Enhancing natural cycles in agro-ecosystems to boost 1 plant carbon capture and soil storage

12 One of society’s greatest challenges is sequestering vast amounts of carbon to avoid 13 dangerous climate change without driving competition for land and resources. Here we 14 assess the potential of an integrated approach based on enhancement of natural 15 biogeochemical cycles in agro-ecosystems that stimulate carbon capture and storage 16 while increasing resilience and long-term productivity. The method integrates plant 17 photosynthesis in the form of (cover) crops and agroforestry which drives carbon 18 capture. Belowground plant-carbon is efficiently stored as stable soil organic carbon 19 (SOC). Aboveground crop and tree residues are pyrolyzed into biochar, which is 20 applied to the soil reducing carbon release through decomposition. Enhanced 21 weathering of basalt powder worked into the soil further captures and stores carbon, 22 while releasing nutrients and alkalinity. The integrated system is regenerative, through 23 enhanced virtuous cycles that lead to improved plant capture, biomass storage and crop 24 yield, the prerequisites for large-scale carbon sequestration along with food security.


26
Human-induced climate change has significant adverse impacts on our environment, 27 economy, and way of life. Reductions of carbon dioxide emissions alone are no longer 28 sufficient to avoid dangerous impacts 1,2 , and capture plus long-term storage of atmospheric 29 carbon will be required. 30 Large-scale carbon sequestration is possible through a range of options, each with its own 31 advantages and drawbacks 3-7 . One family of methods centres on enhancing natural 32 biogeochemical processes. These techniques also called nature-based solutions or geotherapy 33 have positive environmental impacts 3,4,7-9 , and could (partly) pay for themselves by 34 increasing natural capital and agricultural productivity 6 . Examples include (1) boosting the 35 growth and standing carbon stock in plants in cropping and pasture systems through cover-36 and inter-cropping (e.g., agroforestry); (2) re-establishing and/or enhancing soil organic 37 carbon (SOC) stocks 10 ; (3) production of biochar, which is plant biomass transformed at 38 elevated temperatures under oxygen-limited conditions (pyrolysis) into a recalcitrant form 39 that withstands decomposition for many decades/centuries to possibly even millennia 11 ; and 40 (4) increasing the inorganic carbon sink in soils via Mg and Ca silicate weathering by 41 working finely ground rock (basalt) into soils 12 . 42 The combined global carbon sequestration potential of such measures has been estimated at 43 0.3-6.8 Gt C yr -1 13 . The potential of each technique independently has been reported in Smith 44 et al. 13 , who compiled the full range of literature values: 45 (1) Agroforestry: ~0.03-1.55 Gt C yr -1 46 (2) Soil carbon sequestration (SOC): 0.14-1.36 Gt C yr -1 47 (3) Biochar: 0.01-1.80 Gt C yr -1 48 (4) Enhanced rock (basalt) weathering: 0.14-1.1 Gt C yr -1 . 49 3 Large-scale carbon sequestration is an enormous challenge in itself, and doing so without 50 competition for land and resources among different carbon sequestration techniques and with 51 food production is an even greater one [14][15][16] . Here we evaluate an integration of the 52 aforementioned land-based carbon sequestration techniques in agricultural systems on the 53 same land area (Figure 1). This avoids competition for land and resources among drawdown 54 methods, and further helps to build resilient and regenerative agro-ecosystems. Importantly, 55 we contend that interactions between methods and with soil processes can set up synergistic 56 virtuous cycles that further enhance the potential for carbon sequestration. This study aims to 57 (i) discuss the key limitations of individual carbon sequestration techniques by themselves, a 58 prerequisite to maximise their potential; (ii) assess interactions and synergies between the 59 techniques; and (iii) define conditions and strategies that allow for integration and large-scale 60 carbon sequestration in agro-ecosystems. 