PROJENE: High-Resolution Bias-Corrected Climate Projections Dataset (CMIP6) for Northeast Brazil

This dataset provides high-resolution (0.25° x 0.25°) climate change projections developed within the PROJENE project. The files contain daily multidimensional arrays (NetCDF4) derived from a robust selection of Global Climate Models (GCMs) from the Coupled Model Intercomparison Project Phase 6 (CMIP6).

Scientific Relevance for the Brazilian Northeast (NEB):
The Brazilian Northeast is a region distinguished by its strong natural climate variability and high vulnerability to desertification and severe droughts. Generating reliable future scenarios for this region is a major scientific challenge due to the complex local atmospheric dynamics that raw global models often fail to capture. This dataset bridges this gap by providing a tailored, high-resolution, bias-corrected climatological grid. It empowers researchers and policymakers with realistic spatial and temporal bounds for extreme events, crucial for securing water resources, modernizing agricultural planning, and implementing robust adaptation policies in one of the most climate-sensitive regions in the Americas.

Bias Correction Methodology:
To make the raw GCM data actionable for local hydrological and agricultural impact studies, the models underwent rigorous bias correction using the Quantile Delta Mapping (QDM) algorithm.

• Observational Baseline (Calibration): Calibrated against the Brazilian Daily Weather Gridded Data (BR-DWGD) high-resolution observational product, strictly utilizing the historical stationary period of 1961-1995.
• Validation Period: Model performance and bias-correction robustness were independently evaluated via an out-of-sample validation period spanning from 1996 to 2014.
• Key Characteristics of QDM: Unlike standard empirical quantile mapping, QDM is specifically designed to correct systematic biases in the probability distribution of climate variables while strictly preserving the relative changes (the "delta" trends) projected by the original GCMs. This physically consistent approach ensures that future climate anomalies and extreme events (across all probability quantiles) remain true to the GCM's physics while being correctly anchored to the local climatological reality.

Dataset Characteristics:

• Format: NetCDF4 (.nc)
• Spatial Resolution: 0.25° regular grid.
• Temporal Resolution: Daily time series.
• Time Horizons: Historical (1961–2014) and Future Projections (2015–2100) under shared socioeconomic pathways SSP2-4.5 and SSP5-8.5.
• Variables: Precipitation (pr), Maximum Temperature (tasmax), Minimum Temperature (tasmin), Average Temperature (tas — only available through the PROJENE web platform), Surface Downwelling Shortwave Radiation (rsds), Relative Humidity (hurs), and Wind Speed (sfcwind).
• Models: Includes leading individual CMIP6 models along with a Multi-Model Ensemble average.

(Interactive exploration and extraction of this dataset are available via the PROJENE web platform: https://projene.xhidro.com/).