Response to the comments in the Gallagher Review on the work of P. Crutzen et al. on N2O emissions associated with biofuel crop production
A. General comments
The Gallagher Review gives considerable prominence to our work on the emissions of nitrous oxide, N2O, associated with the production of crops used for bioethanol and biodiesel fuel, and uses some very damning language to disparage our findings and conclusions.
The presentation in the Review appears to have been drawn wholly from the review of our work carried out by North Energy Associates, which is available on your website. This analysis by North Energy
- demonstrates a serious lack of understanding of the fundamental scientific arguments we made in our paper in Atmospheric Chemistry and Physics Discussions, 7, 11191-11205, 2007;
- takes material in the paper out of context to suggest we are making inappropriate comparisons with IPCC emission factors;
- completely ignores the final version of our paper, which appeared in Atmospheric Chemistry and Physics, 8, 389-395, 2008 (copy attached), in which we explained in detail how our "top-down" assessment of the magnitude of N2O emissions could be reconciled with the "bottom-up" methodology of IPCC, and in which we also provided calculations of the change in the global warming effect due to N2O if the efficiency of nitrogen fertiliser use should change from 40% to 60%.
We are very surprised that the North Energy review, given these shortcomings, could have been accepted and used so unquestioningly in such a high-profile government document as the Gallagher Review. Our detailed comments below focus on Section 2.2, and particularly on Box 2.1, of the Gallagher Review, and explain where, why and how the statements therein are wrong or misleading. We also provide comments on a proposal in Table 9.1, which is based on these wrong statements in Section 2.2.
We also wish to put on record that the North Energy review acknowledges contributions to its preparation by two ADAS scientists "for providing essential information on the effects of land use change and other considerations on changes in greenhouse gas emissions....." It is unclear whether these "other considerations" included the assessment of our work on N2O, but if that indeed were so we would be particularly concerned, because the position of these ADAS staff is inevitably compromised by their involvement with the UK Home-Grown Cereals Authority's activities promoting the use of cereal crops for bioethanol production.
B. Specific comments on the numbered "Key issues with the methodology" in Box 2.1
Our numbering here follows that in the Box:
1. What the Review calls a "remainder calculation" refers to what we designated as "background" N2O production, and designated as "indirect sources" in IPCC 2006 (and its predecessor document). What is incontrovertible is the rate of increase of N2O in the atmosphere - which is as well documented as that of CO2. As our papers explain, taking into account this increase and the estimated lifetime of N2O in the atmosphere, a value can be calculated for the current annual global emission of N2O. The corresponding value for the pre-industrial era can also be calculated. For this pre-industrial era, a global N2O emission factor was calculated by dividing the annual emission by the estimated annual input of reactive nitrogen; a similar calculation involving the division of the additional N2O emission by the additional anthropogenic input of reactive nitrogen (primarily fertiliser N) gave a very similar emission factor to that for the pre-industrial era. This emission factor is a parametric relationship between the global N2O budget and the budget for N inputs to agriculture, and as we point out in our paper, it is not dependent on detailed knowledge of the terrestrial N cycle.
The match between the natural (pre-industrial) global emission factor and the anthropogenic one is striking, and points strongly to a much larger total N2O emission resulting from the release of reactive nitrogen compounds into the terrestrial environment than that observed as direct emissions from fertilised fields. In Appendix A of our peer-reviewed 2008 paper we gave examples of "secondary" emissions resulting from the complex transformations of N compounds in the various flows within agricultural systems, including dung and urine from livestock fed on N-fertilised grain crops and/or feeds containing N from biological N fixation (regarded as part of the anthropogenic reactive N input). Our calculations thus include all agriculture, including animal husbandry, and contrary to the wording of Point 1 they are therefore not "highly sensitive to the assumed values of individual ‘non-cultivated soil' fluxes", as the remaining global anthropogenic sources of N2O are the relatively small emissions from biomass burning and industrial sources. For these we used the data of Prather et al, published as part of the IPCC 2001 Assessment, which has not been seriously challenged: biomass burning 0.5, and industrial sources 0.7-1.3 Tg N2O-N/yr, compared with the agricultural total of 4.3-5.8 Tg N2O-N/yr.
