A soil–feedstock greenhouse mesocosm matrix for screening carbon dioxide removal via enhanced rock weathering

Enhanced rock weathering (ERW) is a carbon dioxide removal (CDR) strategy in which crushed silicate and carbonate rocks are applied to soils to accelerate mineral weathering, releasing base cations and converting CO2 into dissolved bicarbonate. The ocean is conservatively regarded as the final reservoir for this inorganic carbon. However, mineral dissolution alone does not necessarily translate into CDR. Alkalinity may be neutralised by existing soil acidity, and weathering-derived cations may be retained on soil exchange sites, taken up by plants, incorporated into secondary minerals or leached with non-carbonate anions rather than bicarbonate. Realised CDR is therefore strongly soil-dependent. In order to test this, here we present the design and methodology of GH2025, a greenhouse mesocosm experiment configured to systematically explore soil–feedstock interactions across a broad parameter space. It follows the multi-year Carbon Drawdown Initiative trial GH2023, which established that soil processes strongly influence alkalinity export and thus CDR, with weathering-derived cations largely retained in exchangeable, carbonate, oxide and clay-mineral pools and strong between-combination variability. These processes complicate the quantification of CDR and the development of robust measurement, reporting and verification (MRV) approaches. Identifying the soil–feedstock combinations that deliver efficient and verifiable CDR therefore remains a central challenge for ERW deployment. GH2025 comprises 450 lysimeter pots spanning 29 natural, agricultural and engineered technosol soils and 19 feedstocks representing seven mineralogical and dissolution-kinetic categories, including industrial by-products, carbonates, carbonate–silicate mixes and various types of silicates. The design prioritises breadth over replication, using one pot per soil–feedstock combination and three soil-only controls per soil to maximise parameter-space coverage. Leachate is sampled every six to eight weeks and analysed for total alkalinity, dissolved inorganic carbon, electrical conductivity, pH and dissolved oxygen. Following dismantling, soils will be analysed for pH, effective cation exchange capacity, soil inorganic and organic carbon and operationally defined solid-phase cation pools. The resulting dataset will support machine-learning approaches to identify the soil and feedstock properties governing alkalinity export and CDR, and to predict favourable combinations for scalable deployment.