Successful afforestation in arid and semi-arid environments depends on seedling survival during establishment, yet the relative effects of irrigation method, soil amendment, and species identity remain poorly quantified at equivalent water volumes. We conducted a factorial field experiment in semi-arid Morocco (285 mm long-term mean annual precipitation; 86-174 mm during the study) with 3024 seedlings across seven native and exotic tree species, monitored over 50 months under three irrigation regimes (subsurface drip, tank, rainfed) and three soil amendments (5% biochar, 25% peat moss, unamended control). Water availability dominated establishment success. Relative to drip irrigation, tank irrigation increased mortality risk 2.7-fold despite comparable water volumes, while rainfed conditions increased risk 12.4-fold. Drip and tank irrigation produced similar mean soil water potential but contrasting oscillation amplitudes (2-13 vs. 27-35 kPa), indicating that delivery consistency governed establishment. Most mortality occurred during the first post-planting dry season. Fast-growing exotics (Paulownia, Moringa) showed 16- to 73-fold higher mortality risk and near-complete rainfed mortality, whereas native species (Pistacia, Pinus, Vachellia) maintained 71-85% survival despite severely reduced growth (final heights 40-170 cm lower than under drip irrigation). Tank-drip differences diminished over time for natives, suggesting simpler irrigation may suffice for drought-adapted taxa. Amendment effects were water-dependent: under rainfed conditions, biochar and peat moss reduced mortality risk by 51% and 39%, with biochar effects persisting and peat moss attenuating after year 2; both were negligible under irrigation. Early afforestation success is thus governed primarily by water delivery consistency, modulated by species identity, with amendments playing a secondary, context-dependent role.