Spatial estimates of gravimetric soil carbon (%) for the 0–7.5 topsoil layer, over the land surface of New Zealand. It was derived using gradient boosting regression from a range of environmental classification, national soil classification, and climate layers as explanatory variables.
The model was calibrated on a large number of samples (approx. 17,000 laboratory samples) assembled from a range of different soil science projects, including data in the National Soil Repository Database (NSDR), the Land Management Index (LMI) dataset, and the GNS geochemical baseline dataset.
Estimates of soil carbon are made on a grid where each grid cell has dimensions 100 m x 100 m (1 ha). The value at each grid cell is our best estimate of the average soil carbon percent at the centre point of each grid cell.
Estimates of soil carbon (%) for the surface layer to a depth of 7.5 cm