DCPP contribution to CMIP7 AFT

The newly introduced CMIP7 assessment fast track experiments are intended to support upcoming policy assessments. To ensure an updated initialized climate forecast is available to support these upcoming climate assessments, such as the IPCC AR7 report, DCPP is coordinating a 10-year long initialized prediction using CMIP6-era prediction systems and forcing. 

This usage of CMIP6-era systems is a compromise and reflects the short timeframe involved with the CMIP7 AFT. In other words, the short deadlines associated with the CMIP7 AFT makes the completion of the compulsory, and computationally expensive, hindcast datasets needed for drift correction and calibration of any forecast extremely difficult in the timeframe of the CMIP7 AFT. Therefore, the DCPP CMIP-7 AFT experiment follows the DCPP experimental protocol for CMIP6 (Boer et al., 2016), specifically component B (DCPP-B) of the protocol describing real time forecasts.   

The DCPP contribution to the CMIP7 AFT is as follows: 

  • A 10-year, 10-member, forecast initialised from a 2025 initial condition - the first full forecast year is 2026. 

  • External forcings are to be taken from cmip6 scenario ssp245

  • A CMIP6-era hindcast dataset (dcppA-hindcast) experiment is a prerequisite needed for drift correction and calibration. 

  • The CMIP7 experiment id is dcppB-forecast-cmip6, although modelling centres could also submit their experiment under CMIP6-plus under experiment id dcppB-forecast.     

Note that this experiment is not to be confused with the full DCPP experimental protocol for CMIP7, which is described in the next section.

DCPP experimental protocol for CMIP7

DCPP will be contributing an updated experimental protocol for CMIP7 as part of the CMIP7 community MIPs. As in CMIP6, the DCPP experimental protocol for CMIP7 will have a modular design including the following components:

  • DCPP-A: retrospective predictions, or hindcasts

  • DCPP-B: real-time predictions

  • DCPP-C: process experiments


The protocol will be written up and submitted to the CMIP7 special collection in GMD. However, to aid groups interested in contributing to DCPP experiments in CMIP7 we will add drafts of DCPP-A and DCPP-B as the DCPP panel agrees on them.

Draft protocol

DCPP-A Hindcasts

DCPP-B Forecasts

(coming soon)

Data request 

The data request for DCPP simulations is detailed in the Harmonized Data Request within CMIP7. Specifically, the request for DCPP is contained within the “Multi-annual-to-decadal predictability of the Earth System and risk assessment of climate extremes” opportunity under two variable groups, DCPP_Essential and DCPP_Wider. 

  • The DCPP_Essential variable group contains the core variables needed to provide the baseline assessment of Multi-annual-to-decadal predictions. Thus, the DCPP_Essential variable group has high priority.

  • The DCPP_Wider variable group expands upon the DCPP_Essential variables, designed to improve the understanding of predictability for a wider array of variables, including ocean biogeochemistry, and to better illuminate the processes that govern regional climate and serve as sources of predictability.

DCPP experimental protocol for CMIP6

The experimental protocol for the Decadal Climate Prediction Project (DCPP) contribution and other resources related to CMIP6 can be found here:

https://www.wcrp-esmo.org/projects-and-panels/dcpp/dcpp-cmip6