2. Point 2 says the comparison with the IPCC estimate is "inappropriate since this compares total soil emissions with direct emissions in the IPCC estimate". We assume that "total soil emissions" here presumably means "total emissions, both direct and indirect, whether from soil or from leached or volatilised N". After drawing attention to the disparity between the direct emission from fields and the total emission, we addressed this issue in the very next sentence (2007 paper, p 11194, line 25 onwards): "The large difference between the low yield of N2O in agricultural fields, compared to the much larger average value derived from the global N2O budget, implies considerable "background" N2O production occurring beyond agricultural fields but, nevertheless, related to fertilizer use, from sources such as rivers, estuaries and coastal zones, animal husbandry and the atmospheric deposition of ammonia and NOx". To further clarify this key issue, we added Appendix A to the paper in its final form (as mentioned above), to show the relationship between our method of calculation and that of IPCC 2006. A key feature of this Appendix is the focus on the uncertainty ranges associated with the IPCC default values (e.g. 0.3-3% for the 1% direct emission factor), which often seem to be ignored in this context, and we conclude with the words "Each of the source terms in the bottom-up, IPCC method is very uncertain. However, their sum is not inconsistent with the total derived by the top-down methodology". We submit that this is a fair and objective comparison of the alternative approaches.
3. We reject the criticism that we only considered a limited and unrepresentative list of biofuels. The three examples for which we provided worked calculations (in both the ACPD 2007 version and the ACP 2008 paper) were, respectively: oilseed rape, as the main source of biodiesel in Europe; maize, as the main source of bioethanol in the US; and sugar cane, as the world's biggest source of bioethanol, in Brazil, as a tropical rather than a temperate crop, and the basis of the oldest-established biofuel industry. However, we also performed our calculations for other crops and residues, and these are listed in Table 1 (p. 11205): in addition to rapeseed, maize and sugar cane, there we give figures for wheat, barley and oats, sugar beet leaves, root crops and high- and low-N forages. In the body of the paper we briefly considered oil from oil palm, as well as the "energy crops" switch grass, Miscanthus, eucalyptus, poplar and willow.
Ironically, we simplified Table 1 in the final peer-reviewed 2008 paper (the one apparently not seen by your reviewers), by restricting it to our three main examples. We concluded that as the main purpose of the paper was to establish the magnitude of the global N2O yield from freshly added reactive nitrogen, this was adequate - the calculation for any other crop being readily done on the basis of known values for N content, fertiliser application rates and so on.
4. The estimate of 40% as the global average N use efficiency is soundly based on much exhaustive analysis of agronomic data, and fully referenced in the 2007 paper (Cassman et al, 2002; Galloway et al., 2003; Balasubramanian et al., 2004). However, because some critics argued that the efficiency factor is higher in some environments we also included in Table 2 of the 2008 version a parallel set of calculations for a hypothetical N use efficiency of 60%, as well as the effect of partial replacement of mineral N fertiliser by animal manure, and of the efficient use of byproducts. However, to return to the well-established 40% global average: there is much (and possibly growing) intercontinental trade in agricultural commodities for use in bioethanol and biodiesel production, e.g. US maize and SE Asian palm oil to Europe, determined by price and availability and not by N use efficiency. Thus the production of biofuel is not necessarily based on local sourcing of crops, and use of the global average value therefore seems prudent at the present time.
5. Concerning the issue of coproducts, we stated very clearly in both versions of the paper that it was not our purpose to carry out a full LCA, but rather to highlight the N2O emissions as one important component of such an analysis, which needs to be considered along with, on one side, the fossil fuel input to the global warming balance, and on the other side the creation of useful co-products. However, as mentioned under Point 4, we did include a calculation of the impact of the use of byproducts in our 2008 paper.
C. Comment on final paragraph of Box 2.1
The language used here in the first sentence is offensive and derogatory, and we deplore the fact that it has been included in such a high-profile official document. We believe we have provided sufficient evidence in this response to show how unjustified this description of our work is, and we think a public retraction is necessary.
D. Comment on Table 9.1
The proposals for more research on the effects of different farming practices and cultivation techniques on N2O emissions from soils, contained in the 1st and 3rd proposals on p 74 of the Review will do nothing to resolve the issues raised in our paper, and appear to arise from the lack of understanding of our work manifest in the North Energy review. The large number of such measurements already carried out were subjected to an exhaustive review by Stehfest and Bouwman (2006), and their work contributed to the revised "default emission factor" of 1% used in IPCC 2006. Nonetheless, the recognition that measured values for direct emission factors follow a skewed (log-normal) distribution led to IPCC adopting a "logarithmic" uncertainty range for this default value: from one-third of the value to 3 times the value, i.e. 0.3 to 3 %. This also followed a similar discussion in the IPCC "Good Practice" guidelines in 2000. (The default emission factors for indirect emissions were also given analogous uncertainty ranges in IPCC 2006.)
A consequence is that more conventional field trials in the UK and/or the rest of the EU are merely going to add further unique values (dictated by such factors as weather conditions at the time, as much as by cultivation techniques or other variables) to the population of values reviewed by Stehfest and Bouwman. They will not provide any evidence to support, or to contradict, our assessment of the overall yield of N2O from terrestrial biospheric processes following the introduction of new "fixed N" in the form of fertiliser or by biological N fixation.
Paul Crutzen
Arvin Mosier
Keith Smith
16 July 2008
Last Modified: 09 Feb 2009