Spatially-differentiated atmospheric source–receptor relationships for nitrogen oxides, sulfur oxides and ammonia emissions at the global scale for life cycle impact assessment

Abstract

This paper aims to advance regional worldwide source–receptor relationships, providing fate factors for acidifying and eutrophying air emissions (NO_x, HNO₃, SO₂, SO₄ and NH₃) to be used within life cycle impact assessment. A simulation for the reference year 2005 of the three-dimensional global scale tropospheric GEOS-Chem model was used as the basis of a novel methodological approach to derive source–receptor matrices (SRMs) whose elements are fate factors at a global 2° × 2.5° grid. This new approach makes it possible to assess the impact of transboundary emissions while maintaining regional scale emission differentiation. These 2° × 2.5° grid resolution fate factors were later aggregated at continental and country resolutions using emission weighting. Continental fate factor results showed that 50–70% of nitrogen oxides (NO_x, HNO₃) and sulfur oxides (SO₂, SO₄) and approximately 80% of ammonia (NH₃) emissions will deposit on the same continent. Results showed that the developed fate factor derivation approach was within a ±10% agreement with GEOS-Chem simulations in which fate factors were determined by withdrawing the regional emission inventory over Canada and in ±50% agreement with current state-of-the-art LCIA fate factors (calculated with the European Monitoring and Evaluation Programme (EMEP) model). The SRMs outlined in this paper facilitate further modeling developments without having to run the underlying tropospheric model, thus opening the door to the assessment of the regional life cycle inventories of a global economy.

Introduction

Life cycle assessment (LCA) aims to compile and evaluate the inputs, outputs and potential environmental impacts of a product or service throughout its entire life cycle, from material extraction to end-of-life (Udo de Haes et al., 2002). Inputs and outputs are compiled in the life cycle inventory (LCI) phase and then classified into different impact categories. Among the considered life cycle impact assessment (LCIA) impact categories are acidification and eutrophication, which are related to the atmospheric deposition of inorganic substances such as sulfates and nitrates (Environment Canada, 2004; European Environment Agency, 2005). Evaluating potential impacts requires characterization factors (CF) [Impact*kg emitted⁻¹] that link life cycle inventory results to an impact score that is specific to a given impact category (Udo de Haes et al., 2002). Following the general framework of emission-related impact categories (Udo de Haes et al., 2002), the CF is calculated as the product of an atmospheric fate factor, an exposure factor (or, alternatively, an ecosystem sensitivity factor) and an effect factor, which are each calculated with one or several environmental models describing the biophysical relationships along the cause-effect chains from an emission to a given impact indicator. This paper specifically focuses on the development of atmospheric fate factors.

Atmospheric fate factors [kg_deposited*kg_emitted⁻¹] represent the climatic conditions (e.g. dispersion, atmospheric transport, etc.) and deposition mechanisms between the source and receptor locations (i.e. the source–receptor relationship) in a single fraction. Since there are many receptor elements and/or many sources, source–receptor relationships are written as a source–receptor matrix (SRM) whose elements are fate factors; each line is a source location and each column a receptor location. The main advantage of SRMs is simplicity. For a given source vector, the resulting receptor values can be obtained by a simple matrix-vector multiplication, avoiding the evaluation of the entire numerical model of the transport and dispersion processes (Seibert and Frank, 2003).

SRMs are used in LCIA to support CF development for acidifying and eutrophying chemicals (Bare et al., 2002; Fréchette-Marleau et al., 2008; Huijbregts et al., 2000; Potting et al., 1998; Seppälä et al., 2006; van Zelm et al., 2007). In such cases, SRM derivation relies on a number of regional atmospheric models referring to a specific geographic context—North America or Europe—typically covering the spatial range of the continental scale and with the typical spatial resolutions of countries or regions within countries (scale 50–500 km) (Potting and Hauschild, 2006). Restricted spatial ranges make it possible to include more sophisticated location-specific approaches and thus yield more reliable fate factors (Bare et al., 2002). On the other hand, emissions traveling outside the model's region of interest are not taken into account. Impact assessment is therefore incomplete, consequently diminishing the relevance of the LCA (Hayashi et al., 2004; van Zelm et al., 2007). Furthermore, the models do not facilitate a coherent assessment of life cycle emission inventories occurring outside their geographical context, as is often the case in a global economy (e.g. a North American emission is considered and modeled as a European emission).

Outside LCA, transboundary emissions were analyzed with global models: the spatial resolution was coarser and there was no differentiation between countries or regions. For example, Sanderson et al. (2008) provided continental SRMs for Europe, North America, South Asia and East Asia; Liu et al. (2008) studied source–receptor relationships between East Asian sulfur dioxide emissions and northern hemisphere sulfate concentrations and Kajino et al. (2010) showed that the intercontinental transport of sulfate aerosols could potentially influence regional air quality. Though it has never been explicitly stated, the reason for lowering the spatial resolution is applicability. SRM determination is costly in terms of computing, since simulations must be rerun for each source location or a separate species must be tracked for each source element (Seibert and Frank, 2003).

This paper aims to derive worldwide SRMs for sulfur oxides (SO₂, SO₄), nitrogen oxides (NO_x (i.e., NO, NO₂, NO₃, HNO₂), HNO₃) and ammonia (NH₃) airborne emissions to support LCIA model development at a global scale, while ensuring a relatively high resolution for regional differentiation. More specifically, the objective of this project is to derive SRMs on a 2° latitude × 2.5° longitude grid resolution by using the GEOS-Chem global atmospheric model as a starting point and aggregating SRMs at a coarser resolution (i.e. country and then continental levels).