61 4 Cover crops (the establishment of plants for the purpose of protecting the soil) boost 74 aboveground carbon stocks throughout the year and can increase SOC stocks by 0.1-1 t ha -1 75 yr -1 20,21 . Plant and soil carbon storage increases with plant species-richness due to higher 76 niche partitioning, and thus nutrient and water use efficiencies [22][23][24] . Adding trees to 77 agricultural land and consequently conversion of crop-and grassland into agroforestry, a 78 form of inter-cropping (the integration of at least two plant species in the same area), can 79 increase aboveground biomass more than 10-fold and has been found to increase SOC stocks 80 by 25% and 19% globally, respectively 18,19 . Agroforestry operations can be established and 81 maintained at costs of USD 0.3-20 t -1 CO2 (median USD 2.5 t -1 CO2) 30-33 . 82 Soil organic carbon (SOC) 83 Microorganisms degrade plant carbon (respiring CO2), but also foster conversion into stable 84 forms of SOC 34,35 (Figure 2a). Both processes are affected by the activity, abundance, and 85 community composition of microorganisms and are soil dependent 36 . To achieve long-term 86 sequestration of plant-derived carbon, a simple increase in total SOC content is insufficient. 87 Instead, an increase is needed in persistent SOC stocks, through protection in soil 88 microaggregates (aggregate occlusion), and/or carbon-binding to clay and silt particles 89 (mineral-associated SOC/matrix stabilisation) 17,37 . Therefore, the soil needs to possess sink 90 strength in the form of available minerals or soil aggregation to build stable SOC ( Figure 2

108
During biochar production (pyrolysis), biomass is heated in the absence of oxygen, which 109 directly converts the atmospheric carbon that was captured by plants into a form that is stable 110 for centuries 11 (Figure 2b). The process results in an initial release of ~45% of the plant 111 carbon stored in agricultural and forestry residues (mean over different temperatures) 42 and, 112 hence, in greater carbon emissions in the first few years of biochar production, relative to 113 regular biomass decomposition (negative values in Figure 3). However, over subsequent 114 years, this is offset, as further decomposition emissions are avoided, and net carbon-negative 115 conditions develop. The mean residence time of biochar has been estimated at 500-1000 116 years, several orders of magnitudes greater than that of unpyrolysed biomass 43-45 . Assuming 117 a ~60-times lower degradation rate of biochar than unpyrolysed biomass 44 , biomass 118 pyrolysis becomes net carbon negative after ~3-5 years ( Figure 3). 119 Biochar use is limited by biomass feedstock availability and processing costs. For example, it 120 can be essential to leave crop residues in the field to reduce soil erosion and evaporative 121 losses in water-limited regions 46 . In other cases, some (bioenergy) crop and forestry residues 122 are well suited for biochar production 47 . Globally, wheat, for example, had annual grain 123 6 yields of 0.4-9.4 t ha -1 in 2019 (mean 3.3 t ha -1 ; range/mean of all countries listed in database) 124 48 . With a typical harvest index of 0.5 (50% of biomass in grain, 50% into stem and leaves) 49 , 125 0.4-9.4 t ha -1 of wheat straw residue is produced annually on-farm. Tree plantations can 126 produce ~10-100 t ha -1 of residue over a 30-40 year rotation, equivalent to 0.25-3.3 t ha -1 yr -1 127 50-52 . The biochar yield from woody and grass feedstocks is ~25% on average across different 128 pyrolysis temperatures 53 . Hence, pyrolysis of wheat straw and pine plantation residues 129 produces 0.1-2.4 and 0.06-0.8 t ha -1 yr -1 of biochar, respectively (mean 0.8 and 0.4 t ha -1 yr -1 ). 130 We thus infer that limited on-site availability of biomass residues in agriculture and 131 neighbouring forestry systems will initially enable biochar application rates of ~1 t ha -1 yr -1 , 132 which corresponds to 0.73 t C ha -1 yr -1 at a mean biochar carbon content of 73% 42 . To make 133 more accurate assessments, alternative uses of residues need to be considered locally and 134 biomass availability (e.g., forestry sites) mapped to biochar use (agricultural sites). Estimated 135 CO2 abatement costs using biochar from forestry and agricultural residues are USD 50-300 t -1 136 CO2 (median USD 130 t-1 CO2), which includes costs for feedstock (either collection and 137 transport costs on farm or commercial price), biochar production and application 47,54-56 . 138

