Back to the basic: toward improvement of technoeconomic representation in integrated assessment models

Abstract

With the shift of climate debate from understanding to actions, the use of integrated assessment models (IAMs) is gradually expanding. Since IAMs produce least-cost pathways, technoeconomic parameters constitute one of the basic parameters. Traditionally, IAMs dealt with technologies with slowly-changing, relatively homogeneous manner. Since technologies are rapidly evolving, and the pattern of technological development is regionally heterogeneous, the IAM community must embrace a new strategy to treat their underlying technoeconomic parameters. Here we illustrate such challenges by reviewing the treatment and performance of IAMs with respect to some of the rapidly changing technologies (e.g., solar, wind, and batteries). Our review shows that IAMs have difficulty in updating the cost of the rapidly changing technologies. We then articulate a new strategy, drawing upon the lesson from the current model intercomparison projects and climate sciences. We argue that a loose network of modeling groups across the globe should create a database of technological parameters in a standardized format and standard evaluation tool, perhaps to be facilitated by the IAM Consortium. Such a framework would contribute to the review of the progress toward the Paris Agreement goals.

Appendix: Materials and methods

1.1 Comparison of cumulative capacity and capital costs: Estimated vs. actual

Actual data of power generators were compiled from several sources. Cumulative capacities of PV were taken from the IEA (Boff et al. 2017; Lv Fang et al. 2017; Yamada and Ikki 2017) and Solar Power Europe (SolarPower Europe 2017). Capital costs of PV were taken from the Projected Cost of Generating Electricity (International Energy Agency 2005, 2010, 2015), IEA (Boff et al. 2017; Lv Fang et al. 2017; Yamada and Ikki 2017), IRENA (International Renewable Energy Agency (IRENA) 2012, 2015), and METI (Ministry of Economy Trade and Industry 2013, 2016). In addition, this study referred to the comprehensive database of the Open Energy Information project (National Renewable Energy Laboratory 2015) for the capital costs of power generators in the USA. They classified their cost data by database type into “Engineering,” “Field data,” “Market Analysis,” “Model,” and “Program.” We retrieved data classified as “Engineering,” “Field data,” and “Market Analysis,” as the dataset of historical cost data in the USA. As for the capital cost of PV in the USA, we obtained data from the Open PV projects, which contain more than 760,000 records of historical installation data for PV. Future capital cost targets for PV were gathered from the DOE SunShot Initiative (Department of Energy 2011, 2016) and NEDO PV challenges (New Energy and Industrial Technology Development Organization 2014).

To review the technoeconomic parameter used in IAMs, we checked all databases included in the IPCC AR5 database (The International Institute for Applied Systems Analysis (IIASA) 2014), as well as the SSP scenario database (The International Institute for Applied Systems Analysis (IIASA) 2016), Representative Concentration Pathways Database (The International Institute for Applied Systems Analysis (IIASA) 2009), and Global Energy Assessment (GEA) database (The International Institute for Applied Systems Analysis (IIASA) 2012), and IPCC SR1.5 Scenario Database (IIASA and IAMC 2018) (see Table 1). Since we found only one study, AMPERE, provide capital cost data for country scale, we used the AMPERE database for country-specific capital costs. The cumulative capacity estimated by IAMs was also obtained from AMPERE database.

Historical data of batteries for electric vehicles were taken from several sources, namely, data classified as “Publications,” “News items with expert statements,” and “Additional reported costs for car industry, other” in Nykvist and Nilsson (Nykvist and Nilsson 2015), in addition to the 2016 battery costs reported in IEA (International Energy Agency 2017b) and Schmidt et al. (2017). As the IAM database does not include capital cost parameters for batteries, “estimated” data were collected from data classified as “Future costs estimated in publications” by Nykvist and Nilsson (2015), in addition to data collected by Schmidt et al. (2017) and the Open Energy Information database provided by NREL (National Renewable Energy Laboratory 2015). Future “target” prices for battery costs were taken from Tesla (hybridCARS 2015) and GM (General Motors 2015).

Currency was converted into USD 2005 values by the following two steps (Nykvist and Nilsson 2015). First, we deflated historic prices in local currency using data from the US Bureau of Labor Statistics (US Department of Labor n.d.), and second, we converted prices to USD 2005 value using historical exchange rates from the US Federal Reserve (US Federal Reserve n.d.).

1.1.1 Data availability

The data of figures in this study, except for such aggregated databases as AMPERE and open PV, are given in the Supplementary Information (Supplementary data).