Forest Carbon Sources And Sinks Of The United States

Forest ecosystems, major contributors to the sink of atmospheric carbon (C) in recent decades, play a vital role in the global C cycle. Forests cover about 740 million acres in the U.S., equivalent to one third of the conterminous US area. In spite of the findings by many studies that U.S. forest ecosystems are a C sink in recent decades, the magnitude and distribution of the C sink remain uncertain and controversial. Recent climate variability (e.g. increasing temperature, droughts) and atmospheric composition changes (e.g. nitrogen (N) deposition, rising atmospheric CO2 concentration) along with harvesting, wildfires and insect infestations have significant effects on the US forest C uptake. Forest stand age and disturbances have been recognized as primary drivers of forest structure and its function, and the amount of C in many components of the forest ecosystem are related to forest stand age and disturbances. Thus, it is necessary to evaluate the C sources and sinks, quantify the contributions of these different mechanisms to the C uptake of the conterminous U.S. forests over recent decades, which is prerequisite to its projection into the near future and the right way for future climate policies. The Integrated Terrestrial Ecosystem Carbon Model (InTEC) integrates the effects of both disturbance (such as disturbance events and age effect) and non-disturbance (climate variability, atmospheric CO2 concentration, N deposition) factors on forest C cycles. In this study, we updated InTEC to simulate the U.S. forest C sources and sinks (net biome productivity, or NBP) during the period of 1901-2010 and project the future C sources and sinks based on three climate and disturbance scnerios from 2011 to 2100. The main contents and conclusions are as follows:(1) We updated InTEC model based on forest types, NPP-age curves and temporal integration. We adjusted and calibrated model parameters at 35 AmeriFlux forest sites. Site-by-site comparisons between simulated net ecosystem productivity (NEP) with measured values at Ameriflux forest sites showed that the simulated estimates of NEP in general agree well with the annual data of eddy flux measurements. The model captured 83.2% of the variance in NEP with root mean square error (RMSE) of 102 g C m-1 yr-1 and a slope of 0.77 relative to measured NEP, but in some cases underestimate or overestimate the values. This is a common difference with comparisons between measurements at flux towers representing specific conditions, and estimates at larger scales that represent the heterogeneity of the landscape. State-by-State comparisons between simulated C stock with the US Forest Inventory and Analysis Data showed that the model generally captured the magnitude of C stock in the conterminous 48 states (R2=0.85, RMSE=0.3 Pg C, MBE=0.007 Pg C) and predicted decadal C changes reasonably well. Comparisons at different scales suggested that InTEC can be as an effective tool to simulate forest C sources and sinks of the conterminous US.(2) The updated InTEC model was driven by disturbance and non-disturbance data, forest vegetation information, and soil data to simulate the historical C sources and sinks from 1901 to 2010. Results showed that on average, the US forests was a small sink before the 1940s, and the C sink in US forests increased from the 1950s to the 2000s with an increasing rate of 1.9 Tg C yr-1 The average C sink was 206 Tg C yr-1 with 87% of the sink in living biomass and 13% in soil (broadly defined to include leafy and woody debris) during the period of 1950-2010. The sink of US forests peaked in the late 1970s and early 1980s due to forest regrowth. Increased disturbance events and forest aging reduced the C sink of US forests in the late 1980s and 1990s. The temporal C changes in living biomass were similar to the total NBP while Changes in soil C lag behind that of the living biomass C by about 8 years because of the slow litter transfer from vegetation to soil C pools. The forest living biomass was a C sink from 1940s and reached the peak in the 1970s. The C changes in soil peaked in the 1980s and then became a source after the late 1990s. The temporal progression of the forest sink extended from the Western regions to the Northern regions, and then to the Southern regions. The maximum sink occurred in the Rocky Mountain region from 1940s to 1950s, to the Northeast region in the 1960s, to the Mid-Atlantic region in the 1980s, and then to the Southeast region in the 2000s.(3) A series of scenarios was designed to examine the relative individual and combined effects of non-disturbance and disturbance factors on the conterminous US forest C sources and sinks. Our quantitive analysis of disturbance and non-disturbance factors showed that the contributions of different factors to NBP varied spatially and temporally across the US forests. During the period of 1901-2010. Disturbance factors contributed 82% and non-disturbance factors contributed 18% to the total amount of accumulated NBP. As for the contributions of the individual non-disturbance factors, changes in CO2 accounted for 3%, and changes in N deposition accounted for 10%. Changes in climate reduced accumulated NBP by 1%, and the interaction among these factors contributed 6% of the total NBP accumulation. Disturbance factors contributed about 77% to the total accumualted NBP in the Eastern regions (South and North regions), and 102% to the total accumulated NBP in the Western regions (West Coast and Rocky Mountain regions). N deposition contributed 12%, larger than that of CO2 (3%) and climate change (6%) to the accumulated NBP in the Eastern US. In contrast, CO2 contributed 8%, larger than that of N deposition (3%) and climate (-26%) in the Western US. For the period of 1950-2010, if disturbance factors were omitted, the estimated C sink would be reduced by 95 Tg C yr-1. If non-disturbance factors were omitted, the estimated C sink would be reduced by 50 Tg C yr-1. Non-disturbance effects on the total NBP were smaller before the 1980s but became larger between the 1980s and 1990s relative to disturbance effects.(4) A series of simulations were set based on forest vegetation information, current C sources and sinks, future climate and disturbance scenarios to project the future C source and sinks of the conterminous US forests for the period of 2011-2100. The three climate scenarios included A2 and B2 CGCM2 projections and a control scenario as the continuation of current climate conditions The three disturbance scenarios were included:(1) fixed age and no disturbance (Dconstant), (2) age change and no disturbance (Daging), and (3) age change and disturbance combined (Ddisregion).Results showed that age effects and disturbances played important roles in determining the trend of C sequestration but climate variability had a little impact. The US forests still exhibited as a sink in the end of 21st century for all simulations but the C sink would be larger than the current level if keeping the fixed stand age in 2010 or if stands were disturbance and age were changed, however, the C sink would become smaller or a C source if stands were aging without disturbances. The C sinks under the A2 climate scenario were similar to but larger than those under the B2 climate scenario. Atmospheric CO2 concentration and N deposition were suggested to enhance C sink regardless of disturbance scenarios and the effect of atmospheric CO2 concentration was larger than the effect of N deposition. However, both of these effects would become saturated in different decades in the second 50 years under all disturbance scenarios excluding when keeping fixed stand age in 2010. Disturbances and age effects were suggested to have positive contributions to C sinks during the period of 2011-2050, in contrast, they were suggested to have negative contributions to C sinks.