ABSTRACTTree growth–survival relationships link two demographic processes that individually dictate the composition, structure and functioning of forest ecosystems. While these relationships vary intra‐specifically, it remains unclear how this reflects environmental variation and disturbance. We examined the influence of a 700‐m elevation gradient and an Mw = 6.7 earthquake on intra‐specific variability in growth–survival relationships. We expected that survival models that incorporated recent growth would be better supported than those only using other factors known to influence tree survival. We used a permanent plot network that representatively sampled a monodominant Nothofagus forest in New Zealand's Southern Alps in 1974 and that was remeasured seven times through to 2009. The relationships were assessed using pre‐earthquake growth and survival, pre‐earthquake growth and post‐earthquake survival (0–5 years post‐earthquake), and post‐earthquake growth and survival (5+ years post‐earthquake). Survival was related to growth of 4504 trees on 216 plots using Bayesian modelling. We hypothesised there would be a positive, logistic relationship between growth and survival. Pre‐earthquake, we found a positive, logarithmic growth–survival relationship at all elevations. At higher elevations, trees grew more slowly but had higher survival than trees at lower elevations, supporting our hypothesised demographic trade‐off with elevation. The earthquake altered growth–survival relationships from those found pre‐earthquake and 0–5 years post‐earthquake survival held little relationship with growth. A strong, logarithmic growth–survival relationship developed 5+ years post‐earthquake because of enhanced survival of fast‐growing trees yet low survival of slow‐growing trees. Synthesis. Our findings demonstrate a trend in growth–survival relationships along an elevation gradient. If we assume a gradual climate warming is the equivalent of a forest stand shifting to a lower elevation, then data from our pre‐earthquake period suggest that tree growth–survival relationships at any elevation could adjust to faster growth and lower survival. We also show how these novel growth–survival relationships could be altered by periodic disturbance.