Spatio-temporal mapping reveals distinct phases of soil organic carbon stocks changes across the CONUS over the past 60 years

We provide the spatial distribution of SOC densities in the CONUS region from 1955 to 2014. The dataset contains four soil depths, 0 ~ 15 cm, 15 ~ 30 cm, 30 ~ 60 cm and 60 ~ 100 cm (i.e. 0015, 1530, 3060 and 6000). Each tif file contains twelve bands corresponding to 1955-1959, 1960-1964, ...., 2010-2014. UL and LL are the upper and lower limits of the 95% uncertainty interval, respectively. Abstract: Understanding spatio-temporal patterns of SOC stocks is critical for global climate change mitigation and sustainable soil management. However, none of the existing databases (Harmonized World Soil Database, SoilGrids and some national databases) took into account the spatio-temporal SOC stocks information. Supported by soil samples and environmental covariates, space and time DSM (ST-DSM) has become an important and effective method to reveal the spatio-temporal changes of SOC stocks. Continental United States (CONUS) has abundant and well-documented soil samples with time labels, which laid the groundwork for us to conduct a comprehensive SOC stocks estimation with multi-depth, long-time-series, high-resolution. Specifically, we propose leveraging time-series soil data from WoSIS and ISCN to build a spatio-temporal digital soil mapping (ST-DSM) model based on environmental covariates and machine learning techniques (random forest framework) to generated multi-depth SOC stocks maps for CONUS including 0 ~ 15, 15 ~ 30, 30 ~ 60 and 60 ~ 100 cm depths, at a 250 m spatial resolution and 5-year temporal intervals (1955-1959, 1960-1964, ...., 2010-2014) with predictive uncertainty quantified via Quantile Regression Forest (QRF). Furthermore, we analyze the dynamic trends in SOC stocks across various depths and land uses. The results indicate that over 60 years, overall SOC stocks below 1 m demonstrate a multi-stage change of “rising-declining-rising”. SOC stocks below 1 m rose from 71.98 Pg to 73.49 Pg with an increased rate of 25.18 Tg per year. Surface SOC stocks (0 ~ 15 cm) remained relatively stable, while carbon stocks in subsurface and deep soils (15 ~ 100 cm) showed significant increases, particularly in the 60 ~ 100 cm layer. Across different land uses, the topsoil of cropland may be a source of carbon, while forests are important carbon sinks. Surface SOC stocks (0 ~ 15 cm) in cropland areas significantly decreased and subsurface and deep soil cropland SOC stocks (15 ~ 30 cm, 30 ~ 60 cm and 60 ~ 100 cm) showed a fluctuating rising trend. Forest SOC stocks show a fluctuating rising trend overall. Pastureland and sparse/no vegetation and unmanaged natural grasslands once showed some carbon sequestration potential in the past, but their efficiency has recently begun to decline. This research enhances our understanding of the global carbon cycle and supports the development of effective soil management and climate mitigation strategies.