Ecological indicators of agroecosystem function in the rapidly developing eastern plains of Colombia

Background/Question/Methods: The Orinoco River Basin of eastern Colombia has been undergoing rapid agriculture expansion in recent years and much of the region is expected to transition from extensive pasture systems of naturalized savanna to high input, intensive crop and livestock systems.  To begin to understand the ecological implications of this trend, we evaluated ecosystem functions and soil biodiversity in 5 main land use types: 1) intensive annual cropping systems (soy, rice or maize), 2) recently planted oil palm, 3) newly established rubber plantation, 4) improved pastures (based on establishment and fertilization of introduced species), and 5) naturalized savanna.  Starting in June of 2011, we sampled in 15 replicate sites per land use type (75 production fields in total), distributed across a 150 km transect through the region.  Within each field we evaluated soil chemical fertility, soil physical variables, aggregate morphology, soil macrofauna, aboveground biomass and green house gas emissions.  These data were used to develop sub-indicators of agroecosystem function and biodiversity (e.g., climate regulation, soil water dynamics, biodiversity conservation) with the goal of deriving an overall indicator of environmental function that could be calculated using only a sub-set of the most explanatory variables.

Results/Conclusions: Preliminary results suggest strong differences in the services and biodiversity supported by each land use type.   For example, multivariate comparisons indicate a high degree in separation between annual cropping systems (with relatively poor macrofauna abundance and diversity) as compared to improved pastures, where macrofauna populations were highest (P < 0.001).  Similar multivariate differences were found for soil physical variables, such that annual crops displayed the lowest general soil physical quality, and improved pasture the best.  Soil chemical status on the other hand (largely based on nutrient availability) was highest under annual cropping systems and lowest in the naturalized savannas.  At the same time, initial results indicate that greenhouse gas emissions were highest, and C storage lowest within annual crops (P < 0.05).  Coinertia analyses indicate strong covariation between data sets and provide a preliminary glimpse of the factors driving the observed differences between sites.  Our findings will provide critical information for landscape planning and agricultural development in the region.