Abstract

This chapter reviews the approach and data used to determine the flux of carbon from changes in land use. Both net and gross fluxes are presented, and uncertainties in the data and in the calculated fluxes are discussed. Analyses of land-use change – based on rates of agricultural expansion, logging, and regrowth and their accompanying changes in C/ha – show that terrestrial ecosystems were a net source of approximately 120 Pg C to the atmosphere between 1850 and 1990. In 1990, the net release from changes in land use was approximately 1.7 Pg C, essentially all of it from the tropics.

These estimates of flux are higher than estimates obtained from analyses of data from forest inventories and from inverse calculations with geochemical data. Data from forest inventories show a net accumulation of approximately 0.8 Pg C/yr in northern midlatitude forests, as opposed to a flux of nearly 0 Pg C from changes in land use. Apparently these northern forests are either recovering from harvests more rapidly than they did in the past or accumulating carbon in areas not directly affected by human management. If the analyses of forest inventories are correct, the imbalance in the global carbon equation is approximately half (0.7–0.8 Pg C/yr) what it was when defined on the basis of land-use change. The smaller estimate of missing carbon is unlikely to be found in the trees of midlatitude forests. The additional sink would require a systematic error of about 100% in the observed rates of growth.

Abstract

The various CO2 concentration stabilization profiles used in Intergovernmental Panel on Climate Change (IPCC) analyses are described, and their methods of construction explained. Padé approximant coefficients are given to allow readers to recalculate the profiles precisely. Forward and inverse initialization strategies are discussed, and industrial emissions requirements for the “S” and “WRE” profiles are compared and evaluated. Details of the post-1990 carbon budget breakdown for the specific case of WRE550 are given. Uncertainties in the industrial emissions required for stabilization following any given profile are quantified. Uncertainties considered are those resulting from the effects of: CO2 fertilization formulation (logarithmic versus rectangular hyperbolic); the prescribed future “history” of net deforestation; terrestrial sink specification and the IPCC restriction of this sink to CO2 fertilization only; and ocean flux uncertainties. Sink specification is the greatest source of uncertainty, leading to potential errors in implied emissions of up to ±2 Gt C/yr for WRE550. Errors of this magnitude are equivalent to a misspecification of the stabilization level of approximately ±50 ppmv.

Introduction

The stabilization of future atmospheric CO2 concentration is one of the primary aims of the United Nations Framework Convention on Climate Change (UNFCCC). Article 2 of this convention gives the following objective: “To achieve stabilization of greenhouse gas concentrations … at a level that would prevent dangerous anthropogenic interference with the climate system….”

The sixth annual Global Change Institute (GCI) was held in 1993 in Snowmass, Colorado, to evaluate the state of knowledge of the global carbon cycle. As in previous GCIs, an overarching goal was to increase the interdisciplinary communication between scientists in different disciplines. The 1993 GCI focused on those studying the various facets of the carbon cycle, including emissions of carbon dioxide, carbon in the oceans, the role of terrestrial ecosystems and land use, and measurements of carbon dioxide buildup in the atmosphere.

The goal of the institute was in part scientific, and in part to support the then-ongoing assessment of the carbon cycle by the Intergovernmental Panel on Climate Change (IPCC) (Schimel et al., 1995, 1996; Melillo et al., 1996). The IPCC had assessed the state of knowledge concerning the carbon cycle in its 1990 and 1992 reports (Watson et al., 1990, 1992); however, its 1994 and 1995 reports required a more in-depth analysis. The need for greater depth was driven by the 1992 United Nations Framework Convention on Climate Change (FCCC). Article 2 of the FCCC states as a primary objective that countries should seek to stabilize the concentrations of greenhouse gases in the atmosphere in order to stabilize future climate (within the limits of natural variability).

Abstract

The gradients in partial pressure of CO2 across the air–sea interface provide a starting point for estimating regional and global CO2 fluxes between the atmosphere and ocean. They also are critical constraints on global atmospheric and oceanic models used to infer the land–sea partitioning of CO2 uptake. Here, we assess the factors that contribute to uncertainties in the estimated CO2 fluxes.

