TERRA-PNW - Regional Analysis of NEP PNW Forests

Project Home

Recent Work

Using remote sensing and in situ network observations to evaluate and improve the performance of the Biome-BGC terrestrial carbon cycle model at regional and global scales

Objectives

1 - Investigate the controls on vegetation phenology using remote sensing measurements, surface weather observations, and ecophysiological modeling, to develop an improved and more general model of phenology.

2 - Refine a method to infer canopy resistance to sensible heat flux from remote sensing and surface weather observations, improving model linkages between the carbon, water, and energy budgets.

3 - Improve the model treatment of disturbance history and its influence on carbon allocation, storage and fluxes using intensive observations of NEE and carbon budget components from the Fluxnet eddy covariance measurement network.

4 - Evaluate integrated model carbon cycle dynamics at the global scale through comparison with inverse modeling results constrained by atmospheric CO2 data.

Approach

- select any image to see a larger version -

Our primary goal is to evaluate and improve the ability of the Biome-BGC model to represent the seasonal, interannual, and multi-decadal dynamics of the terrestrial carbon system, which control net ecosystem exchange of carbon (NEE) with the atmosphere. This requires close examination of the processes contributing to NEE: gross photosynthesis, autotrophic (plant) respiration, and heterotrophic respiration (decomposition by microorganisms), and the environmental controls on these processes. We are particularly concerned with terrestrial carbon cycle dynamics that have important interactions with the atmosphere through the surface energy and hydrologic budgets, and with the effects of disturbance history on these interactions. We are using a series of focused investigations to evaluate and improve model components that we have previously identified as critical determinants of terrestrial NEE. In parallel with these detailed process studies, we are producing a global model implementation and analysis designed to assess model behavior as integrated by daily carbon cycle flux components. We focus our efforts in the project on model components and outputs that are especially conducive to study using an innovative combination of remote sensing observations and in situ network observations.

Schematic showing the disturbance sequence and ensembling used to simulate the known history for the Metolius young and old sites. The initial step at both sites is a "spinup" run, which includes average rates of fire and non-fire mortality. This step is used bring soil organic matter and plant biomass pools into steady state with respect to the specified climate and ecophysiological parameters. For both sites, the observed trend in atmospheric CO2 concentration was included as an additional component of historical variation. Because the old site is composed of at least three age classes, at different stages in post-fire development, an ensemble approach was used to represent the contributions of fluxes from subregions at the site with potentially different carbon dynamics.

These advances should improve our ability to predict future terrestrial carbon cycle dynamics, especially as they relate to potential future changes in climate, atmospheric chemistry, and land management practices. Objectives 1 (improved phenology) and 2 (improved surface resistance parameterization) are critical to the community-wide effort to link terrestrial and atmospheric components as complete carbon system models. Objective 3 (investigation of disturbance effects) is a necessary step in identifying and explaining the spatial and temporal patterns of terrestrial sources and sinks of carbon. Objective 4 will be pursued in collaboration with researchers at Scripps Institute of Oceanography, providing additional constraints on the seasonal cycle and interannual variability of NEE at the global scale and over coarse latitudinal and longitudinal zones. Improvements to the model identified under Objectives 1-3 will be tested at the global scale under Objective 4. We have already made substantial progress toward each of these objectives, and we expect that these new investigations will result in a measurably improved description of important processes in the Biome-BGC model.

Simulated trajectory of NEE at old site, showing the influence of disturbance history, interannual variation in climate, and changing atmospheric concentration of CO2. Solid line is with constant atmospheric CO2 at preindustrial level, dashed line is with observed historical CO2. The dashed horizontal line shows zero NEE, with positive values a flux toward the surface (sink). Large excursions below zero occur after fire disturbances due to a sudden drop in NPP while HR stays relatively high. The CO2 source due to combustion is not included in the plotted lines, but the size of these additional large sources is indicated for each fire event. This plot represents the enseble response, so the effect of the second and third fire events is smaller than the first, due to the smaller area affected.

Isolated influence of increasing CO2 on old site NEE (difference between increasing CO2 and constant CO2 runs that are shown in Figure 4.2). The cyclic variation (18 year period) is due to recycling an 18-year surface weather record. The interannual variation seen in Figure 4.2 is reduced, showing that the model process algorithms controlling the respoinse to increasing CO2 act as a low-pass filter for variation in NEE. The timing of fire events is indicated. Even though the 1944 fire affected half as much area as the 1854 fire, the post-fire variation in NEE due to increasing CO2 was nearly twice as large. The first effect of high CO2 post-fire is an increase in NPP while mineral N resources are high (low plant demand). Interannual variation from the surface weather driver is also being amplified as CO2 increases.

