Organic Matter Composition as a Driver of Soil Bacterial Responses to Pig Carcass Decomposition in a Canadian Continental Climate

Abstract

Organic by-products are released into the surrounding soil during the terrestrial decomposition of animal remains. The affected area, known as the Cadaver Decomposition Island (CDI), can undergo biochemical changes that contribute to landscape heterogeneity. Soil bacteria are highly sensitive to labile inputs, but it is unknown how they respond to shifts in dissolved organic matter (DOM) quantity and quality resulting from animal decomposition. We aimed to evaluate the relationship between soil DOM composition and bacterial activity/function in CDIs under a Canadian temperate continental climate. This was studied in soils surrounding adult pig carcasses (n = 3) that were surface deposited within a mixed forested environment (Trois-Rivières, Québec) in June 2019. Using fluorescence spectroscopy and dissolved organic carbon analyses, we detected a pulse of labile protein-like DOM during the summer season (day 55). This was found to be an important driver of heightened soil bacterial respiration, cell abundance and potential carbohydrate metabolism. These bacterial disturbances persisted into the cooler autumn season (day 156) and led to the gradual transformation of labile DOM inputs into microbially sourced humic-like compounds. By the spring (day 324), DOM quantities and bacterial measures almost recovered, but DOM quality remained distinct from surrounding vegetal humic signals. All observed effects were spatially constrained to the topsoil (A-horizon) and within 20 cm laterally from the carcasses. These findings provide valuable insight into CDI organic matter cycling within a cold-climate ecosystem. Repeated CDI studies will however be required to capture the changing dynamics resulting from increasing global temperatures.

Keywords: bacterial metabolism; cadaver decomposition island; carrion; dissolved organic matter; soil biogeochemistry; soil microbial ecology.

Figure 1

Observed seasonal trends in DOC concentration (a) and relative percent contribution of PARAFAC components C1 (b), C2 (c), C3 (d) and C4 (e) in soil of the A‐horizon sampled at varying distances from decomposing pig carcasses (n = 3). Values are expressed as the mean and standard error. Control samples taken outside of the experimental site are included as the furthest distance from the carcass.

Figure 2

PCA biplot of seasonal DOM composition (DOC, FI, HIX, BIX, SR, %C1–%C4) in soil of the A‐horizon sampled at varying distances from decomposing pig carcasses (n = 3). PCA includes 95% confidence ellipses for labeled points that were sampled closest to the carcasses (−20, 0 and 20 cm).

Figure 3

Seasonal trends in bacterial carbon respiration (a) and bacterial cell abundance (b) in soil of the A‐horizon sampled at varying distances from decomposing pig carcasses (n = 3). Values are expressed as the mean and standard error. Control samples taken outside of the experimental site are included as the furthest distance from the carcass.

Figure 4

PCA score (a) and loading (b) plots of seasonal substrate metabolic potential of bacteria in soil of the A‐horizon sampled at varying distances from decomposing pig carcasses (n = 3). Substrates are categorized into their respective chemical classes: carbohydrates (green C1‐C10), amino acids (blue AA1‐AA6), carboxylic acids (orange CA1‐CA7), polymers (purple P1‐P4), amines (gray A1‐A2) and phenolic compounds (yellow PC1‐PC2). PCA score biplot (a) includes 95% confidence ellipses for the labeled points sampled closest to the carcasses (−20, 0, 20 cm).

Figure 5

PLS‐R biplot (a) of bacterial responses (activity and function) in relation to DOM composition in soil of the A‐horizon sampled at varying distances from decomposing pig carcasses (n = 3). Linear regression (b) of PLS‐R t1 and u1 scores categorized by sampling distance and season. Score plots for DOM composition variables of importance to projection (VIP) that contributed to the t1 (c) and t2 (d) components of the PLS‐R model.