Groundwater is a vital ecosystem of the global water cycle, hosting unique biodiversity and providing essential services to societies. Despite being the largest unfrozen freshwater resource, in a period of depletion by extraction and pollution, groundwater environments have been repeatedly overlooked in global biodiversity conservation agendas. Disregarding the importance of groundwater as an ecosystem ignores its critical role in preserving surface biomes. To foster timely global conservation of groundwater, we propose elevating the concept of keystone species into the realm of ecosystems, claiming groundwater as a keystone ecosystem that influences the integrity of many dependent ecosystems. Our global analysis shows that over half of land surface areas (52.6%) has a medium‐to‐high interaction with groundwater, reaching up to 74.9% when deserts and high mountains are excluded. We postulate that the intrinsic transboundary features of groundwater are critical for shifting perspectives towards more holistic approaches in aquatic ecology and beyond. Furthermore, we propose eight key themes to develop a science‐policy integrated groundwater conservation agenda. Given ecosystems above and below the ground intersect at many levels, considering groundwater as an essential component of planetary health is pivotal to reduce biodiversity loss and buffer against climate change.

The regressive evolution of eyes has long intrigued biologists yet the genetic underpinnings remain opaque. A system of discrete aquifers in arid Australia provides a powerful comparative means to explore trait regression at the genomic level. Multiple surface ancestors from two tribes of diving beetles (Dytiscidae) repeatedly invaded these calcrete aquifers and convergently evolved eye-less phenotypes. We use this system to assess transcription of opsin photoreceptor genes among the transcriptomes of two surface and three subterranean dytiscid species and test whether these genes have evolved under neutral predictions. Transcripts forUV,long-wavelengthandciliary-typeopsins were identified from the surface beetle transcriptomes. Two subterranean beetles showed parallel loss of all opsin transcription, as expected under ‘neutral’ regressive evolution. The third speciesLimbodessus palmulaoidesretained transcription of along-wavelength opsin(lwop) orthologue, albeit in an aphotic environment. Tests of selection onlwopindicated no significant differences between transcripts derived from surface and subterranean habitats, with strong evidence for purifying selection acting onL. palmulaoides lwop. Retention of sequence integrity and the lack of evidence for neutral evolution raise the question of whether we have identified a novel pleiotropic role forlwop, or an incipient phase of pseudogene development.

Recent investigation of one of the larger calcretes in the uppermost reaches of the Carey palaeochannel in the Yilgarn region of Western Australia revealed an unprecedented diversity of copepod crustaceans. Twenty-two different species and subspecies, from six copepod families, represent 70% of the previously recorded copepod α-diversity in the whole region, although the area investigated is less than 3% of its surface. The aims of this study were to explore the diversity of the harpacticoid genus Kinnecaris Jakobi, 1972 using both molecular and morphological methods, establish precise species boundaries, find their accurate area of distribution, reconstruct phylogenetic relationships, and explore colonisation pathways. To achieve this we sampled very intensively in the area, as well as in two neighbouring palaeochannels, analysing more than 700 samples from 230 different localities, half of which contained copepods. Seven species are described here as new, five of them from the Yeelirrie palaeochannel (K. esbe sp. nov., K. lined sp. nov., K. linel sp. nov., K. linesae sp. nov., and K. uranusi sp. nov.) and one each from two neighbouring palaeochannels (K. barrambie sp. nov. and K. lakewayi sp. nov.). Parastenocaris jane Karanovic, 2006 from the Pilbara region, along with a newly described third Australian parastenocaridid genus from the Yilgarn, were used as outgroups in our molecular analysis. The COI fragment was successfully PCR-amplified from 12 parastenocaridid specimens using a nested combination of primers. All analyses supported the presence of at least seven genetically divergent lineages, most supported with very high bootstrap values. Three genera formed three separate clades, and the average pairwise distances between Kinnecaris morpho-taxa were found to be very high (8.2-16.8 %), while the highest divergences within morpho-taxa were 0.3%. Some conflict between molecular phylogenies and morphological data was observed when it came to recognizing different groups of species. While morphology indicates that K. esbe, K. linel, and K. uranusi represent a group of very closely related species, supported by a number of synapomorphies, molecular analyses suggest that K. linel and K. uranusi are only remotely related. We argue in favor of morphological data, until more markers can be studied to try to resolve these differences. In Yeelirrie, morphological evidence would suggest a downstream colonisation history in the genus Kinnecaris, where the most plesiomorphic form (K. linesae) lives in the uppermost reaches of the palaeochannel, and the trend in the caudal rami elongation and denser somite ornamentation is obvious downstream the palaeochannel (K. uranusi, K. linel, and then K. esbe), with the only exception being K. lined, which probably represents an independent colonisation event. Parastenocarids are copepods of freshwater origin, and we argure that they can probably disperse downstream during periods of increased rainfall, evolving into separate species in isolated calcrete pockets during periods of increased aridity. Although some of the questions remained unanswered in this study, detailed morphological and molecular observations indicate that we are not dealing with one widely distributed and variable species in the Yilgarn region, but rather with a complex of short range endemics. Areas of distribution for different species range from 30 km to less than 5 km in diameter. Very strong seasonal dynamics in this subterranean community was observed, and this is a novel concept for these ecosystems globally. A key to nine Australian species of Kinnecaris is also included.