Evaluating Soil Microbial Communities and Foliar Nitrogen Across Complex Landscapes: Insights into Terrestrial Biogeochemical Cycles

Abstract

Photosynthesis and decomposition are two fundamental and interconnected components of terrestrial biogeochemical cycles, and large variations in Carbon model projections are due to uncertainties surrounding these parameters. Although foliar Nitrogen and soil microbial activities exert key constraints on plant productivity and decomposition, these variables are seldom included in modeling endeavors because we lack robust methodologies to estimate these parameters across ecosystems. This research demonstrates how advances in remote sensing technologies and machine learning analytical approaches can overcome this limitation and help us understand the distribution and controls of foliar Nitrogen and microbial community biomass and exoenzyme activities across large spatial areas. I used airborne imaging spectroscopy data, provided by The National Ecological Observatory Network (NEON), combined with 475 samples collected across the U.S. to develop generalizable models for the prediction of foliar Nitrogen. Results show higher accuracy (R2 = 0.65) predictions of this key ecosystem parameter across disparate ecosystems than any other existing methodology. Furthermore, many of the wavelength regions identified as important predictors of foliar Nitrogen are associated with regions known to provide information regarding plant growth type and photosynthetic parameters. I then present how foliar Nitrogen influences decomposition dynamics at The Santa Rita Experimental Range (SRER), a dryland site undergoing woody shrub encroachment. In this analysis, I identified the main drivers of soil microbial biomass and exoenzyme activity across plant cover types, and determined that the strength of plant cover effects depends on various state factor controls such as precipitation, topography, and parent material. I used machine learning to link trends in foliar Nitrogen and other remote sensing derived aboveground data products to belowground soil nutrient and microbial community dynamics. This resulted in one of the first high-resolution, complex landscape-scale maps of soil microbial characteristics. These landscape scale predictions of soil microbial communities can help us understand decomposition dynamics across spatial scales that have not previously been possible. These results highlight that high-resolution predictive mapping of foliar Nitrogen and soil microbial biomass and exoenzyme activities can inform key, difficult to measure, constraints on photosynthesis and decomposition in drylands and could also be applied more broadly to other systems.