Can Broad-band Remote Sensing Provide Regional Estimates of Foliar N and Amax?

Abstract:

Foliar nitrogen (N) and photosynthetic capacity (Amax) are two important and tightly coupled variables in forest ecosystems. The strong relationship between foliar N and Amax – at both the leaf and canopy scale – reflects an overall coupling between terrestrial carbon and nitrogen cycling. The ability to use remote sensing to estimate these variables over broad spatial scales would open new doors for terrestrial ecosystem research. For example, reliable broad-scale estimates of foliar N would aid process-based ecosystem modeling efforts, as N is often a key model input parameter. Empirical estimation of foliar N concentrations across a range of biomes and ecosystem types using airborne imaging spectroscopy has been well demonstrated, though nearly all this work has been constrained to localized scales. Multiple studies have shown that common regression analyses used to predict foliar N are largely driven by reflectance over broad portions of the near infrared (NIR) region. Lepine et al. (2016) demonstrated that the relationship between NIR and foliar N could likely be exploited by broad-band sensors to produce continuous regional estimates of foliar N in closed-canopy forests at relatively high spatial resolution (e.g. 30 m, Landsat-8). Here, we examine the potential for regional estimates of foliar N and Amax across the northeastern United States derived from Landsat-8 OLI. We calibrate and validate our predictive model using field-measured whole canopy N estimates from > 200 forested plots in the northeastern U.S. The plots are representative of the closed-canopy temperate forests found throughout the region, and encompass a wide-range of deciduous broadleaf, evergreen needleleaf, and mixed-forest structure and composition. Results are presented with respect to sources of error and implications for other regions.