Soil phosphorus cycling across a 100-year deforestation chronosequence in the Amazon rainforest.

Deforestation of tropical rainforests is a major land use change that alters terrestrial biogeochemical cycling at local to global scales. Deforestation and subsequent reforestation are likely to impact soil phosphorus (P) cycling, which in P-limited ecosystems such as the Amazon basin has implications for long-term productivity. We used a 100-year replicated observational chronosequence of primary forest conversion to pasture, as well as a 13-year-old secondary forest, to test land use change and duration effects on soil P dynamics in the Amazon basin. By combining sequential extraction and P K-edge X-ray absorption near edge structure (XANES) spectroscopy with soil phosphatase activity assays, we assessed pools and process rates of P cycling in surface soils (0-10 cm depth). Deforestation caused increases in total P (135-398 mg kg -1 ), total organic P (P o ) (19-168 mg kg -1 ), and total inorganic P (P i ) (30-113 mg kg -1 ) fractions in surface soils with pasture age, with concomitant increases in P i fractions corroborated by sequential fractionation and XANES spectroscopy. Soil non-labile P o (10-148 mg kg -1 ) increased disproportionately compared to labile P o (from 4-5 to 7-13 mg kg -1 ). Soil phosphomonoesterase and phosphodiesterase binding affinity (K m ) decreased while the specificity constant (K a ) increased by 83%-159% in 39-100y pastures. Soil P pools and process rates reverted to magnitudes similar to primary forests within 13 years of pasture abandonment. However, the relatively short but representative pre-abandonment pasture duration of our secondary forest may not have entailed significant deforestation effects on soil P cycling, highlighting the need to consider both pasture duration and reforestation age in evaluations of Amazon land use legacies. Although the space-for-time substitution design can entail variation in the initial soil P pools due to atmospheric P deposition, soil properties, and/or primary forest growth, the trend of P pools and process rates with pasture age still provides valuable insights.