Carbon dioxide removal (CDR) technology that combines negative greenhouse gas emission pathways alongside substantial emissions reductions is needed to limit climate change. Rock dust amendments to soils is an emerging CDR technology that has the potential to decrease atmospheric CO₂ via enhanced rock weathering (ERW), although there is considerable uncertainty regarding the efficacy of ERW with few multi-year field studies conducted to date. We explored combining finely-ground metabasaltic rock dust with other soil CDR technologies (compost, biochar) to stimulate carbon (C) sequestration and support plant growth. A combined amendment of  ground rock, compost and biochar observed the highest increases in soil C stocks (15.3 ± 6.1 Mg C ha^-1) over three years. All treatments with compost applied alone or in combination increased soil C stocks by 2.0 ± 2.9 to 15.3 ± 6.1 Mg C ha^-1.Ground rock only reduced rates of soil C decreases relative to the control, although still soil C stocks still decreased over time. . Continuous greenhouse gas fluxes showed that ground rock amendments lowered nitrous oxide (N₂O) emissions up to 11.0 ± 0.6 kg CO₂e ha^-1 yr^-1 and increased methane (CH₄) consumption by -9.5 ± 3.5 to -18.4 ± 4.4 kg CO₂e ha^-1 yr^-1 (N₂O and CH₄ reduction after three years: -0.06 ± 0.01 Mg CO₂e ha^-1). Combined compost, ground rock, and biochar yielded the highest estimated net ecosystem benefit of -86.0 ± 7.2 Mg CO₂e ha^-1 after three years. Net benefits (sum of relative changes in soil C stocks, estimated enhanced weathering rates, and annual greenhouse gas emissions) across all amendment treatments included slowing or reversing background losses in soil C stocks (-16.3 ± 5.0 to -50.5 ± 5.3 Mg CO₂e ha^-1). Benefits were dominated by increases in soil organic C stocks, both directly from organic matter amendments and likely increases in belowground plant C allocation. To determine an initial theoretical potential, we scaled to 8% of California rangelands and found supplementing enhanced weathering with compost and biochar could increase the net CO₂e benefit by 43 Mt CO₂e y^-1 relative to enhanced weathering alone. The use of mixed inorganic and organic amendments significantly increased estimated weathering rates and provided additional soil organic C sequestration.

Site description

The field site was located at the University of California Sierra Foothill Research and Extension Center (SFREC) in Browns Valley, California. Soils are derived from Mesozoic and Franciscan volcanic rock and classified as xeric Inceptisols and Alfisols in the Auburn-Sobrante complex (Soil Survey Staff, 2020). Soil depth is approximately 30 cm overlaying unconsolidated rock. The area is classified as oak woodland and rangeland and has been exclusively grazed by cattle since at least 1960. Recent stocking rates were approximately 8 acres per cow. The field sites were not otherwise seeded, irrigated, fertilized, or tilled. The study region has a Mediterranean climate with average annual precipitation of 700 mm with high interannual variability and a mean annual temperature of 16.6 ºC (California Irrigation Management Information System, 2023), but experienced an extended drought throughout the experiment (Liu et al., 2022). The growing season occurs during the cool, wet weather conditions, typically from October to May while summers are hot and dry. The site was dominated by naturalized stands of annual grasses (Ryals et al., 2016).

Experimental design

A control treatment and five combinations of ground rock (GR), compost (CP), and biochar (BC) soil amendments were applied in a randomized block design to account for background spatial variability of the site conditions. This included six 15 x 60 m plots randomly located in each of three blocks, with each block containing a control plot and plots amended with GR or CP only, ground rock plus compost (GR+CP); ground rock plus compost plus biochar (GR+CP+BC) was added in year 2 (Table 1). To measure continuous soil greenhouse gas emissions, a fourth block was established with nine, with three 10 x 10 m plot replicates of control, GR only, and CP only treatments (Figure S1).

Treatments were applied individually and were applied to the soil surface with soil incorporation or tilling. Ground rock for enhanced silicate weathering was finely ground metabasaltic rock sourced as waste material from a regional mining operation (Specialty Granules, Ione, CA). A summary of relevant soil amendment properties, chemistry, total C and N content, and mineralogy are provided in Tables S1-S3. Compost amendments were derived from a mixture of organic yard debris (garden trimmings and wood chips) and cow, chicken, and horse manures and were commercially produced at the West Marin Compost Facility (Nicasio, California) by maturing in watered piles with weekly aeration for approximately three months (Vergara & Silver, 2019). Biochar amendments were purchased commercially (Rogue Biochar, Oregon Biochar Solutions, White City, OR) and made with a feedstock of Douglas fir and pine woody biomass sourced from local logging and mill residues. Compost and biochar amendments were applied once during the study following typical rancher practice in the region, and ground rock was applied annually in 2019, 2020, and 2021 (Table 1). We note that there were no field-scale data available for ground rock amendments to rangelands, so we adopted suggested best practices for cropland applications (Andrews & Taylor, 2019; Larkin et al., 2022)^,. All amendments were applied before the first rainfall in the fall prior to onset of the growing season. 