Basalt weathering
139 Enhanced weathering is the acceleration of the natural process of rock dissolution by 140 crushing Mg-and Ca-rich silicate rocks before application to soil. During weathering, carbon 141 dioxide is captured and initially stored in the form of dissolved bicarbonate (HCO3 -). Further 142 reactions convert the bicarbonate into Ca and Mg carbonates, which deposit in the marine 143 environment where they remain sequestered for millennia 12 . Basalts are the preferred rock 144 types because they are rich in elements beneficial to plant growth (P and K) and contain low 145 concentrations of elements potentially toxic for plants, such as Cr and Ni 12 . 146 7 Actual basalt weathering rates and hence carbon drawdown potential remain uncertain, 147 depending strongly on particle size (limited by grinding cost), climatic and soil conditions, 148 and biological activity 8,57,58 . Water flow is critical because mineral surfaces have to be in 149 contact with water for the dissolution reaction to take place, and disturbed for the reaction to 150 continue 59 . Therefore, wet and warm climates demonstrate the highest weathering rates by 151 far 60,61 . 152 Besides precipitation and runoff, soil hydrology plays a crucial role in mineral weathering 62 . 153 In all climate zones, heavy clays and compacted soils will likely limit the dissolution rates of 154 added basalt severely due to low saturated hydraulic conductivity (poor water flow through 155 soil) and a prevalence of preferential water flow pathways through cracks in soil that 156 minimise interaction with basalt minerals 63,64 . Under natural conditions, flow in soil 157 generally affects only 0.1-10% of the soil matrix 65 , so that most of the available mineral 158 surfaces cannot exchange solutes, which limits dissolution. Poor contact between pore water 159 and mineral surfaces could explain the 2-3 orders of magnitude difference in weathering rates 160 that is measured in field (poor contact) vs. lab (maximum contact) experiments 63,66 . 161 Using sorghum plants and highly controlled experimental conditions with constant irrigation 162 (2,330 mm yr -1 ), drainage, and assuming permanent exposure of mineral surfaces to water, 163 basalt weathering rates were estimated to drive carbon sequestration at 0.63-0.82 t C ha -1 yr -1 164 for 100 t ha -1 basalt application, using a reactive transport model 58 . This is equivalent to 165 around 10% of its total theoretical carbon sequestration potential (~0.08 t C t -1 rock) 67 . 166 Mesocosm studies with wheat and barley, a precipitation of 800 mm yr -1 , and natural 167 processes such as drying cycles, preferential water flow, and mineral precipitation, found the 168 carbon sequestration potential of olivine (more rapid theoretical weathering and more total 169 sequestration potential than basalt) to be much lower, 0.006-0.013 t C ha -1 yr -1 at an Ideally, precipitation is captured in soil through rapid infiltration and high water retention. 212 Soil texture (particle size distribution; sand-silt-clay content), has long been considered the 213 key factor in soil hydrology. In clay-rich soils a low saturated hydraulic conductivity restricts 214 water infiltration and movement within soil. In contrast, saturated hydraulic conductivity is 215 high but water retention is low in sandy soils. Modifying soil texture is challenging because it 216 needs very high application rates of minerals, such as basalt 57 . 217 Biochar application can likely change both soil texture via biochar particle size, and soil 228 structure. Application of <30 t ha -1 of high surface-area biochar can increase hydraulic 229 conductivity in clay-rich soils 92 . While a cumulative biochar application of 10 t ha -1 over 5-230 10 years will only marginally increase the plant-available water content of sandy soil, further 231 application to >30 t ha -1 is expected to substantially increase the water-storage capacity 92 . 