complete measure of the integrated effect on the climate system due to the regional concentration of changes in diabatic forcing. Others have considered using more comprehensive model output to quantify the impact of human disturbance to the climate system (Claussen et al., 2002).
The ocean is the largest heat reservoir in the climate system (Levitus et al., 2000, 2001). Thus, the change in ocean heat storage with time can be used to calculate the net radiative imbalance of the Earth (Ellis et al., 1978; Piexoto and Oort, 1992). In essence, the ocean heat content provides a metric for the integral in time of the TOA radiative forcing. Furthermore, it offers a valuable constraint on the performance of climate models (Barnett et al., 2001). It is not yet standard practice to use ocean heat content observations, which are available for the past 50 years, to validate forced climate simulations. This is in part because there are several open research questions regarding the accuracy with which ocean heat content can be calculated and applied.
It is not clear, for example, that the observation systems have sufficient frequency and spatial coverage to accurately determine the radiative imbalance on an annual basis (on the order of 0.1 W m−2) so as to independently confirm the calculation of radiative imbalance from the changes in ocean heat storage. Another issue is whether the spatial and temporal sampling of the ocean heat content accurately captures the regions and depths at which heat changes are occurring. In particular there could be significant heat storage changes deeper in the ocean that are inadequately monitored by the existing ocean network.
Several estimates of the trend in ocean heat content have been made using the ARGO network of ocean floats, satellite observations of ocean altimetry (Levitus et al., 2000, 2001; Willis et al., 2003), and climate models (Barnett et al., 2001; Crowley et al., 2003). Not all of these studies express the ocean heat content changes in terms of average radiative forcing, although it is straightforward to do so. Pielke (2003) found that for the period 1955-1995 the imbalance was about 0.3 W m−2, with half between the surface and 300 m, and the rest between 300 m and 3 km. He also found large temporal variations in the imbalance with a negative imbalance, for example, in the early 1980s. Willis et al. (2004) used satellite altimetric height combined with about 900,000 in situ ocean temperature profiles to produce global estimates of upper-ocean (upper 750 m) heat content on interannual timescales from mid-1993 to 2002 (see Figure 4-3). Willis et al. calculated a 0.86 ± 0.12 W m−2 warming rate averaged over this period, but with large interannual variability. As seen in Figure 4-3, the ocean warming occurred in the later years of the record with little change in globally averaged ocean heat content prior to 1997.