Reassessing Model Projections for Hot Days from RCPs –

An Aussie Perspective

Dr B Basil Beamish

Hot Day Projections in Australia

At the end of 2015, the CSIRO and Bureau of Meteorology (BOM) released a joint technical report that contained climate model projections for hot days in various locations around Australia. The values are contained in Table 7.1.2 on page 98 of the report for three climate model scenarios, commonly called Representative Concentration Pathways (RCPs). These values are 20-year averages centred on the periods of 2020-2039 (2030 for RCP4.5) and 2080-2099 (2090 for RCP2.6, RCP4.5 and RCP8.5). Hot day model projection values for Broome according to these climate scenarios are contained in Table 1. Each of the model projections indicates a substantial increase in the average number of hot days for the rest of this century. The media most commonly report the higher values to raise the “shock factor” of their articles. Consequently, the general public and politicians are bombarded with what can best be described as disingenuous exaggerations. In isolation as presented in Table 1, there is no real visual context for these values particularly in relation to historical data.

Table 1. Projected average number of hot days (Tmax >35°C)
for Broome, with 10^th and 90^th percentile values shown in brackets
(CSIRO and Bureau of Meteorology, 2015)

The temperature records for Broome are publicly available through the Bureau of Meteorology website. The highest 20-year centred average number of hot days recorded for Broome since 1910 is 63.9 for the years 1944-1963 (centred on 1954) and the annual hot day totals have fluctuated from a high of 90 in 1945 to a low of 20 in 2000. In the most recent year, 2018, there was a total of 55 hot days for Broome, which coincidently happens to be midway between these two extremes.

RCP8.5 The Futuristic Thermal Armageddon or Just Living in Lalaland

The RCP8.5 model projection for Broome is plotted in Figure 1 as a continuation of the actual results from 1910 using the ACORN-SAT maximum temperature data set. The graph also includes the measured and model projections for Global CO₂ concentrations.

It is clear from Figure 1 that the RCP8.5 model projections for Broome are completely unrealistic and the three years of additional data since 2015 confirms this, since the 20-year centred average number of hot days has already begun to fall below the 10^th percentile limit for this scenario. It will require the number of hot days in Broome to average above 60 per year for the next five years for the trend to maintain the lower limit projected from the model. This scenario has occurred in the past, and therefore it will be what happens after this period of time that will be critical in verifying the projected trends from the RCP8.5 model.

Figure 1. Broome annual hot day totals and actual 20-year centred average trend compared to climate model projections for RCP8.5.

RCP4.5 The More Moderate Thermal Armageddon

The RCP4.5 model projection for Broome is shown in Figure 2. There is an initial dramatic increase in the average number of hot days that gradually decreases in the latter years of this century. The current trend of the 20-year centred average for Broome, including the three additional years since 2015, plots within the boundaries of this model. However, this trend is no different to that experienced in previous years on multiple occasions.

By 2030 it will become readily apparent whether this model is valid, as by then the 20-year average number of hot days centred on 2020 for Broome should have exceeded the previous highest value of 63.9 to stay within the 10^th percentile limit (Figure 3).

Figure 2. Broome annual hot day totals and actual 20-year centred average trend compared to climate model projections for RCP4.5.

Figure 3. Broome annual hot day totals and actual 20-year centred average trend compared to climate model projections for RCP4.5 shown on an expanded scale from 2000 to 2050.

RCP2.6 Global CO₂ Emissions Reduction by Human Intervention

The RCP2.6 model projection for Broome shown in Figure 4 is essentially the same as RCP4.5 up to 2030. However, in subsequent years there is a gradual decrease followed by a steady decline in the median projected value for hot days from about mid-century. Consequently, the validity of this model should also be apparent by 2030.

Figure 4. Annual hot day totals and actual 20-year centred average trend compared to climate model projections for RCP2.6.

Placing Model Projections into Perspective

A striking feature of Figures 1, 2 and 4 is that there appears to be no correlation between the Global CO₂ Concentrations and the actual data for Broome from 1910 to the present day. Therefore, past observations do not support the future projections based on increasing CO₂ concentrations.

Climate model projections need to be verified against actual observations before any confidence can be placed in the model results. This takes time. Nevertheless, accountability of model projections should be mandatory given the importance that has been placed on them for policy decision making. The average number of hot days projected for locations around Australia is a good working example for this type of accountability analysis. The hot days trend for Broome from 1910 to present day clearly shows that climate model projections for RCP8.5 are totally unrealistic. As each subsequent year of data becomes available the model projections for RCP4.5 and RCP2.6 can be compared against the measured hot days results. Indications are that by no later than 2030 both these models will be either validated or invalidated. This same accountability analysis can be applied to any of the other locations shown in the CSIRO/BOM report.

References

CSIRO and Bureau of Meteorology, 2015. Climate Change in Australia: Projections for Australia’s Natural Resource Management Regions, Technical Report, CSIRO and Bureau of Meteorology, Australia, 216p.