The rising demand for food, feed, and fuel has put an increasing pressure on agriculture, with agricultural intensification as a direct response. Notwithstanding the higher crop productivity, intensive agriculture management entails many adverse environmental impacts. Intensive agricultural practices take a huge toll on land resources; soil quality and hence the ability to produce biomass in the long term are therefore at stake. In the transition towards sustainable agriculture, a thoughtful selection of farm management practices is required to reduce the environmental impact, thereby ensuring adequate soil quality, and hence long-term productivity. A tool that may assist in the evaluation of sustainable farm management is Life Cycle Assessment (LCA). However, because LCA was traditionally applied to industrial processes, some difficulties (e.g. variability among agricultural systems) and gaps (e.g. impact on long-term productivity) exist when conducting an LCA with respect to the agricultural sector. Therefore, the overall objective of this PhD thesis is to gain insight into some of these complexities and to develop methodologies to face some of the challenges. In this way, we will be able to more comprehensively evaluate the environmental sustainability of agricultural practices, which is an important asset in stimulating sustainable farming.

This PhD thesis starts with a general introduction (Chapter 1). First, a general view on agriculture in Europe is given. Then, the main agricultural policies are discussed. Policy can strongly influence the choices farmers make regarding land use, referred to as land use practices (LUPs) (e.g. crop rotation, fertilization). These LUPs have an impact on the agricultural soil quality, for which soil organic matter (SOM) is a key indicator. Therefore, in the third section, the main characteristics of SOM as well as its relationship with productivity are addressed. In the fourth section, the steps to conduct an LCA and the main gaps currently limiting a thorough evaluation of the environmental sustainability of agricultural systems are discussed. To conclude, in order to comprehensively evaluate the benefits of sustainable LUPs in LCA, one should be able to account for the impact of farm practices on soil quality and the ecosystem services (ES) delivered by an agro-ecosystem. An overview of currently existing tools and their limitations is given.

In Chapter 2, insight is gained into a first complexity related to agricultural LCAs: a large variability exists among agricultural systems due to management decisions or natural conditions. To do so, a case study on grain maize production in Flanders has been performed. Four types of drivers for variability are distinguished: policy, farm management, year-to-year weather variation and innovation. For each driver, scenarios are elaborated using an emission-based (ReCiPe) and a resource-accounting life cycle impact assessment method (Cumulative Exergy Extraction from the Natural Environment, CEENE) to assess the environmental performance. Regarding the scenarios elaborated for the driver policy, which limits fertilization levels in a soil-specific way, the resource consumption is lower for non-sandy soils than for sandy soils. Farm management seems to have less influence on the environmental impact when considering the CEENE only. But choices such as fertilizer type have a large effect on emission-related problems (e.g. eutrophication). In contrast, year-to-year weather variation results in large differences in the environmental footprint. The best environmental performance is obtained by innovation as e.g. plant breeding results in a steadily increasing yield over 25 years. Finally, a comparison between the environmental footprint of grain maize production in Flanders and a generically applied dataset, based on Swiss practices, endorse the importance to compile a local data inventory.

Chapters 3-5 handle about one of the main gaps currently noticed in agricultural LCAs: the poorly integration of the impact of farm management on soil quality, and hence long-term productivity. To simulate the SOC evolution (soil organic carbon, SOM is often measured as SOC) due to farm management and the corresponding response of the biomass productivity on the change in SOC stock, the models RothC and EU-Rotate_N are used, respectively.

First, in Chapter 3, we introduce an indicator called Agricultural Biomass Productivity Benefit of SOC management (ABB_SOC), which, relying on natural resource consumption (e.g. minerals, fossil fuels), enables to estimate the net effect of the efforts made to attain a better soil quality. Hereby the focus is put on SOC. First, a framework to describe the evolution of SOC stock due to farm management decisions is introduced. Then, if the SOC stock is lower than a defined threshold, the extent to which remediation measures are required, is used as a measure for the induced SOC losses. ABB_SOC values are calculated as the balance between the natural resource consumption of the inputs (including remediation efforts) and the desired output of arable crop production systems (i.e. harvested biomass). The developed indicator is applied on several rotation systems in Flanders, comparing different remediation strategies. The results reveal that the generally high benefits offset the effort to remediate SOC levels, but that results depend on the applied remediation actions.

In Chapter 4, we investigate how to take into account the impact of LUPs on soil quality in LCA. Currently, no methods exist that quantify the relationship between the change in SOC and the long-term biomass productivity. We introduce three interdependent indicators, which are situated along the cause-effect chain accounting for the effect of LUPs on soil quality in relation to the area of protection natural resources. At the middle of the cause-effect chain (midpoint), the indicators cumulative SOC deficit (cSOCD) and cumulative biomass productivity loss (cBPL) are proposed to indicate cumulative losses in SOC and yield, respectively, due to choices of LUPs. At the end of the cause-effect chain (endpoint), the indicator additional land requirement (ALR) is introduced to indicate the need of extra agricultural area necessary for provisioning of the lost biomass yield. Characterization factors are developed for a case study in Flanders. Different soil textures, initial SOC stocks, crop rotation systems and LUPs are considered. Because the reference state is based on sustainable LUPs, the developed framework may serve as a decision support tool towards a more sustainable agriculture.

In Chapter 5, the influence policy can have on long-term agricultural resource productivity by stimulating/discouraging farmers to apply certain LUPs, is investigated. We introduce six policy strategies, each characterized by its own mix of LUPs, for the Flemish agricultural sector. Three strategies allow us to evaluate the impact of the European Union Common Agricultural Policy (CAP) in the past, while the others focus on the future: one strategy reflects the potential of the CAP and two other strategies consider the application of compost. The indicators cSOCD and cBPL are used to evaluate the impact of policy on the long-term productivity. To avoid burden shifting, also the resource footprint is calculated. Several farm management systems (FMS) are distinguished, each characterized by a specific combination of farm type, agricultural region, rotation system and manure type. The results highlight the impact of policies on long-term productivity gains. Furthermore, applying extra compost seems to be promising: it results in an increasing resource productivity and reduced resource footprint. As the results differ per FMS, a differentiated approach is advisable when specific LUPs are stimulated in the context of sustainable farming.

A last challenge faced in this PhD thesis is the fact that, traditionally, LCAs only focus on the harvested products, while the supply of other ES is not captured. This contributes to the issue that the lower yields often reported for organic systems might outbalance the positive effect of using more environmental-friendly practices and result in lower environmental burden for conventional than organic products when LCA results are evaluated per product unit. In Chapter 6, we propose an approach to allocate the environmental impact among all delivered ES. This allocation procedure is based on the capacity of an agro-ecosystem to deliver ES, which will be different for conventional and organic systems. In this way, a more unbiased comparison of the environmental sustainability of organically and conventionally produced food is allowed. Using this approach, we demonstrate that for about half of the studied food products (including potatoes, maize), organic farming has clear environmental benefits in comparison to conventional cultivation methods.

Finally, Chapter 7 presents the main conclusions as well as opportunities for future research. Perspectives regarding the development of a more comprehensive soil quality indicator, the enhancement of the evaluation of impact of LUPs on the long-term productivity related to the AoP natural resources, and the increase of applicability regarding the integration of ES in LCA are discussed. To conclude, the work is situated into a broader perspective and deals with agriculture as part of the global ecosystem and the cohesion between policy and agriculture.