Friday, August 26, 2011

Representative Concentration Pathways

Everyone in the climate policy field is familiar with the SRES - Special Report on Emissions Scenarios developed by the IPCC more than a decade ago. It was time for an update and we now have the new RCP's. The primary focus here is on atmospheric concentrations of greenhouse gases and other radiatively active substances. There are four pathways with radiative forcing of 2.6, 4.5, 6, and 8.5 Wm^-2. The new scenarios are being published as a special (open access) issue of Climatic Change. The overview article is authored by Detlef van Vuuren et al. I'll focus here on the related emissions pathways.

The figures compare the new scenarios to each other, the general literature, and the 4 main SRES pathways. The main thing to note is that the 2.6 RCP has much stronger climate action than any of the SRES scenarios on both CO2 and CH4 emissions. For local air pollutants, the new pathways show much lower levels than any of the SRES pathways, especially for sulfur:

This makes sense and is in-line with new understanding on both historical and likely future sulfur emissions. This will make controlling global warming harder than would have been thought with the old scenarios.


  1. I am seriously disturbed by what seem to me to be very large lacunae in this van Vuuren et al paper.your “The representative concentration pathways: an overview”. Like most if not all its predecessors in SRES and AR4, it seems to me the paper treats GHG emissions, mainly CO2, as identical with additions to the atmospheric concentration of CO2, i.e. [CO2].

    For clearly, at no point does the paper display both the emissions and the uptakes that actually generate the “pathways” of NET additions to [CO2]. Leaving out the uptakes, especially whan they are not constant but vary with the actual level of [CO2], is incomplete accounting and misleading.

    In particular, the van Vuuren paper fails to disclose HOW it uses a “selected simple carbon-cycle climate model” to produce “concentration data”.

    It also fails to disclose the “wood harvest” data it uses, given that more than half of all so-called deforestation results in nearly permanent storage of carbon in harvested wood products (my own grandfather’s birthplace’s roof timbers have recently been carbon dated to 1227). I suspect that like Kyoto and Pan, Canadell et al., “Science”, 2011, van Vuuren et al's models implicitly assume all harvested timber is immediately oxidised, but that is not apparent as I sit at my own wooden desk.

    More pertinently, the paper admits it bases itself on Tom Wigley’s MAGICC6 model, which not only ignores all atmospheric water vapor but also assumes that all macro global biotic absorptions of CO2 emissions are based on the micro Michaelis-Menten function, which assumes that all uplifts of CO2 by the global biota conform to an asymptotic hyperbolic function whereby they level off at some point (usually 2000 in MAGICC) such that thereafter ALL emissions necessarily and unavoidably immediately equate to increases in [CO2]. Wigley has a long record of ignoring empirical evidence, of which there is none to support his macro Michaelis-Menten assumption (which is valid for any individual plant, cet. par., but is invalid when it implies there can never again be a net addition to plant life anywhere on this planet).

    This results in the paper's wholly inadequate Section 3.2 Land use, which fails to display BOTH emission AND absorption pathways to yield its claimed NET increases in [CO2], which in the absence of the absorption pathways can only be described as misleading at best. Then the paper states, “the greenhouse gas concentrations in the RCPs closely correspond to the emissions trends (Fig.9)”. But the emissions currently grow at c. 3% p.a. while the atmospheric concentration grows at less than 0.5% p.a. (Source: CDIAC, Marland et. al.), and the correlation between the emissions trends and the [CO2] trends is less than impressive. The actual linear trend in emissions is y = 0.1149x + 3.9085, R² = 0.9758, while that in the changes in the atmospheric concentration of CO2, [CO2], is y = 0.0256x + 0.7621 (R² = 0.4004). If those trends “correspond”, i.e. are the same, then perhaps I like van Vuuren am a Dutchman.
    That means the paper’s RCP trends grossly overestimate the likely level of [CO2] by 2100.

  2. I didn't look yet in any detail at the papers as things are rather busy at the moment, I just wanted to let people know about them. Back when I was doing time series climate modeling I thought based on the empirical time series that uptake of CO2 could be more in the future than most people seemed to think but I didn't have enough knowledge at all of the mechanics to know what is plausible and what is not. As they say they use a carbon cycle model I would be very surprised if they were really ignoring absorption of carbon dioxide, but obviously there are legitimate questions about what is a plausible model. What is "x" and what is "y" in those two regression equations. It's not surprising of course that concentrations grow much slower than anthropogenic emissions...