Short answer
Drought projections with CMIP6 are scenario-based analyses, not exact forecasts. They require climate-model ensembles, consistent historical and future periods, bias correction or downscaling when local interpretation is needed, and careful communication of uncertainty.
Projection workflow
- Define the drought question, region, sector, and decision horizon.
- Select CMIP6 models, scenarios, variables, and historical reference periods.
- Check units, calendars, temporal resolution, and missing data.
- Bias-correct or downscale projections when local or station-scale interpretation is required.
- Calculate drought indicators or anomalies consistently for historical and future periods.
- Summarize ensemble spread instead of relying on one model result.
Scenarios and time periods
CMIP6 projections are commonly analyzed under Shared Socioeconomic Pathway scenarios. A drought study should state the scenario, model list, forcing experiment, historical period, future period, and baseline used to classify drought anomalies. Changing any of these choices can change the conclusion.
| Choice | Why it matters | What to report |
|---|---|---|
| Scenario | Controls future forcing assumptions. | SSP name and period. |
| Model ensemble | Controls uncertainty range. | Model names and number of members. |
| Baseline | Controls drought categories. | Historical reference period. |
| Correction method | Controls local realism. | Bias correction or downscaling method. |
How this relates to DMAP-AI
CMIP6 projection processing should be prepared outside the current DMAP-AI Research Version browser workflow. DMAP-AI can help with SPI-based drought summaries and structured AI interpretation, but the current browser workflow should not be described as producing CMIP6 projection outputs. When projection results are externally prepared, structured metadata can help AI explain model, scenario, baseline, and uncertainty without overstating certainty.
Frequently asked questions
Are CMIP6 drought projections predictions?
No. They are scenario-based projections conditioned on model and emissions assumptions.
Can one CMIP6 model be enough?
Usually no. A multi-model ensemble is preferred because precipitation and drought responses vary substantially among models.
Should future drought categories use the historical baseline?
Often yes, but the baseline choice should be stated clearly because it controls what “normal” means.
Selected references
- Eyring et al. (2016). Overview of the Coupled Model Intercomparison Project Phase 6 experimental design.
- IPCC Sixth Assessment Report climate-projection resources.
- Cook et al. (2020). Twenty-first century drought projections in the CMIP6 forcing scenarios.
- World Climate Research Programme. CMIP6 documentation.