Objectives  Long-term abandonment of paddy fields in the reddish soil region is not only a waste of soil resources but also poses a risk to the environment due to the high levels of active phosphorous (P) fertilization. Here, we studied the changes in soil P compositions of abandoned reddish paddy fields.

Methods The experiment was conducted from 2007 to 2014 in Taoyuan, Hunan Province. Before abandonment, the soil was under an 16-years rice fertilization experiment (1991–2006), including three treatments: no fertilizer control (CK), application of chemical fertilizer nitrogen and potassium (NK) and application of chemical fertilizer nitrogen phosphorus and potassium (NPK). Topsoil samples were collected from the three treatment plots to determine total phosphorus, available phosphorus (Olsen-P), microbial biomass P (MBP), and P fraction.

Results The abandonment of the paddy field reduced TP and Olsen-P content by 5.4%–23.4% and 11.0%–45.4%, respectively. This corresponds to a reduction of 19.3–160.8 mg/kg total-P and 0.7–14.1 mg/kg Olsen-P. The pre-NPK treatment (high P soil) showed the highest (P<0.05) reduction in total-P (23.4%) and Olsen-P (45.4%) content. The P fractions in the treatments were in the order of Residual-P>NaOH-P_o>NaOH-P_i>Sonic-P_i>NaHCO₃-P_o>HCl-P>Sonic-P_o>NaHCO₃-P_i>Resin-P. All the P fractions in the pre-NPK treatment were higher than in pre-CK and pre-NK treatments (P<0.05) except Resin-P and Sonic-P_o. However, there were no significant differences between the P fractions in the pre-CK and pre-NK treatments (P>0.05). Residual-P was stable P fraction, and abandonment did not affect its content. Labile P (Resin-P + NaHCO₃-P_i + NaHCO₃-P_o), moderately labile P (NaHO-P_i + NaOH-P_o + Sonic-P_i + Sonic-P_o), and low active P (HCl-P) showed a downward trend after abandonment. Moderately labile P recorded the highest reduction (decreased by 10.0–100.8 mg/kg), accounting for 51.7%–78.6% of the total -P reduction. This was followed by active P (decreased by 2.8–29.1 mg/kg), accounting for a 14.5%–18.1% reduction in total-P. The moderately labile P included inorganic P (NaOH-P_i and Sonic-P_i) and organic P (NaOH-P_o and Sonic-P_o). Abandonment decreased the content of inorganic P in all abandoned soils and the organic P in the NPK treatment (P<0.05). The abandoned soil maintained a relatively stable and high MBP content (15.1–16.7 mg/kg). For the abandoned soil with low total-P (pre-abandonment CK and NK treatments), the MBP content was 3.0 times that of Olsen-P. P sequestration by weeds ranged from 21.3 to 48.3 kg/hm², explaining 40.6%–54.9% of the total-P loss in the CK and NK abandoned. However, it only explained 14.9% of the total-P loss in NPK treated abandoned soil.

Conclusions Paddy field abandonment decreased the surface soil P pool. Soils with high total-P content (NPK treatment) before abandonment were more sensitive to abandonment than soils with low total-P levels. Further, total-P, Olsen-P, and P fractions decreased after abandonment. The moderately labile P (NaOH-P) contributed the most to the loss of the P pool (accounting for 64.3%), while the decrease of NaOH-P was mainly due to the reduction of inorganic P fractions (NaOH-P_i and Sonic-P_i). P removal by weeds did not explain the loss of soil total-P, and the microbial biomass P was not affected by abandonment. Therefore, increasing soil organic matter might be a effective way to prevent the loss of the P pool.