Review of the U.S. Climate Change Science Program's Synthesis and Assessment Product 3.2, "Climate Projections Based on Emission Scenarios for Long-lived and Short-lived Radiatively Active Gases and Aerosols" (2007)

This chapter provides specific comments on the eight individual chapters of draft Synthesis and Assessment Product (SAP) 3.2. In some cases, these specific comments relate to the overarching comments provided in the previous two chapters of this review. In the other cases, these specific comments are generally minor in nature. The review of each chapter includes a statement that summarizes the committee’s overall thoughts. For some chapters, there are enumerated comments that follow this statement to provide suggested editorial changes or other details for the authors to consider during the revision process.

Chapter 1 needs similar revision to the document as a whole to make it easier to read. The committee is concerned that this chapter is not written so that it can easily be understood by the non-specialist. In particular, readability is impaired by frequent use of acronyms and abbreviations. These concerns are especially relevant to this chapter, as it sets the stage for (and provides a summary of) the other chapters. For example, there are many instances in which undefined acronyms are used or defined at a later point. In general, acronyms are over used and detract from the flow of the material. Certain terms (for example, “very likely”, “likely” etc) are used with the “specific IPCC connotations”. These connotations should be defined, or the text translated so that they are accessible to the non-technical reader, particularly in the “key findings”.

The document should be revised by including explanations and use plain language that make the results more easily interpretable to a non-technical audience. Some specific examples are to explain the following:

• What is an integrated assessment model?

• What is a scenario?

• Explain how IAMs differ from climate models

• Explain why the IAMs differ from each other, and why it is important to use more than one

• Explain why Radiative Forcing is an important concept.

The “Historical Overview” section is useful and well-written. The overview of the IPCC reports (beginning at line 414) could benefit from a very brief statement of what the IPCC is as well as the scope of IPCC assessments (i.e., review of current literature; there is a common misconception that IPCC performs research). At lines 434-441 it could be noted that the models are moving toward finer resolution that can include some topographic features that are important to U.S. climate. Finer resolution in the ocean now allows some important atmosphere-ocean coupling processes such as ENSO to be represented in some AOGCMs (see e.g., van Oldenborgh et al. 2005; Wittenberg et al. 2006).

The three AOGCMs and modeling groups should be briefly introduced in this chapter. Care should be taken to distinguish the AOGCMs from the IAMs, especially for the benefit of non-specialists. Text (perhaps a box) describing these types of models and functions as an introduction to a non-technical reader should be included.

An indication should also be given as to whether the AOGCMs used in the study are appropriate to the task at hand. This need not be a detailed performance evaluation; it would be adequate simply to state that intercomparison studies have shown that the performance of these models is comparable to other state of the art climate models.

Finally, the methodology and its limitations should be made clearer at the outset and should also explain why new emissions scenarios are needed. . .

L428-429: “model” should be plural, “models”

L506: Methane is reactive not only in the troposphere but also in the stratosphere (being a main source of water vapor in the upper stratosphere).

The committee feels that the chapter contains much useful material that serves to fulfill the mandates of the prospectus. It also feels that the chapter can be improved in several regards.

First, the chapter needs revisions to make it easier to read. It also assumes the reader to be a technical expert, and should either have a summary for non-technical reader, or be identified as such at the beginning.

As presently written, the chapter is presented in two parts. Material from the beginning of the chapter to page 39 describes work that examines the climate implications of emissions scenarios developed in SAP 2.1a. Particular emphasis is placed on the combined roles of the Kyoto and non-Kyoto short-lived gases. The chapter notes that all of the scenarios are contained within the range of the scenarios examined by IPCC Working Group I (WGI) in the Fourth Assessment Report (AR4). (Though, one of the attributes of the SAP 2.1a is that it contains stabilization scenarios that fall outside of the range of the SRES scenarios on the low side.) The chapter then argues that because the range of climate scenarios associated with the SAP 2.1a falls within the range of scenarios examined in the IPCC WGI AR4 that the same general conclusions follow. The chapter then proceeds to summarize the IPCC results in the second half of the chapter, Pages 39-51.

The committee recommends that the authors consider revising the chapter to focus more on the first material, moving the summary of results of the IPCC WGI AR4 to a box, and adding a section that identifies the role of the short-lived species that could serve as motivation for and transition to Chapter 3.

The motivation for and important conclusions arising from the section on regional climate models needs to be clarified. The committee speculated that the intent was to show the similarities in surface temperature change and ozone change between the global and regional models, but was left wondering if there was more to the section. The committee agrees that more research is clearly needed to assess if downscaled RCM simulations improve our ability to characterize climate change (lines 1132-1145), but this statement might be better suited for Chapter 4.

