|
|
|
|
|
|
|
(Chapter for proposed book entitled Responses of Northern U.S. Forests to Environmental Change, edited by Richard Birdsey, John Hom, and Robert Mickler) |
|
||||
|
Jennifer C. Jenkins and John D. Aber Complex Systems Research Center, University of New Hampshire, Durham, NH 03824
David W. Kicklighter
|
|||||
Because our understanding of ecosystem dynamics is imperfect, different models may highlight the influences of different environmental factors or feedback mechanisms on NPP. These differences in model structure and assumptions may or may not be important when determining regional NPP estimates or the influence of enhanced CO2 and climate change on future forest productivity. A comparison of model estimates against field-measured data can provide useful information about model accuracy under current conditions (Aber 1997). However, models can estimate very different NPP responses to future climate change even though they may estimate similar NPP under current conditions (VEMAP Members 1995). Comparisons between results generated by models with different underlying principles may help to indicate what differences in model assumptions are important. These differences then suggest lines for further inquiry and point to areas where more experimental data are needed. In this chapter we compare results from two ecosystem process models, PnET-II (Aber and Federer 1992; Aber et al. 1995, 1996) and TEM 4.0 (Raich et al. 1991; McGuire et al. 1992, 1993; Melillo et al. 1993), which are driven by GCM scenarios of potential future climate (temperature, precipitation and irradiance) in order to predict forest productivity in the northeastern region under changing environmental conditions. We highlight those features of the models and input datasets which contribute to differences between model predictions. We then describe state-of-the-art methods to address issues not usually considered when developing regional estimates of forest NPP. Consideration of these additional issues will enable more accurate predictions of forest NPP for the region. Results suggested that climate change and enhanced CO2 are likely to increase forest NPP in this region (Table 1). However, other factors not considered here, such as climatic fluctuations, tropospheric ozone, acid rain, and forest clearing are likely to prevent forest NPP from reaching the much higher levels predicted here. PnET-II parameterizes doubled CO2 as a doubling of water use efficiency (WUE). As a result, the hardwood forest type, which was most prone to water stress, was the most responsive to the CO2 increase (Figure 6a). TEM 4.0 includes belowground N cycling and N availability in its NPP predictions. As a result, temperature increases caused the largest changes in predicted NPP. However, a suitability index, defined as:
The GCM climate data represent equilibrium conditions. To examine the impacts of climate variability of model NPP predictions, we used transient climate data spanning 1950-1995 as input to PnET-II and a transient version of TEM (4.1, see Tian et al. 1997) for Harvard Forest and Hubbard Brook. Climate variability contributed substantially to interannual variability in predicted NPP, and the range in predicted NPP expected as a result of climate variability was smaller than the range that occurred as a result of differences in foliar N (as would be expected on sites with different land use histories). |
|||||
