Effect of phosphorus management in rice-mungbean rotations on sandy soils of Cambodia

Nitrogen (N) and phosphorus (P) deficiencies are key constraints in rainfed lowland rice (Oryza sativa L.) production systems of Cambodia. Only small amounts of mineral N and P or of organic amendment are annually applied to a single crop of rainfed lowland rice by smallholder farmers. The integration of leguminous crops in the pre-rice cropping niche can contribute to diversify the production, supply of C and N, and contribute to soil fertility improvement for the subsequent crop of rice. However, the performance of leguminous crops is restricted even more than that of rice by low available soil P. An alternative strategy involves the application of mineral P that is destined to the rice crop already to the legume. This P supply is likely to stimulate legume growth and biological N-2 fixation, thus enhancing C and N inputs and recycling N and P upon legume residue incorporation. Rotation experiments were conducted in farmers' fields in 2013-2014 to assess the effects of P management on biomass accumulation and N-2 fixation (N-15) by mungbean (Vigna radiata L.) and possible carry-over effects on rice in two contrasting representative soils (highly infertile and moderately fertile sandy Fluvisol). In the traditional system (no legume), unamended lowland rice (no N, + 10 kg P ha(-1)) yielded 2.8 and 4.0 t ha(-1), which increased to 3.5 and 4.7 t ha(-1) with the application of 25 kg ha(-1) of urea-N in the infertile and the moderately fertile soil, respectively. The integration of mungbean as a green manure contributed up to 9 kg of biologically fixed N (17% Nfda), increasing rice yields only moderately to 3.5-4.6 t ha(-1). However, applying P to mungbean stimulated legume growth and enhanced the BNF contribution up to 21 kg N ha(-1) (36% Nfda). Rice yields resulting from legume residue incorporation (green manure use-all residues returned and grain legume use-only stover returned) increased to 4.2 and 4.9 t ha(-1) in the infertile and moderately fertile soil, respectively. The forage legume use (all above-ground residues removed) provided no yield effect. In general, legume residue incorporation was more beneficial in the infertile than in the moderately fertile soil. We conclude that the inclusion of mungbean into the prevailing low-input rainfed production systems of Cambodia can increase rice yield, provided that small amounts of P are applied to the legume. Differences in the attributes of the two major soil types in the region require a site-specific targeting of the suggested legume and P management strategies, with largest benefits likely to accrue on infertile soils.