
Research Background
In arid desert‑oasis regions, agricultural production and ecosystem stability are highly dependent on limited water resources. While farmland shelterbelts mitigate wind erosion, fix sand and improve field microclimate, they compete for water with adjacent crops and alter soil‑water distribution in nearby farmland. Most existing studies evaluate such impacts based solely on soil water content. Knowledge about water‑uptake sources and their overlap between trees and crops remains limited, making it difficult to interpret shifts in soil microbial functions triggered by water competition. How shelterbelt configurations modify crop water‑uptake sources and further affect microbial osmotic adjustment and carbon metabolism still needs clarification.
To fill this research gap, the research team led by Professor Huijie Xiao from Beijing Forestry University took the Populus alba var. pyramidalis‑maize agroforestry system as the research object. By combining hydrogen‑oxygen stable isotope tracing and metagenomic sequencing, the team compared water‑competition patterns under shelterbelts of different stand ages and row configurations, and analysed their correlations with shifts in soil microbial functions. This study provides scientific support for shelterbelt configuration and water‑resource management in arid farmland. The relevant research was published in Applied Soil Ecology.

Figure 1
a) Aerial images of experimental plots obtained by unmanned aerial vehicle (UAV), showing one of four replicates for each plot type (A–F).
b) Schematic diagram showing sampling points and sampling transect layout within each plot.
c) Regression plot of the local meteoric water line (LMWL) and soil‑water line (SWL).
d) Proportions of water sources for shallow‑, middle‑ and deep‑layer soil water calculated from D‑excess values.
e) Flowchart illustrating the overall research framework of this study.
Materials and Methods
Study area and research object: Populus alba var. pyramidalis shelterbelt‑maize agroforestry system in the desert‑oasis zone of Dengkou County, Inner Mongolia.
Experimental design: Six shelterbelt configurations (A‑F, covering different stand ages and row numbers) plus monoculture maize as the control (CK). Each treatment had four replicates, yielding 28 plots in total.
Measured variables: Soil water content, bulk density, pH, soil organic carbon, total nitrogen; hydrogen‑oxygen stable isotopic composition of plant xylem water and different water sources.
Notably, researchers extracted soil water and xylem water from Populus alba var. pyramidalis and maize using the LI-2100 Automated Vacuum Water Extraction System (LICA United Technology Limited, Beijing). Based on ultra‑low‑pressure vacuum distillation and freezing, this equipment enables efficient water extraction and strictly avoids isotopic fractionation caused by evaporation during extraction, guaranteeing high accuracy of hydrogen‑oxygen isotope analysis.
Water‑competition indices: Absolute Competition Intensity (ACI) and Proportional Similarity (PS) were calculated to quantify water competition between shelterbelts and maize.
Soil metagenomic sequencing coupled with KEGG functional annotation was performed to profile microbial functional potential. Differential‑abundance analysis, correlation analysis, partial‑correlation analysis and association‑network analysis were applied to disentangle relationships among shelterbelt configuration, water‑competition intensity and microbial functions.

Figure 2: Soil‑water‑content differences across plots: a) at different soil depths; b) at varying distances from shelterbelts; c) during different maize growth stages.

Figure 3: Maize water‑uptake strategies at three growth stages: a) V6‑V10; b) VT/R1; c) R3‑R4.
Figure 4: a) Spatial distribution of Absolute Competition Intensity (ACI); b) boxplots of Proportional Similarity (PS).

Figure 5: Multi‑group comparisons of gene abundances across plot types and distances from shelterbelts.
Figure 6: a) Correlation fitting between the abundance of four metabolic pathways and water‑related indices; b) multi‑group comparison of pathway abundance; c) two‑group comparison of key metabolic pathways; d) significantly‑differential genes in Glycolysis / Gluconeogenesis and TCA‑cycle pathways.
Key Results
1. Soil water content at 0.3H from shelterbelts was significantly lower than that in monoculture maize plots (CK), whereas no significant difference occurred within the range of 1H‑2H. Such discrepancies were amplified as maize grew. Double‑row and mature shelterbelts generally exerted stronger water‑depletion effects.
2. For plots with shelterbelts, maize gradually shifted its main water‑uptake layer from shallow soil to middle soil as phenology proceeded. During the grain‑filling stage, the contribution of deep‑layer water to maize increased by ~10 % relative to monoculture maize, representing an adaptive response to sustained water competition.
3. Vertical water‑competition layers shifted from shallow soil during the vegetative growth phase to middle soil in reproductive growth; competition involved shallow, middle and deep soil layers at grain‑filling. Double‑row shelterbelts showed higher overall competition intensity than single‑row shelterbelts, while stand age did not yield consistent trends.
4. In shelterbelt‑affected plots, the abundance of betA and betB increased, whereas proB, proC and otsB decreased. These osmotic‑regulation genes were significantly correlated with ACI and PS but showed no significant correlation with instantaneous soil‑water content. This demonstrates that water‑competition patterns exert major control over microbial osmotic‑adjustment strategies.
5. Shelterbelt plots exhibited reduced abundance of Glycolysis / Gluconeogenesis pathways and enhanced abundance of the Tricarboxylic Acid (TCA) cycle. Key genes within these two pathways changed differentially, indicating that microbial carbon‑metabolic potential shifted from rapid carbon utilisation toward long‑term metabolic homeostasis maintenance.
Conclusions
This study reveals that shelterbelts intensify water competition with adjacent crops, driving soil microbes to transition from short‑term stress‑triggered osmotic regulation to long‑term stable adaptation. Microbial carbon metabolism shifts away from rapid carbon input and becomes dominated by TCA‑cycle‑supported homeostasis. Future work should further optimise shelterbelt structural design, combined with staged irrigation and deep‑soil water monitoring, to reconcile crop water demand and soil ecological functions. This supports precise management for sustainable agroforestry systems in arid zones.
Journal: Applied Soil Ecology (Impact factor: 5.6)
Institutions: Beijing Forestry University; Inner Mongolia Dengkou Desert Ecosystem Observation and Research Station, etc.
Study site: Desert‑oasis area of Dengkou County, Inner Mongolia
Instrument: LI-2100 Automated Vacuum Water Extraction System
DOI: https://doi.org/10.1016/j.apsoil.2026.107324