The mechanisms of leaf litter decomposition were examined using a combination of modeling and field experimental approaches. The modeling components utilized previously reported data from the literature to test two hypotheses generated by the Guild Decomposition Model (GDM; Moorhead and Sinsabaugh 2006), a novel model of microbial based litter decay. First, reported rates of lignin decay (dC3/dt) and holocellulose decay (dC2/dt) were examined in aboveground leaf litter of predominately northern conifer forests to test the hypothesis that the rate of lignin decay is a linear function of the lignocellulose index (LCI = lignin/[holocellulose + lignin]). Simulated dynamics of LCI in decaying litter were highly correlated to observed patterns, particularly when water and ethanol soluble litter fractions from model output were pooled with holocellulose fractions (mean R2 = 0.87 ± 0.02, P < 0.01). More detailed analyses of 64 of these studies yielded variable relationships between lignin decay rate and litter LCI; a regression based on pooled data (N = 385; total number of observations) produced a slope and an intercept that were not significantly different from predicted (slope = 2.33, intercept = -0.93). Both site and litter characteristics had significant effects on the proposed LCI threshold for lignin decay (LCI = 0.4), but no effects on slope or intercept, suggesting that the proposed lignocellulose control hypothesis is relatively robust across a range of litter and forest types.
The next analysis examined reports of microbial biomass associated with decaying plant litter (B:C ratios) in both terrestrial and aquatic systems to test the hypothesis that feedback controls link microbial and litter mass dynamics. Results showed that microbial biomass averaged 2.53% of total system organic matter (microorganisms + litter), but was more than twice as large in aquatic (3.43%, N = 632) as terrestrial (1.05%, N = 384) habitats. Data from a subset of 13 more detailed studies that included multiple observations per experiment showed no difference between habitats. While correlations between observed and simulated values of B:C were highly significant (Spearman’s Rho = 0.316, N = 218, P < 0.01) and Partial Mantel analysis of simulated and pooled observed data found that B:C values were related to litter mass loss, initial lignin content and changing lignin content during litter decay, the model explained only a modest fraction of the total variation in observations (R2 = 0.243, P < 0.01). The available data were insufficient to either conclusively validate or refute the modeling approach used by Moorhead and Sinsabaugh (2006) to mechanistically link the dynamics of microbial biomass and litter decay. Although the most detailed studies showed an increase in B:C values in early decay, followed by a decline in later stages of decay, similar to model behavior (Moorhead and Sinsabaugh, 2006), data were insufficient and too inconsistent between studies to clearly elucidate any pattern. For these reasons, it is suggested that greater insight to mechanistic linkages between decomposer microbial communities and litter decay will require more detailed studies that simultaneously monitor changes in both microbial and litter characteristics. This study found that combining data from disparate studies, which did not examine these factors consistently, provided only limited insight.
My experimental studies examined the general, conceptual model of leaf litter decomposition that predicts increasing litter recalcitrance with mass loss, with nutrient limitations often controlling decay rates in early stages and lignin concentrations dominating the late stages. The activities of extracellular enzymes (EEA) responsible for resource acquisition are predicted to track these changes in litter chemistry, with hydrolytic carbon and nutrient acquiring enzyme activities peaking early in decay and oxidative enzymes responsible for degrading recalcitrant compounds peaking late. As the mechanistic driver of these processes, the microbial community is predicted to track the changes in litter chemistry and EEA, with a shift from a copiotrophic (r-selected) community that utilizes the labile litter substrates early in decay to an oligotrophic (K-selected) community that utilizes the recalcitrant substrates late in decay. However, the conceptual model is based on single species litter experiments whereas natural ecosystems often have mixed species litter. The purpose of the field based study was to examine patterns of changing litter chemistry, EEA and microbial community during decomposition of natural leaf mixtures in two oak dominated forests in northwest Ohio.
Over a two-year period, litter decaying in the urban, Stranahan Arboretum revealed rapid loss of soluble compounds but little evidence of a faster relative loss of holocellulose than lignin contrary to predictions. Similarly, EEA indicated a seasonal pattern more strongly related to litter moisture than litter chemistry. Although the microbial community demonstrated a clear transition between early and late community types, there was an unexpected increase in diversity for the fungal community. In contrast, litter decaying in the nearby but rural, Oak Openings Metropark followed expected patterns of change in litter chemistry, with holocellulose decaying more rapidly than lignin. Additionally, EEA was strongly related to the soluble litter fraction and litter nitrate concentration, consistent with the conceptual model for a nitrogen limited site. Similar to the Arboretum, there was a clear transition between early and late community types, but with a decline in community diversity. Spring flooding of the Arboretum study site, its position in an urban location and a dense population of exotic earthworms (Amynthas agrestis) are postulated to have affected litter decay, EEA and microbial community patterns. Overall, the results suggest that the current conceptual model for leaf litter decomposition based on single species litter experiments is adequate to address patterns of decay in more natural mixes of leaf litter, given site specific factors. Although the experimental data supported the conceptual model for litter decay, no support was found for either the proposed lignocellulose control or microbial to litter mass (B:C) feedback control hypotheses generated by the GDM (data not shown). The experimental data were likely not of sufficient quantity or resolution to provide a conclusive test for simulated predictions.