Da. Mcgrath et al., Nitrogen and phosphorus cycling in an Amazonian agroforest eight years following forest conversion, ECOL APPL, 10(6), 2000, pp. 1633-1647
Commercial plantation agroforestry systems have emerged as a promising Amaz
onian land use with the potential to reduce soil degradation, improve livin
g standards, and decrease pressures on remaining forested areas. While it i
s generally accepted that tree-based agroecosystems have more closed nutrie
nt cycles that help conserve soil productivity, few studies of nutrient dyn
amics in Amazonian agroforestry systems exist, making it difficult to asses
s the extent to which this land use-is sustainable in the region's weathere
d soils. We quantified net primary productivity (NPP) and the stocks and fl
uxes of carbon (C), nitrogen (N), and phosphorus (P) over one year in a com
mercial plantation agroforestry system in Acre, Brazil, to determine how or
ganic matter and N and P dynamics change when native Amazonian forest is co
nverted to perennial crops, and how nutrient cycling in tree-based agroecos
ystems compares with other Amazonian land uses. We used a participatory app
roach that involved farmers throughout the research process, with the aim o
f stimulating a community dialogue on the importance of nutrient cycling dy
namics to sustained agricultural production. As part of this process, manag
ement recommendations to protect and enhance nutrient cycling in a manner p
ractical for resource-limited households were formulated by farmers during
focus group discussions. Construction of a nutrient budget revealed that th
e system's largest elemental stocks were located in the top 20 cm of soil.
By the ninth year following establishment, agroforest productivity was high
, with annual NPP approaching 3000 g.m(-2).yr(-1). Approximately 80% of the
annual N and P requirements were taken up from soil stocks. Nearly 40% of
the annual aboveground P requirement was allocated to reproductive tissues,
half of which was removed with the harvest of agroforest products. Removal
of P was half as much as would be expected for a first-year annual crop gr
own in a shifting cultivation system, while N export was similar to that of
annual crops. Inputs of N and P to the agroforest were negligible, and res
orption of N and P from agroforest leaves was comparable to rates cited for
mature forest growing on similarly nutrient-poor soils. Return of P to the
soil in litterfall exceeded that reported for mature Amazonian forests, de
spite similar rates of litter production and N flux, rendering agroforest P
-use efficiency considerably lower than that reported for native vegetation
growing on equally P-poor soils. Overall, NPP and rates of P cycling in th
e 8-yr-old agroforest were high relative to Amazonian forests, but low soil
contents of total N and extractable P, as well as moderate rates of N and
P removal, suggest that this land use will require inputs of both nutrients
to sustain system productivity in the future. Discussions with farmers ind
icated that planting and pruning N-fixing legumes was a feasible way to add
N, but the efficient use of scarce inorganic amendments would require furt
her research.