Robertson, G. P., K. M. Klingensmith, M. J. Klug, E. A. Paul, J. C. Crum, and B. G. Ellis. 1997. Soil resources, microbial activity, and primary production across an agricultural ecosystem. Ecological Applications 7:158-170.

Citable PDF link: https://lter.kbs.msu.edu/pub/2807

The degree to which soil resource availability is linked to patterns of microbial activity and plant productivity within ecosystems has important consequences for our understanding of how ecosystems are structured and for the management of systems for agricultural production. We studied this linkage in a 48-ha site in southwest Michigan, USA, that had been cultivated and planted to row crops for decades. Prior to seeding the site to genetically identical soybean plants (Glycine max) in early spring, we removed soil samples from ,600 locations; plant biomass was harvested from these same locations later in the season. Soil samples were analyzed for physical properties (texture, bulk density), chemical properties (moisture, pH, total C, total N, inorganic N), and biological attributes (microbial biomass, microbial population size, respiration potential, and nitrification and N-mineralization potentials). Plant analyses included biomass and C and N contents. Soil resource variability across this long-cultivated site was remarkably high, as was variability in microbial activity and primary productivity. In almost all cases variability exhibited a strong spatially explicit structure: for most properties and processes >50% of sample variance was spatially dependent at a scale of 5-60 m. Exceptions included microtopography, soil pH, and inorganic P, which were spatially dependent across the entire 1-1200 m range of separation distances examined in this study, and the culturable-bacteria population, which was not spatially autocorrelated at any scale examined. Both topographic relief and soil pH exhibited strongly nested structures, with autocorrelation occurring within two (topography) or more (pH) distinct ranges. Multiple regression analysis showed surprisingly little correlation between biological processes (soybean productivity, soil N turn-over, soil respiration), and static soil properties. The best predictor of soybean biomass at late reproductive stages (r2 = 0.42) was a combination of nitrate N, bulk density, inorganic P, N-mineralization rates, and pH. Overall, results suggest a remarkable degree of spatial variability for a pedogenically homogeneous site that has been plowed and cropped mostly as a single field for >100 yr. Such variability is likely to be generic to most ecosystems and should be carefully evaluated when making inferences about ecological relationships in these systems and when considering alternative sampling and management strategies.

Associated Datatables:

  1. Baseline Spatial Variability Study - Elevation
  2. Baseline Spatial Variability Study - Nitrogen Mineralization
  3. Baseline Spatial Variability Study - Soil Total Carbon and Nitrogen
  4. Baseline Spatial Variability Study - Soil Physical Properties
  5. Baseline Spatial Variability Study - Soil Microbial Biomass/Respiration
  6. Baseline Spatial Variability Study - Seed Bank
  7. Baseline Spatial Variability Study - Soybean Biomass and Carbon/Nitrogen Content

Download citation to endnote bibtex

Sign in to download PDF back to index