Scott V. Ollinger, John D. Aber, Marie-Louise Smith, Julian Jenkins Peter B. Reich, Rita Freuder, Mary E. Martin
In the northeastern United States, forest carbon and nitrogen dynamics have been investigated at a variety of spatial and temporal scales using a number of different approaches. Regionally, the PnET forest ecosystem model has been run at 30 arc-second resolution (approximately 1 km) in conjunction with a Geographic Information System data base to estimate regional patterns of forest productivity and runoff. Predicted net primary production (NPP) ranged from 700 to 1450 g m-2 yr-1 with a regional mean of 1084 g m-2 yr-1. Validation at a number of locations within the region showed close agreement between predicted and observed values. Disagreement at one site was proportional to differences between foliar N concentrations measured at the site and the value used in the model.
Spatial patterns in NPP followed regional patterns of precipitation and growing degree days, depending on the degree of predicted water versus energy limitation within each forest type. Randomized sensitivity analyses indicated that NPP within hardwood and pine forests was limited by variables controlling water availability to a greater extent than foliar nitrogen, suggesting greater limitations by water than nitrogen for these forest types. In contrast, spruce-fir NPP was not sensitive to water availability and was highly sensitive to foliar N, indicating greater limitation by available nitrogen.
At the landscape scale, patterns of forest growth and N cycling have been investigated by combining extensive field measurements with hyperspectral remote sensing and modeling. This work has been conducted in the White Mountain National Forest under the White Mountain MAPBGC project (Mapping and Analysis of Productivity and Biogeochemical Cycling). Field data have shown strong relationships among canopy nitrogen, aboveground net primary production and rates of N cycling in soils, indicating that C and N cycles in northeastern forests are tightly linked. The ability to detect canopy nitrogen with a high spectral resolution remote sensing instrument provides a means of extrapolating these trends from individual plots to the greater White Mountain region. The detailed foliar N data this effort has produced also provides a means of modeling forest growth dynamics with a higher degree of confidence than is possible at broader scales. This has been demonstrated through validation exercises that show improved model accuracy when detailed, plot-level input data are available.
The PnET models have also been used for analyses of multiple environmental stress effects at individual research sites and across broader temporal scales. This is made possible by recent model developments that include the physiological effects of tropospheric ozone, CO2 and N deposition. Results suggest that historical increases in atmospheric CO2 and N deposition have increased forest NPP, but the magnitude of this increase has been substantially offset by concurrent increases in ozone pollution.