4. Discussion

#### 4.1. Novelty and limitations of the proposed method

Our method assumes that agricultural LULUC is a consequence of increasing demand for agricultural products and thus for land. We derive robust and globally consistent emission shares and emission factors based on the dynamic development of agricultural production, expressed in increases of produced (and net-imported) energy equivalents rather than on static, absolute shares of production (e.g., exported energy quantities as such). This focus on dynamic developments has the advantage of capturing the trends triggering LULUC impacts, but it also requires up-to-date information on rapidly changing global agricultural developments, making it difficult to extend the method to geographical entities smaller than countries (i.e., the level at which statistics are usually available; see [18]).

On the one hand, the method illustrated here is predicated on the principle of assigning an environmental burden (LULUC emissions) to an increase in commodity consumption, i.e., to the importing country, whose increased demand for the commodity is seen as causing the burden. On the other hand, one could also argue that it is the producer, not the consumer, who decides to satisfy a perceived demand, and therefore, the LULUC emissions should be assigned to the country of origin. Applied to LULUC, this shifted perspective would mean that exportrelated LULUC emissions are still assigned to the producing country. Hence, no "iLULUC emissions pool" would be necessary. A compromise approach would be to evenly divide the LULUC emissions from imports and exports between producer and consumer. Mathematically, this would correspond simply to cutting the size of the iLULUC emissions pool in half.

In most countries, the larger part of increased food and feedstuff production is for domestic purposes. Thus, most of a given country's LULUC emissions (globally approximately 70%) are assigned to the domestic territory. The remaining roughly 30% are exported or imported and are thus assigned to a global iLULUC emissions pool. In many countries though, LULUC from import increases accounts for more than half of the net LULUC (hatched areas in Figure 1).

For some countries, CO2 emissions from LULUC could be overestimated because not all LULUC is linked to infrastructure, settlements, and agriculture, but also to, e.g., mining. The relatively undetailed allocation on the basis of the increase or decrease in areas for infrastructure, settlements, and agriculture introduces uncertainty. So far, the model also ignores the role of intensification as a cause of net export increases without causing LUC. Further studies could add such elements to the model, which is crucial for a correct assessment and allocation of agricultural LULUC emissions.

As stated above, emission shares are allocated in proportion to the energy content of agricultural product groups (based on their LHVs). As has long been debated (e.g., in LCA [24, 25]), allocation could also be based on commodity prices, but for the purposes of this study, the required data were not available. Such an economic aggregation would emphasize the role of monetary drivers for cultivation and agricultural management decisions, but on the other hand, it would be subject to confounding factors such as currency exchange rate fluctuations and fluctuations of auxiliary material prices (fuels, fertilizers, and pesticides).

Uncertainties may be introduced by input data from [18] concerning areas, yields, national consumption, or traded amounts. These data are reported by the national statistical authorities. In addition, the aggregation of single commodities into product groups such as "cereals" causes uncertainties, as different commodities within a group (e.g., types of cereal grains) will have different LHVs, which even further vary under practical conditions. For example, for the average LHV of the product group "cereals," we used the LHV of the globally dominant cereal commodity wheat as a default value. A comparison of the wheat LHV with the actual weighed average of the US cereal grain production mix shows a difference of 1.9% between the default value and the actual mix (US Department of Agriculture's statistical data sets for the years 1998–2000 and 2007–2009; http://quickstats.nass.usda.gov/). Additional uncertainty originates from the conversion of volume-based production information (bushels) to mass-based production data, as well as from the variability of published LHV values for grains.

4. Discussion

product (kg LULUC-CO2 kg�<sup>1</sup> product).

84 Land Use - Assessing the Past, Envisioning the Future

4.1. Novelty and limitations of the proposed method

(i.e., the level at which statistics are usually available; see [18]).

Our method assumes that agricultural LULUC is a consequence of increasing demand for agricultural products and thus for land. We derive robust and globally consistent emission shares and emission factors based on the dynamic development of agricultural production, expressed in increases of produced (and net-imported) energy equivalents rather than on static, absolute shares of production (e.g., exported energy quantities as such). This focus on dynamic developments has the advantage of capturing the trends triggering LULUC impacts, but it also requires up-to-date information on rapidly changing global agricultural developments, making it difficult to extend the method to geographical entities smaller than countries

Figure 4. Average net LULUC emissions of specific livestock product groups with relatively high emissions per kg of

On the one hand, the method illustrated here is predicated on the principle of assigning an environmental burden (LULUC emissions) to an increase in commodity consumption, i.e., to the importing country, whose increased demand for the commodity is seen as causing the burden. On the other hand, one could also argue that it is the producer, not the consumer, who decides to satisfy a perceived demand, and therefore, the LULUC emissions should be assigned to the country of origin. Applied to LULUC, this shifted perspective would mean that exportrelated LULUC emissions are still assigned to the producing country. Hence, no "iLULUC emissions pool" would be necessary. A compromise approach would be to evenly divide the LULUC emissions from imports and exports between producer and consumer. Mathematically,

this would correspond simply to cutting the size of the iLULUC emissions pool in half.