We estimate measurement precision in pCO2 to be ±2 μatm, and extrapolation of the data to regions with no measurements yields uncertainties of ±0.8 μatm. The short duration of spring blooms in the North Atlantic diminishes the uncertainties arising from sparse seasonal coverage in the measurements. We estimate an oceanic uptake of 0.3 Gt C/yr due to spring blooms in the North Atlantic. It is difficult to quantify the extent to which pCO2 gradients may change by correcting the pCO2 measurements to skin instead of bulk temperatures, as skin–bulk temperature differences may be positive (negative) with strong surface heating (cooling), or may vanish under high wind conditions. Uncertainties in fluxes associated with gas exchange rates cannot be separated from the yet unknown flux contributions from the covariance between high-frequency wind and pCO2 fluctuations.

In addition to expanded spatial coverage, high-resolution and high-frequency sampling of the meteorology, hydrography, and carbon system in the atmospheric and oceanic boundary layers at a few locations is needed for improving estimates of air-sea CO2 fluxes.

Abstract

Extensive sampling networks have been established to determine the space–time distributions of greenhouse gases so that these data can be used to provide information about the sources and sinks. However, the problem of deducing sources and sinks from concentration data is an ill-conditioned (and often underdetermined) problem and as such is subject to large amplification of errors in observations or models. Various techniques that have been introduced to address this problem are reviewed. Particular attention is given to techniques of Bayesian synthesis inversion that can provide estimates of the uncertainty for the sources that are deduced.

The Context

The atmospheric budgets of radiatively active gases such as carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O) remain subject to very considerable uncertainty. This uncertainty exists despite a wide range of observational and theoretical approaches that have been used in attempts to resolve the ambiguities. One particularly important approach has been to estimate the strengths of the various source and sink processes by using trace gas concentrations from global sampling networks. The principle is that the spatial distributions of concentrations constrain the spatial distributions of sources and sinks. These constraints in turn imply constraints on the possible source and sink processes.

In principle, determining sources and sinks from the spatial distributions of trace gas concentrations (technically called inversion) provides a “snapshot” of the distribution of sources, regardless of the processes that produce the sources.

Summary

Interest in the carbon cycle has increased because of the observed increase in levels of atmospheric CO2 (from ∼280 ppmv in 1800 to ∼315 ppmv in 1957 to ∼356 ppmv in 1993) and because the signing of the UN Framework Convention on Climate Change has forced nations to assess their contributions to sources and sinks of CO2n, and to evaluate the processes that control CO2 accumulation in the atmosphere. Over the last few years, our knowledge of the carbon cycle has increased, particularly in the quantification and identification of mechanisms for terrestrial exchanges, and in the preliminary quantification of feedbacks.

The Increase in Atmospheric CO2 Concentration Since Pre-Industrial Times

Atmospheric levels of CO2 have been measured directly since 1957. The concentration and isotope records prior to that time consist of evidence from ice cores, moss cores, packrat middens, tree rings, and the isotopic measurements of planktonic and benthic foraminifera. Ice cores serve as the primary data source because they provide a fairly direct and continuous record of past atmospheric composition. The ice cores indicate that an increase in CO2 level of about 80 ppmv paralleled the last interglacial warming. There is uncertainty over whether changes in CO2 levels as rapid as those of the 20th century have occurred in the past.

Since 1988, the Office for Interdisciplinary Earth Studies of the University Corporation for Atmospheric Research (UCAR) has run a series of annual Global Change Institutes (GCIs) on a range of topics under the broad theme of global environmental change. Participants in each GCI have come from a wide range of disciplines, including those peripheral to the main topic, in order to stimulate discussion and to ensure a multidisciplinary perspective. All GCIs have been highly successful and have led to important and useful proceedings volumes.

The sixth annual Global Change Institute was held over July 18–30, 1993, in Snowmass, Colorado. The topic of the institute was the carbon cycle. As a unique feature, the GCI focused on the practical problems of projecting future concentration changes for given emissions, estimating emissions for prescribed concentration profiles, and assessing the uncertainties in these calculations. Much of the discussion still involved processes, but the viewpoint fostered was as much that of the user of carbon cycle model output as of the “pure” scientist. Over the past decade, there has been a trend toward applied or socially relevant science. With the concern over future climatic change resulting from increasing greenhouse gas concentrations, and with the central role that CO2 plays in this problem, there is no area of science in which cognizance of the social and policy implications is more important than in carbon cycle research.

