Remote Sensing of Foliar %N across Broad Spatial Scales Using Data from Multiple Platforms

The concentration of nitrogen (N) in foliage is central to numerous biogeochemical, physiological and ecological processes operating at a wide range of scales, and serves as a useful indicator of ecosystem behavior. Despite growing understanding regarding the role of foliar N in ecosystems, application of this knowledge in regional- to global-scale analyses has lagged, in part because we lack a reliable means of extending N concentration measurements to broad-scale spatial patterns. At relatively fine spatial scales (~100-1000 km2), foliar N estimation has been repeatedly demonstrated using high spectral resolution remote sensing instruments. The potential for extending N estimation to sites that have remote sensing measurements but little or no field measurements has also been demonstrated via a generalized, multi-site equation that incorporates field and imaging spectrometer data to estimate foliar N across a wide range of forest conditions. Still, N detection efforts such as these have been limited to small landscapes because presently available imaging spectrometers have swath widths in the range of 10 km or less. There are at least two potential solutions to this problem. The first is development of a space-based imaging spectrometer capable of providing regional to global coverage. Although planning for such instruments is underway (e.g. HyspIRI), it will be years before data become routinely available. A second possibility is through further investigation of whether some level of foliar N estimation may be possible by integrating foliar chemistry and other important plant traits with spectral features available from existing broad-band sensors. Here, we report progress on examining these potential solutions. First, we present results from refining generalized methods for estimating foliar N by including optically important plant traits—e.g. LMA, LAD, water content—in iterative regression equations with whole canopy spectral reflectance over an expanded range of vegetation conditions. Second, we examine whether these generalizeable methods can be easily translated from imaging spectrometers to broad-band sensor data by quantifying the influence of spectral, spatial and radiometric resolution on foliar N estimation. Our work is based on an integration of a large data set consisting of intensive field measurements from North American forests, imaging spectrometry scenes from AVIRIS, and broad-band sensor data from Landsat and MODIS.