Agroecology and regenerative farming

It is more honest to start this subject with the definition, because there isn’t one. In most corporate and certification documents “regenerative agriculture” is specified as a list of practices — reduced or zero tillage, cover crops, residue retention, extended rotations, livestock integration — and sold as an outcome: restored soil health, higher carbon stock, drought resilience. Practice and outcome are not tightly coupled. The same operation on a temperate loam and on a dryland sand produces different, sometimes opposite, results. While the definition is practice-based and the claim is outcome-based, the question “does regenerative farming work?” has no resolvable form: it compares incomparable bundles on incomparable soils.

What actually happens in the soil#

The primary mechanism is structure. Soil aggregates are bound by microbially derived polysaccharides, fungal hyphae and root exudates; tillage mechanically breaks those bonds and exposes physically protected organic matter to mineralisation. Continuous cover and living roots maintain aggregation, and a stable aggregate is above all a preserved pore network. Hence the most reproducible effect of the practice bundle: higher infiltration rate and water-holding capacity, less runoff and erosion. This is measurable faster and more reliably than carbon, and on many sites it is the water response rather than the carbon response that carries the economics — a field that accepts a downpour instead of shedding it enters the dry phase of the season with its own stored moisture.

The second mechanism is nitrogen. Legume cover crops fix atmospheric nitrogen in symbiosis with rhizobia; non-legumes scavenge residual nitrate that would otherwise leave in drainage and return it to organic form. But that nitrogen is released as residues mineralise, which is to say on temperature and moisture rather than on crop phenology. Synchronising supply with demand is an unsolved agronomic problem, and it is why a transition rarely permits an immediate full withdrawal of mineral fertiliser.

The yield gap#

The evidence base is heterogeneous but not empty. A global meta-analysis of no-till (Pittelkow et al., Nature, 2015), covering more than 5,000 paired observations, found an average yield decline of about 6% where no-till was adopted in isolation; combined with residue retention and rotation the penalty nearly disappeared, and in dry rainfed systems yields rose by roughly 7%. The conclusion is neither “it works” nor “it doesn’t”, but “it works as a system and fails as a single practice”. For organic farming — an adjacent but distinct category — meta-analyses report a persistent gap on the order of 19–25%.

The practical implication for anyone assessing a specific farm is to ask not about the label but about three measurable quantities: infiltration rate, temporal soil cover, and the nitrogen balance. They can be checked in the field, and they answer the question the practices are adopted for. The word “regenerative” cannot be checked; it describes an intention, not a state of the soil.