Soil and aboveground biomass analysis

To quantify changes in total soil C and N content, samples were collected prior to amendment application and annually at the end of each of three growing seasons. Soil was collected from five random locations within each plot using hand augers (6 cm diameter). Soil was collected from 0-10 cm, 10-20 cm, and 20-30 cm depths. Samples were air-dried, sieved to < 2 mm, visible roots removed, and a subsample ground to a fine powder. Samples were then analyzed in duplicate for total C and N on a elemental analyzer (CE Elantech, Inc., Lakewood, New Jersey). To quantify changes in soil inorganic C, air-dried soil samples were analyzed on an soliTOC cube (Elementar, Ronkonkoma, NY) using temperature ramping method DIN 19539 which can separately calculate total organic C (TOC), residual organic C (ROC), and total inorganic C (TIC).

Subsamples were analyzed for soil pH, gravimetric soil moisture, and extracted with potassium chloride (KCl) to quantify inorganic N pools within 24 h of sample collection. Soil pH was determined by vortexing a 1:1 soil to water solution (5 g soil to 5 ml deionized water) for 1 minute then measuring the solution pH after 10 minutes (McLean, 1982). Soil moisture was determined gravimetrically by weighing fresh soil, oven drying for 24 hours at 105 °C, reweighing the dried soil, and calculating the difference as percent soil moisture.

Nitrate (NO₃^-) plus nitrite (NO₂^-) and ammonium (NH₄⁺) were measured after extraction of 15 g of field-fresh soil in 75 ml of 2 M KCl solution³⁶. Soil KCl extracts were stored at -20 ^ºC until colorimetrically analyzed using a discrete analyzer (Model AQ300, Seal Instruments, Mequon, WI). Bulk density was sampled at 0-10, 10-20, and 20-30 cm in each plot prior to amendment application and after three growing seasons using a 6.35 cm diameter bulk density corer (Table S9). Subsamples were dried at 105 ^ºC to a constant weight. Bulk density was calculated as the rock-free dry volume of the total soil core.

Aboveground vegetation was sampled at peak standing crop for biomass production at the end of each growing season (n = 9 per plot). In annual grasslands, peak standing crop is generally considered an index of net primary productivity (Chiariello, 1989). Live aboveground plant tissue was clipped from randomly located 20 cm diameter rings and collected into pre-dried and weighed paper bags. Samples were dried at 65 ºC and subsequently weighed for biomass.  Aboveground biomass C content was assumed to be 50% of the oven dried mass.

Soil pore water analyses

In Fall 2021, soil lysimeters (Soil Water Sampler, Soil Moisture Corp., Santa Barbara, CA) were installed at 30 cm depth to sample soil pore water following rain events. Lysimeter sampling was attempted approximately 24 h after each rain event, but sample collection was not always possible due to low solute volumes. Sample collection occurred from November 2021 following the first rain event through January 2022 until collection was impossible due to decreases in soil moisture content. Following sample collection, dissolved organic carbon (DOC) and dissolved inorganic carbon (DIC) were analyzed with a combustion oxidation nondispersive infrared absorption method on a total dissolved C analyzer (varioTOC, Elementar, Hanau, Germany), with DIC calculated as the difference between acidified and non-acidified samples. Sample acidification to pH < 2 was performed by adding a small amount of pure hydrochloric acid (Tisserand et al., 2024). To quantify another metric of silicate weathering, soil lysimeter subsamples were also analyzed for calcium (Ca), magnesium (Mg), and sodium (Na) in triplicate via inductively coupled plasma optical emission spectroscopy (ICP-OES; Model Optima 5300 DV, Perkin Elmer, Waltham, MA).

