EGU2020-5601
https://doi.org/10.5194/egusphere-egu2020-5601
EGU General Assembly 2020
© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

The outstanding 2019 Heatwaves in Central Europe – driving mechanisms and soil-atmosphere feedbacks

Ricardo Trigo1, Pedro Sousa1, David Barriopedro2, Ricardo García-Herrera2,3, Carlos Ordóñez2, and Pedro Soares1
Ricardo Trigo et al.
  • 1Universidade de Lisboa, FCiências.ID - Associação para a e Investigação e Desenvolvimento de Ciências, Instituto Dom Luiz (IDL), Lisbon, Portugal (rmtrigo@fc.ul.pt)
  • 2Instituto de Geociencias, IGEO (CSIC-UCM), Madrid, Spain
  • 3Departamento de Física de la Tierra y Astrofísica, Facultad de Ciencias Físicas, Universidad Complutense de Madrid, Madrid, Spain

In the current study, we analyzed the two outstanding heatwaves (HWs) that affected Europe in summer 2019. The events occurred in late June and late July and were record-breaking, although peak temperatures were observed in distinct areas. During the June HW the highest temperatures were recorded in SE France, when the country registered for the first time temperatures above 45ºC. The July HW made thermometers cross the psychological barrier of 40ºC for the first time in Belgium and the Netherlands, breaking all-time records in widespread areas of Central Europe.

We detected that a subtropical ridge fostering warm advection from lower latitudes was a common feature for both HWs. However, we have also found distinct mechanisms shaping the two HWs. While the June HW was predominantly characterized by the intrusion of a vertically homogenous air mass of Saharan origin, surface processes and upward transport of sensible heat were pivotal for the July HW. Our results suggest that the intensity and extension of the June HW contributed to soil desiccation, which together with the persistence of dry and clear sky conditions during early July led to an amplification of the surface temperature anomalies during the late July HW. This is supported by a flow analogue exercise, showing amplified surface heating for flow analogues of the July HW when they are preceded by short-term dry soil moisture conditions, like those caused by the June HW. In turn, we show that, in agreement with the long-term regional warming, soil desiccation during the June 2019 event was larger than it would have been in the recent past (assessing 1984-2018 versus 1950-1983). Finally, we compared the spatio-temporal distribution of summer temperature for 2019 and the previous record-breaking summer 2003. Results show that an outstanding warming fingerprint (circa +1.5ºC in summer daily maximum temperatures averaged over Europe) has been superimposed on the relatively larger magnitude of the August 2003 HW (with respect to the climatology at that time), thus explaining the exceptionality of the record-breaking values observed in 2019.

This work was supported by national funds through FCT (Fundação para a Ciência e a Tecnologia, Portugal) under project IMPECAF (PTDC/CTA-CLI/28902/2017).

How to cite: Trigo, R., Sousa, P., Barriopedro, D., García-Herrera, R., Ordóñez, C., and Soares, P.: The outstanding 2019 Heatwaves in Central Europe – driving mechanisms and soil-atmosphere feedbacks, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-5601, https://doi.org/10.5194/egusphere-egu2020-5601, 2020

Comments on the presentation

AC: Author Comment | CC: Community Comment | Report abuse

Presentation version 1 – uploaded on 30 Apr 2020
  • CC1: Comment on EGU2020-5601, Linda van Garderen, 04 May 2020

    Dear mr. Trigo,

    Thank you for your abstract and presentation.Your work is very interesting.

    I have a question concerning the analogues you use in Figure 6. It is not clear to me how you created them. Did you randomly add or diminish soil moisture in a model ore did you group wet and dry years from NCEP/E-OBS? Are you creating dry and wet summer climatologies or is this still applicable for temperature extremes only?

    Would you be able to explain why the RMSE is somewhat larger for wet years then dry years? Could this be method based. Do these outputs fall within climate variability?

    Thank you very much.

    Linda van Garderen

    • AC1: Reply to CC1, Pedro M. Sousa, 04 May 2020

      Dear Linda, thank you for your comment

      Wet and dry conditions are defined using data from the NCEP reanalysis, as summer days of the 1950-2018 period with 15-day mean regional anomalies for the previous [-15,-1] day interval laying above the 66.6th percentile and below the 33.3rd percentile of the climatological distribution, respectively. In that way, soil moisture departures of a given analogue day represent previously accumulated values and are not the direct response to the actual atmospheric circulation conditions. To avoid effects of long-term trends that may further complicate the causality of the relationships between soil moisture and temperature, all fields were detrended by removing the regional mean linear trend over the considered domain. We have considered two sets of distributions, constrained and unconstrained by the atmospheric circulation referred as “flow-conditioned” and “random”, respectively.

