SOC-MRV-UncertaintyCredit

About Dataset

SOC-MRV-UncertaintyCredit simulates 150 matched treatment/control field pairs under the full published 5-pool Rothamsted Carbon Model structure (Decomposable Plant Material, Resistant Plant Material, Microbial Biomass, and Humified Organic Matter kept as genuinely separate pools with their own officially-published decomposition rate constants, plus Inert Organic Matter), stepped monthly with a simulated seasonal temperature/rainfall cycle -- matching RothC's native time resolution. Matched treatment/control pairs support a genuine difference-in-differences additionality estimate, the actual accounting standard real MRV protocols use, rather than crediting raw treatment-arm sequestration alone. Within-field spatial heterogeneity is modeled with a spatially correlated Gaussian random field across 10-12 soil-core sampling points per field, not independent per-core noise. Direct lab-measured soil cores are sparse (3 of 5 monitoring years), while remote-sensing proxy channels (SAR backscatter, vegetation indices) are available every year, mirroring the real cost structure of operational SOC-MRV. Even the "ground truth" lab measurements are not treated as error-free: each batch of cores is analyzed by one of four simulated laboratories with its own systematic bias and precision, grounded in real inter-laboratory ring-trial findings for soil-carbon measurement. The dataset's central analytical contribution is two distinct, sequential real-world discounts on carbon-credit economic value: a statistical measurement-uncertainty conservativeness deduction, and a separate permanence/reversal-risk buffer-pool contribution modeled on real current carbon-market policy. In this dataset's own generated data, treatment fields gain a mean of over 1 tonne of carbon per hectare over 5 years while control fields lose carbon, and the two compounding discounts retain only about 30% of the raw estimated credit value at standard statistical confidence.