From a global perspective, livestock products seem not to lead to particularly high LULUC emissions. However, the resulting numbers for nlk,p (see Eq. (14) and Table 2) are to some extent misleading, as they are based on production and net import increases. Those increases were rather low for livestock products over the observed period (e.g., in Brazil in Table 1), but arable land is increasingly used for livestock feed production, i.e., cereals or by-products from oil crops (oil cakes or solvent-extracted meal). The real LULUC emissions from livestock products are therefore likely to be higher than the numbers obtained with this method. Consequently, a part of the emissions linked to, e.g., oil crops have actually to be allocated to livestock products.

A limitation of our approach is that it does not consider historically grown and established bilateral trade connections between countries. For example, when the US corn is explicitly produced for the Chinese market, then US LULUC emissions end up in the global pool and obliterate the fact that China alone would be responsible for the LULUC change emissions. However, the focus of the study was the construction of a global iLULUC emissions pool in order to account for the changing global interrelationships of the agricultural commodity marketplace.

national iLULUC emissions may be derived from it and represent the LULUC emissions

Consequences from Land Use and Indirect/Direct Land Use Change for CO2 Emissions Related…

http://dx.doi.org/10.5772/intechopen.80346

87

Our results account for the allocation of emissions to specific product groups consumed in a country in proportion to their corresponding energy content on an LHV basis. This allows for the aggregation of agricultural product group data on different spatial levels, and it provides a more detailed focus compared to generic agricultural land-related emission estimates. With this approach, 3150 new results from 175 countries are provided with the respective indirect (LU)LUC effects. The results vary substantially between nations, with clear differences between producing and exporting countries versus importing countries. A similar differentia-

LUC-related GHG-accounting should rest on a well-documented computational basis as a prerequisite for a fair differentiation of "LULUC-emitting/exporting nations" versus "LULUC-importing nations" on the one hand and between (LU)LUC-driving product groups versus product groups with little or no effects on LULUC emissions on the other. Further work should address the

The authors are grateful to Dr. Christian Schader, Dr. Rainer Weisshaidinger, Theresia Markut, and Dr. Matthias Meier from the Research Institute of Organic Agriculture (FiBL) for their valuable input and for many stimulating discussions. We gratefully acknowledge support

Stefan J. Hörtenhuber1,2\*, Michaela C. Theurl1,3, Gerhard Piringer2 and Werner J. Zollitsch<sup>2</sup>

1 Research Institute of Organic Agriculture (Forschungsinstitut für biologischen Landbau

2 Department of Sustainable Agricultural Systems, BOKU—University of Natural Resources

3 Institute of Social Ecology (SEC), BOKU—University of Natural Resources and Life Sciences,

inherent in the traded products.

Acknowledgements

Conflict of interest

Author details

FiBL), Vienna, Austria

Vienna, Vienna, Austria

The authors declare no conflict of interest.

and Life Sciences, Vienna, Vienna, Austria

tion applies to specific product groups within a country.

validation and improvement of the model and its input data.

from the Austrian Science Fund (FWF) project GELUC (P29130-G27).

\*Address all correspondence to: stefan.hoertenhuber@fibl.org

#### 4.2. Direct (LU)LUC emissions versus results of the proposed method

Some studies (e.g., [1, 26]) computed direct LU emissions and dLUC emissions for specific oil crops from specific countries, e.g., Brazil and Argentina, and for the import mix of such crops used, e.g., in Austria [26]. For the latter, our results are comparable to those for Germany, as most oil crops imported into Austria are transported through Germany and they are influenced in both countries by the European markets.

For the example of oil crops, i.e., the basis for vegetable oils and by-products (mainly feed), which are consumed in Austria, the method proposed here assigns 1.99 kg CO2 to 1 kg of product. Most of the oil crops or their products are imported into Austria and, in addition, no dLUC emissions are relevant for domestic oil crops. Thus, LULUC emissions are sourced exclusively from contributions to the iLULUC pool. Based on market information (e.g., Refs. [27, 28]), 50% each of the oil crops are estimated to come as soybeans from North America (no dLUC emissions) and South America. The resulting level of 1.61 kg of dLULUC emissions is in line with the 1.99 kg CO2 stated above. The emissions are linked to imports from Brazil, which show 3.097 kg dLUC-CO2 per kg of soybeans and LU-related emissions of 0.019 kg LU-CO2 per kg of soybeans [1]. Together, dLULUC accounts for 3.22 kg CO2 per kg of Brazilian soybeans, which is comparable to the 3.66 kg CO2 derived with the method presented herein. It has to be noted that d(LU)LUC emission factors cannot be directly compared to the iLULUC emission factors presented here. While dLULUC estimates are close to the numbers from the presented method in specific cases such as of Austria, dLUC emission factors alone are insufficient and should be replaced or accompanied by emission factors which consider iLULUC effects in LCAs and carbon footprints.
