In situ edaphic factors affecting seed germination and seedling emergence of three framework species of Acacia were investigated with the intent of developing fundamental and scalable restoration capacity for Arabian dryland restoration. Direct seeding represents the most efficient means to restore vegetation at the landscape scale and this study provides insight into edaphic and ecological limitations, as well as effective protocols governing the use of native seeds for restoration in hyper-arid environments. The study was conducted in extant Acacia woodland habitat on conserved land (Thumamah Nature Park) in close proximity to Riyadh, Saudi Arabia. Broad-scale direct seeding using un- and pretreated Acacia gerrardii, A. tortilis, and A. ehrenbergiana seed, and two seed burial depths were implemented across three sites with distinct soil surface characteristics. Eight weeks post-sowing, random samples for each species × seed treatment × burial depth combination were excavated, sieved, and categorized as follows: failed to germinate, germinated but died prior to emerging, or successfully emerged. We show that germination and emergence of Acacia gerrardii, A. tortilis, and A. ehrenbergiana were driven by a three-way interaction among species, site, and seed burial depth. Treating seed with the signaling compound Moddus did not have a definitive effect, positive or negative, on any of the species investigated. Acacia gerrardii was the only species that exhibited widespread emergence, though emergence was not consistent across sites or burial depths. Germination was highest in disturbed soil (up to 69% for A. gerrardii), but very few (<2%) successfully emerged; a greater proportion of germinants in sandy soil emerged (up to 44% for A. gerrardii) even though the overall germination was less. Though species-dependent, a 2-cm sowing depth was most effective in sand; while in disturbed soil, sowing depths of 1 and 2 cm were comparable; and no germination was observed in gravelly clay soil. Sandy soil exhibited rapid water infiltration (107.6 mm min−1), and post-sowing surface crusting was a non-factor (0.44 kg cm−2). Disturbed soil exhibited moderate water infiltration (1.46 mm min−1) and post-sowing surface crusting was double that of sand (0.88 kg cm−2) and restrictive on seedling emergence. Gravelly clay exhibited extremely poor water infiltration (0.12 mm min−1), and surface crusting was severe (4.49 kg cm−2) and an order of magnitude greater than sand. The medium-coarse sand fraction, a key driver of the observed soil surface processes, was greatest in sand (55%) and significantly less and uniform in the disturbed (22%) and gravelly clay (22%) soils. Our findings demonstrate that soil surface characteristics and associated processes can dictate ecological processes at depths as shallow as 1–2 cm, and that soil crusts that slow water infiltration and impede seedling emergence rapidly reconstitute after disturbance; both are important considerations for restoring dryland vegetation.