Oral

IAMAS

JMP10 - Ocean-Atmosphere Mechanisms of Climate Variability, Change and Predictability

Exploring Ocean-Driven Multi-year Predictability of Terrestrial Ecosystem Components

The demand for improved near-term prediction and projection of terrestrial ecosystem components has been growing to support disaster management and adaptation strategies. However, estimating their predictability and identifying sources of predictability on multi-year time scales remains challenging. Here, we explore the multi-year predictability of key terrestrial ecosystem components, such as soil moisture, Gross Primary Productivity (GPP), total soil carbon, and burned area, mainly driven by ocean variation. A set of Earth system model simulations based on Community Earth System Model version 2 (CESM2) is utilized, including 50-member uninitialized runs with historical external forcings and 20-member ocean data assimilation runs (ODA) from 1951 to 2021. The ODA runs incorporate observed three-dimensional ocean temperature and salinity to constrain the coupled climate system, providing realistic representations of ocean variability. Our results show that the key terrestrial ecosystem variables, particularly GPP, are predictable for up to 1 to 3 years in many parts of the globe, contributed by large-scale ocean variation as well as external forcings. By applying Singular Value Decomposition (SVD) analysis on the sea surface temperature (SST) and terrestrial ecosystem variables, we further show that the essential sources of their predictability on multi-year time scales include Atlantic multi-decadal variability, El Nino-Southern Oscillation, and tropical trans-basin variability through atmospheric teleconnections. These findings highlight the importance of ocean-atmosphere interactions in shaping ecosystem processes on multi-year timescales, with implications for improving long-term predictions of carbon cycles.