Life-Cycle Assessment (LCA) for Food

Life-cycle assessment turns a food claim into a bounded accounting question: what process, what product, what impact, what allocation rule, and what evidence?

If you have ever seen two honest-looking food footprint numbers disagree by half, you have met the reason life-cycle assessment matters. The number is not only about the crop. It is about the boundary around the crop: whether the calculation stops at the farm gate, includes processing and cold storage, counts land-use change, credits co-products, or treats soil carbon as temporary.

Life-cycle assessment (LCA) does not remove judgment. It makes the judgment visible.

Definition

Life-cycle assessment is the ISO 14040 and ISO 14044 method family for estimating the environmental burdens of a product, service, or production system across a defined life cycle. In food, the common boundary is cradle-to-gate: seed, fertilizer, energy, feed, machinery, field emissions, harvest, and delivery to the processor or farm gate. A wider cradle-to-grave study also counts processing, packaging, retail, cooking, waste, and disposal.

Every LCA has four parts. First, it defines the goal and scope: what question is being answered and for whom. Second, it builds a life-cycle inventory: all measured or modeled inputs and outputs. Third, it converts that inventory into impact categories such as climate change, eutrophication, acidification, land occupation, water use, particulate matter, or energy demand. Fourth, it interprets the result, including uncertainty and sensitivity.

The functional unit is the denominator. It may be one kilogram of tomatoes, one liter of milk, one serving, one calorie, one gram of protein, or one dollar of farm output. Change the functional unit and the ranking can change. A system that looks efficient per kilogram may look worse per gram of protein. A low-yield crop can look good per hectare and poor per unit of food.

Allocation is the other hinge. If a dairy farm produces milk, calves, cull cows, manure, and possibly energy from anaerobic digestion, the study has to decide how burdens are split. Economic allocation, mass allocation, energy allocation, and system expansion can all be defensible. They don’t give the same answer.

Why It Matters

Food-system arguments are full of footprint numbers. A buyer compares field lettuce with greenhouse lettuce. A vertical-farm investor asks whether shorter transport offsets purchased light. A beef program claims soil carbon changes the emissions balance. A brand chooses between glass, plastic, and refill packaging. A policy team compares methane, land use, water use, and nutrient runoff across supply chains.

Without LCA discipline, those conversations collapse into single-number advocacy. With it, the reader can ask sharper questions. What is the boundary? What functional unit? Which impact categories? Were field emissions measured or modeled? Was land-use change included? How was uncertainty treated? Did the study credit soil carbon, and if so, for how long?

For regenerative agriculture, LCA is useful because it keeps soil-health claims from floating outside the rest of the system. A cover-crop or grazing shift may improve soil organic carbon, infiltration, and resilience while changing fuel use, seed cost, yield, labor, feed demand, or manure emissions. The net result depends on which burdens are counted and how long the change persists. A Soil Carbon MRV Pipeline can verify a stock change; LCA asks what that change does to the product footprint.

For controlled-environment agriculture, LCA is often the first serious filter. A greenhouse may use less land and water than an open-field comparator while adding heating, cooling, screens, substrates, fertigation equipment, and packaging burdens. A vertical farm compresses land use further but buys nearly all of its light and climate control. If the electricity mix, crop yield, waste rate, and retail distance don’t support the claim, the footprint story won’t hold.

How It Shows Up

Poore and Nemecek’s global food dataset. The 2018 Science paper assembled a large farm-to-retail dataset across thousands of producers and showed why averages can mislead. Beef from one production system can carry a very different greenhouse-gas, land, and water profile from beef in another. The same spread appears inside crops, dairy, and aquaculture. One global table does not settle every decision. It shows how strongly production method, geography, yield, and supply-chain stage shape the result.

A greenhouse versus field lettuce comparison. A procurement team comparing field lettuce, heated greenhouse lettuce, and indoor vertical-farm greens needs more than water-use claims. The LCA has to count yield per square meter, electricity, heating fuel, nutrient losses, packaging, refrigeration, spoilage, transport distance, and the electricity grid. Field production uses more land and water per kilogram; a heated facility may spend the advantage in energy. The comparison is only useful when the same functional unit and boundary apply to all three.

A regenerative beef claim with soil carbon included. Some grazing LCAs change direction when modeled or measured soil carbon is counted. That does not make the carbon claim wrong, and it doesn’t make it automatically saleable. It means the result depends on the carbon method, baseline, depth, permanence period, animal performance, methane assumptions, and allocation to co-products. A reader should treat “carbon-negative beef” as a claim requiring a full method note, not as a slogan.

A USDA LCA Commons inventory. Public inventory datasets matter because they let different analysts start from common process data rather than private spreadsheets. Shared data does not make the result neutral by itself. Analysts still choose the boundary, region, functional unit, allocation rule, impact method, and scenario. But it lowers the fog around the calculation and makes review possible.

Caveats and Open Questions

LCA is not a moral ranking of foods. It is an accounting method tied to a stated question. A study designed to compare climate impacts per kilogram will not answer whether a crop improves local livelihoods, protects farmworker safety, builds regional food capacity, or supports biodiversity. Those may matter. They need their own evidence.

The method is also vulnerable to false precision. A result reported as 1.37 kilograms of carbon dioxide equivalent per kilogram of product may rest on generic fertilizer factors, modeled nitrous oxide, estimated yields, old electricity data, and assumed transport distances. The decimals can outpace the evidence. A credible study reports ranges, sensitivity, and the assumptions that move the result.

Soil carbon remains one of the hardest inputs. If a study credits regenerative management with carbon storage, it has to state the baseline, depth, sampling or modeling method, duration, reversal risk, and uncertainty. Short-term soil gains can be real and still not permanent enough to offset fossil energy use forever.

CEA studies have their own weak spots. Commercial vertical farms and greenhouses rarely publish full operating data. Researchers often work from modeled facilities, pilot data, or operator disclosures that can’t be independently audited. Energy mix, crop density, HVAC load, waste rate, and lighting efficacy move the result quickly. A 2026 comparison should not reuse a 2018 grid factor unless the location and power contract still match.

Finally, LCA can be misused as procurement theater. A buyer can choose a narrow boundary that favors the product already preferred, ignore inconvenient impact categories, or compare unlike functional units. The defense is not to reject LCA. The defense is to read the method.

Related Articles

Sources

• ISO 14040 and ISO 14044 define the life-cycle assessment principles, framework, requirements, and guidelines used by serious product-footprint studies.
• Poore and Nemecek’s 2018 Science article, “Reducing food’s environmental impacts through producers and consumers,” is the canonical large comparative food-system LCA dataset.
• USDA’s LCA Commons provides public life-cycle inventory data for agricultural and food-system processes in the United States.
• FAO’s State of Food and Agriculture 2023 and 2024 reports show how life-cycle evidence feeds into true-cost and hidden-cost accounting for agrifood systems.
• Barbosa and colleagues’ 2015 lettuce comparison illustrates how hydroponic and conventional production comparisons depend on land, water, energy, boundary, and yield assumptions.
• Stanley and colleagues’ Michigan beef-finishing LCA shows how soil-carbon assumptions can change the interpretation of grazing-system emissions.