1. The Andes are predicted to warm by 3–5 °C this century with the potential to alter the processes regulating carbon (C) cycling in these tropical forest soils. This rapid warming is expected to stimulate soil microbial respiration and change plant species distributions, thereby affecting the quantity and quality of C inputs to the soil and influencing the quantity of soil-derived CO2 released to the atmosphere.2. We studied tropical lowland, premontane and montane forest soils taken from along a 3200-m elevation gradient located in south-east Andean Peru. We determined how soil microbial communities and abiotic soil properties differed with elevation. We then examined how these differences in microbial composition and soil abiotic properties affected soil C-cycling processes, by amending soils with C substrates varying in complexity and measuring soil heterotrophic respiration (RH).3. Our results show that there were consistent patterns of change in soil biotic and abiotic properties with elevation. Microbial biomass and the abundance of fungi relative to bacteria increased significantly with elevation, and these differences in microbial community composition were strongly correlated with greater soil C content and C:N (nitrogen) ratios. We also found that RH increased with added C substrate quality and quantity and was positively related to microbial biomass and fungal abundance.4. Statistical modelling revealed that RH responses to changing C inputs were best predicted by soil pH and microbial community composition, with the abundance of fungi relative to bacteria, and abundance of gram-positive relative to gram-negative bacteria explaining much of the model variance.5. Synthesis. Our results show that the relative abundance of microbial functional groups is an important determinant of RH responses to changing C inputs along an extensive tropical elevation gradient in Andean Peru. Although we do not make an experimental test of the effects of climate change on soil, these results challenge the assumption that different soil microbial communities will be ‘functionally equivalent’ as climate change progresses, and they emphasize the need for better ecological metrics of soil microbial communities to help predict C cycle responses to climate change in tropical biomes.

Summary Tropical ecosystems play a key role in the global carbon cycle, but their response to global warming is not well understood. Altitudinal gradients offer the unique possibility of undertaking in situ experimental studies of the influence of alterations in climate on the carbon (C) cycle. In a soil‐translocation experiment we took replicate soil cores at 3030 m, 1500 m, 1000 m and 200 m above sea level along an altitudinal gradient in tropical forest in Peru, and exchanged (i.e. translocated) them among these sites to observe the influence of altered climatic conditions on the decomposition of soil organic matter under natural field conditions. Soil respiration rates of the translocated soil cores and adjacent undisturbed soils were measured twice a month from April 2007 to October 2007. The temperature sensitivity of heterotrophic respiration in each core was examined using a Lloyd & Taylor function and a simple modified third‐order polynomial fit. Calculated Q10 values decreased with decreasing altitude using both mathematical functions (2.53–1.24 according to the Lloyd & Taylor function, and 2.56–0.63 using the polynomial fit). Soil organic C‐stocks increased markedly and linearly with altitude, but surprisingly the average total soil respiration rate did not vary significantly with altitude along the transect (3.98–4.31 μmol CO2 m−2 s−1). This implies an increase with elevation of absolute C allocation to below‐ground allocation.

[1] To simulate the effect of temperature on soil respiration rates, we translocated soil cores among four sites (3030, 1500, 1000, and 200 m asl) along an altitudinal tropical forest gradient in the Peruvian Andes, traversing a difference in mean annual temperature of 13.9°C. Rates of total (R s ) and heterotrophic (R sh ) respiration were measured twice a month from April 2007 to March 2009 and additionally for full 24 h periods. The diurnal range in R s increased with altitude; this variation was mainly root and litter derived, whereas R sh varied only slightly over full 24 h periods. Although mean annual daytime R s rates were not significantly different among the four sites (4.45-4.05 mmol CO 2 m −2 s −1 ), the annual amount of respired C decreased with increasing altitude from 1639 g C m −2 yr −1 at 200 m asl to 1064 g C m −2 yr −1 at 3030 m asl. The contribution of R sh to R s was not correlated with elevation and ranged from 25% to 60%. The temperature dependence of R s was lower at the midelevation sites (Q 10 of 2.07 and 2.94 at 1500 and 1000 m asl, respectively) than at the highest and lowest sites of the gradient (Q 10 of 4.33 and 6.92 at 3030 and 200 m asl, respectively). The temperature sensitivity of R sh was higher for the sites at 3030 and 200 m asl and increased with time, i.e., with the loss of the most labile C pools.

Tropical soils contain huge carbon stocks, which climate warming is projected to reduce by stimulating organic matter decomposition, creating a positive feedback that will promote further warming. Models predict that the loss of carbon from warming soils will be mediated by microbial physiology, but no empirical data are available on the response of soil carbon and microbial physiology to warming in tropical forests, which dominate the terrestrial carbon cycle. Here we show that warming caused a considerable loss of soil carbon that was enhanced by associated changes in microbial physiology. By translocating soils across a 3000 m elevation gradient in tropical forest, equivalent to a temperature change of ± 15 °C, we found that soil carbon declined over 5 years by 4% in response to each 1 °C increase in temperature. The total loss of carbon was related to its original quantity and lability, and was enhanced by changes in microbial physiology including increased microbial carbon‐use‐efficiency, shifts in community composition towards microbial taxa associated with warmer temperatures, and increased activity of hydrolytic enzymes. These findings suggest that microbial feedbacks will cause considerable loss of carbon from tropical forest soils in response to predicted climatic warming this century.

No abstract