EGU General Assembly 2020
© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Temperature monitoring in mountain regions using reanalyses: Lessons from the Alps

Simon C. Scherrer and Sven Kotlarski
Simon C. Scherrer and Sven Kotlarski
  • Federal Office of Meteorology and Climatology MeteoSwiss, Climate Division, Zurich-Airport, Switzerland (

The monitoring of near-surface temperature is a fundamental task of climatology that remains especially challenging in mountain regions. Here we assess the regional monitoring capabilities of modern reanalysis products in the well-monitored northern Swiss Alps during the last 20 to almost 60 years. Monthly and seasonal 2 m air temperature (T2m) anomalies of the global ERA5 and the three regional reanalysis products HARMONIE, MESCAN-SURFEX and COSMO-REA6 are evaluated against high quality in situ observational data for a low elevation (foothills) mean, and a high elevation (Alpine) mean. All reanalysis products show a good year-round performance for the foothills with the global reanalysis ERA5 showing the best overall performance. The high-resolution regional reanalysis COSMO-REA6 clearly performs best for the Alpine mean, especially in winter. Most reanalysis data sets show deficiencies at high elevations in winter and considerably overestimate recent T2m trends in winter. This stresses the fact that even in the most recent decades utmost care is required when using reanalysis data for near-surface temperature trend assessments in mountain regions. Our results indicate that a high-resolution model topography is an important prerequisite for an adequate monitoring of winter T2m using reanalysis data at high elevations in the Alps. Assimilating T2m remains challenging in highly complex terrain. The remaining shortcomings of modern reanalyses also highlight the continued need for a reliable and dense in situ observational monitoring network in mountain regions.


How to cite: Scherrer, S. C. and Kotlarski, S.: Temperature monitoring in mountain regions using reanalyses: Lessons from the Alps, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-8945,, 2020

Comments on the presentation

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

Presentation version 1 – uploaded on 14 Apr 2020
  • CC1: Comment on EGU2020-8945, Julien Beaumet, 04 May 2020

    Interesting and inspiring work. I tried to look at this surprising negative winter trend in my RCM simulation (MAR RCM run at 7 kms and driven by ERA-20C) which unfortunately stops in 2010. Here is the result for 1988-2010. It seems somehow a bit underestimated in high elevations as well but it seems not so bad. I have a simulation driven by ERA5 about to finish, so it will be interesting to see the added value of the downscalling for this 1988-2012 cooling trend. Looking forward to some exchanges in the chat.

    Julien Beaumet

    • AC1: Reply to CC1, Simon Scherrer, 04 May 2020

      Dear Julien

      Looks pretty good indeed. Since assimilation of T2m seems to be tricky for reanalysis, RCM runs with good snow schemes and a good underlying larger scale reanalysis (like ERA5) might be an option to test. It would certainly be interesting to compare high-res RCM runs with modern high resolution regional reanalyses. Also important to note: All regional reanalyses I validated were run with the coarse ERA-interim as input.  

      Best, Simon