Ecosystem carbon storage does not vary across a 5.2°C mean annual temperature gradient in Hawaiian tropical montane wet forests

Background/Question/Methods

Tropical forests are a key component of the global carbon (C) cycle. Although they occupy only ~12% of Earth’s land surface, tropical forests account for ~40% of terrestrial net primary production and ~25% of C stored in living biomass. Given their disproportionate influence on global C cycling, it is critical to understand how ecosystem C storage in tropical forests will respond to future warming. A global meta-analysis of tropical forest C cycling (Raich et al. 2006) found no net change in total ecosystem C storage with increasing mean annual temperature (MAT), but a shift in the relative distribution of C from detritus to live biomass with increasing MAT. However, this study combined data from sites spanning wide ranges in biotic and abiotic conditions, potentially masking direct effects of temperature. We established a highly constrained MAT gradient on the Island of Hawaii consisting of nine tropical montane wet forest plots spanning 5.2°C MAT (13.0-18.2°C) with constant vegetation, disturbance history, soil type, geology, and soil water balance. In each plot, we used a combination of direct measurements and allometry to estimate the C content of living biomass (foliage, stems, and roots) and detritus (coarse woody debris, forest floor, and soil organic matter to a depth of 1 m).

Results/Conclusions

Total ecosystem C storage did not vary with MAT (adj. R² = 0.02, P = 0.32), which is consistent with results from Raich et al. (2006). In contrast to Raich et al. (2006), however, we found no change in the relative amount of C stored in detritus vs. live biomass with increasing MAT. Across the nine plots, detritus and living biomass averaged 59 ± 3% and 41 ± 3% of total ecosystem C storage, respectively.  Because our gradient did not vary with respect to factors other than MAT, these findings suggest that the relative distribution of C pools in tropical forests may be less sensitive to rising MAT than indicated by previous studies. The apparent stability of C pool distribution in forests with rising MAT is critical given that the majority of C across this gradient is in the form of soil organic C, which is less sensitive to disturbance-induced losses than aboveground C pools. Our findings also provide needed detail on how individual ecosystem C pools will respond to warming, which will better inform predictions of how tropical wet forest ecosystems will respond to rising MAT.