52
A.D. Del Genio
in the vertical profile of water vapor, changes in the spatial distribution of clouds and cloud
properties, and changes in surface properties.
The next question is whether changes in these quantities can be monitored with current instrument capabilities. To define accuracy requirements, one can take two approaches: (1) By
how much do climate models or theory predict that a given parameter is going to change over
several decades? (2) By how much would a parameter have to change to cause a significant
TOA radiative flux change? To answer the latter question, consider that the current atmospheric loading of greenhouse gases since the Industrial Revolution has been sufficient to cause
a net radiative imbalance of 1.5-2.5 Wm- 2 to date (IPCC, 1992). Thus, one might argue that
a change in flux roughly an order of magnitude smaller than this, 0.25 Wm- 2 , is significant
and defines a useful accuracy for monitoring instruments. (By comparison, over the next 20
years the expected additional climate forcing due to increasing greenhouse gas concentrations
is 1 Wm- 2 ). One can calculate the required change in a given parameter using a I-dimensional
radiative transfer model analogous to that used in a GCM.
To answer the former question, we must resort to theory and existing climate model simulations.
Typical predicted warmings over the next 20 years in transient climate change GCM simulations
are of order 0.7°C (Hansen et aI., 1988). We can use the Clausius-Clapeyron equation to
estimate associated changes in water vapor, given constant relative humidity (keeping in mind
that upper troposphere changes will in general be larger than lower troposphere changes because
warming increases with height in climate change scenarios). Cloud height changes can be
estimated from the dependence of convective penetration depth on surface temperature (Del
Genio, 1993a). The temperature dependence of adiabatic liquid water content, which can be
derived from simple thermodynamic arguments (Betts and Harshvardhan, 1987) can be used
to infer changes in cloud optical thickness in the absence of aerosols. Furthermore, if one
assumes constant droplet number concentration, an estimate of cloud particle size changes can
be made as well. In the presence of indirect aerosol effects, changes in both quantities might
be much larger. Changes in cloud cover are still not understood fundamentally, but we can use
the results of climate change simulations (cf. Fig. 2.7) to estimate plausible regional changes.
Aerosol changes can be estimated empirically from current, albeit crude, assessments of the
anthropogenic effect over the past century (Charlson et aI., 1992).
Table 2.2 shows the predicted changes in each of these quantities over the next 20 years, as well
as the more restrictive estimate of changes required to produce a 0.25 Wm- 2 perturbation of
the TOA radiation balance. A more complete list and discussion of the basis for the predicted
changes can be found in Del Genio (1993b) and Hansen et al. (1995b). The table gives us
a rough idea of how accurately we need to monitor various parameters to actually detect a
climate change and attribute it to a particular combination of forcings and feedbacks.
Some of the changes are quite small and present an imposing challenge for satellite instruments
and retrieval algorithms. But the numbers in the table do not refer to absolute accuracies.
Instead they denote the precision with which changes in these quantities must be measured, a
less formidable demand on observing systems.
How can these requirements be met? If one starts with the general statement that all climate
forcings and feedbacks operate via modification of either the solar or terrestrial radiation spectrum, then the requirement is for monitoring instruments that cover the solar and thermal
spectra with sufficient sensitivity to detect these small changes. Hansen et al. (1995b) have
proposed a small satellite mission, Climsat, that includes a Michelson interferometer and scanning polarimeter for this purpose. Space-proven predecessors of both instruments have flown
on planetary missions, and each is self-calibrating, a crucial element for monitoring of decadal
trends (Hansen et aI., 1995b). Combined with monitoring of stratospheric aerosols, water vapor
and spectral solar irradiance by other spacecraft, as well as ground-based monitoring of ozone
A.D. Del Genio
in the vertical profile of water vapor, changes in the spatial distribution of clouds and cloud
properties, and changes in surface properties.
The next question is whether changes in these quantities can be monitored with current instrument capabilities. To define accuracy requirements, one can take two approaches: (1) By
how much do climate models or theory predict that a given parameter is going to change over
several decades? (2) By how much would a parameter have to change to cause a significant
TOA radiative flux change? To answer the latter question, consider that the current atmospheric loading of greenhouse gases since the Industrial Revolution has been sufficient to cause
a net radiative imbalance of 1.5-2.5 Wm- 2 to date (IPCC, 1992). Thus, one might argue that
a change in flux roughly an order of magnitude smaller than this, 0.25 Wm- 2 , is significant
and defines a useful accuracy for monitoring instruments. (By comparison, over the next 20
years the expected additional climate forcing due to increasing greenhouse gas concentrations
is 1 Wm- 2 ). One can calculate the required change in a given parameter using a I-dimensional
radiative transfer model analogous to that used in a GCM.
To answer the former question, we must resort to theory and existing climate model simulations.
Typical predicted warmings over the next 20 years in transient climate change GCM simulations
are of order 0.7°C (Hansen et aI., 1988). We can use the Clausius-Clapeyron equation to
estimate associated changes in water vapor, given constant relative humidity (keeping in mind
that upper troposphere changes will in general be larger than lower troposphere changes because
warming increases with height in climate change scenarios). Cloud height changes can be
estimated from the dependence of convective penetration depth on surface temperature (Del
Genio, 1993a). The temperature dependence of adiabatic liquid water content, which can be
derived from simple thermodynamic arguments (Betts and Harshvardhan, 1987) can be used
to infer changes in cloud optical thickness in the absence of aerosols. Furthermore, if one
assumes constant droplet number concentration, an estimate of cloud particle size changes can
be made as well. In the presence of indirect aerosol effects, changes in both quantities might
be much larger. Changes in cloud cover are still not understood fundamentally, but we can use
the results of climate change simulations (cf. Fig. 2.7) to estimate plausible regional changes.
Aerosol changes can be estimated empirically from current, albeit crude, assessments of the
anthropogenic effect over the past century (Charlson et aI., 1992).
Table 2.2 shows the predicted changes in each of these quantities over the next 20 years, as well
as the more restrictive estimate of changes required to produce a 0.25 Wm- 2 perturbation of
the TOA radiation balance. A more complete list and discussion of the basis for the predicted
changes can be found in Del Genio (1993b) and Hansen et al. (1995b). The table gives us
a rough idea of how accurately we need to monitor various parameters to actually detect a
climate change and attribute it to a particular combination of forcings and feedbacks.
Some of the changes are quite small and present an imposing challenge for satellite instruments
and retrieval algorithms. But the numbers in the table do not refer to absolute accuracies.
Instead they denote the precision with which changes in these quantities must be measured, a
less formidable demand on observing systems.
How can these requirements be met? If one starts with the general statement that all climate
forcings and feedbacks operate via modification of either the solar or terrestrial radiation spectrum, then the requirement is for monitoring instruments that cover the solar and thermal
spectra with sufficient sensitivity to detect these small changes. Hansen et al. (1995b) have
proposed a small satellite mission, Climsat, that includes a Michelson interferometer and scanning polarimeter for this purpose. Space-proven predecessors of both instruments have flown
on planetary missions, and each is self-calibrating, a crucial element for monitoring of decadal
trends (Hansen et aI., 1995b). Combined with monitoring of stratospheric aerosols, water vapor
and spectral solar irradiance by other spacecraft, as well as ground-based monitoring of ozone