NPP and NEE during regrowth following clearcutting, projected to the year 2040 using the same 18-year surface weather record, and increasing CO2 according to the IS92a scenario. The trees at the young site are observed to have a higher ratio of fine root : leaf annual carbon allocation (FR alloc) than trees at the old site. Only by shifting from a high to a low fine root allocation in the early stages of stand development could the model produce rates of carbon accumulation that agreed with intensive measurements at the sites. This suggests an important new parrameter that should be included in the model in order to produce accurate estimates for NEE over regions where forest recovery following large-scale disturbance is important.

project page last updated 8-23-2001

web contact

Our primary goal is to evaluate and improve the ability of the Biome-BGC model to represent the seasonal, interannual, and multi-decadal dynamics of the terrestrial carbon system, which control net ecosystem exchange of carbon (NEE) with the atmosphere. This requires close examination of the processes contributing to NEE: gross photosynthesis, autotrophic (plant) respiration, and heterotrophic respiration (decomposition by microorganisms), and the environmental controls on these processes. We are particularly concerned with terrestrial carbon cycle dynamics that have important interactions with the atmosphere through the surface energy and hydrologic budgets, and with the effects of disturbance history on these interactions. We are using a series of focused investigations to evaluate and improve model components that we have previously identified as critical determinants of terrestrial NEE. In parallel with these detailed process studies, we are producing a global model implementation and analysis designed to assess model behavior as integrated by daily carbon cycle flux components. We focus our efforts in the project on model components and outputs that are especially conducive to study using an innovative combination of remote sensing observations and in situ network observations.
		Schematic showing the disturbance sequence and ensembling used to simulate the known history for the Metolius young and old sites. The initial step at both sites is a "spinup" run, which includes average rates of fire and non-fire mortality. This step is used bring soil organic matter and plant biomass pools into steady state with respect to the specified climate and ecophysiological parameters. For both sites, the observed trend in atmospheric CO2 concentration was included as an additional component of historical variation. Because the old site is composed of at least three age classes, at different stages in post-fire development, an ensemble approach was used to represent the contributions of fluxes from subregions at the site with potentially different carbon dynamics.



	Simulated trajectory of NEE at old site, showing the influence of disturbance history, interannual variation in climate, and changing atmospheric concentration of CO2. Solid line is with constant atmospheric CO2 at preindustrial level, dashed line is with observed historical CO2. The dashed horizontal line shows zero NEE, with positive values a flux toward the surface (sink). Large excursions below zero occur after fire disturbances due to a sudden drop in NPP while HR stays relatively high. The CO2 source due to combustion is not included in the plotted lines, but the size of these additional large sources is indicated for each fire event. This plot represents the enseble response, so the effect of the second and third fire events is smaller than the first, due to the smaller area affected.

	Isolated influence of increasing CO2 on old site NEE (difference between increasing CO2 and constant CO2 runs that are shown in Figure 4.2). The cyclic variation (18 year period) is due to recycling an 18-year surface weather record. The interannual variation seen in Figure 4.2 is reduced, showing that the model process algorithms controlling the respoinse to increasing CO2 act as a low-pass filter for variation in NEE. The timing of fire events is indicated. Even though the 1944 fire affected half as much area as the 1854 fire, the post-fire variation in NEE due to increasing CO2 was nearly twice as large. The first effect of high CO2 post-fire is an increase in NPP while mineral N resources are high (low plant demand). Interannual variation from the surface weather driver is also being amplified as CO2 increases.

	NPP and NEE during regrowth following clearcutting, projected to the year 2040 using the same 18-year surface weather record, and increasing CO2 according to the IS92a scenario. The trees at the young site are observed to have a higher ratio of fine root : leaf annual carbon allocation (FR alloc) than trees at the old site. Only by shifting from a high to a low fine root allocation in the early stages of stand development could the model produce rates of carbon accumulation that agreed with intensive measurements at the sites. This suggests an important new parrameter that should be included in the model in order to produce accurate estimates for NEE over regions where forest recovery following large-scale disturbance is important.