Land use classification and steady-state assumptions for assessing land use effects on soil carbon stocks: A New Zealand case study

Land-use classification is important to forecasting changes in soil organic carbon (SOC) stocks in greenhouse gas inventories and informing policies to lower future emissions. Two large New Zealand (NZ) datasets were analysed to evaluate the importance of land-use and soil classification for predicting land use effects on SOC stocks (0–30 cm) and to test steady state assumptions under high producing grasslands (HPG) and annual croplands (ACL). A more detailed land-use type (LU-type) classification explained more variability (≈28%) in SOC stocks in HPG and ACL than the broader land-use category classification (≈21%) used to derive NZ’s soil C inventory. Additional variation (≈25%) was explained by NZ soil order. LU-type explained more SOC variability than soil order in 0–15 cm soils, whereas soil order was more important in 15–30 cm soils. Overall, SOC stocks under dairy (110 Mg C/ha) were 10% higher than sheep/beef pasture (100 Mg C/ha) and have increased in Brown, Pallic and Gley soils under dairy since pre-intensification (1990s), in line with large gains in pasture productivity. Predicted post-intensification loss of SOC from ACL (22.6%) was greater than previously reported (≈10%) and much higher in volcanic (31–53%) than sedimentary (8–12%) soils. Predicted SOC losses were greater under vegetable (31%) than arable (18%) cropping, consistent with differences crop types, tillage frequency and fallow periods. The time required to reach steady-state SOC following land-use change (LUC) was > 20 years and differed between soil types and land-use practices.