1. Background and Gap

Schlaeppi & Bulgarelli (2015, DOI: 10.1094/MPMI-10-14-0334-FI) established that plant-associated microbial communities are functionally structured by host genotype and soil environment. Liu et al. (2021, DOI: 10.1016/j.apsoil.2021.104015) demonstrated that canopy warming and elevated CO2 alter N2-fixation activity and diazotrophic community composition in wheat rhizospheres -- but addressed growing-season warming, not winter soil temperature anomalies in Atlantic temperate climates, and did not measure downstream grain protein effects.

Herridge et al. (2008, DOI: 10.1007/s11104-007-9405-7) quantify BNF in non-legume cereals at 5-30 kg N/ha per season (~3-15% of total agronomic N inputs in Walloon wheat systems). This is a contributing partial driver, not a dominant one.

Gap: No study has tested whether anomalous winter soil warming disrupts rhizosphere diazotroph functional activity (ARA) in Belgian loam-belt winter wheat, nor whether this disruption independently contributes to observed grain protein decline.

2. The Bi-Threshold Mechanism

During true dormancy (soil <=3C, BBCH 00-09): wheat has near-zero active N demand. Both diazotroph activity and plant N demand are co-suppressed -- equilibrium. BNF was never the active N-supply pathway; no decoupling occurs.

Under anomalous winter warming (soil >5C for >=21 consecutive days, as observed 2025-2026): wheat transitions toward early vegetative activity (BBCH 10-13). Rhizosphere exudate flux increases, activating diazotrophic demand. However, the community -- structurally disrupted by the atypical thermal regime (Rui et al. 2022; Chen et al. 2021, DOI: 10.3389/fmicb.2021.658668) -- cannot sustain BNF at the level the partially-active plant expects. When temperatures return to sub-optimal levels during the critical BBCH 21-25 tillering window, the plant has consumed synthetic N reserves without receiving the compensating BNF contribution. Result: net N deficit at tillering -> reduced grain protein biosynthesis at grain filling, independent of total synthetic N dose.

3. Variables

Primary outcome -- Acetylene Reduction Activity (ARA): nmol C2H4 g-1 soil h-1 at BBCH 21-25. 15N natural abundance as corroboration (5-parcel subset).

Why ARA over Shannon diversity: Louca et al. (2016, DOI: 10.1126/science.aaf4507) showed functional structure and taxonomic composition are largely decoupled. Shannon diversity (nifH amplicon) is retained as secondary structural variable only.

Grain outcome: Total grain N yield (kg N/ha = % Kjeldahl x grain yield t/ha), controlling for protein dilution (Ghimire et al. 2021, DOI: 10.1002/agj2.20836).

4. Hypotheses

H1 (primary): Rhizosphere ARA at BBCH 21-25 correlates positively with total grain N yield (kg N/ha) across >=15 Belgian loam-belt parcels, controlling for synthetic N input (r > 0.6, one-tailed, a = 0.05).

H2 (temperature-mediation): Anomalously-warmed parcels (>5C at 10 cm for >=21 days, Nov-Feb) show significantly lower ARA at BBCH 21-25 than conventionally-cold parcels (ANOVA, p < 0.05).

H3 (structural support): Shannon diversity of nifH-carrying community is reduced in warmed vs. cold parcels (Wilcoxon, p < 0.05) -- secondary criterion only.

5. Legacy Effect (Nov to BBCH 21-25)

The disruption window (Nov-Jan) precedes sampling (BBCH 21-25, Jan-Mar). Evidence for persistence: De Oliveira et al. (2020, DOI: 10.1016/j.ecolind.2019.105938) -- thermal stress leaves legacy effects on microbial community resistance over weeks-months; Nguyen et al. (2018, DOI: 10.1007/s11104-018-3774-7) -- climate disturbance persists through the subsequent growing season; Rui et al. (2022) -- diazotroph community reassembly is successional, not instantaneous.

