Search

In recent years many Life Cycle Assessment (LCA) studies have been conducted to quantify the environmental performance of products and services. Some of these studies propagated numerical uncertainties in underlying data to LCA results, and several applied Global Sensitivity Analysis (GSA) to some parts of the LCA model to determine its main uncertainty drivers. However, only a few studies have tackled the GSA of complete LCA models due to the high computational cost of such analysis and the lack of appropriate methods for very high-dimensional models. This study proposes a new GSA protocol suitable for large LCA problems that, unlike existing approaches, does not make assumptions on model linearity and complexity and includes extensive validation of GSA results. We illustrate the benefits of our protocol by comparing it with an existing method in terms of filtering of noninfluential and ranking of influential uncertainty drivers and include an application example of Swiss household food consumption. We note that our protocol obtains more accurate GSA results, which leads to better understanding of LCA models, and less data collection efforts to achieve more robust estimation of environmental impacts. Implementations supporting this work are available as free and open source Python packages.

The limitations of the static nature of Life Cycle Assessment (LCA) are well known. To overcome the loss of temporal information due to the aggregation of flows in the Life Cycle Inventory (LCI), several dynamic LCA methodologies have been proposed. In this paper we present a new generic and operational methodology for dynamic LCA that allows for the introduction of temporal information in both in the inventory and the Life Cycle Impact Assessment (LCIA) phases. The method makes use of graph traversal and convolution to calculate the temporally differentiated inventory, and makes it possible to use several types of dynamic impact assessment. We describe our method and apply it to a cradle-to-grave dynamic LCA of a glued laminated timber (glulam) product. We also test the sensitivity of the global warming results to temporal explicit LCI data. There is a considerable difference in outcome between the static and dynamic approaches. We have implemented our framework in the free and open source software Temporalis that is fully operational and can be used with existing LCA databases., (Copyright © 2018. Published by Elsevier B.V.)

The future environmental impacts of battery electric vehicles (EVs) are very important given their expected dominance in future transport systems. Previous studies have shown these impacts to be highly uncertain, though a detailed treatment of this uncertainty is still lacking. We help to fill this gap by using Monte Carlo and global sensitivity analysis to quantify parametric uncertainty and also consider two additional factors that have not yet been addressed in the field. First, we include changes to driving patterns due to the introduction of autonomous and connected vehicles. Second, we deeply integrate scenario results from the IMAGE integrated assessment model into our life cycle database to include the impacts of changes to the electricity sector on the environmental burdens of producing and recharging future EVs. Future EVs are expected to have 45-78% lower climate change impacts than current EVs. Electricity used for charging is the largest source of variability in results, though vehicle size, lifetime, driving patterns, and battery size also strongly contribute to variability. We also show that it is imperative to consider changes to the electricity sector when calculating upstream impacts of EVs, as without this, results could be overestimated by up to 75%.

The goal of this study was to identify drivers of environmental impact and quantify their influence on the environmental performance of wooden and massive residential and office buildings. We performed a life cycle assessment and used thermal simulation to quantify operational energy demand and to account for differences in thermal inertia of building mass. Twenty-eight input parameters, affecting operation, design, material, and exogenic building properties were sampled in a Monte Carlo analysis. To determine sensitivity, we calculated the correlation between each parameter and the resulting life cycle inventory and impact assessment scores. Parameters affecting operational energy demand and energy conversion are the most influential for the building's total environmental performance. For climate change, electricity mix, ventilation rate, heating system, and construction material rank the highest. Thermal inertia results in an average 2-6% difference in heat demand. Nonrenewable cumulative energy demand of wooden buildings is 18% lower, compared to a massive variant. Total cumulative energy demand is comparable. The median climate change impact is 25% lower, including end-of-life material credits and 22% lower, when credits are excluded. The findings are valid for small offices and residential buildings in Switzerland and regions with similar building culture, construction material production, and climate.

Land use is one of the main drivers of biodiversity loss. However, many life cycle assessment studies do not yet assess this effect because of the lack of reliable and operational methods. Here, we present an approach to modeling the impacts of regional land use on plants, mammals, birds, amphibians, and reptiles. Our global analysis calculates the total potential damage caused by all land uses within each WWF ecoregion and allocates this total damage to different types of land use per ecoregion. We use an adapted (matrix-calibrated) species-area relationship to model the potential regional extinction of nonendemic species caused by reversible land use and land use change impacts. The potential global extinction of endemic species is used to assess irreversible, permanent impacts. Model uncertainty is assessed using Monte Carlo simulations. The impacts of land use on biodiversity varied strongly across ecoregions, showing the highest values in regions where most natural habitat had been converted in the past. The approach is thus retrospective and was able to highlight the impacts in highly disturbed regions. However, we also illustrate how it can be applied to prospective assessments using scenarios of future land use. Uncertainties, modeling choices, and validity are discussed.

