B046-0002
Isolating the impacts of soil moisture initialization on land carbon fluxes at the sub-seasonal to seasonal scale

Thursday, 10 December 2020
Poster
Eunjee Lee, Universities Space Research Association / NASA Goddard, Greenbelt, MD, United States, Randal D Koster, NASA Goddard SFC, Greenbelt, MD, United States, Lesley E Ott, NASA Goddard Space Flight Center, Greenbelt, MD, United States, Jana Kolassa, USRA/GESTAR, Columbia, United States, Sarith P P Mahanama, NASA/GSFC, Greenbelt, MD, United States and Kristi R Arsenault, SAIC, Greenbelt, MD, United States
Abstract:
Soil moisture, through its impact on transpiration, plays a significant role in controlling the productivity of terrestrial vegetation. Initial soil moisture conditions in a seasonal forecast may therefore affect the forecast of land carbon uptake. Here we investigate the relative impacts of soil moisture initialization and carbon reservoir initialization on forecasts of land carbon fluxes at the sub-seasonal to seasonal (S2S) scale. The bias-corrected, retrospective meteorology of NASA Global Modeling and Assimilation Office (GMAO)’s S2S ensemble forecast was used to force the stand-alone Catchment-CN model and thereby estimate terrestrial carbon responses out to nine lead months. Our results show that soil moisture initialization is a major contributor (approximately 44%) to the high global carbon uptake forecast skill seen during the first three lead months. The carbon reservoir initialization explains roughly another half of the monthly carbon forecast skill during this period and becomes relatively more important at longer leads (while the overall forecast skill decreases after the 3rd lead month), suggesting a slower but longer-lasting influence of carbon reservoir initialization on carbon fluxes. Our results highlight the significance of a good soil moisture initialization for improved forecasts of carbon fluxes at leads of several months, further support for the usefulness of assimilating the satellite-based soil moisture information into terrestrial biosphere models and of a short term carbon forecast to understand current events with often a lag in the availability of flux estimates.