Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
4 Development of Monitoring and Early Detection Strategies CURRENT STATUS Attempts have been made to detect an atmospheric warming caused by the past increase of CO2 concentrations in the atmosphere (Madden and Raman- athan, 1980; Hansen et al., 1981; Kukla and Gavin, 1981; Wigley and Jones, 1981). However, such a warming has not been unequivocally identified, perhaps for the following reasons: 1. As discussed in the section Role of the Oceans, the full manifestation of the CO2-induced warming may be delayed because of the large thermal inertia of oceans. 2. The CO2-induced warming may be masked by climate changes caused by other factors such as the secular variations of atmospheric aerosols and solar irradiance. 3. Since the records from instrumented observations of the past climate variation are available over a relatively short period, it is difficult to obtain a long enough record of the natural variability of climate to establish the statistical significance of a CO2-induced signal. DETECTION STRATEGIES It has been suggested that the long-term variation of global-mean (or hemispheric-mean) atmospheric temperature has been influenced by the changes in insolation and atmospheric aerosol concentrations (e.g., Budyko, 1969; Hoyt, 1979; Hansen et al., 1981; Gilliland, 1982). Climate variations 61
62 CARBON DIOXIDE AND CLIMATE: A SECOND ASSESSMENT due to these non-CO2 influences must be quantified to the extent possible in order to permit the climate changes attributable to CO: to be identified. It is therefore advisable to monitor the temporal variations of atmospheric aerosol concentration and spectra of solar and terrestrial radiation at the top of the atmosphere and at the Earth's surface. In addition, it is desirable to monitor the concentrations of some minor atmospheric constituents, such as fluoro- carbons, methane, and nitrous oxide, that may contribute to future changes of the atmospheric temperature as discussed in the section Trace Gases Other Than CO2. Other candidates for monitoring include planetary albedo as influenced by deforestation and desertification. On the basis of these mea- surements, one should be able to distinguish the climate changes attributable to changes in these factors and thereby facilitate the detection of a CO2- induced climate signal. A comprehensive set of variables should be monitored in order to discriminate CO2-induced changes from changes in climate caused by other factors. These variables should include CO2 concentration in the atmosphere, the solar irradiance, the spectral distribution of solar and terrestrial radiation (at the top and bottom of the atmosphere), and concentrations of aerosol and minor constituents in the atmosphere. Other variables derivable from existing records and observing programs may provide indications of CO2 effects. These include total precipitable water and the diurnal range or temperature. For the earliest detection of a CO2-induced climate signal, it is desirable to monitor a set of variables that reveals CO2-induced changes at the earliest possible time. Preliminary attempts to identify such indicators have already been made by Madden and Ramanathan (1980) and Wigley and Jones (1981). For example, these studies noted that the zonal-mean surface air temperatures in summer, except at very high latitudes, have relatively large signal-to-noise ratios. Also, zonal-mean summer temperatures in the stratosphere and mesosphere are expected to show large CO2-induced cooling, whereas the natural temperature variability is relatively small. However, since the radiation balance of these regions is largely independent of the troposphere, observations of stratospheric and mesospheric cooling would only confirm that CO2 concentrations had increased and that an approximately correct radiation model had been used. Moreover, changes in other constituents, e.g., ozone, could confuse the signal. The effects of changing CO2 concentrations on atmospheric radiation transfer may also be monitored by looking for systematic trends in satellite remote temperature sounding data, which depend on infrared radiance from atmospheric CO2. Another possible candidate for consideration is the temperature of deep ocean layers (see the section Monitoring Ocean Climate Response). One of the meteorological variables that are useful for monitoring the past and future climate change is the global-mean (or hemispheric-mean) surface
Development of Monitoring and Early Detection Strategies 63 air temperature of the atmosphere. This variable has been used in some attempts to detect the CO2 climatic signal partly because it has a large signal- to-noise ratio. It is expected that the CO2-induced temperature change is positive in most of the troposphere, whereas the natural temperature variation changes sign from one geographical location to another. Therefore, the signal-to-noise ratio for an area-mean temperature should increase as the area for the averaging increases. One can also introduce indices that are better suited for the early detection of CO2-induced climate change. One example of such an index is the weighted mean global (or hemispheric) mass integral of the atmospheric temperature. The weighting factor may be defined such that it is small in the regions of large natural temperature fluctuation and large in the regions where the CO2- induced temperature change is expected to be large from model experiments. Obviously, one can devise many other similar indices. A set of indices that have a large signal-to-noise ratio should be identified and monitored. From the preceding discussion, it is clear that the early detection of the CO2-climate signal requires not only a prediction of the CO2-induced climate change but also a knowledge of the natural climate variabilities. Therefore, it is necessary to determine (from the past climatic records) the variability of relevant climatic variables such as temperatures of the atmosphere and oceans. For example, some of the important variables requiring improved determination are hemispheric- and global-mean surface air temperatures. The present information on the temporal variation of these quantities may contain inaccuracies of a substantial magnitude (Damon and Kunen, 1976, 1978; Carter, 1978; Barnett, 1978). Emphasis should be placed on the compilation and analysis of past climate data to acquire more reliable reconstructions of past variations of climate on a variety of space scales. Finally, it should be noted that a major workshop convened by the Department of Energy in June 1981 addressed the problem of early detection. Its report was not available in time for the panel's consideration, but it will clearly shed much light on the problem. MONITORING OCEAN CLIMATE RESPONSE Operational monitoring of the ocean's response to climate change requires observation of sea-surface temperature, water-mass parameters, and sea-ice extent. The problem with sea-surface temperature measurements is their large fluctuations due to mesoscale eddies and observational errors. However, changes in the heat content of the ocean may possibly be detectable over periods of a decade or longer. Water-column observations have the potential of being able to monitor
64 CARBON DIOXIDE AND CLIMATE: A SECOND ASSESSMENT directly the temperature and salt response of the ocean for a range of isopycnals, all with differing time responses. Of particular importance to this monitoring will be the wind-driven gyres, which have a decadal time scale (Jenkins, 1980). The monitoring of potential temperature and salinity changes on isopycnals in the wind-driven gyres may provide an early indication of climatic change. The gradients of potential temperature and salinity along isopycnals are weak in the wind-mixed gyres; the only signature of the mesoscale eddy field would be that of small fluctuations along isopycnals. These measurements thus have an inherently large signal-to-noise ratio. Estimates based on a series of four cruises in the eastern North Atlantic subtropical gyre show that potential temperature can be spatially determined in the presence of eddies to an error less than 0.02Â°C for any cruise and a scatter less than 0.1Â°C for four cruises extending over 2 years. The difficulty in interpreting any water-mass indicators of climatic varia- bility is that poorly understood salinity changes associated with evaporation and precipitation may accompany these changes. It is unlikely, however, that temperature and salinity changes should coincide to produce no apparent change on any isopycnal. Measurements of varying quality and geographical distribution exist that date back to the 1920's. The extent of sea ice has been shown by Manabe and Stouffer (1980) to respond sensitively to CO2-induced climatic changes in models. Paleoclimatic reconstructions of sea-ice extent in the southern hemisphere by Hays (1978) and correlations between northern hemisphere ice extent and surface tem- peratures (e.g., Vinnikov and Groisman, 1981; Vinnikov et al., 1980) seem to indicate a strong direct relationship between them. Indeed, reductions in sea-ice extent in both hemispheres have recently been noted from satellite observations (Kukla and Gavin, 1981). Since sea-ice extent is easily and routinely measured from satellites, it should be a good parameter for mon- itoring CO2-induced climatic changes.