from climate etc.
Author: Frank Boss
Neither trend analysis nor model observation comparisons support the conclusions of the attribution study, which found:
“Combined changes are twice as likely and 7% more intense due to human-induced climate change”.
Starting on September 11, heavy rainfall occurred in parts of Austria, Poland and the Czech Republic. The first assessment noted widespread record rainfall due to a “Vb weather condition,” named after the historical classification of the direction tracked by a European low pressure field. In Vb weather conditions, the low pressure area moves to the Mediterranean and then northeast, usually ending in the Baltic Sea region of Europe. Vb conditions are often associated with heavy rainfall and flooding events in Central and Eastern Europe, such as in 1997 (Oder) and 2002 (Elbe).
However, a few days later, an “attribution study” emerged. The core message about the incident (cited in the media) is:
“Combined changes are twice as likely and 7% more intense due to human-induced climate change”.
To assess the robustness of this claim, the full text of the attribution study was downloaded.
Meteorological classification of relevant events involves a variety of atmospheric dynamic characteristics. The triggering event is an “Arctic outburst,” involving the extreme northward shift of the Intertropical Convergence Zone (ITCZ). To make matters worse, there is a stable blocking high-pressure field to the north of the area, so the precipitation zone is relatively stationary and cannot move northward to the Baltic Sea as usual.
The key question is whether the thermodynamic factors that contributed to the events described (related to warming due to “climate change”) can actually be quantified with some robustness as claimed in attribution studies.
The attribution study describes the trend analysis of observational data (E-Obs.) and (weather) model observational reanalysis data (ERA5) for the period 1950 to 2023 (2024). The data used are available through the KNMI Climate Explorer, allowing the evaluation of these figures. This study uses the GMST GISS dataset to describe the link between heavy rains in central Europe and global warming. Attribution research states:
“All data sets show similar trends in the region, with increasing trends…” (see Section 3.1)
The same data set is used here, but averaged over the relevant European regions rather than globally. The average temperature anomalies in the 20°W-25°E region from 1950 to 2023; 35°N-65°N are shown in the figure below. The region's land area (which is warming faster than the oceans) is larger than the global average land area of about 30%.
Figure 1: European regional temperature time series (GISS). This image was produced by KNMI Climate Explorer.
(Not too surprising) Observation: From about 1950 to 1981, there was no increase in temperatures. Anthropogenic warming manifested in average temperatures began around 1981, not 1950.
In order to calculate the trend of “RX4days” precipitation (that is, the accumulation of precipitation in 4 days), the ERA5 data from 1950 to 2024 was recalculated:
Figure 2: Outstanding events in September 2024 are clearly visible. It gives a positive trend slope (“one-year trend”) for 1981-2024 (green) and a zero trend slope for 1981-2023 (black). This graph was generated using ChatGPT.
As noted in the attribution study, the Ordinary Least Squares (OLS) trend (blue) from 1950-2024 is robustly positive (p=0.025). However, it does not mention that this trend becomes insignificant when calculated from the late 1960s to 2024. One would expect that the trend tilts towards being completely insignificant in 2024 (p=0,32) and zero in 1981-2023 (black). Based on these findings, OLS trends through 2024 may be more the result of internal variability. During the 1950-1981 period when there was no warming (see Figure 1), the most positive trend slope (orange) for RX4day is 2 times steeper than the 1981-2024 period when forced warming is observed.
This study evaluates climate models for attribution analysis. Many models belong to the CMIP6 series. It is known that these models face great difficulties with atmospheric dynamics due to low resolution. Multi-model averages show that the study region (46°N- 52°N; 11°E- 24°E) has no skill in the spatial correlation of model observations (E-Obs.) of precipitation.
Figure 3: Spatial correlation of CMIP6 multi-model mean precipitation with warm season observations (E-Obs.) from 1975-2023. This image was produced by KNMI Climate Explorer.
Based on model-real-world comparisons, meaningful correlations should be a prerequisite for attributing anthropogenic warming simulated in CMIP6 models to significant extreme precipitation events.
In Table 4.1 of the attribution study, the models were evaluated and some (only a few) were labeled “good” in terms of precipitation. The model number “IPSL-CM6A-LR” is marked as “reasonable”. The spatial correlation of E-OBS observations during the Vb event observation months 1950-2023 is shown below, also for the “good” model “EC Earth 3”, both below 20% and indistinguishable from random noise:
Figure 4: There are no skills in the selected research model (white indicates zero correlation). This image was produced by KNMI Climate Explorer.
Neither “IPSL-CM6A-LR” (left) nor “EC Earth3” (right), which were labeled “Good” in the study, have any tricks up their sleeves in terms of spatial correlation of precipitation with the real world. Neither does model MPI-ESM1-2LR (study “reasonable” in Table 4.1), but is not shown here.
Finally, attributing extreme precipitation events to climate change seems questionable when using CMIP6 models. Warming oceans are certainly responsible for more evaporation and more rainfall, although the increase in precipitation with warming is only a fraction of the increase in evaporation.
However, the influence of atmospheric dynamics is overwhelming, hampering the attribution of single extreme weather events based on thermodynamic arguments.
in conclusion
Upon closer inspection, neither the trend analysis nor the comparison of model observations supports the conclusions of the attribution study.
The problem of inadequate attribution research on extreme weather events is not limited to extreme precipitation. As this recent article by Roger Pielke Jr explains, attribution studies of extreme weather events of all types are often highly questionable and appear to be motivated more by “political” purposes And not for scientific purposes.
Relevant
