here, we evaluate the dynamics of morphological disparity and extinction selectivity across the Permian-Triassic mass extinction (PTME). We developed a new approach based on the extraction of morphological features from fossil images using a deep learning model (DeepMorph) and applied it to a high-resolution temporal dataset encompassing 599 genera across six marine clades. While five clades experienced strong taxonomic losses, we detected different patterns and magnitudes of morphological change during the PTME, including evidence of selective extinction of specialized forms in some clades. The presence and intensity of morphological selectivity likely reflect the variations in environmental tolerance thresholds among different clades. In clades affected by selective extinctions, the intensity of diversity loss promoted the loss of morphological disparity. Conversely, in clades experiencing non-selective extinctions, the magnitude of diversity loss had a negligible impact on disparity. Besides, We conducted a further study on the recovery processes of three groups affected by selective extinction during the Olenekian to Middle Triassic, encompassing 654 genera. We find that morphological recovery patterns may constrain the potential for taxonomic diversity to rebound. Our results highlight that the PTME had heterogeneous morphological selective impacts across clades, offering new insights into how mass extinctions can reshape biodiversity and ecosystem structure.

Morphological evolution, Disparity, Deep learning, Selectivity, Permian-Triassic mass extinction