DCENT-I: A Globally Infilled Extension of the Dynamically Consistent ENsemble of Temperature Dataset

The Dynamically Consistent ENsemble of Temperature – Infilled (DCENT-I) is a spatially complete version of DCENT, providing global and regional monthly temperature fields at 5°×5° resolution from 1850 to the present. Project Site: https://dcent-i.github.io DCENT-I is produced by infilling land-air and sea-surface temperatures from DCENT using ordinary kriging with anisotropic and heterogeneous kernels. Over sea-ice areas, air-temperature anomalies are incorporated to ensure complete coverage. It consists of 200 ensemble members that sample the full range of uncertainties, including those from bias adjustments and the infilling process. Compared with the original DCENT ensemble, infilling increases the global mean surface temperature (GMST) warming estimate by 0.08 °C [0.05, 0.11] (95 % confidence interval), in agreement with independent model-based assessments. Spatially, DCENT-I exhibits century-scale warming patterns similar to other datasets, including stronger continental than ocean warming and stronger Arctic than mid-latitude warming. It also improves the reconstruction of early El Niño events (e.g., 1887) and maintains realistic El Niño variability throughout the record. DCENT-I is available in NetCDF format in this repository and at CEDA, and in Zarr format on Google Cloud. For accessing DCENT-I data from Google Cloud, see the Demo_DCENT-I_google_cloud.ipynb notebook in this repository.