In the preceding sections the characteristics of the SRES scenarios are summarized in terms of scenario driving forces such as population, economic development, resources, technology, land-use changes, and other factors. The scenarios were designed in such a way as to deliberately span a wide range, reflecting uncertainties of the future, but not cover the very extremes from the scenario literature concerning driving forces. A distinguishing feature of the SRES scenarios is that various driving-force variables are not combined numerically (or arbitrarily), but instead try to reflect current understanding of the interrelationships between important scenario driving forces. For instance, according to the literature review of Chapter 3 it would be rather inconsistent to develop scenarios of rapid technological change in a macro-economic and social context of low labor productivity and stagnant income per capita. Scenario storylines were the method developed within SRES to help guide the scenario quantifications and to assure scenario consistency in terms of the main relationships between scenario driving forces.
The different quantifications discussed in the previous sections demonstrate
that even if scenarios share important main input assumptions in terms of population
and GDP growth, considerable uncertainty and scenario variability remains. For
instance, features of technological and land-use changes can be interpreted
quite differently within the framework of different models, even if they conform
to the overall conceptual description and "scenario logic" described in a particular
scenario family. In some instances the broad outlines of scenario driving forces
were not followed entirely in particular scenario quantifications, but alternative
scenario interpretations were submitted. These highlight important scenario
uncertainties or express scientific disagreements within the writing team, as
for future labor productivity growth and economic "catch-up" possibilities of
currently developing countries. These alternative scenario interpretations and
different model quantifications are presented here to reflect the SRES Terms
of Reference for an open process and the use of multiple modeling approaches,
even if this necessarily increases complexity and reduces simplicity and transparency
in discussion of a large number of scenario quantifications.
Spatially explicit data on socio-economic activity is sparse. The reason arises mainly in that Systems of National Accounts and similar socio-economic statistics are available only at high levels of spatial aggregations defined by administrative boundaries (countries, provinces, or regions). As a result, gridded emission inventories largely rely on estimations of current population density distributions (e.g. Olivier et al., 1996) and modeling approaches to date have also relied exclusively on rescaling future socio-economic activities (economic output, energy use, etc.) based on current or future population density distribution patterns (see e.g., Sørensen and Meibom, 1998; Sørensen et al., 1999). Population density is, however, not necessarily a good indicator for spatial patterns of socio-economic activities (Sutton et al., 1997). At higher levels of spatial aggregation, it is estimated that about two billion people remain outside the formal economy, most of them in rural areas of developing countries (UNDP, 1997). At lower levels of spatial aggregations, locations such as airports, industrial zones, and commercial centers have low resident population densities, but high levels of economic activity. Also, with increasing urbanization future population distribution will be markedly different from present ones (HABITAT, 1996).
Night satellite imagery from the US Air Force Defense Meteorological Satellite Program (DMSP) Operational Linescan System (OLS) offers an interesting alternative based on direct observations. Early nighttime lights data were analyzed from analog film strips (Croft, 1978, 1979; Foster, 1983; Sullivan, 1989). Digital DMSP-OLS data have recently become available with global coverage (Elvidge et al., 1997a, 1997b, 1999). Nocturnal lighting can be regarded as one of the defining features of concentrated human activity, such as flaring of natural gas in oil fields (Croft, 1973), fishing fleets, or urban settlements (Tobler, 1969; Lo and Welch, 1977; Foster, 1983; Gallo et al., 1995; Elvidge et al., 1997c). Consequently, extent and brightness of nocturnal lighting correlate highly with indicators of city size and socio-economic activities such as GDP, and energy and electricity use (Welch, 1980; Gallo et al.; 1995; Elvidge et al., 1997a).
Figure 4-13 (bottom panel) shows a 1995/1996 night-luminosity map of the world developed by National Oceanic and Atmospheric Administration's National Geophysical Data Center. The map was derived from composites of cloud-free visible band observations made by the DMSP-OLS (see Elvidge et al., 1997b; Imhoff et al., 1997). The DMSP-OLS is an oscillating scan radiometer that generates images with a swath width of 3000 km. The DMSP-OLS is unique in its capability to perform low-light imaging of the entire earth on a nightly basis. With 14 orbits per day, the polar orbiting DMSP-OLS is able to generate global daytime and nighttime coverage of the Earth every 24 hours. The "visible" bandpass straddles the visible and near-infrared (VNIR) portion of the spectrum. The thermal infrared channel has a bandpass that covers 10-13 µm of the spectrum. Satellite altitude is stabilized using four gyroscopes (three-axis stabilization), a starmapper, Earth limb sensor, and a solar detector. Image time series analysis is used to distinguish lights produced by cities, towns, and industrial facilities from sensor noise and ephemeral lights that arise from fires and lightning. The time series approach is required to ensure that each land area is covered with sufficient cloud-free observations to determine the presence or absence of VNIR emission sources.