232 More available water can increase plant growth, which in turn helps to retain and re-circulate 233 water locally (transpiration instead of runoff) 93 , and to improve the contact between water 234 and minerals and, hence, the mineral weathering rate. 235 Aboveground plant carbon sequestration efficiency 236 Producing biochar from aboveground plant residues in high-biomass systems is key because 237 it has a higher CSE than natural biomass decomposition on a century timescale. Optimising 238 the biochar production system for maximum (stable) carbon yield decreases carbon losses 239 further, and significantly improves the CSE. The carbon sequestration potential of woody 240 biochar per unit biomass input can be increased by up to 45% by spraying low levels (2%) of 241 alkali (and earth alkaline) metals onto the biomass, such as potassium or sodium 53 , or by 242 incorporating wood ash 94 (Figure 2c). A significant part of basalt comprises alkali and earth 243 alkaline metals ( Table 2) that could also have the potential to catalyse biochar formation 244 when incorporated into the biomass before pyrolysis, which in addition increases the nutrient 245 content of biochar, providing further benefits for plant growth and carbon sequestration. 246 Biochar could be produced from biomass harvested from matured agroforestry systems, 247 ( Figure 4d), which are estimated to ultimately provide up to 10x higher biomass yields than 248 simple cropping and pasture systems 19 with estimates ranging from 0.3-15 t C ha -1 95 . Light 249 limitation can cause tree growth to decline with age, so that tree pruning and thinning 250 stimulate higher growth rates 96,97 . The conversion of harvested agroforestry residues 251 (average ~15% tree biomass pruning/thinning assumed per year) into biochar could make 252 0.05-2.3 t ha -1 yr -1 biomass available that supports the production of 0.01-0.6 t ha -1 of biochar 253 12 per year (~0.01-0.4 t C ha -1 yr -1 ), which is in addition to crop/forestry residue biochar. Other 254 options to obtain biomass for biochar production on-farm are setting aside land for tree 255 plantations 98 or fast growing bioenergy crops, harvesting woody weeds, which can yield up 256 to 44 t ha -1 of biomass 99 , or increasing straw residues by planting crop varieties with lower 257 harvest indices. 258 Dividing aboveground tree and shrub residues from agroforestry systems into N-rich green 259 material and carbon-rich woody debris could further increase the CSE and improves N 260 management (Figure 4b, d). During pyrolysis, N is mostly lost or converted into an 261 unavailable form 100 , so that only the N-poor biomass fraction should be used for biochar 262 production. Green, N-rich biomass is (biologically) converted into SOC more efficiently 263 (higher CSE) than N-poor biomass 34,37,39 , which makes N-rich (composted) biomass ideal for 264 building up SOC and providing N to plants (Figure 4b). 265 Importantly, more N is needed in the formation of mineral-associated SOC, relative to less 266 persistent forms of SOC (aggregate carbon); more available N in soil increases SOC stability 267 101 . Consequently, the SOC pool is typically higher and more stable under N-rich plant 268 species, e.g. legumes and N-fixing trees, than under N-poor species, which highlights the 269 value of N-rich biomass as cover or inter crop 39,102,103 (Figure 4b, d). results in a net agronomic advantage 77 . In addition, a higher carbon allocation in 294 rhizodeposits can result in enhanced nutrient supply from microorganisms, since 295 rhizodeposits directly feed microorganisms in exchange for nutrients 110 . Nurturing healthy 296 soils by investing energy and resources belowground will bring benefits that allow farming 297 systems to maintain yields in a changing climate, in stark contrast to a system purely focused 298 on short-term optimization of carbon allocation into grains (Figure 1). 299  During this process biochar remains stable 117 . In fact, (bio)char in rivers and marine 322 sediments have a residence time of thousands of years 119,120 . Therefore, the capacity to store 323 carbon in the form of biochar in the environment is likely unlimited (Table 1, Figure 2). 324

325
