In our review of IPCC AR1 (1990) on radiative forcing, I noted that the logarithmic relationship and 4 wm-2 values were attributed to: Hansen et al (1988), which in turn cited Lacis et al 1981; and Wigley (1987) which is not presently available to me (or to Wigley himself) and appears not to have been peer-reviewed (FWTW). Feedback analysis primarily relied on Cess et al 1989. I’ll examine those references at some point, but today I’ll continue the review through two supplements to IPCC (1990), published in 1992 and 1994.
Climate Change 1992
A few pages (6-8,45-47,57-59, 63-70) of Houghton et al, Climate Change 1992: The Supplementary Report to the IPCC Scientific Assessment are now online at Google here, but the sections of interest here are not.
Chapter A2 (Radiative Forcing) starts by focusing on ozone and sulphate aerosols, saying that there have bee “significant advances in our understanding” of their impact on radiative forcing. CO2 results are not mentioned in the Executive Summary of this chapter. On page 53, it re-capitulates the definition of radiative forcing, but I did not locate any fresh analysis of CO2. Their analysis of aerosols is primarily of sulphate aerosols, where they report negative sulphate aerosol forcing, stating:
A very important implication of this estimate is that the net anthropogenic radiative forcing over parts of the NH during the past century is likely to have been substantially smaller than was previously believed.
They note high geographic variability of sulphate aerosol forcing, with their Figure A2.3 showing high values only over North America and Europe-Asia.
Climate Change 1994:
Radiative Forcing of Climate Change and an Evaluation of the IPCC 1992 IS92 Emission Scenarios
Again a few pages of this report (1-27, 157-160, 167-170, 190), this time including a little bit of interest in this post, are online here . This report, published in June 1995, contains a discussion of the “saturation” argument that one sometimes sees by “skeptics” and which was the subject of a contemporary exchange (1994-early 1995) in Spectrochimica Acta between “skeptic” Jack Barrett and, on the other side, John Houghton and Keith Shine (among others). I don’t think that the saturation argument has any merit, but I’m not especially impressed with how IPCC handled it either.
First they spend a couple of pages distinguishing instantaneous forcing from no-feedback forcing – a distinction that is not material to the derivation of 4 wm-2 or the logarithmic relationship. They report results for doubled CO2 from a 1-D model of Rind and Lacis 1993 (presumably a descendant of the 1-D Lacis et al 1981 model cited in AR1):
(170) The utility of radiative forcing will be the subject of Section 4.8, but it is useful here to illustrate the potential importance of the adjustment process using 1-dimensional radiative-convective results reported by Rind and Lacis (1993) see Table 4.1. …For a doubling ..of CO2 the stratosphere cools so that the adjusted forcing is about 6% less than the instantaneous.
Table 4.1 Surface temperature changes using a a 1-D radiative -convective model (Lacis and Rind 1993)… (Note that surface temperature changes would be about 1 to 4 times larger if climate feedbacks were included – see IPCC 1990). The changes in radiative properties used here are meant to be illustrative and do not necessarily represent actual or projected changes. Results below are for doubling CO2 from 300-600 ppmv.
Item Impact Surface temperature change (ΔT_0) 1.31 K Radiative forcing at tropopause, instantaneous (ΔF_i) 4.63 wm-2 Radiative forcing at tropopause, no feedback (ΔF_o) 4.35 wm-2 Climate sensitivity parameter, instantaneous (λ_i) 0.28 K/ wm-2 Climate sensitivity parameter, no feedback λ_o 0.30 K/ wm-2
They introduce section 4.2 (Greenhouse Gases) as follows:
(p. 171:) Our understanding of the enhanced greenhouse effect is dependent on two broad areas. First we need to understand the fundamental radiative properties (“spectroscopy”) of the gases involved. Next, these spectroscopic data need to be included in radiative transfer models to calculate radiative forcing due to changes in gas concentration, for a given atmospheric profile of temperature, water vapour and other trace gases, and cloudiness. In this section the discussion will concentrate on the radiative forcing of greenhouse gases as a result of direct emission of that gas. This is referred to as the direct greenhouse forcing.