Figures 2.1, 2.2, 2.3, 2.4: Show the three SAP 2.1a reference cases as well as the stabilization cases. Also note that in addition to the SRES cases that a commitment run is also shown. (The latter is what allows the assertion that the IPCC work contains the SAP 2.1a.)

Figures 2.1, 2.2, 2.3, 2.4: In general it would be good to provide a table for year 2100 values. This would make it easier for the reader to get a sense of the absolute differences between cases.

L27642-643: How can the lower bound of the 5-95% range (i.e., 0.19 m) be less than the minimum of the entire range (0.28 m)? Please clarify.

L28655-658: ENSO and the AMOC are two different phenomena. Lumping them together risks creating a misleading association in the minds of non-experts.

L29688-690: This is right for one model, MERGE. It is a formal optimization model. The other two models are recursive. However, the two recursive models adopted two assumptions that resulted in results being similar to that of a formal optimization model. Both assumed that all regions of the world and all economic activities faced the same price of carbon, though each model adopted their own treatment of the non-CO₂ greenhouse gases. Only MERGE was a true optimization frame. The MiniCAM adopted the assumption that the price of carbon rose at the rate of interest plus the rate of net loss of carbon from the atmosphere to the ocean-terrestrial system. This is consistent with intertemporal cost optimization for carbon. The IGSM used a similar assumption, namely that the price of carbon rises at the rate of interest. For the purposes of this report it should be adequate to simply state that, “All of the groups developed pathways to stabilization targets designed around economic principles. However, each group used somewhat different approaches to stabilization scenario construction.”

L30694-695: “...trajectories ... were produced...”. Would a non-expert know what this phrase means?

L30696: Change “optimization process” to “scenario definitions”.

L31724: What are “F-gases”?

L32739: Note that the MiniCAM uses MAGICC as its representation of carbon cycle and the atmosphere.

L34786:Some discussion of the methodology employed to link MERGE output to MAGICC, particularly in the carbon cycle is needed. The MERGE model appears to have adjusted its ocean to reproduce essentially the same behavior as the other two models’ combined ocean and terrestrial system models. This sparks the question of how was this case run so that it is true to the underlying MERGE approach?

L35805-806: committee is skeptical that EPPA runs a 200-year trajectory to stabilization. This assertion should be checked.

L35800-816: It is important to note that in general all of the models hit their targets with their own carbon cycle and atmosphere models. Thus, failure to hit targets in MAGICC is not the same thing as failing to meet the stabilization target for the model. The methodology employed to get the radiative forcing and transient temperature changes were to have each model use the MAGICC atmosphere. However, the models did hit the target using their own atmospheric representations. Thus, when results differed when all used the MAGICC atmosphere, this would seem to be a reflection of underlying uncertainty in the carbon cycle models, which, as we all know, is substantial.

L36823-825: It seems odd to replace the BC and OC from the MiniCAM and MERGE models with arbitrary trajectories.

L36836: What does “later for MiniCAM” mean? The comment is unclear.

L37842-843: This statement should be checked. Is it really true that SO₂ emissions are invariant across the scenarios as indicated? How could SO₂ emissions not vary with dramatic reductions in fossil fuel use? EPPA assumes lots and lots of CO₂ capture and storage, which means almost complete clean up of S, so SO₂ emissions should decline as the stabilization level tightens.

L 371132-1145: Downscaled information is necessary for more than air quality; in particular, it is needed for hydrologic and agricultural uses (among many others).

The committee thought that Chapter 3 was the most substantive. They believed that this chapter should more clearly identify what the major take-home messages are and should also consider including additional analysis of the mechanisms involved.

The authors’ main point, that the short-lived greenhouse gases are important factors in projections of future climate, is well supported. However, the climate models do not use consistent forcing scenarios for the short-lived species, nor do they use consistent natural emissions of primary aerosols and ozone and aerosol precursors or consistent removal mechanisms for the short-lived species. This makes comparison of the model results challenging.

Additional discussion of the difference between uncertainties in processes and uncertainties in future emissions is needed. Uncertainties in chemical and physical processes represent the state of our current knowledge. The fact that one modeling group chooses to include a process while another group chooses not to shows that our knowledge about short-lived species is still evolving. Eventually, with further research, uncertainties in chemical and physical processes can be ironed out. Uncertainties in future emissions, however, will never be completely erased. What modelers can do is choose consistent emission scenarios to bracket possible future outcomes.

