The fix isn't "ban soil carbon credits" -- the France and pan-European feasibility studies confirm real sequestration is achievable in the right soil-climate-management combinations. The fix is forcing crediting math to track the actual decay curve instead of an assumed flat one, and building the re-measurement into the credit price rather than treating it as an optional add-on nobody wants to pay for.

Mechanism: require every soil-carbon crediting protocol seeking registry approval to file a plot-specific "decay-adjusted crediting curve" instead of a flat annual rate, built from three inputs already used piecemeal across the literature I cited: (1) baseline SOC stock and soil texture class (from existing soil surveys, which most regions already have at coarse resolution), (2) a saturation-timeframe estimate drawn from the nearest applicable published range (10-100 years, narrowed by climate zone) rather than an unstated house assumption, and (3) mandatory re-sampling checkpoints at years 3, 7, and 12 of any crediting period, with credits issued in tranches tied to verified measurement at each checkpoint rather than upfront based on year-one projections.

Where full soil-core re-sampling is too costly for smallholder programs, pool the cost: fund shared regional soil-testing labs (similar to agricultural extension co-ops) that serve multiple farms in a crediting program, cutting per-farm MRV cost through shared fixed infrastructure instead of shrinking sampling frequency.

Pair this with a "true-up ledger" requirement: if year-7 or year-12 measurement shows sequestration below the tranche already sold, the registry must publicly flag the shortfall against that specific credit batch, and buyers holding those credits get notified their claimed tonnage was revised downward -- turning "genuinely uncertain, default to optimism" into a policy that carries a visible cost when it turns out to be wrong, rather than a policy with no downside for the party who profited from the optimistic estimate.

For the farmer-side equity problem, tie program enrollment to a land-tenure and multi-year-commitment feasibility screen upfront (not to exclude the farmers who need the income most, but to route farmers with insecure tenure or thin capital toward shorter-duration, lower-crediting-period products, priced accordingly, instead of the same 15-20 year product sold to farmers who can actually sustain the management practices that long) and pair enrollment with the same shared regional testing infrastructure so smallholders aren't individually bearing MRV costs that make the program uneconomical for them specifically.