Their section 4.2.1 ( Spectroscopy) introduces the HITRAN and GEISA databases, mentioning “uncretainties” in the water vapor continuum, without elaboration:
In the thermal infrared (approximately 4-500 μm) molecules absorb and emit radiation by changing the energy with which they vibrate and/or rotate the wavelengths of the vibration/rotation transitions occur over narrow spectral intervals. Laboratory and theoretical studies are required to determine the wavelengths, strengths and widths of these transitions”¬⤮
The main databases of spectral parameters of atmospheric gases are subject to periodic update; the two main catalogues, HITRAN (Rothman et al, 1992) and GEISA (Husson et al., 1992) have both been substantially revised since IPCC (1992). These revisions are based on improvements to both laboratory measurements and theoretical techniques… A detailed assessment of the effect of these revisions has not yet been reported. Fomin et al. (1993) report that the net irradiance at the tropopause, evaluated for a mid-latitude clear-sky atmosphere, changes by less than 1% between using the 1986 and 1992 versions of HITRAN. It seems unlikely that the effect of these revisions on radiative forcing will be greater than 5% but there is a need to assess both the effect of these changes and the potential impact of remaining uncertainties.
Continuum absorption, especially by water vapour. Is also of importance in calculating radiative forcing. In spite of recent progress in describing the effect, further theoretical and laboratory investigations are required to resolve remaining uncertainties.
Next section 4.2. Calculating the Radiative Forcing , sketches how the HITRAN line strength information gets translated into models: essentially, for computational reasons, the infrared spectrum is divided into “bands” and an average absorption is assigned to each band. Bands may be “narrow” (at that time about 10 cm-1) or “broad”. IPCC 1994:
A whole hierarchy of different radiative transfer schemes are available to compute the radiative forcing, ranging from line-by-line models through to so-called wide-band models (Ellingson et al 1991). In addition the results from such models can be represented by relatively simple empirical formulae, such as those presented in IPCC (1990) and Shi (1992) to allow rapid and reasonably accurate computation of forcing.
Many details of the radiation schemes (such as methods of handling clouds, spectral overlap of gases, treatment of water vapour continuum) affect the radiative forcing and are not handled in the same way by all schemes (see Ellingson et al 1991)
The ultimate test of such models is their ability to reproduce observed irradiances, given the observed state of the atmosphere. Until recently, the quality of observations was generally inadequate to assess the models; now high-quality experimental data (see e.g. Ellingson et al. 1992) are becoming available and should provide valuable checks on the realism of radiative models.
The radiative forcing due to changes in concentration of CO2, CH4, N2O, CFC-11 and CFC-12 presented in IPCC (1990) have been re-assessed by Kratz et al (1993) and Shi and Fan (!992), although neither set of authors accounted for stratospheric adjustment …
Despite the general agreement between the earlier IPCC reports and more recent calculations, a similar agreement is not found for radiative transfer calculations performed in GCMs used for climate prediction. Cess et al (1993) report an intercomparison of instantaneous clear-sky radiative forcing due to a doubling of CO2 from 15 different GCMs. The results deviated by as much as 20% from reference line-by-line calculations; some of the difference was due to neglect of some minor absorption bands, and in particular those near 10 μm, The size of the deviation should not be taken to indicate the uncertainty in calculating the CO2 forcing; it is more indicative of the weakness in the radiative transfer schemes used in some GCMs.
The Ellingson et al 1991 study mentioned here is an interesting one. Ellingson et al did an intercomparison of GCMs noting a wide variation in infrared codes, many of which he said were simply incorrect. Regardless of whether their code was erroneous or not, the GCMs in the survey all agreed quite closely on the impact of doubled CO2, provoking an arch comment from Ellingson et al about “tuning”.
Box: THE GREENHOUSE EFFECT OF INCREASED CONCENTRATIONS OF CARBON DIOXIDE
Next IPCC 1994 has an interesting discussion of how the enhanced greenhouse effect actually works – which, to my knowledge, is by far the most explicit such discussion in the entire IPCC corpus, which is, in part, a response to an exchange between skeptic Jack Barrett and John Houghton in Spectrochimica Acta in 1994. I’ve repeated it in full below.
