ABSTRACT
Potassium (K) nutrition of rice-wheat (RW) systems of the Indo-Gangetic Plain (IGP) of South Asia is important because of its role in productivity and the large quantities of this macronutrient that are extracted by such intensive cropping systems. Field experiments on the RW cropping sequence were conducted at three locations in Bangladesh with three soil types. Two fertilizer doses—farmers' practice (FP) and soil-test based (STB)—of nitrogen (N), phosphorus (P), K and other nutrients were combined with mungbean or maize as a third crop. The objective of the experiments was to detect K deficiency, if any, in rice, wheat, mungbean, and maize, and to compare the FP- and STB-based sequences in terms of the K nutrition of those crops and the apparent K balance in soil. Frequent K deficiency was observed in rice and wheat at all sites, especially at Ishwordi, while maize was less affected and mungbean not deficient. There was a significant effect of fertilizer on K uptake by maize, mungbean, and rice, but little effect of the retention of mungbean residue on K uptake by crops at any site. Mean annual system-level K uptake was greatest at Ishwordi (126–239 kg ha−1) and least at Joydebpur (64–116 kg ha−1). The majority of K uptake was in straw and the proportion in grain varied little across sites (range: 11%–29%). There were large negative apparent K balances in all treatments at all sites (range: −25–212 kg ha−1), with the greatest at Ishwordi and the smallest at Joydebpur. Soil K balance responded differently to the retention of residues across soils, and positive effects could be observed on clayey soils. Long-term experiments will be required to monitor soil and plant K dynamics under various fertilizer and residue management of crops in RW systems of the IGP.
ACKNOWLEDGMENTS
The study was part of the collaborative research project, “Water and Nutrient Management for Sustainable Rice-Wheat Cropping Systems in Bangladesh and Australia,” jointly conducted by the University of Melbourne, Bangladesh Agriculture Research Institute (BARI), Bangladesh Rice Research Institute (BRRI), and CSIRO. The study received funding support from the Australian Centre for International Agricultural Research (ACIAR) as a bilateral research grant for Bangladesh. We thank the many researchers and technicians at BARI and BRRI in Bangladesh who took part in various discussions and helped with various aspects during the project period. We also thank Mark Conyers, Robert Edis, and Himanshu Pathak, who offered valued suggestions on an earlier draft that have improved the manuscript.
Notes
1 Source: Rice (CitationYoshida, 1981); wheat (CitationReuter et al., 1997); maize (CitationRoberts and Rhee, 1993); mungbean (CitationReuter et al., 1997).
1Least significant difference at α = 0.05.