Dietz et al. (2000) performed illustrative simulations of possible future light patterns. Considering the high correlation between observed radiance-calibrated night satellite imagery data and economic activity variables like GDP, these simulations indicate future spatial patterns of economic activity. As a basis of the simulation, Dietz et al. (2000) rescaled present night luminosity patterns using global and regional GDP growth patterns of the preliminary A1B marker scenario reported in the SRES open-process web site combined with a simple stochastic model of spatial evolution and interaction. An illustrative simulation for the year 2070 is given in Figure 4-13 (top panel). The resultant changes in spatial light patterns indicate socio-economic activities and provide useful information for infrastructure planning, such as expansion of gas and electricity networks. When combined with topographical information, like latitude, the data can also be used as input to climate impact and vulnerability assessments (e.g., extent of socio-economic activities that may be affected by sea-level rise).
Figure 4-13: Radiance calibrated lights obtained from night satellite imagery. Situation in 1995/1996 (bottom panel) and illustrative simulation for the SRES A1 scenario's implied GDP growth for 2070 (top panel). Color codes refer to radiance units (DN), where radiance = DN3/2 ×10 -10 W/cm 2 per sr/µm (Watts per square centimeter per steradian per micrometer, the brightness units to which the US Air Force Defense Meteorological Satellite Program (DMSP) Operational Linescan System (OLS) is calibrated; it normalizes for the bandpass (m) and solid angle of the optics (cm 2 /sr)).
To guide readers through the different driving-force assumptions that characterize the various scenarios, Tables 4-2 and 4-3 give an overview of the SRES scenario set. They classify scenarios that share important input assumptions (harmonized scenarios share global population and GDP assumptions) from scenarios that offer alternative quantifications. Table 4-4 summarizes the main quantitative scenario descriptors for each of the four SRES scenario families of the "harmonized" scenario category. Here an attempt is made to link this information with the resultant scenario outcomes (emissions) that are discussed in more detail in Chapter 5.
In Chapter 5, an additional, complementary scenario classification scheme to that used in this chapter is presented and focuses on driving forces. Scenarios are classified according to their cumulative carbon emissions (1990 to 2100, all sources), the best single quantitative indicator available to compare emission scenarios that portray widely different dynamics and different combinations and magnitude of a variety of emission categories. Four categories of cumulative emissions, Low (<1000 GtC), Medium-Low (1100 to 1450 GtC), Medium-High (1450 to 1800 GtC), and High (>1800 GtC) are presented. Table 4-20 links the scenario overview from Tables 4-2 and 4-3 with this information to guide readers through the differences in scenarios.
|Table 4- 20: Overview of SRES scenarios categorized into the four scenario families and associated scenario groups (four for the A1 family, combined into three in the SPM one for each of the other scenario families). The scenarios are classified as "harmonized" and "other" scenarios with respect to whether they share harmonized input assumptions on global population and GDP growth (see also Tables 4- 1 to 4- 4). A second layer of classification relates to scenario outcomes in terms of cumulative emissions (see Chapter 5). Four categories are distinguished: Low (< 1100 GtC), Medium- Low (1100- 1450 GtC), Medium- High (1450- 1800 GtC), and High (> 1800 GtC).|
|Marker||A1B- AIM||A2- ASF||B1- IMAGE||B2- MESSAGE|
|Globally||A1C- AIM||A1G- AIM||A1B- ASF||A1T- AIM||A2- MESSAGE||B1- AIM||B2- AIM|
|Harmonized||A1C- MESSAGE||A1G- MESSAGE||A1B- IMAGE||A1T- MESSAGE||B1- ASF||B2- MARIA|
|Scenariosa||A1G- MiniCAM||A1B- MARIA||B1- MESSAGE||B2C- MARIA|
|A1B- MESSAGE||B1- MiniCAM|
|A1B- MiniCAM||B1T- MESSAGE|
|Other Scenarios||A1C- MiniCAM||A1v1- MiniCAM||A1T- MARIA||A2- AIM||B1- MARIA||B2- ASF|
|A1v2- MiniCAM||A2G- IMAGE||B1High- MiniCAM||B2- IMAGE|
|A2- MiniCAM||B2- MiniCAM|
|A2- A1- MiniCAM||B2High- MiniCAM|
|Scenario outcomes: Cumulative emissions, GtC 1990- 2100|
Classification according to cumulative CO2 emissions (Chapter 5):
Table 4-20 indicates that in most cases there is an easily discernable direct connection between main scenario characteristics of a particular scenario family or scenario group and the resultant outcomes in terms of cumulative emissions. For instance, in the high GDP, high energy demand scenario family A1, all scenarios within the two scenario groups that are fossil fuel and technologies intensive (A1C and A1G combined into A1FI in the SPM) result in high cumulative carbon emissions. Conversely, cumulative emissions of the "efficiency and dematerialization" (without additional climate initiatives) scenario family B1 are generally in the "low" emissions category, but two model quantifications indicate medium-low emissions. For the scenario family B2, outcomes in terms of cumulative carbon emissions can also be related clearly to scenario characteristics. One group of scenarios (which includes the B2 marker) adopts an incrementalist perspective of technological change ("dynamics as usual") applied to medium levels of population and GDP (and resultant energy demand) and results in medium-low cumulative carbon emissions. Another group of scenarios explored the sensitivity of a gradual return to coal-based technologies (B2C-MARIA, B2-ASF), in one case combined with higher energy demand than in the other scenarios (B2High-MiniCAM); and results in higher cumulative emissions (Medium-High category in Table 4-20).
