IPCC Fourth Assessment Report: Climate Change 2007
Climate Change 2007: Working Group I: The Physical Science Basis Interannual Changes in the Carbon Cycle Interannual changes in global fluxes

The atmospheric CO2 growth rate exhibits large interannual variations (see Figure 3.3, the TAR and http://lgmacweb.env.uea.ac.uk/lequere/co2/carbon_budget). The variability of fossil fuel emissions and the estimated variability in net ocean uptake are too small to account for this signal, which must be caused by year-to-year fluctuations in land-atmosphere fluxes. Over the past two decades, higher than decadal-mean CO2 growth rates occurred in 1983, 1987, 1994 to 1995, 1997 to 1998 and 2002 to 2003. During such episodes, the net uptake of anthropogenic CO2 (sum of land and ocean sinks) is temporarily weakened. Conversely, small growth rates occurred in 1981, 1992 to 1993 and 1996 to 1997, associated with enhanced uptake. Generally, high CO2 growth rates correspond to El Niño climate conditions, and low growth rates to La Niña (Bacastow and Keeling, 1981; Lintner, 2002). However, two episodes of CO2 growth rate variations during the past two decades did not reflect such an El Niño forcing. In 1992 to 1993, a marked reduction in growth rate occurred, coincident with the cooling and radiation anomaly caused by the eruption of Mt. Pinatubo in June 1991. In 2002 to 2003, an increase in growth rate occurred, larger than expected based on the very weak El Niño event (Jones and Cox, 2005). It coincided with droughts in Europe (Ciais et al., 2005b), in North America (Breshears et al., 2005) and in Asian Russia (IFFN, 2003).

Since the TAR, many studies have confirmed that the variability of CO2 fluxes is mostly due to land fluxes, and that tropical lands contribute strongly to this signal (Figure 7.9). A predominantly terrestrial origin of the growth rate variability can be inferred from (1) atmospheric inversions assimilating time series of CO2 concentrations from different stations (Bousquet et al., 2000; Rödenbeck et al., 2003b; Baker et al., 2006), (2) consistent relationships between δ13C and CO2 (Rayner et al., 1999), (3) ocean model simulations (e.g., Le Quéré et al., 2003; McKinley et al., 2004a) and (4) terrestrial carbon cycle and coupled model simulations (e.g., C. Jones et al., 2001; McGuire et al., 2001; Peylin et al., 2005; Zeng et al., 2005). Currently, there is no evidence for basin-scale interannual variability of the air-sea CO2 flux exceeding ±0.4 GtC yr–1, but there are large ocean regions, such as the Southern Ocean, where interannual variability has not been well observed. Interannual variability in regional fluxes, atmospheric inversions and bottom-up models

Year-to-year flux anomalies can be more robustly inferred by atmospheric inversions than mean fluxes. Yet, at the scale of continents or ocean basins, the inversion errors increase and the statistical significance of the inferred regional fluxes decreases.[5] This is why Figure 7.9 shows the land-atmosphere and ocean-atmosphere flux anomalies over broad latitude bands only for the inversion ensembles of Baker et al. (2006), Bousquet et al. (2000) and Rödenbeck et al. (2003b). An important finding of these studies is that differences in transport models have little impact on the interannual variability of fluxes. Interannual variability of global land-atmosphere fluxes (±4 GtC yr–1 between extremes) is larger than that of air-sea fluxes and dominates the global fluxes. This result is also true over large latitude bands (Figure 7.9). Tropical land fluxes exhibit on average a larger variability than temperate and boreal fluxes. Inversions give tropical land flux anomalies of the order of ±1.5 to 2 GtC yr–1, which compare well in timing and magnitude with terrestrial model results (Tian et al., 1998; Peylin et al., 2005; Zeng et al., 2005). In these studies, enhanced sources occur during El Niño episodes and abnormal sinks during La Niña. In addition to the influence of these climate variations on ecosystem processes (Gérard et al., 1999; C. Jones et al., 2001), regional droughts during El Niño events promote large biomass fires, which appear to contribute to high CO2 growth rates during the El Niño episodes (Barbosa et al., 1999; Langenfelds et al., 2002; Page et al., 2002; van der Werf et al., 2003, 2004; Patra et al., 2005).

