Nutrient availability in soil samples from shifting cultivation sites in the semi-arid Caatinga of NE Brazil

Soil fertility differed substantially among three neighboring sites representing different stages of a shifting cultivation cycle on a deeply weathered Oxisol on the Chapada de Araripe in the semiarid region of NE Brazil. Samples taken from a native thorn forest area (Caatinga), from a recently slashed and burnt area (Burnt) and from a site abandoned after five years of manual cultivation to cassava (Abandoned), were compared using subtractive and additive fertilization greenhouse trials and incubations for C and N mineralization. The effect of burning was to increase soil nutrient content relative to the Caatinga area samples, followed by decreases of N and organic P by 20%: available P by 70%, and exchangeable bases by 55%, on average, after abandonment. The unfertilized control of the missing-element trial gave the lowest dry matter (DM) yields, but still those of the Burnt were twice as large as those of the Abandoned area samples. The relative yield ratios were not changed by nutrient addition, although total yields were increased by a factor of 5. In all three areas, the most severe nutrient limitation was of P, followed by N. The N deficiency somewhat increased relative to P in the Abandoned samples. Addition of 25 mg P kg(-1) soil produced 4 to 5-fold dry matter increases in the Caatinga and 7 to 8-fold increases in the Burnt area samples. Further increasing P levels improved yields only marginally and only at high additions of N and K. In samples from the abandoned area, 25 mg P kg(-1) not even doubled the dry matter yield, and no additional effects were observed for higher N and K rates. Carbon mineralization during the 12 week incubation followed the order Caatinga > Abandoned > Burnt, whereas the order for N removed in 13 leachings was Caatinga > Burnt > Abandoned. Accordingly, organic matter mineralized from the abandoned soil had a higher C:N ratio. In addition to the nutrient limitations of this Oxisol, samples from the Abandoned area exhibited further limitations resulting in yield reductions, which were not explained by any of the factors examined. This indicates that alternatives to long-fallow shifting cultivation will require further refinement of methodologies to measure and predict soil quality. (C) 1997 Elsevier Science B.V.