IntroductionLocal biodiversity is increasing in many temperate and subtropical waters due to climate change. It is often caused by shifting fish distributions, thus the biodiversity gradient, from lower to higher latitudes. However, these shifts in distributions do not occur uniformly across all species. Consequently, communities are not only shifting their spatial distributions, but species compositions are also changing. We investigated spatiotemporal differences in the compositions of fish species in the bays of the northeastern Gulf of Mexico and identified species that contribute to the temporal changes.MethodsWe used fish count data collected using gillnets in eight major bays, encompassing over 600 km of coastline, during spring and fall seasons from 1982 to 2019. The nonmetric multidimensional scaling (NMDS) on the Bray-Curtis dissimilarity index among species composition vectors was used to detect the differences in species composition, and the similarity percentages (SIMPER) were used to determine the contribution of species to the differences. ResultsThe result shows there was a gradual change in species composition in all bays over the years, and the composition was different among bays and seasons. The species contributing to the temporal changes included those that are expanding (e.g., Common snook, Centropomus undecimalis, and Smallscale fat snook, C. parallelus) as well as retracting (e.g., Southern flounder, Paralichthys lethostigma, and Spanish mackerel, Scomberomorus maculatus) their distributions toward the north. The species observed only in recent years in these bays tended to have a preference for warmer water (e.g., Gulf pipefish, Syngnathus scovelli, and Chain pipefish, S. louisiana). DiscussionThe results are consistent with the potential effects of climate change. However, the salinity of the bays in the study area generally exhibits an increasing trend from the northern to southern bays. The spatial salinity gradient has a substantial impact on species compositions, indicating that species distributions are influenced by multiple environmental conditions. This complexity makes our ability to accurately predict future species compositions under changing environmental conditions challenging.