Ecosystem carbon balance of temperate forests differing in elevation and nitrogen availability

This PhD thesis addressed the carbon (C) balance of temperate deciduous forests across natural gradients of temperature and nitrogen (N) availability, the major drivers of net primary production (NPP) and the soil C balance. A mean annual temperature difference of 6 K across a 1200 m change in elevation from the Swiss Plateau to the Central Swiss Alps, and the presence or absence of the N2-fixing tree species Alnus glutinosa or Alnus incana within each elevation, offered the framework (1) to test the hypothesis that cumulative annual soil respiration (Rs) at contrasting temperatures reflects the difference in the production of short-lived biomass in the longer run; (2) to test whether or not high rates of N inputs increase the rate of C-cycling by accelerating both, NPP and Rs; (3) to analyze the composition of soil organic C pools and their contribution to Rs in response to elevation and N-availability; and (4) to test whether high-N-input Alnus forests stimulate soil N transformations, resulting in enhanced N2O emissions.
Forest soil respiration reflects plant productivity across a temperature gradient in the Alps
In forests, the biomass components of interest for Rs are those undergoing rapid recycling, i.e. litter production by the canopy, understory and fine root system. Despite the 6 K difference in temperature and a difference in the length of the growing season of three months moving from the high to the low elevational sites, total annual litter production did not change. This is surprising, in view of the estimated doubling of annual wood increment from high to low elevation, that largely resulted from the difference in the length of the growing season. Stem growth and total NPP signals almost disappeared when expressed per day available for growth. Although following temperature variability throughout the seasonal course, cumulative annual Rs did not differ across elevations on a full-year basis. Within each elevation, the short-term temperature response of Rs (Q10) was close to 2, which is in accordance to the often assumed more than doubling in respiration for a 10 K warming. However, when calculated across sites, i.e. from high to low elevation, the apparent Q10 dropped to ~1.2, implying a down-regulation of Rs at higher temperatures, close to homeostasis. In other words, across the sites adapted to different temperatures, temperature was not exerting a strong net influence on Rs. Adopting a simple C budget that assumes 50% of total Rs is derived from autotrophic root respiration, we arrive at c. 40% of the respiratory soil C release from concurrent litter production, both for high- and low-elevational sites. Whereas the remaining unaccounted 10% are a reasonable estimate for root exudates and mycorrhizal consumption in temperate forests. Cumulative annual soil CO2 release thus largely reflected the input of labile C to soil, and not temperature per se. Climatic warming of the past decades most probably was slow enough, so that metabolism could track it, causing no significant deviation from a thermal equilibrium at our test sites. These results caution against expectations of strong positive effects of climatic warming on Rs.
Does nitrogen input enhance respiratory carbon release from temperate forest soils?
In forests with the symbiotically N2-fixing genus Alnus forest, we found that biological N2-fixation enhanced total litter production (the sum the above- and belowground forest litter components) and facilitated higher Rs at low elevation only. At high elevation, enhanced N input was associated with lower litter production and lower Rs, compared to non N2-fixing stands.
Hence, Rs remained in proportion to forest litter production, irrespective of the effect of N-availability or site temperature. Annual litter C production and annual soil C release via Rs correlated well (r2 = 0.85), at a ratio of litter C to annual C release of 2.4:1. Thus, total plant litter C explained ~ 40% of Rs. Assuming a ~ 10% contribution of C input to soils by root exudates and micorrhizal C consumption, the balance of C input in soil through NPP and the C output from soils via Rs would become closer to 2:1. Hence, autotrophic and heterotrophic soil respiration, contribute similar fractions to total Rs. In conclusion, the results of this comparison lines up with the results of chapter 2, suggesting that Rs exhibits a rather robust relationship to substrate availability, rather than showing direct responses to N availability and temperature. These findings offer no justification for modelling Rs by assuming either N or temperature to exert direct (independent) effects on Rs. In the long run though, the soil C pool could still be affected by small deviations from this relationship, provided element stoichiometry permits.
Soil organic carbon pools and their contribution to soil respiration
In this study we compared C release from forest C pools by means of a 600-day incubation, that revealed information about the contribution of three soil C pools to total Rs, and how these respiratory components related to the size of the respective C pools. We identified three heterotrophic sources sources of C for Rs: rapidly cycling rhizosphre-derived C with a maximum turnover time of a few days, labile soil C turning over within a few weeks to months, and recalcitrant soil C with a residence time of several years. We found that more than 90% of total Rs is explained by the sum of autotrophic (root) respiration, rhizosphere-derived C and the contribution of soil C fractions that decompose within a growing season. The soil C fractions turning over within this relatively short time, however, accounted for a minor part (<5%) of total soil C. In other words, Rs was directly linked to short-lived NPP. The contribution of recalcitrant C to total Rs in contrast, was virtually negligible (1-7%), although this fraction accounted for the major part of total soil C (>95%). We thus conclude that Rs, in essence emerging from C pools turning over within days to months, cannot increase under warming to any significant extent independently from NPP. However, for long-term changes in soil C, small warming-induced changes of old soil C fractions can potentially become relevant.
N2 fixation by Alnus tree species enhances forest soil N2O emissions
The comparison of fluxes of greenhouse gases from soils of N2-fixing and adjacent control forest plots clearly revealed in enhanced soil N transformation and subsequent N2O emissions in high N-input Alnus stands. Concurrently, productivity data, recorded for the study presented in chapter 3, indicated that all Alnus stands reached N saturation. Obviously, the high N inputs with N2-fixation turned N-cycling into an open cycle, with higher losses of gaseous N to the atmosphere. While the N2O emissions were generally higher at lower elevations than in the 6K colder forest stands in the Alps, the effect of N2 fixation was also more pronounced at lower elevation. Yet, with an increase in N2O losses of 330% and 250% in Alnus stands relative to controls though, the effect was considerable at both elevations. While soil CH4 exchange did not show a conclusive effect of N2 fixation, soils at the high elevation were CH4 sources and soils at low elevation were CH4 sinks over a complete season. The findings of this study suggest that N losses from soils to the atmosphere in form of N2O can substantially increase, once the biological demand for N reached saturation.
The main conclusions that can be drawn from this thesis are: (1) cumulative annual C release from forest soils largely reflects the input of labile C to soil, and not temperature per se; (2) this robust relationship of Rs to substrate availability also holds across forest sites with high N-inputs, such as under N2-fixing trees; and (3) N2O losses from forest soils to the atmosphere can substantially increase, once the biological demand for N reached saturation.