Phosphorus status and cycling in native savanna and improved pastures on an acid low-P Colombian Oxisol

On acid low-phosphorus (P) Colombian Oxisols, improved pastures with acid-s oil-tolerant grass and legume varieties have increased beef production by a factor of 10 to 15 with only modest P fertilizer inputs. This indicates th at the efficiency of P fertilization could be greater than is commonly expe cted on such strongly P-sorbing soils. To understand the effect of improved pastures on P cycling and availability, we estimated P budgets, and charac terized soil P by sequential fractionation, isotopic exchange and biologica l activity measurements on soil samples from unfertilized native savanna, a nd fertilized improved grass-only (Brachiaria decumbens cv. Basilisk) and g rass-legume (B. decumbens + Pueraria phaseoloides, Kudzu) pastures establis hed in 1978 on a medium-textured isohyperthermic, tropeptic haplustox. Comp arison of calculated P budgets, based on inputs and exports, with total soi l P contents showed that fertilization, as part of the improved pasture man agement, had resulted in a measurable increase of total P in the surface 0- 20 cm soil layer of nearly 30 mg kg(-1) or about 20% over the savanna level . Sequential soil P fractionation of different seasonal samplings indicated that grass-legume maintained higher organic and available inorganic P leve ls with less temporal variation than the two other types. The linkage of or ganic P and available P was also reflected in soil biological activity. Est imates of P in microbial biomass and phosphatase activity were significantl y higher in grass-legume than grass-only and savanna. The improvement in so il P availability, as measured by solution P concentration, P sorption and exchangeable P, was much greater in grass-legume than in grass-only. With c omparable fertilizer inputs and greater product exports, improved P availab ility in grass-legume cannot be due to differences in budgets but can be at tributed to changes in the overall biological activity in the soil-plant sy stem caused by the presence of legumes in the vegetation cover. Total C, or ganic P content and macrofaunal activity were all significantly higher in g rass-legume soils. Greater turnover of organic litter in grass-legume may p rovide for steadier organic P inputs and, therefore, higher P cycling and a vailability.