Inversions robustly attribute little variability to ocean-atmosphere CO2 flux (±0.5 GtC yr–1 between extremes), except for the recent work of Patra et al. (2005). This is in agreement with ocean model and ocean observations (Lee et al., 1998; Le Quéré et al., 2003; Obata and Kitamura, 2003; McKinley et al., 2004b). However, inversions and ocean models differ on the dominant geographic contributions to the variability. Inversions estimate similar variability in both hemispheres, whereas ocean models estimate more variability in the Southern Ocean (Bousquet et al., 2000; Rödenbeck et al., 2003b; Baker et al., 2006). Over the North Atlantic, Gruber et al. (2002) suggest a regional CO2 flux variability (extremes of ±0.3 GtC yr–1) by extrapolating data from a single ocean station, but McKinley et al. (2004a,b) model a small variability (extremes of ±0.1 GtC yr–1). The equatorial Pacific is the ocean region of the world where the variability is constrained with repeated ∆pCO2 observations (variations of about ±0.4 GtC yr–1; Feely et al., 2002), with a reduced source of CO2 during El Niño associated with decreased upwelling of CO2-rich waters. Over this region, some inversion results (e.g., Bousquet et al., 2000) compare well in magnitude and timing with ocean and coupled model results (Le Quéré et al., 2000; C. Jones et al., 2001; McKinley et al., 2004a,b) and with ∆pCO2 observations (Feely et al., 1999, 2002).

Figure 7.9

Figure 7.9. Year-to-year anomalies in ocean-atmosphere and land-atmosphere CO2 fluxes (GtC yr–1) from interannual inversion ensembles covering the past 20 years or so, grouped into large latitude bands, and over the globe. Three different inversion ensembles from Bousquet et al. (2000), Rödenbeck et al. (2003a) and Baker et al. (2006) are shown. For each flux and each region, the anomalies were obtained by subtracting the long-term mean flux and removing the seasonal signal. Grey shaded regions indicate El Niño episodes, and the black bars indicate the cooling period following the Mt. Pinatubo eruption. Slowdown in carbon dioxide growth rates during the early 1990s

The early 1990s had anomalously strong global sinks for atmospheric CO2, compared with the decadal mean (Table 7.1). Although a weak El Niño from 1991 to 1995 may have helped to enhance ocean uptake at that time, inversions and O2:N2 and δ13C-CO2 atmospheric data (Battle et al., 2000) indicate that the enhanced uptake was of predominantly terrestrial origin. The regions where the 1992 to 1993 abnormal sink is projected to be are not robustly estimated by inversions. Both Bousquet et al. (2000) and Rödenbeck et al. (2003b) project a large fraction of that sink in temperate North America, while Baker et al. (2006) place it predominantly in the tropics. Model results suggest that cooler temperatures caused by the Mt. Pinatubo eruption reduced soil respiration and enhanced NH carbon uptake (Jones and Cox, 2001b; Lucht et al., 2002), despite lower productivity as indicated by remote sensing of vegetation activity. In addition, aerosols from the volcanic eruption scattered sunlight and increased its diffuse fraction, which is used more efficiently by plant canopies in photosynthesis than direct light (Gu et al., 2003). It has been hypothesised that a transient increase in the diffuse fraction of radiation enhanced CO2 uptake by land ecosystems in 1992 to 1993, but the global significance and magnitude of this effect remains unresolved (Roderick et al., 2001; Krakauer and Randerson, 2003; Angert et al., 2004; Robock, 2005). Speed-up in carbon dioxide growth rates during the late 1990s

The high CO2 growth in 1998 coincided with a global increase in CO concentrations attributable to wildfires (Yurganov et al., 2005) in Southeast Asia (60%), South America (30%) and Siberia (van der Werf et al., 2004). Langenfelds et al. (2002) analyse the correlations in the interannual growth rate of CO2 and other species at 10 stations and link the 1997 to 1998 (and the 1994 to 1995) anomalies to high fire emissions as a single process. Achard et al. (2004) estimate a source of 0.88 ± 0.07 GtC emitted from the burning of 2.4 × 106 ha of peatland in the Indonesian forest fires in 1997 to 1998, and Page et al. (2002) estimate a source of +0.8 to +2.6 GtC. During the 1997 to 1998 high CO2 growth rate episode, inversions place an abnormal source over tropical Southeast Asia, in good agreement with such bottom-up evidence. The relationship between El Niño and CO2 emissions from fires is not uniform: fire emissions from low productivity ecosystems in Africa and northern Australia are limited by fuel load density and thus decrease during drier periods, in contrast to the response in tropical forests (Barbosa et al., 1999; Randerson et al., 2005). In addition, co-varying processes such as reduced productivity caused by drought in tropical forests during El Niño episodes may be superimposed on fire emissions. From 1998 to 2003, extensive drought in mid-latitudes of the NH (Hoerling and Kumar, 2003), accompanied by more wildfires in some regions (Balzter et al., 2005; Yurganov et al., 2005) may have led to decreased photosynthesis and carbon uptake (Angert et al., 2005; Ciais et al., 2005b), helping to increase the atmospheric CO2 growth rate.

  1. ^  In other words, the model bias has only a small influence on inversions of interannual variability. These interannual inversion studies all report a random error and a systematic error range derived from sensitivity tests with different settings. Bousquet et al. (2000) used large regions and different inversion settings for the period 1980 to 1998. Rödenbeck et al. (2003) used one transport model and inverted fluxes at the resolution of the model grid for the period 1982 to 2002, with different inversion settings. Baker et al. (2006) used large regions but 13 different transport models for the period 1988 to 2002.