The centennial legacy of land-use change on organic carbon stocks of temperate agricultural soils

Land-use change (LUC) in agricultural settings is common both historically and under more recent climate-smart agriculture guidelines aimed at reducing the impact of agriculture on the climate. Recognising that conversion of perennial grasslands to annual cropland results in a large, but potentially reversible, loss of soil organic carbon (SOC) such guidelines often call for increasing the overall area under grassland. To date, the magnitude and direction of SOC change following LUC has been fairly well accounted for, but the time it takes to reach a new SOC equilibrium is not well understood. While broad scale emission reporting best practices (e.g. IPCC) suggest that SOC equilibrium is reached approximately 20 years after LUC, there is a growing body of knowledge that supports a centennial timescale in temperate or boreal climates. With data from the first German Agricultural Soil Inventory alongside extensive per-site land-use histories, we established SOC change timelines that show that not only does SOC take much longer than 20 years to reach equilibrium but it reaches equilibrium at vastly different rates depending on the direction of LUC. Sites converted from cropland to grassland took 83 years (95 % CI: 79 to 90 years) to reach SOC equilibrium whereas sites converted from grassland to cropland took 180 years (95 % CI: 151 to 223 years). In order to map the effects of historic LUC on SOC stocks in temperate agroecosystems with similar grassland and cropland SOC stocks to Germany, we applied these timeline models to comparable sites from the HILDA+ global LUC database (Winkler et al., 2020); a global reconstruction of annual land use and land cover at a 1 km spatial resolution from 1899 to 2019. Compiled from a range of open data sources (remote sensing, reconstructions, and census data) the HILDA+ dataset offers insights into relatively fine scale LUC dynamics that follow known socioeconomic drivers over the past 120 years. Using this dataset, we determined that 112 Million ha, or 3.5 % of the total agricultural area worldwide, was comparable to German agriculture in terms of SOC and growing conditions, and 11 % of that land (12.6 Million ha) had undergone LUC from cropland to grassland or vice versa since 1899. After accounting for duration on a per-cell basis, areas having undergone LUC from cropland to grassland (7.5 million ha) accounted for a 86.2 million Mg C increase over the past 120 years. Conversely, areas having undergone LUC from grassland to cropland (5.1 million ha) accounted for a -55.0 million Mg C decrease in SOC. The overall net increase of 31.2 million Mg C corresponds to about 5‰ of total SOC stocks across all agricultural land with pronounced regional differences. We conclude that land-use change histories of at least one century should be considered when interpreting present-day, and predicting future, SOC dynamics. Winkler, K., Fuchs, R., Rounsevell, M. D. A., & Herold, M. (2020). HILDA+ Global Land Use Change between 1960 and 2019 [dataset]. PANGAEA. https://doi.org/10.1594/PANGAEA.92184