Ecologists have found the distributions of many groups of organisms to be elevationally stratified. Consequently, various taxa (or species) have been proposed as indicators for inclusion within long‐term monitoring programmes to quantify the ecological impacts of future climatic change. Ideal indicators should be restricted to a particular elevational range (i.e. have high specificity) and be readily detectable across space and time (i.e. have high fidelity). This, however, has not been rigorously tested for elevational studies. We employed a spatially and temporally replicated sampling design to test the utility of tree, ant, and canopy and understorey moth species as indicators of elevation within continuous subtropical rainforest of eastern Australia. Using the classical indicator value protocol, we tested (i) whether the number of indicator species (all taxa) found in the observed data was significantly greater than the number obtained by chance; (ii) whether the indicator species (ants and moths) identified from one sampling season responded to elevation in a similar way in samples obtained from other seasons; and (iii) whether the indicator species (ants) identified from one elevational transect responded to elevation in a similar way in a nearby transect that incorporated similar elevational ranges. All groups had significantly greater numbers of indicator species than expected by chance. Temporal fidelity of moth indicator species was lower than that of ants as the suite of moth indicator species showed high seasonal variation. In contrast, ants showed high spatial and temporal fidelity. Most ant indicator species were, however, indicative of low and mid‐elevations, and only one species was indicative of the highest elevation, suggesting their relatively low conservation significance in relation to climate warming in this region. It is essential that we understand how spatial and temporal variation affects the distributions of different taxonomic groups when incorporating multiple taxa for long‐term monitoring programmes.