Substrate Quality and Quantity Dominate Temperature Effects on Microbial Community Temperature Response

Alster et al (2020) proposed that 'ΔC‡ P' and Topt are the temperature response traits most relevant to test hypotheses related to the thermal adaptation of microbial communities. We hypothesized that Topt and 'ΔC‡ P' would both increase with warming due to a thermal selective pressure leading to increased abundance of microbes producing warm-adapted (rigid) enzymes to constrain reaction rates when temperature increases (Hochachka & Somero, 2002), in line with an attenuating effect of thermal adaptation on soil C losses (Bradford et al, 2019). However, Topt and 'ΔC‡ P' did not vary with MET, so we must reject this hypothesis.

The variables with the most influence on the parameters derived from MMRT was C concentration. However, soil C:N ratios and N concentrations were positively correlated with C concentration, and so we are unable to distinguish between the influence of these variables on the MMRT parameters. Nonetheless, our findings can explain the positive influence of C concentration on the respiration rate at 25°C (R25) and 'ΔH‡ T0', a parameter related to the magnitude of change in respiration rates. Substrate deprivation leads to decreased metabolism and respiration in heterotrophic microbes (Bradford, 2013). As a result, the response of microbial respiration to temperature is dependent on substrate availability, with an increase in the response as C concentration increase and substrates become more abundant (Davidson & Janssens, 2006). Furthermore, higher substrate availability and respiration rate are positively correlated with microbial biomass (Allison et al, 2010; Bradford, 2013), so, an increase in microbial biomass with higher C concentration and substrate availability in the treatments at our site would also explain increases in R25 and 'ΔH‡ T0'.

Topt and 'ΔC‡ P' were also influenced by C concentration. Again, this effect could be attributed to the correlation with C:N ratios (or N concentrations). Both a positive influence of C:N ratios and C concentration on Topt and 'ΔC‡ P' can be supported theoretically. Indeed, when substrates are accessible by microbes, there is a positive relationship between the recalcitrance of substrates (of which C:N is often taken as an indicator) and the temperature sensitivity of their decomposition (Conant et al, 2011; Davidson & Janssens, 2006; Fierer et al, 2005). Decreases in the substrate quality due to the depletion of labile substrates have been proposed as an explanation for compositional shifts in microbial communities and increases in temperature sensitivities that occur as an indirect consequence of warming (Bai et al, 2017; Karhu et al, 2014; Pold et al, 2015).

Moreover, the temperature response of microbial decomposition is constrained by both biological enzymatic reaction and chemical reactions regulating substrate exchange between the solid and aqueous phases of the soil (Conant et al, 2011). Numa et al (2021) and Schipper et al (2019) argued that the resulting temperature response would be a combination of an Arrhenius-driven temperature response of sorption/desorption and diffusion and an MMRT-driven response of the biological process. If C concentration at our site were positively related to physicochemical protection of C substrate (Kirschbaum et al, 2020), an increasing contribution of Arrheniusdriven reactions could have resulted, and therefore a lower observed 'ΔC‡ P' (Schipper et al, 2019). Having disentangled the effects of warming from those of substrate quality and quantity (MET varied independently from C concentration and C:N ratios), our data suggest that shifts in substrate quality and/or quantity may exert a selective pressure greater than that for temperature in the composition of communities with distinct temperature response parameters and relative temperature sensitivities.