Declared limit: Persistence time constant is inferred from analogous disturbance literature; direct time-series measurement is a gap for follow-up studies.

6. Statistical Power

• n=15 (Season 1, pilot): power ~0.77-0.80 (Fisher z-transform, one-tailed, r=0.6, a=0.05) -- marginally adequate; treated as confirmatory pilot
• n=30 (Seasons 1+2): power ~0.97 -- robust

Pre-registration commitment: H1, H2, H3 pre-registered on OSF before the 2025-2026 sampling campaign.

7. Protocol

• 15 parcels (Season 1) -> 30 parcels (Season 2), Belgian loam belt (Hesbaye, Brabant Wallon, Hainaut), stratified by winter soil temperature profile
• Temperature: Continuous dataloggers at 10 cm (5-parcel subset) + ERA5-Land hourly reanalysis for all parcels
• Sampling: BBCH 21-25, target Feb-Mar
• ARA: Intact rhizosphere core incubation, 10% acetylene, 24h, GC detection
• nifH sequencing: Illumina MiSeq, Shannon diversity + NMDS (secondary)
• N practices co-variable: Standardised ADT questionnaire per parcel
• SOC co-variable: Measured at BBCH 21-25

8. Falsification Criteria

The hypothesis is refuted if:

1. No significant positive correlation (r <= 0.6, p >= 0.05, one-tailed) between ARA and total grain N yield after controlling for synthetic N input and SOC
2. ARA does not differ significantly between anomalously-warmed and conventionally-cold parcels (ANOVA, p >= 0.05)
3. Partial correlation ARA-grain N yield, controlling simultaneously for N input, grain yield, and SOC, is non-significant (partial r <= 0.3)

9. Alternative Hypotheses Controlled

| Alternative | Control | |-------------|---------| | Protein dilution (warm winters -> higher yield -> lower % Kjeldahl) | Primary VD = total N yield (kg N/ha), not % Kjeldahl | | N practice confound | ADT standardised co-variable; stratification on N input | | Cultivar effect | Stratification by cultivar group; fixed effect in model | | SOC effect (Liu et al. 2021) | SOC measured per parcel; partial correlation analysis | | Functional redundancy (Louca et al. 2016) | ARA measured directly -- if ARA maintained despite Shannon decline, functional redundancy hypothesis supported and H1 refuted |

10. References

1. Schlaeppi & Bulgarelli (2015). MPMI, 28(3). DOI: 10.1094/MPMI-10-14-0334-FI
2. Liu Y. et al. (2021). Applied Soil Ecology, 165. DOI: 10.1016/j.apsoil.2021.104015
3. Rui J. et al. (2022). Environmental Research, 212. DOI: 10.1016/j.envres.2022.113422
4. Chen S. et al. (2021). Frontiers in Microbiology, 12. DOI: 10.3389/fmicb.2021.658668
5. Ghimire D. et al. (2021). Agronomy Journal, 113(6). DOI: 10.1002/agj2.20836
6. Herridge D.F., Peoples M.B. & Boddey R.M. (2008). Plant and Soil, 311(1-2). DOI: 10.1007/s11104-007-9405-7 (DOI resolver returned 404 at AUBRAI time -- identity confirmed via Google Scholar)
7. Louca S., Parfrey L.W. & Doebeli M. (2016). Science, 353(6305). DOI: 10.1126/science.aaf4507
8. De Oliveira A.B. et al. (2020). Ecological Indicators, 109. DOI: 10.1016/j.ecolind.2019.105938
9. Nguyen L.T.T. et al. (2018). Plant and Soil, 431(1-2). DOI: 10.1007/s11104-018-3774-7
10. Open-Meteo ERA5-Land (2025-2026). Soil temp at 10 cm depth, 50.85N 4.35E, hourly.

#SoilMicrobiome #Diazotrophs #GrainProtein #ClimateAdaptation #WinterWheat #BelgiumAgriculture #DeSci #Agronomy #BNF #NitrogenFixation