The authors need to emphasize that the magnitude and signs of effects of the short-lived species on climate may be totally different using different projected emissions in the same models. Also, the committee thinks that following the A1B emission scenarios for the precursors of short-lived species all the way out to 2100 may result in unrealistically large surface concentrations of pollutants. The committee recommends inclusion of a figure showing the monthly mean surface ozone, BC, and OC at 2100. A caveat should then be added that such large abundances are not likely to be tolerated.

The committee recommends the addition of a table that includes descriptions of each of the models, including resolution, inputs, reactive chemical mechanisms, emissions assumptions, removal mechanisms, and residence times. In the accompanying discussion, sufficient detail should be provided for each experiment regarding what radiatively active species are predicted (emissions) vs. those prescribed (concentrations) and how they vary temporally and spatially so that the reader can understand exactly what was done. Discussion of this table should include some analysis of the differences between model results produced by the different parameterizations.

A graphic comparison of the temperature response to short-lived species vs the response to long-lived species should be presented. In this way, readers can appreciate 1) the contribution of the short-lived species to future climate change and 2) the similarities (or differences) of the responses to the short-lived vs long-lived species. Past work investigating the climate response to heterogeneous forcing should be discussed.

Many of the plots showing future changes provide only the annual mean. Because the short-lived species have large seasonal trends, plots showing seasonal forcings and temperature responses are essential. Much information could be lost in the mean. This is true of course for the surface temperature response, but also especially important for the response of precipitation to changes in aerosol. It could be that the seasonal precipitation response would have much greater statistical significance than the annual mean precipitation response.

Regional changes, particularly surface temperature, appear to be important, and the committee recommends that considerably more attention, discussion, and analysis be paid to this, including a comprehensive treatment of uncertainty. For example, a summertime 2°C increase over the central United States by 2100 would have large consequences for both human health and the economy.

Results on how temperature responded to changes in short-lived species would be greatly strengthened by additional sensitivity studies that could help to establish causes and mechanisms. For example, in the GISS model, how much warming did the declining trend in the indirect effect contribute to the climate response and where? How would the GISS results differ if dust had not been permitted to take up sulfur dioxide? Determining the relative importance of these and other processes to the climate response would help prioritize the gaps in our knowledge. In addition, a discussion of how the system might respond to controls on short-lived species and the possible feedbacks, and what the impact of climate changes might be on short-lived species would be helpful. At this point, there is sufficient information from present study and previous ones to get an approximate idea of what the feedbacks and control sensitivities are on the system to get a first order estimate of what controls on short-lived species and their precursors might do to climate. While the committee notes that the present document would benefit from these additional analyses, it may not be feasible given time and potential monetary constraints; in such a case, a recommendation for future analyses should be included in Chapter 4.

Discussion and citation of previous studies is insufficient. The authors need to show that their work builds on what has already been done. Also, citing previous work will enrich the study by making clear where various model agreements and disagreements lie, and will help clarify how robust the current findings are.

All of the methods used to calculate statistical significance should be described in detail either within this chapter or in an appendix.

The authors should emphasize that the impacts of climate change on the short-lived species were not included in this exercise, except for the methane/isoprene simulations. The authors should refer to other studies that show the relative importance of these climate impacts, and briefly describe how including such impacts might affect their results.

It is not clear how the model simulations were set up, and why the authors made the choices they did. How were the time-slice monthly chemical fields of ozone and aerosol implemented in the transient climate simulations? Were the future composition simulations performed with present-day climate? How were the effects of long-lived species implemented in the models, as forcings or concentrations? The description of ensemble runs needs clarification for the lay audience.

The methane text leads to many questions. Was three years a sufficient length of time to calculate the methane response to changing climate and chemistry? How much did OH concentrations further decline when the biogenic emissions of methane and isoprene were permitted to interact with the changing climate? What chemical mechanism was used for isoprene oxidation? (The choice of mechanism could make a difference in the outcome. Given that the fate of isoprene oxidation products is a major issue among air quality modelers, this has importance. See Wu et al., 2007.) Was OH also allowed to respond to changing water vapor concentrations? How did changes in NOx emissions impact OH? This section also neglects much previous work looking at the effect of changing emissions and/or climate on methane abundances, e.g., Wild et al., 2001; Wigley et al., 2002; Stevenson et al., 2006.

The methane text should not be a box, but a separate section. The result here, that including biogenic chemistry-climate impacts increases methane concentrations and thus climate forcing, has importance and should be included in the chapter summary.

