Abstract
Although emerging technologies like carbon capture and storage and advanced nuclear are expected to play leading roles in greenhouse gas mitigation efforts, many engineering and policy-related uncertainties will influence their deployment. Capital-intensive infrastructure decisions depend on understanding the likelihoods and impacts of uncertainties such as the timing and stringency of climate policy as well as the technological availability of carbon capture systems. This paper demonstrates the utility of stochastic programming approaches to uncertainty analysis within a practical policy setting, using uncertainties in the US electric sector as motivating examples. We describe the potential utility of this framework for energy-environmental decision making and use a modeling example to reinforce these points and to stress the need for new tools to better exploit the full range of benefits the stochastic programming approach can provide. Model results illustrate how this framework can give important insights about hedging strategies to reduce risks associated with high compliance costs for tight CO2 caps and low CCS availability. Metrics for evaluating uncertainties like the expected value of perfect information and the value of the stochastic solution quantify the importance of including uncertainties in capacity planning, of making precautionary low-carbon investments, and of conducting research and gathering information to reduce risk.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Both approaches also suffer from the curse of dimensionality. For dynamic programs, the curse results from the number of states and dimensionality of the state and control spaces. For stochastic programs, it is due to the number of stages and scenarios.
As pointed out by Weitzman (2009), uncertainties like the climate sensitivity parameter inherently have diffuse distributions. Future realizations of parameter values, particularly those outside of the range of experience, are not adequately covered in past observations, which makes it challenging to learn limiting tail behavior through induction using finite historical samples.
In order to retain some degree of probabilistic dependence while avoiding pitfalls of conditional elicitations, it is preferable to circumvent the issue by explicitly modeling the cause of the dependency. For instance, if future costs of nuclear power plants are correlated with CCS-equipped coal facilities, this dependency may be caused by construction cost inflation, which can be incorporated in the model as an extra parameter.
Although the dynamic programming setting is more appropriate when decision-dependent uncertainties play prominent roles, stochastic programs can be formulated to accommodate such structures.
Deterministic energy-economic and integrated assessment models are implicitly solving the expected value problem.
In many applications, information is neither complete nor perfectly accurate, so the expected value of imperfect information is less than the EVPI.
These metrics can be defined and used in alternative modeling frameworks for uncertainty analysis beyond the two-stage stochastic programming approach discussed here.
Online Resource 1 provides more detail about how stochastic MARKAL determines optimal strategies.
Results for the public acceptance of CO\(_\textrm{2}\) storage and natural gas price uncertainties are discussed in Online Resources 3 and 4, respectively.
Values are expressed in billions of dollars with a five percent discount rate.
The climate change uncertainty is considered less controllable than the others in the analysis, since the randomness originates primarily from incomplete and imperfect understandings of the climate system itself. New information can reduce uncertainty, but there are a limited number of interventions (e.g., low-risk geoengineering, costless ambient air capture) that can influence this outcome.
Since R&D investments do not guarantee specific outcomes, there will always be an aleatoric component to R&D-influenced uncertainties. Thus, the VOC assumption of perfect intervention, where the control action specifies a single outcome with certainty, is an upper bound.
This scenario would occur if backlash in the US in the wake of Fukushima became similar to Germany, where the government announced that 17 nuclear reactors would be taken offline within the next 11 years.
The outdated solar assumptions in the EPA database suggest that rapidly-developing technologies require more frequent elicitations to survey the current expert state of knowledge and expectations about future developments to avoid assumption drag.
