ESCI Datasets and Scientific Publications

Datasets created for the ESCI project will be published as they become available and quality-control/assurance has been completed. Where these have been published in the peer-reviewed literature, links to the articles will also be provided.

Heat Case Study

The draft (i.e. not fully quality checked) data developed for the May 2020 discussion draft of the Heat Case Study are available on request. Time series data for the South New South Wales region have been derived from the CCAM regional downscaling representation of six models from the CMIP5 dataset, under a high emissions pathway (RCP8.5). These data comprise time-series (in 30-minute time steps from 1980-2040) for the following metrics:

1. Air temperature (2m) (K)
2. Global Horizontal Irradiance (W/m2)
3. Direct Normal Irradiance (W/m2)
4. Wind speed at 10m (m/s)
5. Wind speed at 150m (m/s)
6. Wind speed at 250m (m/s)

This dataset is available (in Excel format) on request by contacting the project coordinator.

A detailed explanation of the approach used to develop these datasets is published in the following peer-reviewed journal article.

Huang, J., Jones, B., Thatcher, M., and Landsberg, J. (2020). Temperature impacts on utility-scale solar photovoltaic and wind power generation output over Australia under RCP 8.5. Journal of Renewable and Sustainable Energy 12, 046501. doi: 10.1063/5.0012711.

Abstract

Climate change has the potential to impact the generation of renewable energy significantly subject to location and equipment specifications. As the penetration of renewable energy in the energy systems keeps increasing, this impact needs be systematically assessed so that investment and reliability information is accurate. Australia represents an ideal study case characterized by its frequency of extreme weather events and the recent and planned growth in the renewable energy sector. In this study, we model and quantify the long-term temperature de-rating impact of utility-scale solar photovoltaic and wind power generation over Australia. Using climate projections simulated by six Global Circulation Models and the CSIRO's Cubic Conformal Atmospheric Model, we analyze half-hourly time series of key weather variables such as temperature, surface solar irradiance, and wind speed for 1980–2060 at two sites where variable renewable generators are located, or are likely to be located in the future based on the current Integrated System Plan by the Australian Energy Market Operator. We also built power conversion models for the temperature de-rating of solar and wind power with added focus on high temperature scenarios. We found that the general temporal trends in annual solar and wind power generation due to climate change are small, being at the order of 0.1% of their average production per decade. However, for peak temperature events, which coincide with the peak power demand and, generally, high prices, the temperature de-rating impact can be much more substantial and disruptive.

Access the journal article here (a subscription is required to access the full text): https://aip.scitation.org/doi/10.1063/5.0012711

Regional Maps of Extreme Temperature Exceedance Probability

These data have been developed in collaboration with the electricity sector to provide concise fit-for-purpose information to assist in numerous risk management processes and have been referenced in the AEMO 2020 ISP . Extreme heat is a major vulnerability of the electricity sector, and climate projections for temperature have a high degree of confidence.

Figure 1 shows the projected change in global average surface temperature under various emissions scenarios; from RCP2.6, a scenario with very low to net-negative global emissions, to RCP8.5, a high global emissions scenario. In the next 20 to 40 years, temperature is projected to increase regardless of scenario, consistent with increasing concentrations of atmospheric greenhouse gases. The greatest scenario variation is apparent in the latter half of the century. While this study focusses on the high emissions scenario (RCP8.5), the implications for lower emissions scenarios are likely to be similar in the next 10 to 20 years.

Figure 1. Projected global average surface temperature change (relative to 1986-2005). Source: IPCC 2013 Fig. SPM.7a (https://www.ipcc.ch/report/ar5/wg1/)

A 10% probability of exceedance (POE) temperature is a temperature threshold that should be exceeded, on average, once every ten years. As the climate changes, the threshold is projected to rise, requiring consideration by electricity sector decision makers. Such 10% POE thresholds are used in numerous existing forecasting, planning and asset management decision-making processes in the sector.

While climate models project increases in temperature with a high level of agreement and confidence, the regional impacts vary and there is increased uncertainty as spatial resolution increases. Following best practice in climate risk analysis, a large ensemble of climate models has been collated to explore the uncertainty in regional impacts by comparing different results from different climate models. The 10^th, 50^th and 90^th percentiles were calculated for the range of different ensemble members to form the lower, middle, and upper range in projected changes. Maps of the ensemble ranges are shown in Table 1, below.

Table 1. Maps of historic and projected 10% POE for Australia

Percentile of Ensemble Range	2000-2019	2020-2039	3040-2059
90th percentile (upper)	N/A
50th percentile (mid)/ Observed 2000-2019
10th percentile (low)	N/A

Download all maps, including for 1990-2009 in a single zip file (1.4 MB)

Users should explore the range of uncertainty for regions of interest. For example, Figure 2, below demonstrates interpretation of the information for Mildura, Victoria. The 10% POE temperature over the last two decades has been 44 degrees Celsius, while climate projections show that this could range from 44 to 48 degrees in 2040-2059 subject to the degree that global warming impacts this particular region.

Figure 2. Interpretation of 10% POE data for Mildura

Methods

The 1-in-10 year probability of exceedance (POE) daily maximum near surface air temperature plots were generated using a combination of observed datasets and climate model projections. The 1-in-10 year probability was calculated with a Generalised Extreme Value L-movements approach. The periods 1990-2009 and 2000-2019 were based on the Australian Gridded Climate Data (AGCD) provided by the Australian Bureau of Meteorology. The future climate projections were based on an ensemble of downscaled climate simulations for the RCP8.5 emission pathway, with the downscaling models described below. The change in the future climate projections of 1-in-10 year POE daily maximum temperature from 2020-2039 and 2040-2059 were calculated relative to 2000-2019. This change was then added to the observed 1-in-10 POE daily maximum temperature data from the AGCD for 2000-2019 so that the future projections are consistent with the historical observations. The 10th, 50th and 90th percentiles of the 1-in-10 year POE for daily maximum temperatures were calculated for the range of different ensemble members (i.e., different combinations of downscaling techniques, downscaling models and host global climate models), to form the lower, middle and upper range in projected changes. Several downscaling techniques were employed. Bureau of Meteorology Atmosphere Regional Projections for Australia (BARPA) downscaled ACCESS1-0. Quantile Matching of Extremes (QME) from the Bureau of Meteorology downscaled ACCESS1-0, CNM-CM5, GFDL-ESM2M and MIROC5 global climate models. The Conformal Cubic Atmospheric Model (CCAM) from CSIRO Oceans and Atmosphere downscaled ACCESS1-0, CNRM-CM5, CanESM2, GFDL-ESM2M, MIROC5 and NorESM1-M. NARCLiM v1.0 data was not used due to only being available for the 2020-2039 time period, although NARCLiM v2.0 data will be included in a future update.

Page updated 2^nd October 2020