Automated soil and greenhouse gas flux measurements 

Surface fluxes of N₂O, CH₄, and CO₂ were measured continuously during three complete growing seasons (October to June) from 2019 to 2022 using an automated chamber system. The system consisted of nine opaque, automated gas flux chambers (eosAC, Eosense, Nova Scotia, Canada) plumbed to a multiplexer (eosMX, Eosense, Nova Scotia, Canada). Each chamber was randomly deployed in the 10 x 10 m replicate, and plants were included in the soil collars and only trimmed to ensure complete chamber closure. The multiplexer allowed for dynamically signaled chamber deployment and routed gases to a cavity ring-down spectrometer (Model G2508, Picarro Inc., Santa Clara, CA, USA). Chambers were measured sequentially approximately every two hours, with  a 10 min sampling period with a 1.5 min flushing period before and after each measurement. Chamber volumes were used to calculate the minimum detectable flux of 0.004 nmol N₂O m⁻² s⁻¹, 0.02 nmol CO₂ m⁻² s⁻¹, and 0.004 nmol CH₄ m⁻² s⁻¹ (Nickerson, 2016).

Flux calculations and analyses were first performed using Eosense eosAnalyze-AC v. 3.7.7 software, then data quality assessment and control were subsequently performed in R (RStudio, v.1.1.4633). Fluxes were removed from the final dataset if they were associated with erroneous spectrometer cavity temperature or pressure readings or if any gas concentrations were negative, corresponding to instrument malfunction. Fluxes were also removed if the chamber deployment period was less than 9 min or greater than 11 min, indicative of chamber malfunction. Calculated linear and exponential fluxes were compared using estimate uncertainty to estimate ratios, and in cases where both the linear and exponential models produced high uncertainty, the individual flux was eliminated from the dataset. Data filtering removed 8.7% of flux measurement periods, generating a final dataset of 47,762 simultaneous flux measurements of CO₂, N₂O, and CH₄. Daily mean flux values across years and treatments are presented in Figure S1. To convert flux measurements to CO₂e, we used the IPCC AR5 100-year GWP values of 28 CO₂e for CH₄ and 298 CO₂e for N₂O (Myhre et al., 2013).

Estimation of weathering rates and net ecosystem C benefits 

Bicarbonate production via enhanced weathering of basaltic minerals is largely dependent on Ca²⁺ and Mg²⁺ concentrations. As a conservative estimate to account for potential CO₂ production via carbonate (CaCO₃, MgCO₃) precipitation, we assume 1 mol of CO₂ is converted to bicarbonate (HCO₃^-) per Ca²⁺ and Mg²⁺ produced via weathering (Beerling et al. 2018). The weathered metabasaltic material in this study was applied annually for three years at a rate of 37.8 Mg ha^-1 yr^-1 and contained 4.25-6.46% Ca and 1.02-4.14% Mg by mass (Table S2). During the study period, annual rainfall in this location varied from 266.7 to 548.2 mm yr^-1 (1 mm is equivalent to 10000 L ha^-1), with 266.7 mm yr^-1 of rainfall during the lysimeter observation period(California Irrigation Management Information System, 2023). To calculate net rates of enhanced weathering, all rainfall was assumed to become soil water with similar DIC concentrations to our lysimeter observations. As this approach likely overestimates solute movement through the soil values should be treated with appropriate caution. We used the 2021-2022 value as an estimate for weathering rates during the other two growing seasons; since 2021-2022 was the lowest rainfall year of the study these should be considered conservative estimates.

We estimated the net ecosystem C benefit (reduction of CO₂ or CO₂e added or increased removal of CO₂ or CO₂e from the atmosphere) of each treatment after three growing seasons. This was done by combining annual changes in aboveground biomass, net changes in soil C stocks from the initial baseline, estimates of annual enhanced weathering rates, and measured or estimated changes in annual N₂O and CH₄ fluxes in CO₂e from the individual compost and ground rock treatment flux measurements. Changes in soil C stocks within blocks were calculated as the difference in soil C stocks at each depth (0-10 cm, 10-20 cm, and 20-30 cm) from prior to treatment application in Fall 2019 and the end of third growing season in Spring 2022 and converted to CO₂e values. As a first approximation of the changes in N₂O and CH₄ fluxes for combined treatments, we calculated the summed differences from both compost and/or ground rock only treatments relative to the control. It is important to recognize that these values may not capture the synergistic effects of combined amendments and are used here to capture likely order-of-magnitude scale impacts. Emissions associated with amendment production, transport, and application were not included, thus reported values should be considered net C benefit estimates from within ecosystem boundaries.

Calculations and statistical analyses

All statistical analyses were performed using JMP Pro 15 (SAS Institute Inc., Cary, NC). Differences in treatment year total C and inorganic soil C stocks, aboveground biomass, soil characteristics (soil moisture, soil pH, mineral N), lysimeter cation concentrations, and annual CO₂, CH₄, and N₂O fluxes were analyzed with one-way ANOVAs followed by post-hoc Tukey tests. Figures were created using ggplot2 in R (Wickham, 2009). Values reported in the text are means ± standard errors unless otherwise noted.