Multi-agency workshop on S2S predictability
On timescales of two weeks to a few months, slowly varying land conditions provide a source of memory that can potentially influence atmospheric behavior and near-surface meteorology, thus enhancing the skill to forecast them. Alongside oceanic and stratospheric influences, observations and reanalyses indicate land processes as major sources of subseasonal-to-seasonal (S2S) predictability, including extremes like droughts and heatwaves.
In the forecasting community, the S2S window is held as a “predictability desert”: the lead time is long enough for the memory of the atmospheric initial state to dissipate, yet too short for slowly varying ocean boundary conditions to exert a strong influence. Additionally, forecast skill is limited by model errors in land processes and land-atmosphere coupling, which can translate into sizable S2S biases. Sparse observations of key land states further limit initialization, verification, and model improvement. Addressing these gaps requires better process-level understanding, improved assimilation of relevant observations at the land-atmosphere interface, and enhanced representation of land–atmosphere coupling within S2S forecast systems.
To tackle these challenges, a three-day community workshop organized by National Science Foundation National Center for Atmospheric Research (NSF NCAR), National Oceanic and Atmospheric Administration (NOAA), National Aeronautic and Space Administration (NASA), and Department of Energy (DOE) was convened at NSF NCAR in June 2025 to discuss how land processes and land-atmosphere interactions contribute to S2S predictability. Focused themes included source of S2S predictability, initialization and data assimilation, artificial intelligence/machine learning (AI/ML) applications, process- and application-oriented metrics, and coordinated modeling strategies. For more information about the workshop, look out for a workshop report (He et al. 2025), which has been accepted for publication in the Bulletin of the American Meteorological Society. The report summarizes the workshop discussion and future actions to (1) develop a set of process- and application-oriented metrics for assessing S2S prediction skills across modeling centers, and (2) develop experimental protocols for coordinated experiments to quantify the role of land-atmosphere processes in S2S predictability.
He, C., L. N. Zhang, L. R. Leung, J. H. Richter, C. Bassett et al. 2025: Harnessing Land-Atmosphere Interactions to Enhance Subseasonal-to-Seasonal Predictability. Bulletin of the American Meteorological Society, accepted.
