Soil Organic Carbon Distributions in Tidal Wetlands of the Northeastern USA

This dataset provides estimates of soil organic carbon (SOC) in tidal wetlands for the northeastern United States. The data cover the period 1998-2018. Northeastern U.S. tidal wetlands and bordering areas were harmonized from government agencies [U.S. Department of Agriculture - Natural Resources Conservation Service (USDA-NRCS), National Cooperative Soil Survey (NCSS), USDA-NRCS - Rapid Carbon Assessment (RaCA), U.S. Environmental Protection Agency - National Wetland Condition and Assessment (EPA-NWCA)] and published studies. Point data for carbon stocks (in kg m-2) at four soil depths (0-5, 0-30, 0-100, and 0-200 cm) are included. SOC for the four depths was predicted for eight regional zones using regression models driven by environmental covariates. Two methods were used to estimate parameters for these models, a Random Forest (RF) Ranger method and a Quantile Regression Forest (QRF) model. The distribution of SOC was predicted for tidal wetland cover types mapped by Correll et al. (2019). Predictions and uncertainties are available at a 3 m resolution.

本数据集提供了美国东北部潮汐湿地土壤有机碳（SOC）的估算值。数据覆盖了1998年至2018年期间。通过对美国农业部自然资源保护服务局（USDA-NRCS）、国家合作土壤调查（NCSS）、USDA-NRCS快速碳评估（RaCA）以及美国环境保护局国家湿地状况与评估（EPA-NWCA）等政府机构及已发表的研究成果进行整合，对美国东北部潮汐湿地及其周边地区进行了数据融合。数据中包含了四个土壤深度（0-5厘米、0-30厘米、0-100厘米和0-200厘米）处的碳储存点数据（单位：千克每平方米）。利用由环境协变量驱动的回归模型，对八个区域区域的四个深度处的SOC进行了预测。采用随机森林（RF）Ranger方法和分位数回归森林（QRF）模型两种方法估算这些模型的参数。利用Correll等人（2019年）所绘制的潮汐湿地覆盖类型图，预测了SOC的分布。预测结果及不确定性以3米分辨率为准。