Soil Microbiomes Across Depth and Ecosystems in Dubai, UAE: Potential Environmental Signatures for Forensic Geolocation.

Soil microbial communities exhibit strong sensitivity to environmental gradients, yet their distribution across depth and land-use types in hyper-arid environments remains poorly characterised. Using whole-genome shotgun metagenomics via Oxford Nanopore Technologies long-read sequencing, we profiled soil microbial communities across six contrasting land-use sites in Dubai, UAE: urban, industrial (two locations), marine, desert and agricultural, where each sampled at three depth intervals (0-25 cm, 25-50 cm and 50-100 cm). Marine soils exhibited extreme salinity (EC 23.7-30.7 dS m-1) and the highest organic matter content (1.19%-1.76%), while desert soils were nutrient-poor with minimal salinity. Actinomycetota and Pseudomonadota co-dominated across all sites, collectively accounting for 77%-96% of classified sequences. Actinomycetota prevailed in undisturbed desert horizons (up to 53.4%), while Pseudomonadota dominated nutrient-enriched environments, reaching 69.4% at industrial sites. A notable compositional reversal was observed in the desert deep horizon (50-100 cm), where Pseudomonadota increased to 56.8%, departing from the expected oligotrophic depth gradient. PERMANOVA confirmed land use as the primary driver of community composition (p = 0.001), with depth exerting a secondary but significant effect (p ≤ 0.01). NMDS ordination revealed strong site-specific clustering, with each environment harbouring a distinctive microbial fingerprint with promising forensic geolocation potential.