Researchers at the University of Geneva have used a statistical learning framework to analyze 1,550 scenarios that describe what PV might achieve in terms of globally deployed volumes and generating capacity by 2050.
“These scenarios are not predictions, so they cannot be assessed in terms of accuracy. Instead, they show alternative plausible pathways,” researcher Marc Jaxa-Rozen told pv magazine.
The group assessed scenarios from scientific studies published after 2010, including 1,360 outlooks from the Intergovernmental Panel on Climate Change (IPCC) and 190 non-IPCC scenarios. They found that the first ones tend to be more conservative about the future growth of photovoltaics.
“However, it is important to keep in mind that the IPCC scenarios are not intended as forecasts of technology growth,” Jaxa-Rozen said.
The scientists found that the analyzed scenarios diverge significantly in forecasting solar uptake.
“Some uncertainty in scenarios until now was attributed to techno-economic parameters. For a given energy model, different cost assumptions can more than double projected PV generation,” they said. “Across multiple models with consistent parametric assumptions, structural choices on power system modeling can similarly shift PV generation by a factor of two.”
Machine learning
The scientists analyzed the forecasts via spectral clustering and topic modeling. Their proposed statistical learning framework considers the types of organizations behind the scenarios, the models that were used, and a series of structural assumptions.
“The results indicate that general properties of the models and publications are associated with a large portion of the variation in projected PV outcomes,” the academics said. “On the basis of these properties alone, the scenarios can be classified into quintiles in terms of PV capacity growth with 73% accuracy.”
The research group used XGBoost, a decision-tree ensemble machine-learning algorithm that uses a gradient-boosting framework. They analyzed a subset of 1,392 scenarios based on three groups of indicators, including the share of PV in total primary energy demand (TPED), the TPED itself, and total energy-related CO2 emissions.
Their analysis showed that European organizations tend to predict a significantly higher compound annual growth rate (CAGR) than their Asian and North American counterparts. They found that academic assessments tend to consider the widest range of uncertainties, while scenarios issued by corporate organizations usually forecast higher CAGR. They also found that scenarios based on analytical methods have higher CAGR than optimization or simulation models, while the non-IPCC scenarios have a higher mean CAGR than IPCC scenarios.
“This difference equates to an installed PV capacity that is on average higher by a factor of 3.7 in non-IPCC scenarios by 2050,” the researchers said. “The difference between IPCC and non-IPCC scenarios is more pronounced at the near-term 2030 horizon and it is equivalent to an average installed capacity higher by a factor of 4.5 in non-IPCC scenarios.”
Regional patterns
The scientists also used the analytical framework to evaluate a subset of 116 scenarios outlining PV cost assumptions. The analysis was based on the costs of utility-scale solar, and not on the levelized cost of energy (LCOE). It showed a general trend between projections for higher capital costs and lower projected CAGRs, but with substantial variations caused by other scenario factors. It also showed that IPCC scenarios tend to report higher PV costs and that differences emerge in regional PV deployment patterns.
The Swiss group also noted that international organizations tend to predict stronger PV deployment in Asia, relative to the region's TPED, but comparatively low levels of PV development in Africa. The IPCC scenarios show a wide range of regional outcomes, including more optimistic pathways for PV growth in Africa and some pessimistic pathways in Asia.
“The main design focus for IPCC scenarios is on the type and ambition of climate mitigation policies, and technology has been somewhat less of a focus,” Jaxa-Rozen said. “The detailed modelling of PV technology, grid integration, and PV-specific policies is relatively recent in the IPCC scenarios, and remains an active direction of research.”
The scientists presented their findings and described the analytical framework in “Sources of uncertainty in long-term global scenarios of solar photovoltaic technology,” which was recently published in Nature Climate Change.
This content is protected by copyright and may not be reused. If you want to cooperate with us and would like to reuse some of our content, please contact: editors@pv-magazine.com.
By submitting this form you agree to pv magazine using your data for the purposes of publishing your comment.
Your personal data will only be disclosed or otherwise transmitted to third parties for the purposes of spam filtering or if this is necessary for technical maintenance of the website. Any other transfer to third parties will not take place unless this is justified on the basis of applicable data protection regulations or if pv magazine is legally obliged to do so.
You may revoke this consent at any time with effect for the future, in which case your personal data will be deleted immediately. Otherwise, your data will be deleted if pv magazine has processed your request or the purpose of data storage is fulfilled.
Further information on data privacy can be found in our Data Protection Policy.