The methane section could end with a brief description of other chemistry-climate feedbacks that could play a major role in the future climate. Processes involved in the feedbacks include: lightning NOx emissions, land cover change, changes in convection and transport, and changes in absolute humidity.

The current practice to include tropospheric chemistry in global models has a common problem in methods, i.e., the simplified representation of subgrid-scale processes (e.g., fast chemistry affecting species from nitrogen oxides and isoprene to ozone, nucleation of aerosols). The authors are encouraged to make a comment on the consequences of using coarse-grid models to describe fine-scale chemistry.

The bullets at the beginning of the chapter could be revised to ensure that key points are highlighted. The first bullet in the Introduction and Key Findings section, line 1328, is awkward. Bullet 2, line 1333, infers that short-lived species are emitted when actually some of the most important (ozone and sulfate) are formed in the atmosphere.

Line 1500. It’s more appropriate to use “amount of sulfate and ammonia^” instead of “amount of sulfate”. Note that the added detail that a lognormal distribution is assumed for all aerosols is not needed for this audience.

Line 1572: Why would the treatment of natural and biomass burning emissions affect sulfur dioxide emissions to such a large extent?

Lines 1606-1608. Nitrate can be a dominant component of aerosol during the winter, and may therefore play an important role in climate at that time of year. Therefore, the reviewers are not convinced that nitrate has a “minimal effect” and that it doesn’t matter that only GISS includes nitrate aerosol. Further, as sulfate concentrations decline, and ammonia increases (as estimates suggest will be the case), nitrate may become an even bigger player.

Many of the Figure captions are not clear. E.g., in Figure 3.2, what is being shown here in what units, and for what time period? “Other” in the NCAR bar should be defined differently in the labels. Most of the captions are not “stand-alone.” The reader needs to burrow through the text to know what is going on in this plot.

Table 3.3. The term “model production efficiency” is confusing, since it resembles the well-known but differently defined term ozone production efficiency. The reviewers suggest employing a different term, such as burden-emission ratio (BER). Alternatively, what would be lost if the authors instead just looked at species lifetimes?

Also in Table 3.3, the GFDL ozone production efficiency declines dramatically between 2000 and 2030 (7.19 to 2.24). Why is this? In the A1B future, volatile organic compounds go up and NOx goes down, which would typically lead to a higher OPE. The authors need to explain this large jump.

Line 1650. This paragraph describes several trends in aerosol production efficiencies. The authors need to attempt to explain the reasons for these trends. Could trends in wet and dry deposition of sulfate, nitrate, BC, OC and other aerosols and precursor gases matter?

Line 1651 What does “are more effective” mean? Please clarify.

Paragraph beginning with Line 1720. The reviewers are surprised that OC is not considered a more major player. Could the authors comment on this small contribution?

A better figure and table of the radiative forcing of each of the different short-lived greenhouse gases for each model would be helpful (Figure 3.3)

Color-coding the tables to emphasize the sign and magnitude of the differences in burden and emissions would be helpful.

Consistent scales and map projections would be helpful in Figures 3.4 and 3.7.

Figure 3.4. The reviewers suggest including maps of radiative forcing from both long-lived and short-lived greenhouse gases for the same time periods.

Lines 2176-2177. The authors state that uncertainties in socio-economics dominate “uncertainties in physical sciences.” Here chemical mechanisms should also be mentioned because of important chemical reactions for sulfate and ozone formation.

Line 2248-2250. This sentence concerning the possible impact of future air-policy decisions in Asia on U.S. climate change is loaded with importance and needs more discussion.

Table 3.8 The table of radiative forcing impacts from regional sector perturbations is interesting, but needs more explanation in the caption. What are the perturbations? Give them in the table caption or footnote.

Lines 2276-80: use 0.01 W rather than 10 mW (and not 10 MW!)

Paragraph beginning with line 2293. This paragraph assumes that reducing surface transportation emissions of short-lived species and their precursors is done by reduced fuel consumption. This is not necessarily the case. Indeed, some controls might increase fuel consumption.

