Reading marine diets with stable isotopes: what the tissue records

A chemical ratio carries ecological context

Marine ecologists measure stable carbon and nitrogen isotope ratios in animal tissues, commonly written δ¹³C and δ¹⁵N and reported in per mille (‰) relative to standards. The notation shows how a sample’s heavy-to-light isotope ratio differs from a reference; it is not a percentage of the animal’s carbon or nitrogen. Because tissues incorporate material from food, they preserve evidence about feeding and habitat that can complement direct observation. [1]

The signals answer different questions but are not unique labels. Carbon can distinguish energy pathways or habitats when sources have different signatures. Nitrogen often rises between food and consumer, so it can help estimate relative trophic position. Both also depend on the local baseline and transformations during growth and metabolism. A point on a graph is evidence to interpret, not a species-level diet identification. [1]

Start with the baseline, not the predator alone

A consumer’s value is meaningful only against plausible sources from the same food web. A coastal predator feeding on benthic prey and one feeding on pelagic prey may inherit different carbon signals because the underlying production pathways differ. Nitrogen comparisons also need a reference organism or resource whose trophic position is understood. If the baseline changes across locations or seasons, a predator’s isotope shift might reflect a different starting food-web signature rather than a change in what it ate. [1]

A useful design measures candidate prey and baseline organisms alongside consumers, and records where and when each sample was collected. Samples should match the study’s scale: prey values from a distant region or another season may poorly represent food available to the predator. In a North Pacific reference compilation, the authors gathered published values across whales, seals, sea otters, fishes, sharks and rays, and cephalopods, but described the overview as comprehensive rather than exhaustive. Such compilations can guide comparisons; they cannot supply every local baseline or erase the need to inspect original sample context. [1]

For more field-method readings and research explanations, browse the library.

Tissue is a clock as well as a sample

Different tissues incorporate dietary material at different rates. A fast-turnover tissue can reflect a relatively recent period, while a slow-growing structure may preserve a longer history. The North Pacific compilation notes that blood can turn over on the scale of weeks, whereas bone may integrate values over years. Comparing a predator’s bone with a prey’s blood as if they represented the same time window can create a mismatch even if both laboratory measurements are accurate. Tissue type belongs in the interpretation and in the caption of a comparison. [1]

Tissue also affects the offset between diet and consumer, known as a trophic discrimination factor. Rule-of-thumb factors of about 3.4‰ for nitrogen and 1‰ for carbon per trophic transfer are not universal. A 2023 vertebrate meta-analysis found that discrimination varies with tissue, trophic level and diet source. Use a factor measured for a similar system where possible; otherwise show how the conclusion changes under reasonable alternative assumptions instead of presenting one calculated trophic level as exact. [2]

A mixing model estimates contributions under assumptions

When researchers have several candidate food sources, isotope mixing models estimate how combinations of those sources could produce the consumer’s measured values. They compare plausible diets and quantify uncertainty, but depend on which sources and discrimination factors are included. Isotopically similar prey may be indistinguishable. If an important source is missing, a model may assign its signal to the available alternatives. The result is conditional on design, not an observation of meals. [2]

Consider a hypothetical coastal fish whose tissue values fall between sampled kelp-associated and plankton-based prey. A mixing model might estimate contributions from both pathways, but the estimate is only informative if the sampled prey represent what that fish could consume, the tissues and collection periods are comparable, and plausible discrimination factors are included. If the model’s range remains broad, that uncertainty is part of the ecological result: the measurements do not resolve a single precise menu.

Prey signatures can vary without a diet switch

A consumer’s isotope values can shift because a prey species changes its own isotope signature with body size, place or ocean conditions. A NOAA-hosted study of Antarctic krill found variation associated with krill size and chlorophyll conditions. Its modeling showed that this variation could shift predator values by as much as 2.4‰ even without a change in the species composition of the diet. That result is specific to the krill system studied, but it demonstrates a general interpretive problem: a stable-isotope difference does not automatically mean an animal switched prey. [3]

Researchers can reduce ambiguity by sampling prey across relevant places, seasons and size classes and pairing isotope data with other evidence. Stomach contents identify recent ingestion; cameras document behavior; fatty-acid or molecular methods add other dietary evidence. Each has its own time window and biases, so agreement across methods is more persuasive than one measure alone.

A careful reading keeps the conclusion proportional

When reading an isotope study, check the baseline organisms and area, consumer and prey tissue types, sample dates and locations, and model assumptions. Then ask whether the conclusion concerns broad trophic position, relative habitat use, or exact prey proportions. These are different levels of inference; exact proportions need especially strong sampling and model support.

Stable isotopes integrate feeding information that can be difficult to observe directly, especially for mobile or deep-diving animals. Their strength is comparison across appropriately sampled organisms and food webs. A tissue value also combines processes and time rather than naming a meal. Read each ratio with its baseline, tissue and uncertainty, and it becomes useful ecological evidence rather than a shortcut to a diet label. [1] [2]