Abstract

This chapter assesses the state of the art in modeling the physical conditions of the global ocean. Results of some new simulations are now becoming available for use by carbon cycle models for the case of modern climatic forcing. Future improvements in computer power should allow realistic physical and biogeochemical simulations for other climatic conditions, such as paleoclimates and anthropogenic climate change.

Introduction

About 10 years ago, Charles D. Keeling stopped by my office at the National Center for Atmospheric Research (NCAR) and inquired about the availability of ocean simulation fields for modeling the carbon cycle of the global ocean. I told him that, unfortunately, there was as yet no global model representation of ocean circulation that was physically realistic enough to be used for the purpose he envisioned. The global models of the time had been run for only short periods from simple initial conditions, and/or their grids were very coarse (approximately 5 degrees). I did recommend that he contact Jorge Sarmiento at the Geophysical Fluid Dynamics Laboratory (GFDL) in Princeton, New Jersey, to obtain their new modeling results for the North Atlantic, which were being produced on a relatively fine grid for that time (approximately 1 degree of latitude and longitude). Even those results had not achieved the same length scales as those for currents in the real ocean, which are typically small fractions of a degree in longitude.

Abstract

Carbonate reefs and platforms have accumulated CaCO3 at a rate of 8–9 × 1012 mol/yr over the last few million years. The Holocene rate of shallow water CaCO3 deposition is approximately 17 × 1012 mol/yr. In order for the shallow water CaCO3 flux to maintain its long-term average deposition rate, it must decline to below 8 × 1012 mol/yr during glacial intervals. Shallow water CaCO3 sediments represent a large, dynamic carbon reservoir that rapidly affects the alkalinity of the surface ocean and hence the CO2 content of the atmosphere. Shallow water carbonate deposition, while probably an important constraint on paleoatmospheric CO2 concentrations, can only slightly influence the anthropogenically driven buildup of atmospheric CO2.

Introduction

In order for the record of atmospheric CO2 changes contained in glacial ice (Neftel et al., 1982; Barnola et al., 1987) to reflect changes in the deposition of marine carbonate, shallow water deposition during interglacial intervals must be significantly higher than the long-term average, and the global weathering of near–sea level carbonates must contribute an increased flux of dissolved calcium carbonate (CaCO3) to the oceans during glacial low stands (Milliman, 1974). Review of research on the global Holocene shallow water carbonate flux reveals that the size of the shallow water carbonate reservoir is not well constrained.

Abstract

The CO2 fertilizing effect on vegetation growth arises from primary effects of CO2 concentration on photosynthetic CO2 fixation, suppression of photorespiration (and possibly of dark respiration), and reduction in stomatal conductance. These mechanisms increase the efficiency of use of growth-restricting inputs of light, water, and nitrogen in the formation of dry matter. It is of critical significance that the C:N ratio of plant tissues varies considerably when CO2 is varied. The relative response of plant stand seasonal growth to high CO2 is typically similar to that which would be calculated on basic photosynthetic biochemical and stomatal diffusion grounds. Researchers are still determining the full extent of various negative and positive feedbacks and other factors that accentuate or attenuate the propagation of this primary response into the size of live and dead C pools. However, on the basis of present evidence it seems unjustified to assume that all such modifiers act to annul completely the primary stimulus of high CO2 in terms of increase in C pool sizes. Indeed, the likely magnitude of the CO2 fertilizing effect is such that it can comfortably account for the “missing carbon sink” of approximately 1–2 Gt C/yr, according to several independent terrestrial C cycle models.

For modeling the response of net primary production (NPP) to CO2, there are five approaches. The once common approach of assuming a flat response (i.e., nonresponse) above the preindustrial CO2 concentration of 280 μmol/mol is highly unlikely to be correct.

Abstract

The characteristics of a new vegetation production model, integrated with a novel approach to nitrogen uptake from the soil, is described and assessed. The model shows an effective capacity to predict plant and vegetation processes scaling up from the leaf, through the canopy to the biome scale. When the model is used to assess the impact of CO2 enrichment and global climate change, it becomes clear that global warming dominates the net primary production (NPP) responses of major biomes. In general terms, a global warming of 3°C and a 10% increase in precipitation is likely to cause reductions in NPP. However, this reduction might be offset by a doubling in atmospheric CO2.