Equally discernable in Table 4-20 is the wide range in cumulative carbon emissions that characterize the various scenario groups within the A1 scenario family. By design, the different scenario groups within this family explored the implications of different directions of technological change, ranging from carbon-intensive developments (A1C and A1G, combined into A1FI in the SPM) to decarbonization (A1T), with the "balanced" technology development scenario group taking an intermediary position. Different developments concerning fossil or non-fossil resource and technology availability in a less populated but affluent and thus high energy demand world (such as A1) can lead to widely different outcomes in terms of cumulative emissions, with a range as wide as that spanned by all four scenario families together. Technology can thus be as important a driving force as population and GDP growth combined. In other words, very different emissions outcomes are possible for future worlds that otherwise share similar developments of main driving forces such as population and economic growth and high rates of technological change.
However, areas of overlap and uncertainties of scenario outcomes (cumulative emissions) occur even for scenario quantifications that share otherwise similar assumptions for the main scenario drivers. Not surprisingly, differences in quantifications are largest within the A1 "balanced" technological progress scenario group, which includes the A1B marker scenario. Most model interpretations result in cumulative carbon emissions within the Medium-High category (1450-1800 GtC). However, there are also scenario quantifications in which technological change tilts more in the direction of the A1C (A1-ASF) or A1T (A1-MARIA) scenario groups that favor fossil (coal) or post-fossil (nuclear, renewables, and biomass) technologies, respectively. This leads to very wide differences in cumulative emissions, from the Medium-Low through to the High categories. A similar range of scenario outcomes between Medium-High to High categories also characterizes the A2 scenario family that otherwise describes an entirely different world (high population and comparatively low per capita income compared to low population with high per capita income for the A1 scenario family; see Table 4-4). Departing from the main scenario characteristics of the A2 scenario family in terms of population and income in direction of lower values (such as in the A2-A1-MiniCAM scenario) could even yield emissions in the Medium-Low category. Thus, the A2 scenario family also indicates that a wide range of emissions outcomes is possible for any given development path of main scenario driving forces, such as population and income per capita.
Finally, the categorization of scenarios in terms of their (cumulative carbon) emission outcomes illustrates that similar emission outcomes could arise from very different developments of main scenario drivers. For instance, High category cumulative emissions could arise from scenarios of low population growth, combined with high incomes (and energy use) and globalized technological developments that favor accessibility and economics of fossil fuels (coal, unconventional oil and gas; e.g., A1C and A1G scenario groups). Alternatively, similar High category cumulative emissions could also arise from scenarios of high population growth combined with slower per capita income growth and more regionally oriented technology development trends (scenario family A2). A comparison of the B1 and A1T scenario groups (see Table 4-20) also confirms this conclusion. Both scenarios explore pathways that reduce current income disparities between regions. They indicate that such a tendency does not necessarily lead to high emissions, but could be achieved with Low to Medium-Low category cumulative emissions (as scenario groups A1C and A1G also indicate that High category emission pathways are possible).
Perhaps the most important conclusion from the SRES multi-model, open process, and the large number of scenarios it has generated is the recognition that there is no simple, linear relationship between scenario driving forces and outcomes or between emission outcomes and scenario driving forces. High or low population scenarios need not automatically lead to high or low emissions; similar statements also hold for economic growth and for closing regional income gaps.
Other reports in this collection