“It is sometimes stated that, because there is so much CO2 already in the atmosphere, it is “saturated” and extra CO2 can have no additional greenhouse effect. The purpose of this box is to present a simplified explanation as to why this is a misconception.
When gases are present in small concentrations (the halocarbons are examples), the radiative effect of a gas is almost linear in concentration: doubling the concentration of, for example, CFC-11 will approximately double its greenhouse effect. This is not the case for greenhouse gases in larger concentrations, for well-understood reasons (see e.g. Goody and Yung, 1989). Doubling the concentration of CO2 from its present day concentrations leads to a 10-20% increase in the total greenhouse effect due to CO2 ”¬’ this effect is well-understood and has been included in climate models for several decades.
Figure 4.1 illustrates aspects of the increased greenhouse effect of CO2 using a detailed radiative transfer model (see footnote below for details ). Figure 4.1a shows the spectral variation in the net infrared irradiance (“flux”) at the tropopause in Wm-2/cm-1. The shape of the curve is dictated by two factors. First, the Planck function determines the maximum amount of energy that can be emitted at a given wavelength and temperature. At typical atmospheric temperatures, the maximum lies between 10 and 15 μm; at wavelengths short than 5 μm little can be emitted. The second factor is the absorbing properties of the atmosphere, which is dictated by the presence of greenhouse gases (of which water vapour is the most important, followed by CO2) and clouds.
If the troposphere were transparent to infrared radiation, then the irradiance reaching the tropopause would be the same as that leaving the surface. However, greenhouse gases and clouds absorb the radiation emitted by the surface over a range of wavelengths they emit energy in all directions, but since the temperature generally decreases with altitude in the troposphere, less on average is emitted upwards than is absorbed from below, so less reaches the tropopause. The downward radiation from the stratosphere into the troposphere is also an important factor.
CO2, like many other gases absorbs and emits radiation by changing the energy at which it vibrates and rotates, The wavelengths of CO3 absorption are grouped into bands (see e.g. Goody and Yung, 1989) with a strong absorption band centred near 15 μm (Figure 4.1b) The centre of the 15 μm absorption band is so strong that the radiation reaching the tropopause comes from very close to the tropopause and more importantly the CO2 in the stratosphere emits as much downwards as the troposphere emits upwards – the net irradiance is thus very close to zero (see Fig 4.1a – note the feature at about 10 μm is mainly due to ozone.)
Figure 4.1c shows the modelled effect of an instantaneous change in CO2 on the net irradiance at the tropopause. (The change in concentration between 1980 and 1990 is chosen for the purposes of illustration.) All other factors, such as cloudiness and temperature are held fixed. This plot indicates significant change in irradiance. At the centre of the 15 μm band the increase in CO2 concentration has almost no effect – the CO2 absorption is indeed saturated at these wavelengths. Away from the band centre CO2 is less strongly absorbing so that an increase in CO2 concentration does have an effect, The net irradiance at the tropopause decreases – this corresponds to positive radiative forcing that would tend to warm the climate system. As more and more CO2 is added to the atmosphere, more of its spectrum will become saturated – but there will always be regions of the spectrum, which remain, unsaturated and capable of enhancing the greenhouse effect if CO2 concentrations are increased. An example is the 10 μm band system. As shown in Figure 4.1b, it is about 1 million times weaker than the peak of the 15 μm band, but its contribution to the irradiance change in the lower frame is much higher than might be anticipated; as CO2 concentrations increase, the 10 μm band would increase in its importance relative to the 15 μm band.
The saturation effect is partly responsible for the fact that CFC molecules are bout 10,000 times more effective at enhancing the greenhouse effect than molecules of CO2. However, for every extra CFC molecule in the atmosphere since pre-industrial times, there are around 70,000 more CO2 molecules – thus, the relative weakness of CO2 per molecule is more than compensated by the large absolute increase in the number of CO2 molecules in the atmosphere.
