Stabilization of new carbon inputs rather than old carbon decomposition determines soil organic carbon shifts following woody or herbaceous vegetation transitions

Abstract

Background and aims

Although numerous studies have quantified the effects of land-use changes on soil organic carbon (SOC) stocks, few have examined simultaneously the weight of carbon (C) inputs vs. outputs in shaping these changes. We quantified the relative importance of soil C inputs and outputs in determining SOC changes following the conversion of natural ecosystems to pastures or tree plantations, and evaluated them in light of variations in biomass production, its quality (C:N) and above/belowground allocation patterns.

Methods

We sampled soils up to one-meter depth under native grasslands or forests and compared them to adjacent sites with pastures or plantations to estimate the proportion of new SOC (SOC_new) retained in the soil and the decomposition rates of old SOC (k _SOC-old ) based on δ ¹³C shifts. We also analyzed these changes in the particulate organic matter fraction (POM) and estimated above and belowground net primary production (ANPP and BNPP) from satellite images, as well as changes in vegetation and soil’s C:N ratios.

Results

The conversion of grasslands to tree plantations decreased total SOC contents while the conversion of forests to pastures increased SOC contents in the topsoil but decreased them in deep layers, maintaining similar soil stocks up to 1 m. Changes in POM were less important and occurred only in the topsoil after cultivating pastures, following SOC changes. Surprisingly, both land-use trajectories showed similar decomposition rates in the topsoil and therefore overall SOC changes were not correlated with C outputs (k _SOC-old ) but were significantly correlated with C inputs and their stabilization as SOC_new (similar results were obtained for the POM fraction). Pastures although decreased ANPP (as compared to forest) they increased belowground allocation and C:N ratios of their inputs to the soil, probably favoring the retention and stabilization of their new C inputs. In contrast, tree plantations increased ANPP but decreased BNPP (as compared to grasslands) and scarcely accumulated SOC_new probably as a result of the high C retention in standing biomass.

Conclusions

Our results suggest that SOC changes are mainly controlled by the quantity and quality of C inputs and their retention in the soil, rather than by C outputs in these perennial subtropical ecosystems.

Appendix

Table 4 List of abbreviations

Table 5 Estimates of productivity from radiation interception and vegetation indexes measured by MODIS satellites, at sites with natural vegetation (forest and grasslands) and pastures or tree plantations

Plant production values estimated from remote sensing in our work were similar than those measured by other studies. NPP of subtropical grasslands varies from 6 to 16 Mg ha⁻¹ year⁻¹ (Gomez and Gallopin 1991; Grace et al. 2006; Sundaravalli and Paliwal 2000), while forest NPP varies between 20 and 25 Mg ha⁻¹ year⁻¹ (Grace et al. 2006; Ito and Oikawa 2004; Malhi et al. 2009), although in a global review Clark et al. (2001) observed extreme values between 3.4 to 43.4 Mg ha⁻¹ year⁻¹. On the other hand in plantations, NPP of Pinus elliotti and P. taeda in other regions are between 10 and 20 Mg ha⁻¹ year⁻¹(Clark et al. 2001; Grace et al. 2006; Ito and Oikawa 2004; Litton et al. 2007), but ANPP in similar sites of the province of Misiones was reported around 34.8 ± 1.5 Mg ha⁻¹ year⁻¹(Pérez et al. 2006); while NPP measurements of tropical pastures are between 28 and 40 Mg ha⁻¹ year⁻¹(Boddey et al. 2004; Fisher et al. 1997; Rezende et al. 1999)