1. The mean and variance of ecological variables are dependent on sampling attributes such as the coverage of environmental heterogeneity (sampling extent) and spatial scale. Trophic niche width is often approximated by bulk tissue stable isotopes of C and N, i.e. the population isotopic niche. However, recent studies suggest that environmental heterogeneity (experienced by individuals) may be more important in defining the isotopic niche width than trophic variability. We hypothesised that isotopic niche width will increase monotonically with spatial scale, largely produced by environmental variation, e.g. nutrient source. 2. To refine this hypothesis, by describing the shapes of isotope scaling curves, we explored a previously published dataset describing three Chilean intertidal species representing different feeding guilds (grazing snails, suspension feeding mussel). We tested these hypotheses on a new, larger dataset describing three functionally-analogous intertidal species from Northern Ireland. We generated isotopic variance-area curves from a spatially-explicit bootstrap and investigated the scale-dependency of environment-isotope relationships, including wave exposure and sub-habitat heterogeneity. 3. Spatial scale explained 50% of the variance in population isotopic niche widths (bivariate C-N ellipse area) by simple, non-linear relationships. Finer scales (< 1 to 10 km lag) accounted for most variance. Scale dependence was strong for ẟ15N variance, of which > 40% was explained by modelling linear coefficients. A ẟ15N baseline gradient, or isoscape, dominated ẟ15N variance scaling patterns, from sheltered, terrestrially-influenced embayments to exposed, pelagic-dominated coastline. Consumer ẟ13C variance had a weaker scale-dependence, plateauing at mesoscales (> 20 km lag). 4. We show that isotopic niche width is strongly dependent on sampling spatial extent, which controls the environmental heterogeneity experienced by individual consumers. Environmental heterogeneity must be accounted for before isotopic niche width can be considered to accurately represent trophic niche width. Studies conducted at different spatial scales are likely to identify different environment-isotope relationships. 5. We recommend that spatial scale should be incorporated into sampling designs explicitly, easiest by maintaining a consistent lag distance or area within which populations are sampled. Identified isoscapes can be de-trended, where necessary.