The title provides a nice paradigm for the chapter as it suggests that issues, opportunities and recommendations will be discussed. However, few issues, opportunities or recommendations are apparent in the organization or presentation of the chapter material. The committee offers the following suggestions for reorganization within the chapter:

1. Introduce this chapter with a restatement of what the scope of this SAP is, why the scope has been so defined (what was seen to have highest priority

and why; what it was possible to do at the time), and what is not being addressed in this SAP;

1. Avoid jargon and acronyms; use more functional descriptions of models (with model names in parentheses). Otherwise, the text is seen as inaccessible and thus detail-dense information is largely lost;

2. Refer back to new table(s) added to chapter 1 (Introduction) in which the different model configurations are described; possibly add figures to clarify steps taken/model process and use examples from chapter 3 (perhaps even show one of the figures (e.g. 3.1 bottom) again in this chapter) to highlight findings/conclusions drawn in this SAP;

In addition the chapter would be improved by the addition of a section recapitulating the highpoints of the study. Some suggestions are.

1. The SAP model scenarios for long-lived species produce projections that are within the IPCC range, although it should be noted that the SAP response range tends to be lower than all but the IPCC “commitment scenario”.

2. The most important uncertainties in characterization in short lived species were found to be emissions and the indirect effect.

3. Part of the reason for the different emission inventories used here and in the IPCC studies was that the Integrated Assessment Models did not recognize that these species were necessarily important when the scenarios were first constructed. Clarification of the challenges associated with emissions projections (not a simple matter of improving quantitative skill, as these are a function of difficult-to-anticipate socioeconomic choices) should also be included;

4. Natural aerosols are also important and their emissions differed greatly between the models, with consequences to the role of anthropogenic emissions. The inconsistencies between models should be reconciled in future studies.

5. Calculation of the indirect effect is potentially the single most important deficiency in the study. The modeling community as a whole cannot as yet produce a credible characterization of the climate response to aerosol/cloud interactions. All models (including those participating in this study) are currently either ignoring it, or strongly constraining the model response. Using a box to highlight this issue is recommended. Additional referencing of work on aerosol indirect effects is also recommended.

6. The results suggest that the short-lived species do matter to the climate in the long term (e.g. out to 2100). The presence of radiatively active short-lived species can significantly change the regional surface temperature

response (for example over the continental US). It is noteworthy and surprising however that the response location is not local to the forcing.

1. The 3 model frameworks participating in the study produced different outcomes. Each model represents a thoughtful, but incomplete characterization of the driving forces and processes that are believed to be important to the climate. Much work remains to be done before there can be confidence that the climate response to short-lived species is well understood.

At present, a list of the priorities and opportunities for future work is not presented in the chapter. This is unfortunate; because the result of this report clearly illustrated that there are many needs for future research.

Additional regional modeling studies could provide information on local effects of short-lived species on regional climate. For example, the local impacts of aerosol forcing on surface temperature and precipitation could be significant. In addition, there is a need for modeling studies with finer resolution models, both at regional and global scales, to determine the resulting impacts on derived effects from short-lived emissions

This SAP examines only a subset of processes controlling short-lived species and their interactions with clouds. Other processes might be important but have not been addressed such as ice clouds and their interactions with short lived species and the climate system

There is evidence that future biomass burning and land cover change could have a large effect on the climate response.

To conclude the document, a reflective assessment of the product would be useful. For example, – what lessons were learned during these experiments, what would be done differently if the experiments were to be repeated? How should other experiments be set up to answer the key questions generated by this study?

Page 15 of the executive summary introduces a different chapter 4 than exists and raises expectations regarding a long list of other potentially important short-lived species and anthropogenic impacts (land use change, reactive nitrogen deposition and ecosystem responses, changing VOC emissions, changing oxidant and SOA formation).

A more-detailed discussion of the impact of fire (biomass burning) on aerosols and hydroxyl is needed. A concern was raised that 2-3 year runs are not long enough to capture the impact of ENSO-related fires in the tropics (like the 1997-1998 Indonesian fires). This may be an issue where current capabilities restrict a thorough treatment as part of this effort. If so, this should be stated in Chapter 3, and this issue should be raised as a future need in Chapter 4;

L2642: This line seems to all of a sudden pop in here, yet is potentially rather important in regards to results in Chapter 3. Given that NH3 comes from a rather large number of processes, some rather disconnected to N2O production, this seems a stretch.

L2750 “five different RCMs” should be “six different RCMs…”

L 2755-2756 The sentence beginning “Future IPCC…” should be corrected to “The IPCC A1B scenario is used in this intercomparison study.” (i.e., only A1B will be used and not A2)

L 2790: “The future sources of most of

L 2800: They need to be more specific as to what is meant by biofuel, and also need to show it is CO2 neutral. In many cases, it is not.

L 2841: This last sentence is rather weak and equivocating. From what is presented, reducing NOx will reduce tropospheric ozone. Reducing tropospheric ozone should reduce radiative forcing. The report should have a strong ending.