[A-CG37] Biogeochemical cycles in Land Ecosystem

[ACG37-P07] On the factors affecting reproducibility of GPP estimated by the
Earth System Model MIROC-ES2L

*Akira Kurisu¹, Toru Nozawa², Manabu Abe³, Tomohiro Hajima³ (1.Faculty of Science, Okayama University, 2.Graduate School of Natural Science and Technology, Okayama University, 3.Japan Agency for Marine-Earth Science and Technology )

Keywords:Earth System Models, GPP

Earth System Models (ESM) that include biogeochemical cycles in land and
marine ecosystems have come to be used for global warming simulations.
Reproducibility of the biogeochemical cycles in the ESM, however, does
not seem to be sufficient. For example, the MIROC-ES2L, one of the ESM
developing in Japan, has some problems on the reproducibility of the
Gross Primary Production (GPP); the terrestrial GPP has very little
seasonality over the tropics. In this study, focusing on the African
continent, we investigated that biases (defined as differences between
the MIROC-ES2L outputs and satellite observations) in what kind of
variables are related to the biases in the GPP.

We analyzed datasets of historical runs done by the MIROC-ES2L and those
obtained by satellite and in-situ observations. The MIROC-ES2L data has
horizontal resolutions of 2.8 degrees in longitude times 2.8 degrees in
latitude. The horizontal resolutions of the observation datasets are
converted into the same ones of the MIROC-ES2L. We investigated
relationships between biases in the GPP and those in Leaf Area Index
(LAI), surface downward shortwave radiation, surface temperature,
surface soil moisture content, and precipitation.

Latitude-time distributions of the GPP biases over the African continent
show that they are larger in June-October on the south of the Equator
whereas larger in December-March on the north of the Equator. Comparing
geographical distributions of the GPP biases with those of the LAI ones,
both biases have the same sign in June-October on the south of the
Equator and in December-March on the north of the Equator. On the other
hand, both biases have the opposite sign in June-October on the north of
the Equator and in December-March on the south of the Equator.
Therefore, the GPP biases seems to be largely explained by the LAI
biases, especially in December-March. However, in June-October over the
tropical area, the GPP biases cannot be explained by the LAI biases
only. The GPP biases in that case might be largely related to the other
climatic variables in addition to the LAI biases.