On a global scale, plant growth is limited by P and N, although K can also limit productivity 326 25,121 . N for crop growth can be provided by microorganisms that live in natural symbiosis 327 with plants, but P, K, and other nutrients are non-renewable and depleted in many ecosystems 328 25,121 . Basalts contain mineral nutrients in relevant quantities to satisfy plant demand, and 329 therefore can (partly) replace conventional fertiliser application; on average basalts from four 330 continents contained 0.2% P, 0.7% K, 5.3% Ca, and 3.7% Mg (Table 2) 122-125 . 331 Basalt application at 10 t ha -1 provides 10-50 kg P ha -1 and 20-430 kg K ha -1 (Table 2). 332 Typical recommendations (depending on soil type, existing soil nutrients, etc.) are 40 kg P 333 ha -1 and 133 kg K ha -1 for winter wheat, and 26 kg P ha -1 and 50 kg K ha -1 for improved rice 334 varieties 126 . This demonstrates that basalts can theoretically supply sufficient K and P to 335 compensate for nutrients that are removed with the harvest. However, not all of the K and P 336 is immediately plant available 58,71 , and further research is needed to establish basalt-based 337 nutrient supply in the short (immediate plant uptake), medium (one growing season), and 338 long term (several growing seasons). 339 Basalts (and biochar) also contain Ca and Mg that can neutralise acidic soil 71,127 . Biochar and 340 rock dust application at rates of 1 t ha -1 and 10 t ha -1 , respectively, supplies calcium carbonate 341 equivalent to 0.8-3.6 t ha -1 of lime; woody biochar provides ~0.06 t ha -1 128 and basalt 0.8-3.6 342 t ha -1 (Table 2). At such proposed application rates, the pH in soils of most textures and CECs 343 will likely increase to 5.5-6.5, the ideal pH for most plants 129,130 . Yet, the response of soil pH 344 to biochar and basalt application is slower than that to conventional lime addition because of 345 a lower solubility 8,131 . Still, silicate rocks can be a sustainable lime replacement that avoid 346 the CO2 emissions associated with lime production and application 132 . 347 In semi-arid and arid areas rehydration strategies that supply water to plants will result in 348 additional plant carbon and SOC accumulation 133 and likely basalt weathering. Given that 349 severe droughts accelerated by climate change already affect many areas around the world, 350 and are predicted to intensify and spread geographically 134 , the development of efficient 351 rehydration strategies will be key to climate change adaption and ecosystem resilience. In our proposed method increasing plant carbon capture and growth in agricultural systems 366 with (perennial) ground cover and partial tree canopy cover is the first step (Figure 1). 367 Improvements to water and nutrient supply enhance long-term soil properties and plant 368 growth. The extra plant biomass then is managed through efficient conversion into biochar 369 and stable SOC (Figure 2c). This enables virtuous cycles that further capture and storage 370 water and carbon (Figure 1). 371

372
Various unanswered issues arise as key future research questions, such as the weathering rate 373 and plant nutrient-provision potential of basalt and the degree to which a specifically 374 designed and regeneratively managed landscape can increase water use efficiency. Field trials 375 and demonstration sites across climate and soil types are urgently needed to establish 376 guidelines toward optimised carbon sequestration in productive agro-ecosystems. 377 Even more importantly, gaps between disciplines need to be bridged. First, to facilitate 378 adoption of these concepts in practice, novel soil models with measurable soil carbon pools 379 140,141 and improved representation of soil structure and associated hydrologic responses 26 380 need to be integrated into crop growth models, and calibrated to local conditions. Prediction 381 tools will increase confidence in long-term sequestration benefits, which is required to garner 382 further support from industry, government, and farmers. Second, application of biochar and 383 basalt in different proportions and compositions needs to be incorporated into the models and 384 combined with techno-economic analyses and decision-support tools. Detailed landscape 385 mapping and analysis will allow further fine-scale modelling of nutrient and water flows and 386 help in determining the ideal placement of trees and establishment of rehydration strategies in 387 water-limited environments. Such fine-scale modelling and adaptations in heterogenous 388 landscapes are essential for tailored approaches with respect to local to regional scale soil and 389 climate conditions, which form the corner stone of successful implementations that safeguard 390 our climate, environment, and food production.