Reconstruction of past terrestrial carbon storage in the Northern Hemisphere from the Osnabrück Biosphere Model and palaeodata

Abstract

Until now the reconstruction of past carbon storage from data has often been done by using modern carbon databases The results are likely a rough approximation of the reality, and can be improved by the use of biosphere models. These models usually need to be parameterized by a large number of environmental inputs, which are often not available from palaeodata The empirical Osnabriick Biosphere Model (OBM) needs as input only 3 environmental parameters, easily derivable from pollen data. We adapted it to reconstruct the past terrestrial carbon storage from palaeodata. Sensitivity experiments performed by uniformly decreasing the mean annual temperature, average annual precipitation and/or CO2 concentration suggest that temperature and CO2 concentration affect the carbon storage more than does precipitation. The use of the only palaeodata available at a global scale shows that only a weakening of CO2 fertilization must be invoked to reconstruct the Last Glacial Maximum (LGM) carbon storage, which is an intermediate situation between the no-fertilization effect assumed by previous studies and the fertilization effect based on modernist empirical equations. The terrestrial carbon storage in the Northern Hemisphere for this period ranged from 910 to 1270 Pg, which represents an increase of 330 to 710 Pg (a planetary increase of 470 to 1014 Pg from the LGM to the present). This result IS similar to our previous reconstruction and agrees broadly with the values estimated on the basis of 0.32%0 change obtained for global deep ocean 613C.

Simulating changes in the terrestrial biosphere during the last glacial/interglacial transition

“…Thus, their β value left no space for temperature driven metabolic changes. Only a weak CO 2 fertilisation was also proposed by a former simulation study (Peng et al, 1995) and model sensitivity to pCO 2 is of similar magnitude (case I) than in a DGVM study on the variability of the last millennium (Gerber et al, 2003). The temperature dependent metabolic control parameters in our study are smaller than in previous ones.…”

“…In this context it should be noted that in a modelling study the terrestrial carbon stock was insensitive to a 30% reduction in global precipitation and temperature effects and moderate pCO 2 fertilisation were identified as most likely drivers of G/IG changes (Peng et al, 1995). This implies that regional differences in precipitation and not global average changes are of importance on paleoclimatic time scales.…”

From static biogeographical model to dynamic global vegetation model: a global perspective on modelling vegetation dynamics

“…However, these models usually need to be parameterized by a large number of environmental inputs, which are not often available from palaeoecological data (Peng et al, 1998). An alternative method is the use of an empirical biospheric model, such as the Osnabrück Biosphere Model (OBM) (Esser 1987(Esser , 1991 or CARAIB biosphere model (Warnant et al, 1994), or statistical models (Peng et al, , 1995a. These models only need three environmental parameters as input, which are easily derivable from palaeoecological data (Peng et al, 1998) or from GCM simulations (François et al, 1998(François et al, , 1999.…”

“…These models only need three environmental parameters as input, which are easily derivable from palaeoecological data (Peng et al, 1998) or from GCM simulations (François et al, 1998(François et al, , 1999. For these reasons, biosphere models have been widely used to estimate past terrestrial carbon dynamics in response to past climatic changes (Peng et al, , 1995aLautenschlager, 1993, François et al, 1998;Peng et al, 1998).…”

“…An understanding of effects of past and possible future climate change on terrestrial ecosystems and global carbon cycles will require a clear picture of how vegetation changed in the past and may change in the future (Prentice et al, 1991;Overpeck et al, 1992;Peng et al, 1995aPeng et al, , 1998Cramer et al 1999). The distribution of potential…”