Comprehensive sensitivity analysis is a significant tool to interpret and improve life cycle assessment (LCA) models, but is rarely performed. Sensitivity analysis will increase in importance as inventory databases become regionalized, increasing the number of system parameters, and parametrized, adding complexity through variables and nonlinear formulas. We propose and implement a new two-step approach to sensitivity analysis. First, we identify parameters with high global sensitivities for further examination and analysis with a screening step, the method of elementary effects. Second, the more computationally intensive contribution to variance test is used to quantify the relative importance of these parameters. The two-step sensitivity test is illustrated on a regionalized, nonlinear case study of the biodiversity impacts from land use of cocoa production, including a worldwide cocoa products trade model. Our simplified trade model can be used for transformable commodities where one is assessing market shares that vary over time. In the case study, the highly uncertain characterization factors for the Ivory Coast and Ghana contributed more than 50% of variance for almost all countries and years examined. The two-step sensitivity test allows for the interpretation, understanding, and improvement of large, complex, and nonlinear LCA systems.

We describe a new methodology for performing regionalized life cycle assessment and systematically choosing the spatial scale of regionalized impact assessment methods. We extend standard matrix-based calculations to include matrices that describe the mapping from inventory to impact assessment spatial supports. Uncertainty in inventory spatial data is modeled using a discrete spatial distribution function, which in a case study is derived from empirical data. The minimization of global spatial autocorrelation is used to choose the optimal spatial scale of impact assessment methods. We demonstrate these techniques on electricity production in the United States, using regionalized impact assessment methods for air emissions and freshwater consumption. Case study results show important differences between site-generic and regionalized calculations, and provide specific guidance for future improvements of inventory data sets and impact assessment methods.

Food consumption represents the dominant exposure pathway of the general public to pesticides. In this paper, we characterize the lifelong cumulative human health damage from ingestion of pesticides contained in fruits and vegetables in Switzerland and the United States. We evaluated pesticide residues in 62,151 food samples. Chemical specific concentrations were combined with pesticide emission data and information on country-specific diets and chemical toxicity to assess the human health impacts of 51 food commodities and national average diets. Furthermore, a list of characterization factors for pesticide ingestion via food was calculated for use in life cycle impact assessment. On average, the Swiss population takes in via food ingestion 0.41g of every 1kg of pesticide applied during agricultural cultivation. The corresponding value in the United States is 0.51. Intake fractions based on experimental monitoring data were compared with outputs from the USEtox model for life cycle impact assessment of toxic substances. The modeled intake fractions were underestimated by up to two orders of magnitude. However, even when using the monitored residue concentration data, the absolute health damage via fruits and vegetable ingestion was small: The potential lifelong damage of pesticides is estimated to be only 4.2 and 3.2 min of life lost per person in Switzerland and the United States, respectively. The results of this study indicate that pesticide intake due to the ingestion of fruits and vegetables consumed in Switzerland and the United States does not lead to significant human health damages.

Life cycle assessment (LCA) studies have shown that site-dependent impact assessment for categories like acidification and eutrophication give more accurate and realistic results than site-generic assessments. To date, existing geography-specific, or regionalized, impact assessment factors have not been applied to LCA databases and software tools. We describe a simple, generic methodology to couple existing regionalized characterization factors with large life cycle inventory databases. This approach allows for detailed geographic life cycle impact assessment results. Case-study results for European country-specific electricity mixes are calculated using the Ecoinvent 2.01 database and the EDIP 2003 and Accumulated Exceedance impact assessment methods and CASES project external energy cost characterization factors. In most cases, regionalization shows different total scores, different processes of high importance, and varying geographic distributions of environmental impacts. As the methodology requires no additional input other than the geographic information already in existing LCA databases, it can be used routinely. Better and more consistent geographic information in life cycle inventory databases and impact assessment methods, tailored to the specific spatial range of all environmental effects considered, would be beneficial.