References
Birge J, Louveaux F (2011) Introduction to stochastic programming, 2nd edn. Springer, New York
Clarke L, Edmonds J, Jacoby H, Pitcher H, Reilly J, Richels R (2007) Scenarios of greenhouse gas emissions and atmospheric concentrations. Tech. rep., US Climate Change Science Program and the Subcommittee on Global Change Research
Dantzig G (1955) Linear programming under uncertainty. Manage Sci 1(3–4):197–206
Dooley J, Dahowski R, Davidson C, Bachu S, Gupta N, Gale J (2004) A CO\(_\textrm{2}\)-storage supply curve for North America and its implications for the deployment of carbon dioxide capture and storage systems. In: Proceedings of the seventh international conference on greenhouse gas control technologies
Dupacova J, Hurt J (2002) Stochastic modeling in economics and finance. Springer
FEEM (2012) Technology elicitations and modeling project (TEaM). http://www.feem.it/getpage.aspx?id=4278&sez=Research...18&padre=70&sub=86
Fishbone L, Abilock H (1981) MARKAL, a linear-programming model for energy systems analysis: technical description of the BNL version. Int J Energy Res 5(4):353–375
Ha-Duong M, Grubb M, Hourcade J (1997) Influence of socioeconomic inertia and uncertainty on optimal CO\(_\textrm{2}\)-emission abatement. Nature 390:270–273
Hu M, Hobbs B (2010) Analysis of multi-pollutant policies for the U.S. power sector under technology and policy uncertainty using MARKAL. Energy 35(12):5430–5442
IEA (2007) World Energy Outlook 2007. Tech. rep., International Energy Agency, OECD/IEA
Infanger G (1994) Planning under uncertainty: solving large-scale stochastic linear programs, 1st edn. The Scientific Press Series
Infanger G (1999) GAMS/DECIS user’s guide. http://www.gams.com/dd/docs/solvers/decis.pdf. Accessed 8 Aug 2013
IPCC (2005) Special report on carbon capture and storage. Tech. rep., Intergovernmental Panel on Climate Change
IPCC (2007) Contribution of working groups I, II, and III to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Tech. rep., Intergovernmental Panel on Climate Change
Kalvelagen E (2003) Two-stage stochastic linear programming with GAMS. Tech. rep., Amsterdam Optimization Modeling Group LLC
Kann A, Weyant J (2000) Approaches to performing uncertainty analysis in large-scale energy/environmental policy models. Environ Model Assess 5(1):29–46
Kanudia A, Loulou R (1998) Robust responses to climate change via stochastic MARKAL: the case of Québec. Eur J Oper Res 106(1):15–30
Kanudia A, Loulou R (1999) Advanced bottom-up modelling for national and regional energy planning in response to climate change. Int J Environ Pollut 12(2–3):191–216
Kanudia A, Shukla P (1998) Modelling of uncertainties and price elastic demands in energy-environment planning for India. Omega 26(3):409–423
Manne A, Richels R (1992) Buying greenhouse insurance: the econmoic costs of CO\(_\textrm{2}\) emission limits. MIT Press
Morgan M, Henrion M (1990) Uncertainty: a guide to dealing with uncertainty in quantitative risk and policy analysis. Cambridge University Press, Cambridge
OTA (1984) Nuclear power in an age of uncertainty. Tech. Rep. OTA-E-216, US Congress, Office of Technology Assessment
Rai V, Victor D, Thurber M (2010) Carbon capture and storage at scale: lessons from the growth of analogous energy technologies. Energy Policy 38(8):4089–4098
Roe G, Baker M (2007) Why is climate sensitivity so unpredictable? Science 318(5850):629–632
Ryan S, McCalley J, Woodruff D (2011) Long term resource planning for electric power systems under uncertainty. Tech. rep., Lawrence Berkeley Laboratory for Computation of the Next Generation Grids
Shapiro A, Dentcheva D, Ruszczyński A (2009) Lectures on stochastic programming: modeling and theory. SIAM
Shay C (2008) EPA US nine region MARKAL database. Tech. rep., US Environmental Protection Agency, Air Pollution Prevention and Control Division, National Risk Management Research Laboratory
Stern N (2006) The economics of climate change: the Stern review. Cambridge University Press, Cambridge
Vattenfall/McKinsey (2007) Global mapping of greenhouse gas abatement opportunities up to 2030. Tech. rep., Vattenfall/McKinsey
Webster M, Santen N, Parpas P (2012) An approximate dynamic programming framework for modeling global climate policy under decision-dependent uncertainty. Comput Manage Sci 9(3):339–362
Weitzman M (2009) On modeling and interpreting the economics of catastrophic climate change. Rev Econ Stat 91(1):1–19
Weitzman M (2012) GHG targets as insurance against catastrophic climate damages. J Public Econ Theory 14(2):221–244
Acknowledgements
J.E. Bistline would like to acknowledge support by the William K. Bowes, Jr. Stanford Graduate Fellowship. J.P. Weyant’s participation in the research reported here was supported by the US DOE, Office of Science, Office of Biological and Environmental Research, Integrated Assessment Research Program, Grant No. DE-SC005171.
Author information
Authors and Affiliations
Corresponding author
Electronic Supplementary Material
Below is the link to the electronic supplementary material.
Rights and permissions
About this article
Cite this article
Bistline, J.E., Weyant, J.P. Electric sector investments under technological and policy-related uncertainties: a stochastic programming approach. Climatic Change 121, 143–160 (2013). https://doi.org/10.1007/s10584-013-0859-4
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10584-013-0859-4