Figure 4.1: An illustration that additional amounts of CO2 in the atmosphere do enhance the greenhouse effect- the details of the calculations are given in the footnote to the box. (a) Net infrared irradiance (Wm-2/cm-1) at the tropopause from a standard radiative transfer code using typical atmospheric conditions; (b) Representation of the strength of the spectral lines of CO2 in the thermal infrared: note the logarithmic scale [not bolded in original]. (c) Change in net irradiance at the tropopause (in Wm-2/cm-1) on increasing the CO2 concentration from its 1980 to 1990 levels, whilst holding all other parameters fixed. Note the change in irradiance at the wavelength of maximum absorption, as shown in (b) is essentially zero, while the most marked effects on the irradiance are at wavelengths at which CO2 is less strongly absorbing. Note:The radiative transfer calculations were performed using the standard narrow band code of Shine (1991), which has a spectral resolution of 10 cm-1. The atmosphere used is Northern Hemisphere mean for January, including the effect of clouds, water vapour, ozone carbon dioxide and a number of other gases, although the above argument is non sensitive to the details. The spectrum in the middle frame is the sum of the line strengths in each of the 10 cm-1 intervals at a temperature of 250 K, using the HITRAN database (Rothman et al., 1992); the units are cm-1/(kg m-2). In the lower frame, the irradiance change is the instantaneous forcing.
I’m not going to comment on this argument right now as I want to collate a couple of other thoughts first, but will return to it.
Climate Change 1995 (SAR)
Next came the IPCC Second Assessment Report (SAR) also published in 1995. SPM here ; some sections from Google online here (1-73, 117- 119, 127, 142-148) with a bit of the radiative forcing chapter online (65ff). They re-state their definition of radiative forcing:
(109ff:) the detailed rationale for using radiative forcing was given in IPCC (1994). .. The definition of radiative forcing adopted in previous IPCC reports (1990, 1992, 1994) has been the perturbation to the net irradiance (in Wm-2) at the tropopause after allowing for stratospheric temperatures to re-adjust (on a time-scale of a few months) to radiative equilibrium, but with the surface and tropospheric temperature and atmospheric moisture held fixed …
Their discussion of greenhouse gases is shorter than IPCC (1994); much of the emphasis at this time seems to be on the non-CO2 impacts. IPCC section 2.41 (Greenhouse gases):
Estimates of the adjusted radiative forcing due to changes in the concentration of the so-called well-mixed greenhouse gases (Co2, CH4, N2O and the halocarbons) since pre-industrial times remain unchanged from IPCC (1994); the forcing given there is 2.45 Wm-2 with an estimated uncertainty of 15%. CO2 is by far the most important of the gases, contributing about 64% of the total forcing. .
The basic understanding of the ways in which greenhouse gases absorb and emit thermal infrared radiation (e.g. Goody and Yung, 1989) is supported by abundant observations of the spectrally resolved infrared emission by the clear-sky atmosphere (e.g. Kunde et al., 1974; Lubin 1994)). Barrett (1995)’s suggestion that greenhouse gases are unable to emit significant amounts of infrared radiation is contradicted by these observations.
Again, no readers should doubt that greenhouse gases can emit significant amounts of infrared radiation; downwelling IR radiation can be measured and is not a theoretical construct.
From our limited objective here, nowhere so far has there been a thorough explication by IPCC of how AGW actually works. The Box in IPCC (1994) is the longest discussion in the corpus, but it’s hardly a thorough exposition. I presume that IPCC’s position is that providing such an explication is baby food – however, I think that that’s an unjustifiable position for them to take, as there are many earnest readers, including many CA readers, who looked to IPCC for an explication of how AGW worked in some detail. Or if they weren’t going to do so themselves, they could have cited texts that did so to IPCC satisfaction in their list of references, so that people inquiring into the topic would not be thrashing.
Rind, D., and A. Lacis (1993), The role of the stratosphere in climate change, Surv. Geophys., 14, 133-165. Abstract