Introduction

The qualitative and quantitative natures of the past and current terrestrial carbon pools have been described by other contributors to this volume (Houghton, Gifford, Schlesinger, Hall, and Keeling) and recently by Sundquist (1993). This chapter aims to describe a global-scale use of plant physiological mechanisms for investigating the influences of changes in climate and CO2 on photosynthesis, transpiration, and nitrogen uptake. These mechanisms can be scaled to indicate the changes in carbon fluxes to and from vegetation with a changing climate and atmospheric concentration of CO2.

Abstract

Soils hold one of the largest near-surface pools in the global carbon cycle, containing at least 1,500 Pg C in organic forms, with a large proportion of this amount lying near the surface. Largely as a result of the human disturbance of soils, especially in cultivation, 36 Pg C was lost from soils between 1860 and 1960, with a current rate of loss of approximately 0.8 Pg C/yr. Thus, the loss of carbon from soils is a significant component of the biotic flux of CO2 to the atmosphere The soil carbon pool does not appear likely to house the missing sink. In fact, as a result of global warming, substantial amounts of CO2 are likely to be lost from soils.

Introduction

Soils hold one of the largest near-surface pools in the global carbon cycle, containing at least 1,500 Pg C in organic forms. Although some fractions of soil organic matter are very old, the global mean residence time for organic carbon in soils is approximately 30 years. The soil carbon pool is large and dynamic; increases or decreases in the amount of carbon in soils could have significant effects on the concentration of CO2 in the atmosphere (Trumbore et al., 1996). A large literature shows that human activities – especially cultivation – reduce the pool of carbon in soils and that most of this carbon is transferred to the atmosphere.

Abstract

This chapter surveys the literature regarding potential future fossil fuel carbon emissions in the absence of explicit control policies. We have assembled 30 base cases and uncertainty analysis trajectories from 18 separate analyses of fossil fuel carbon emissions for comparison to the Intergovernmental Panel on Climate Change (IPCC) 1991 Integrated Analysis of Country Case Studies. We discuss global forecasts of fossil fuel carbon emissions and associated energy consumption, regional forecasts of fossil fuel carbon emissions and associated energy production and consumption, analyses that have explicitly explored the uncertainty associated with global energy and fossil fuel carbon emissions, and differences in key assumptions among various base cases.

Introduction

In our survey of the literature on potential future fossil fuel carbon emissions in the absence of explicit control policies, we have assembled 30 base cases and uncertainty analysis trajectories from 18 separate analyses (Table 14.1, column 2) of fossil fuel carbon emissions for comparison to the Intergovernmental Panel on Climate Change (IPCC, 1991). A list of the studies, dates of publication, and models used is given in Table 14.1. Six of these trajectories have been drawn from the results of the 12th Energy Modeling Forum, “Global Climate Change: Energy Sector Impacts of Greenhouse Gas Emission Control Strategies” (EMF-12), and reflect a comparison of base cases with some standardization of assumptions. We have made no attempt to create an assessment of models. Several thorough literature reviews already perform that function.

Abstract

Simple ocean carbon cycle models are constructed to reflect the interaction between the atmospheric and oceanic components of the global carbon cycle. In this chapter, (1) a two-box ocean model is used to demonstrate principles used in constructing simple ocean carbon cycle models, (2) a variety of simple ocean carbon cycle models are described, and (3) results of various models are shown and compared. Physical transport of carbon in simple ocean models is not based on first principles, but is accomplished using parameterizations calibrated with carbon isotopes and/or other tracer fields. Well-calibrated simple ocean carbon cycle models may yield CO2 absorption predictions that are more accurate than predictions based on ocean general circulation models. Nevertheless, ocean general circulation models may be required to estimate the impact of climate and ocean circulation feedbacks on CO2 fluxes between the ocean and atmosphere.

Introduction

The activities of Homo sapiens have been significantly perturbing the global carbon cycle for the past several hundred years. From 1750 to 1990, the burning of fossil fuels and the clearing of forests has released approximately 380 Gt C as CO2 into the atmosphere, but only approximately 160 Gt C remains there (Sundquist, 1993); the rest has been absorbed by the oceans and terrestrial ecosystems. Ocean models indicate that the oceans have absorbed approximately 140 Gt C, leaving approximately 80 Gt C unaccounted for (perhaps indicating an additional sink in terrestrial ecosystems).