Impacts of forest-to-agriculture conversion on aboveground and soil carbon and nitrogen stocks along a bioclimate gradient in Costa Rica

Tropical forest-to-agriculture conversions contribute to ∼20% of contemporary global C emissions. Current global C flux estimates assume C is lost from vegetation and soils over time due to this conversion, and that C stocks in agricultural lands are fairly uniform across the tropics. Global C stock and flux estimates may contain large errors if these assumptions are incorrect. I asked the following research questions: (1) Do aboveground and soil C and N stocks and their net changes due to forest-to-agricultural conversion differ among life zones and agricultural land use types in Costa Rica?; (2) Do aboveground and soil C and N stocks and their net changes due to forest-to-pasture conversion differ by pasture age?; and (3) Can life zone specific-estimates reduce potential biases and uncertainty of regional and global C flux estimates for forest-to-agricultural conversions? I sampled soil and aboveground C and N stocks at 62 agricultural sites in Costa Rica stratified by land use and life zone and compared them to mature forests. Ecosystem and soil C and N stocks varied more by life zone than by land use. Net soil C and N stock changes due to forest-to-pasture conversion differed by life zone; soil stocks decreased in the Tropical dry forest and Lower Montane rain forest life zones and increased in the other life zones. Generally, ecosystem C and N stocks in pastures decreased with pasture age, but the effect of age was smaller than life zone and conversion effects. My estimates of Costa Rican ecosystem C stocks in agricultural lands were 2-fold greater and my ecosystem C flux estimates due to forest-to-agricultural conversions were 8-19% lower than Intergovernmental Panel on Climate Change (IPCC)-endorsed estimates. Therefore to reduce uncertainty and bias, incorporation of ecosystem C stock variability related to life zone into regional and global models is more important than including variability related to pasture age or land use type. Overall, my results confirm that forest-to-agriculture conversions represent a significant alteration to global C cycles, and contribute to reducing uncertainty in the magnitude of such changes.