Background, aim and scope The production of alumina involves its extraction from bauxite ore using sodium hydroxide under high temperature and pressure. This process yields a large amount of residue wastes, which are difficult to revegetate due to their inherent hostile properties-high alkalinity and sodicity, poor water retention and low nutrient availability. Although phosphorus (P) is a key element limiting successful ecosystem restoration, little information is available on the availability and dynamics of P in rehabilitated bauxite-processing residue sand (BRS). The major aim of this experiment was to quantify P availability and behaviour as affected by pH, source of BRS and di-ammonium phosphate (DAP) application rate. Materials and methods This incubation experiment was undertaken using three sources of BRS, three DAP application rates (low, without addition of DAP; medium, 15.07 mg P and 13.63 mg N of DAP per jar, 100 g BRS; and high, 30.15 mg P and 27.26 mg N per jar, 100 g BRS), and four BRS pH treatments (4, 7, 9 and 11 (original)). The moisture content was adjusted to 55% water holding capacity and each BRS sample was incubated at 25àfor a period of 119 days. After this period, Colwell P and 0.1 M H2SO4 extractable P in BRS were determined. In addition, P sequential fractionation was carried out and the concentration of P in each pool was measured. Results and discussion A significant proportion (37% recovered in Colwell P and 48% in 0.1 M H2SO4 extraction) of P added as DAP in BRS are available for plant use. The pH did not significantly affect 0.1 M H2SO4 extractable P, while concentrations of Colwell P in the higher initial pH treatments (pH 7, 9 and 11) were greater than in the pH 4 treatments. The labile fractions (sum of NH4Cl (AP), bicarbonate and first sodium hydroxide extractable P (N(I)P)) consisted of 58-64% and 70-72% of total P in the medium and high DAP rate treatments, respectively. This indicates that most P added as DAP remained labile or moderately labile in BRS, either in solution, or in adsorbed forms on the surface of more crystalline P compounds, sesquioxides and carbonate, or associated with amorphous and some crystalline Al and Fe hydrous oxides. In addition, differences in the hydrochloric acid extractable P and the residual-P fractions among the treatments with and without DAP addition were relative small comparing with other P pools (e.g., NaOH extractable P pools), further indicating the limited capacity of BRS for fixing P added in Ca-P and other most recalcitrant forms. Conclusions P availability in the original BRS without addition of DAP was very low, mostly in recalcitrant form. It has been clearly demonstrated that significant proportions of P added as DAP could remain labile or moderately labile for plant use during the rehabilitation of bauxite-processing residue disposal areas. There was limited capacity of BRS for fixing P in more recalcitrant forms (e.g., Ca-P and residual-P). Concentrations of most P pools in BRS increased with the DAP application rate. The impact of the pH treatment on P availability varied with the type of P pools and the DAP rate. Recommendation and perspectives It is recommended that the development of appropriate techniques for more accurate estimation of P availability in BRS and the quantification of the potential leaching loss of P in BRS are needed for the accurate understanding of P availability and dynamics in BRS. In addition, application of organic matters (e.g., biosolids and biochar, etc.) to BRS may be considered for improving P availability and buffering capacity.