Spatio-Temporal Estimation of Renewable Energy Potential in Built Environments using Big Data

To achieve the goals of the Paris Agreement on climate change, governments around the world must shift to renewable energy (RE) to drastically reduce their CO2 emissions within the next 10-30 years. The development of effective strategies for RE integration into national energy systems requires assessing the spatial and temporal availability of these resources at country scale. This spatio-temporal availability is described by the technical RE potential, defined as the maximum electrical or thermal energy that can be obtained from a given technology. While most previous technical potential assessments have used empirical models and geospatial tools to estimate RE potential, the increasing opportunities provided by Big Data and Machine Learning approaches, which extract additional information from large building and environmental datasets, are largely under-exploited. Integrating such data-driven methods with existing models enables national-scale RE resource assessments with higher precision and at higher spatial and temporal resolution than ever before.This work proposes Big Data approaches to estimate renewable energy potentials at national scale, with application to Switzerland. Focusing on the built environment, the technical potentials of Rooftop Solar Photovoltaics (RPV) and shallow Ground-Source Heat Pumps (GSHPs) are assessed at the spatial resolution of individual buildings and at hourly (RPV) or monthly (GSHP) temporal resolution. The proposed approaches combine large-scale data processing and Machine Learning with novel computational methods and state-of-the-art models of RE technologies. Furthermore, the quantification and propagation of uncertainties is addressed for RPV. The results are publicly available datasets of RE potentials, which enable spatio-temporal assessments of energy systems with a high share of renewables. At national level, the estimated RPV potential for Switzerland is 25+/-9 TWh, which may increase up to 39 TWh for more optimistic scenarios for solar PV panel installation, or beyond 60 TWh if rooftop-integrated PV is used. The shallow GSHP potential is estimated at up to 97 TWh, of which 20% is located in urban areas. I show that this potential may increase by over 50% through seasonal regeneration from space cooling. Spatial mismatches between supply and demand may be reduced by district heating networks (DHN).These findings have several implications for the Swiss Energy Transition for 2050: (i) To achieve the national target of 34 TWh from PV systems, 50-55% of all nearly flat (< 20° tilt) or south-facing surfaces must be exploited, whereby roofs with a high annual yield should be prioritised; (ii) GSHPs may deliver 30-70% of the heat demand of the targeted 1.5 million heat pumps and contribute significantly to the expansion of DHN in urban areas; (iii) these fractions may be increased through seasonal regeneration, for example from space cooling or excess solar thermal generation. The proposed approaches are transferable to RE potential assessments in other regions or countries, in particular by adapting them to open and cross-country input data, as well as to related RE technologies. The proposed data-driven methods and the publicly available datasets can be used by researchers, practitioners and policy makers to assess future decarbonisation pathways. I hope that this work will contribute to achieving national CO2 emission targets, moving one step closer to the goals of the Paris Agreement.