Climate change is altering precipitation patterns, increasing drought and flood frequencies. Water significantly impacts soil carbon dynamics, making it crucial to understand how moisture disturbances affect carbon availability and fluxes. Antecedent moisture conditions strongly influence soil CO2 fluxes, a phenomenon amplified during rewetting (the "Birch effect"). The mechanisms behind this effect are complex, involving physicochemical and biochemical processes that increase microbial oxidation of soil organic carbon (SOC). Soil solution ionic strength varies greatly with moisture content, affecting carbon molecule desorption from minerals. Soil moisture also influences microbial activity; drought causes osmotic stress, while saturation improves SOC and nutrient access. This research aims to understand the impacts of extreme soil water conditions (drought and flood) on soil carbon cycling across diverse soils, attributing these impacts to microbial, chemical, or physical soil traits. Two hypotheses were tested: (i) drought will increase complex aromatic carbon species and osmoprotectant expression; (ii) extended saturation will increase motility factor expression.

Previous research highlights the significant impact of antecedent moisture conditions on soil carbon dioxide (CO2) fluxes, particularly the increased fluxes observed during rewetting events, known as the "Birch effect." The mechanisms behind this are complex and not fully understood. Studies have shown that varying moisture conditions lead to different chemical environments at the pore scale, affecting carbon molecule interactions with minerals and their solubility. Drought stresses soil microbes through osmotic stress, while saturation enhances their access to SOC and nutrients. The connectivity of the soil pore network plays a crucial role in microbial access to resources, with reduced connectivity in field-moist soils limiting exchange. Existing literature suggests that microbial communities are adapted to their specific environments, and thus, the responses to drought and flood are likely to be site-specific.

Soil cores were collected from three locations with differing climates and soil properties: a silty soil from Alaska, a clay loam soil from Washington (subject to frequent inundation), and a sandy soil from Florida (subject to drying-wetting fluctuations). Cores were randomly assigned to one of three treatments: field-moist (control), drought (air-dried to constant weight), and flood (saturated). Incubations were conducted at 21°C in the dark for 30 days. Two additional controls were included: baseline field-moist and time-zero saturation. CO2 and CH4 production were monitored. Pore water was extracted at different suctions to represent different pore sizes. Dissolved organic carbon (DOC) was analyzed using combustion catalytic oxidation. Molecular composition of DOC was characterized by electrospray ionization coupled with Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR-MS). DNA and RNA were extracted to analyze metagenomes and metatranscriptomes, respectively. Soil properties (pH, electrical conductivity, nutrients, particle size distribution, and pore size distribution) were also determined. Statistical analyses included ANOVA, MANOVA, PERMANOVA, LEfSe, and DESeq2. The data were analyzed using R.

Drought had a stronger effect on soil responses than flooding. Alaska soils showed increased pore-water DOC under drought, while Florida soils showed a decrease. Washington soils showed increased DOC across all treatments, suggesting a sampling effect. Molecular diversity of DOC decreased during incubation except for drought-incubated Alaska soils. FT-ICR-MS analysis revealed site-specific differences in DOC composition. Alaska soils were dominated by lignin-like molecules, Florida soils by amino sugars and carbohydrates, and Washington soils by protein- and lipid-like molecules. The Alaska soils responded most strongly to drought, showing increased oxidized molecules. Florida soils showed loss of simple aliphatic peaks and an increase in complex molecules across treatments. Washington soils showed minimal changes in DOC composition. Metatranscriptome analysis showed site-specific responses to treatments. Drought increased expression of sporulation and osmoprotectant synthesis genes in the Alaska and Washington soils. Flood also increased sporulation gene expression. Drought decreased gliding motility genes in Alaska soils but increased flagellar motility genes. Washington soils showed high gliding motility gene expression even under drought due to high residual water content. Soil texture and disturbance history significantly influenced the responses. Alaska soils (silt loam) showed greater potential for SOC destabilization under drought. Florida (sandy) soils showed microbial adaptation to moisture extremes. Washington (clay loam) soils, with high initial moisture, showed minimal response to drought and flood.

The findings partially support the initial hypotheses. Drought increased sporulation and osmoprotectant expression, but didn't uniformly increase complex aromatic compounds. The study highlights the significant influence of soil texture and environmental history on soil responses to moisture fluctuations. Soil texture affects pore size distribution and water availability, influencing chemical responses. Environmental conditions and disturbance history drive microbial responses. The concept of ecological stress is relevant, as changes in the microenvironment may not always be stressful for an adapted system. The varying responses across sites underscore the challenges of generalizing soil responses to moisture extremes.

This study demonstrates the complexity of soil responses to drought and flooding, influenced by soil texture, pore network structure, and microbial community adaptation. Drought had a greater impact than flooding. Future research should investigate the specific mechanisms driving the observed site-specific responses and consider a wider range of soil types and environmental conditions.

The relatively small sample size and the focus on three specific sites might limit the generalizability of the findings. The study did not explicitly account for factors such as mineralogy and litter quality, which could influence soil responses. Laboratory incubation may not perfectly replicate field conditions.