The distributions of ectothermic marine organisms are limited to temperature ranges and oxygen conditions which support aerobic respiration, quantified within the Metabolic Index (MI) as the ratio of oxygen supply to metabolic oxygen demand. However, the utility of MI at local scales and across heterogeneous environments is unknown, yet these scales are often where actionable management decisions are made. Here we test if MI can delimit the entire distribution of marine organisms at local scales (10 km) using the endemic reef fish, Chrysoblephus laticeps, which is found in the highly heterogeneous temperature and oxygen environment along the South African coastal zone. In laboratory experiments we find a bi-directional (at 12 C) hypoxia tolerance response across the temperature range tested (8 to 24 C), permitting a piecewise calibration of MI. We then project this calibrated MI model through temperature and oxygen data from a high spatial resolution ocean model to quantify various magnitudes of MI across space and time paired with complementary C .laticeps occurrence points. Using random forest species distribution models, we quantify a critical MI value of 2.78 below which C. laticeps does not persist and predict current and future distributions of C. laticeps in line with already observed distribution shifts. Overall, we find that C. laticeps’ distribution is limited by increasing temperatures towards its warm edge but by low oxygen availability towards its cool edge, which is captured within MI at fine scales and across heterogeneous oxygen and temperature combinations – supporting MI’s application to make local-scale predictions for local management solutions.