Date Published: July 7, 2017
Publisher: Public Library of Science
Author(s): Amy L. Kaleita, Linda R. Schott, Sarah K. Hargreaves, Kirsten S. Hofmockel, Dafeng Hui.
Soil microbial communities are structured by biogeochemical processes that occur at many different spatial scales, which makes soil sampling difficult. Because soil microbial communities are important in nutrient cycling and soil fertility, it is important to understand how microbial communities function within the heterogeneous soil landscape. In this study, a self-organizing map was used to determine whether landscape data can be used to characterize the distribution of microbial biomass and activity in order to provide an improved understanding of soil microbial community function. Points within a row crop field in south-central Iowa were clustered via a self-organizing map using six landscape properties into three separate landscape clusters. Twelve sampling locations per cluster were chosen for a total of 36 locations. After the soil samples were collected, the samples were then analysed for various metabolic indicators, such as nitrogen and carbon mineralization, extractable organic carbon, microbial biomass, etc. It was found that sampling locations located in the potholes and toe slope positions had significantly greater microbial biomass nitrogen and carbon, total carbon, total nitrogen and extractable organic carbon than the other two landscape position clusters, while locations located on the upslope did not differ significantly from the other landscape clusters. However, factors such as nitrate, ammonia, and nitrogen and carbon mineralization did not differ significantly across the landscape. Overall, this research demonstrates the effectiveness of a terrain-based clustering method for guiding soil sampling of microbial communities.
Microbial biomass and activity are both critical to, and sensitive indicators of, ecosystem health . In production agricultural systems, microbial functions are especially important due to the high nutrient demands of plants.
Correlation analysis indicated that both terrain and soils attributes have similar trends to microbial indicators (Table 2). Overall, microbial indicators were most strongly correlated with the EMI data, with correlations ranging from 0.39 to 0.83 in magnitude. Specific carbon and nitrogen mineralization were negatively correlated with EMI, while the other indicators were positively correlated with EMI. Microbial indicators were also correlated, to a somewhat lesser degree, with elevation and slope. Plan curvature as an individual index was not correlated with any of the microbial indicators. Overall, specific N mineralization showed the least strong correlation with any of the landscape data, with no correlation coefficient greater than 0.39 in magnitude.
These data suggest that the sampling design for microbial monitoring should consider within-field landscape variability, and that the clustering approach is useful for doing so. While the single BMUs did not necessarily represent the microbial activity for their respective clusters, the soil analysis data showed statistically distinct microbial activity among clusters with different terrain and soils characteristics. The topography data used to generate the clusters can be obtained from high resolution LiDAR, increasingly available through regional government investment, or other surveys.