**Background:** Floodplain ecosystems are highly productive but increasingly degraded by climate change and human activities. Shrub vegetation restoration, particularly with Tamarix chinensis Lour. (northern Tamarisk), is used to combat soil erosion and land degradation. Shrubs can create 'fertile islands' beneath their canopies, enriching soil nutrients and altering microbial communities. However, the effects of shrub patch size and sampling location on soil microbial community composition in the middle Yellow River floodplain were not well understood. This study aimed to determine how microbial community composition varies with shrub patch size and sampling location, and to identify the underlying mechanisms influencing microbial responses to T. chinensis plantations.
**Methods:** The study was conducted at the Yellow River Floodplain Ecosystems Research Station in Xingyang County, Henan Province, China (34°59′65″ N, 113°25′05″ E). The region has a mean annual temperature of 14.3°C, mean annual precipitation of 645.5 mm (65% falling July–September), and sandy soil with pH ~8.10. A randomized block design with four 20 m × 20 m sites was used, selecting three shrub patch sizes: small (<100 cm canopy diameter), medium (100–200 cm), and large (>200 cm). Twelve T. chinensis plants were selected (four per size category). Soil samples (0–15 cm depth) were collected from two locations: inside-canopy (mid-radius) and outside-canopy (1.5× canopy radius). Four subsamples per radius were composited, yielding 24 mixed samples. Soil physicochemical properties (SWC, pH, SSC, SOM, TN, AP, particle size) were measured using standard methods. Microbial community composition was characterized via phospholipid fatty acid (PLFA) analysis, identifying total PLFAs, bacteria, fungi, Gram-positive (GP) bacteria, Gram-negative (GN) bacteria, and arbuscular mycorrhizal fungi (AMF). Two-way ANOVA, one-way ANOVA, paired t-tests, Pearson correlations, and redundancy analysis (RDA) were performed.
**Key Results:** Shrub basal diameter, height, and canopy diameter significantly increased with patch size (p < 0.001). Soil salt content (SSC), soil organic matter (SOM), total nitrogen (TN), and available phosphorus (AP) varied significantly with shrub patch size across both sampling locations (all p < 0.05). In inside-canopy soils, SSC, SOM, TN, and AP increased by 86.51%, 75.57%, 38.12%, and 39.03%, respectively, from small to large patch size. In large patches, inside-canopy SSC, SOM, and TN were 155.47%, 80.33%, and 43.17% higher than outside-canopy soils. AP was 20.23%, 24.50%, and 54.64% higher inside-canopy than outside for small, medium, and large patches, respectively (all p < 0.05). Soil pH decreased with patch size; SWC was not affected. Microbial PLFAs increased with patch size: in inside-canopy soils, total PLFAs, bacterial, GP bacterial, GN bacterial, fungal, and AMF PLFAs increased by 41.07%, 43.61%, 29.33%, 53.72%, 28.58%, and 127.66% from small to large patches (all significant). In outside-canopy soils, corresponding increases were 28.77%, 27.09%, 13.32%, 27.46%, 39.04%, and 132.43% (all p < 0.05). Fungal and AMF PLFAs significantly differed between inside- and outside-canopy soils across all three patch sizes. Fungal PLFAs were 33.42%, 10.69%, and 23.38% higher inside-canopy for small, medium, and large patches; AMF PLFAs were 80.66%, 49.71%, and 76.94% higher inside-canopy. The fungi:bacteria ratio decreased with patch size inside-canopy but increased outside-canopy. Pearson correlations showed total, bacterial, fungal, GP, GN, and AMF PLFAs were positively correlated with clay, silt, SSC, SOM, TN, and AP, and negatively correlated with sand and pH. RDA revealed that soil physicochemical properties explained 76.10% of variance in inside-canopy soils (axis 1: 68.67%; axis 2: 4.69%) and 48.50% in outside-canopy soils (axis 1: 34.78%; axis 2: 9.26%). SOM was the most significant variable, explaining 61.90% of PLFA variation in inside-canopy soils, followed by AP.