The White Mountain MAPBGC project was originally conceived as a means of developing spatial data sets of important ecosystem properties that are not readily attainable via traditional approaches (Figure 1). In related research that involved regional modeling with a forest ecosystem modeling known as PnET (PnET model web site), it became evident that the ability to accurately simulate current ecosystem dynamics was limited to some degree by our knowledge of current forest nitrogen status. Present-day patterns of N cycling across complex landscapes are influenced by a variety of factors that are difficult or impossible to characterize quantitatively.

One example involves the effects of historical disturbance and human land use. Most forests in the northeast region have, at one time or another, been cleared for agriculture, harvested for timber products, or burned following timber harvests where the remaining slash provided high fuel loads. Such disturbances have left an imprint on present-day forest composition and biogeochemical status, but understanding their specific effects across large areas is extremely difficult, given the lack of information regarding the timing, severity and consequences of individual disturbance events.

As an alternative, the MAPBGC project has been investigating the degree to which present-day productivity and nutrient status can be inferred through patterns of forest foliar chemistry (nitrogen, lignin & cations). A large body of literature has shown linkages between foliar chemistry and various ecosystem processes (photosynthesis, decomposition, etc.) and recent advances with hyperspectral resolution remote sensing make detection across large, heterogeneous landscapes possible. To investigate this potential, MAPBGC research has proceeded along two parallel lines: 1) studying basic relationships between leaf chemistry and forest C and N dynamics and 2) expanding the capacity for remote-detection of leaf biochemistry (Figure 2).

Results from field studies indicate a tight coupling of C and N cycles in White Mountain forest ecosystems, as seen through strong relationships between canopy chemistry, measured aboveground forest productivity and soil N cycling variables (Figure 3 & Figure 4, Ollinger et al. 2002 (view MS as pdf), Smith et al. In Review).

The project's remote sensing component has involved image data from a number of airborne and satellite sensors, but the primay instrument for foliar chemistry mapping has come from NASA's AVIRIS program, and mode recently, from the hyperion satellite. AVIRIS is a 224-band high spectral resolution (~10 nm) instrument that is flown on an ER-2 aircraft (Figure 5) at an altitude of 20,000 m. This provides a spatial resolution of ~18 m, although low-altitude flights have also provided data at 3-4 m resolution.

In 1997, the ER-2 conducted a cloud-free overpass of the White Mountain National Forest, which provided high-quality image data for the entire region (54 scenes). Within two weeks of the overflight, plot-level foliar data were collected from over 100 plots, which allowed calibration of image data to measured canopy chemistry (Figure 6 & Figure 7). Given the strong linkages between foliage and ecosystem processes observed in the field data, this allowed us to use AVIRIS-based canopy chemistry to estimate variables such as productivity (Figure 8) and soil C:N ratios (Figure 9) across the region.

Other MAPBGC analyses include field studies to investigate the long-term effects of disturbance on N cycling (e.g. Goodale and Aber 2001) and a spatial model of soil mineralogy and cation composition (S.W. Bailey et al.). More recently, we've begun to expand the approach to other forest types in other regions. In early 2001, field data collection was conducted at the Bago Forest Research Station in New South Wales Australia to support validation efforts for NASA's Hyperion sensor (view slide show). Over the next several years, this work will also be expanded to a series of FLUXNET research sites where spatial patterns of canopy nitrogen will be used to examine observed patterns of net carbon exchange.