      Regarding the differences in RMSE you pointed out, they could in fact reflect methodological issues (limited sampling) or feedbacks of soil moisture deficits in the atmospheric circulation anomalies. In any case, note that this difference is relevant for “flow-conditioned” distributions (dark grey). This should indicate that atmospheric environments leading to dry analogues resemble more each other, while there is a larger variability in large-scale forcing environments leading to wet conditions in this region.

      We hope this helps answering your question. Any further questions, don’t hesitate!

      Best regards

  • CC2: Comment on EGU2020-5601, Francisco Pastor, 04 May 2020

    Dear Dr. Sousa

    Congrats on your good and interesting work. I especially like that you focus on regional land-soil effects on HW. I'd like to know if you think this is a result of an individual event or if we could use soil moisture as a proxy of the possible summer temperature regimes. I am interested in the Valencia region in Spain, as we provide HW forecast to the regional health authorities, where we have had a very wet spring and would like to know your opinion. This high soil moisture, can weaken HW episodes coming in the next months? For sure there will be heat events but maybe they will be not so extreme as in the case of dry soils. Thanks and best regards

    • AC2: Reply to CC2, Pedro M. Sousa, 04 May 2020

      Dear Francisco, thank you for your comments

      As you correctly say, soil moisture contribution to enhancing or moderating surface temperature anomalies during Heatwave events (HWs) is just a factor “downstream” of the process leading to their occurrence. Naturally, large-scale circulation forcing is “upstream” and is the factor that triggers HWs occurrence, which can result in stronger/weaker observed anomalies depending on soil moisture conditions. Evidently this process is not linear, and is also highly regionally dependent. In this sense, yes we believe it can be used as a proxy in some degree to estimate HW severity potential (always depending if a favorable synoptic environment occurs) by using regional data and observations, for example to calibrate statistical models.

      In the particular case you refer, current high soil moisture values should in theory diminish potential for very large anomalies, in the case a HW occurs. But as in the 2019 example discussed here, if favorable synoptic conditions prevail, successive HW events and/or atmospheric forcing may lead to progressive soil desiccation towards summer, regardless of relatively wet conditions in early spring.

      Best regards

      • CC4: Reply to AC2, Francisco Pastor, 04 May 2020

        Dear Pedro, I completely agree with you. Although we now have high soil moisture, it can disappear if synoptic forcing prone to heat waves occurs. It can at least be an indicator of potential HW intensity in early summer.
        Thanks and best regards

        • AC4: Reply to CC4, Pedro M. Sousa, 04 May 2020

          Yes, that particular case will be interesting to follow this summer!
          Best regards

  • CC3: Comment on EGU2020-5601, Chiem van Straaten, 04 May 2020

    Dear authors,

    Thank you for this work. Could you explain the method you used for Figure 6? Are these flow analogues defined on Z500 or temperature analogues? Your additional subsetting based on dry vs wet conditions seems to result in different flow patterns / analogues (panels a,b).

    Kind regards,

    Chiem van Straaten (KNMI / VU Amsterdam)

    • AC3: Reply to CC3, Pedro M. Sousa, 04 May 2020

      Dear Chiem, thank you very much for your feedback

      Flow analogue days are defined from their root-mean-square differences (RMSD) with respect to the actual Z500 anomaly field at the time of the HW event over the [35°–65°N, 10° W–25 °E] region. For each day of the considered HW event, the search of flow analogues was restricted to the -30 to +30 days around the corresponding calendar day, excluding the year of occurrence of the HW. Analogue days are used to reconstruct the target field by randomly picking one of the 20 best flow analogues for each day of the HW event. This process was repeated 5000 times to derive flow-conditioned distributions. To test whether the dynamics played a significant role in the reconstructed anomalies of the target field, unconditional distributions were also retrieved by repeating the whole process with a random selection of days (instead of restricting the search to days with similar flow configurations).

      In the analogues exercise from Fig.6, we reconstructed the maximum 2 m temperature anomalies expected from the circulation during the July HW, distinguishing between analogue days preceded by dry or wet conditions over central Europe. Wet and dry conditions are defined as summer days of the 1950-2018 period with 15-day mean regional anomalies for the previous [-15,-1] day interval laying above the 66.6th percentile and below the 33.3rd percentile of the climatological distribution, respectively.

      In this sense, we did in fact derive two subsets based on wet VS dry soil conditions. And yes, we use both Z500 and temperature series, to specifically attribute flow analogues and HW days, respectively.

      While both wet and dry analogues present above average Z500 and temperatures, dry ones are more intense. In fact, their difference suggests that dry pre-conditioning might also amplify circulation anomalies and not only surface temperature anomalies. But here we must be careful, since our results do not specifically address that causality, and may also reflect differences in the nature of the cases contained within each analogues subset (despite the confidence and consistency of the “bootstrapping” method).

      Hope this explanation is clear, any additional question, don’t hesitate.

      Best regards