Continuous high-resolution midlatitude-belt simulations for July–August 2013 with WRF

Schwitalla, Thomas; Bauer, Hans-Stefan; Wulfmeyer, Volker; Warrach-Sagi, Kirsten

doi:https://doi.org/10.5194/gmd-10-2031-2017

Articles | Volume 10, issue 5

https://doi.org/10.5194/gmd-10-2031-2017

© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 3.0 License.

https://doi.org/10.5194/gmd-10-2031-2017

© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 3.0 License.

Articles | Volume 10, issue 5

Model evaluation paper

|

31 May 2017

Model evaluation paper |

| 31 May 2017

Continuous high-resolution midlatitude-belt simulations for July–August 2013 with WRF

Thomas Schwitalla, Hans-Stefan Bauer, Volker Wulfmeyer, and Kirsten Warrach-Sagi

Download

Final revised paper (published on 31 May 2017)
Supplement to the final revised paper
Preprint (discussion started on 29 Aug 2016)

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'Comments to gmd-2016-195', Anonymous Referee #1, 09 Nov 2016
- AC1: 'Reply to comments of Referee 1', Thomas Schwitalla, 24 Feb 2017
RC2: 'Reviewer comments', Anonymous Referee #2, 25 Nov 2016
- AC2: 'Reply to comments of Referee 2', Thomas Schwitalla, 24 Feb 2017

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

AR by Thomas Schwitalla on behalf of the Authors (11 Mar 2017) Author's response Manuscript

ED: Referee Nomination & Report Request started (16 Mar 2017) by Simon Unterstrasser

RR by Anonymous Referee #1 (04 Apr 2017)

Suggestions for revision or reasons for rejection

General Comments
The paper has significantly improved. In particular, the comparison with the two-sigma ranges from the ECMWF analysis fields demonstrates that (1) the authors have successfully applied the WRF model in the belt set up (also in a convection permitting configuration) and (2) the model gives reasonable results that are lying within a realistic bandwidth most of the time (when the model exceeds this bandwidth the authors try to give reasonable explanations, like for the Western Pacific region). However, this does not solve the concerns about the large-scale decoupling/internal variability, but since the presented work is introduced as a pilot study and since the overall afford for doing thorough investigations based on long term evaluation simulations is way too high, the presented evaluation can be seen as a successfully passed plausibility check.
The authors are encouraged to provide a more complete picture of the belt configuration and to discus its strengths and weaknesses in a direct manner. Especially, its relationship to internal model variability and its consequences for the model’s comparability with observational data in the face of short-term (2 months) simulations should be addressed. This does not belittle the authors’ work, but would demonstrate their attempt in taking care of their article to avoid potentially misguided interpretations. A clear and direct presentation of the facts would furtherly increase the quality of the paper.

Specific comments
Page 3 line 3-4: The secondary circulations described in Becker et al. (2015) are the result of lateral boundaries in general, not only in west-east direction. The manuscript is very suggestive at that point. Especially, since Zagar et al. (2013) have clearly demonstrated the sensitivity of the model results with respect to the meridional extension of the belt (see their figure 4 and figure 5).
Page 6, line 6: As a consequence of Reynolds averaging which is the basis for a discrete differences scheme, model output on single grid cells can only be interpreted as grid cell averages. Mapping one model grid onto another by means of a bilinear interpolation associates the centre of the grid cells with their values. This might artificially introduce variability. To circumvent this effect, model output on different grids are remapped onto a common grid first by means of a conservative method (e.g. Kotlarski et al., 2014; Katragkou et al., 2015). Although the ECMWF forecast model is based on a discretisation scheme that turns the governing equations into truncated spherical harmonics, the area assumption also holds for the forecast products (http://www.ecmwf.int/sites/default/files/elibrary/2015/16559-user-guide-ecmwf-forecast-products.pdf - page 15).
Katragkou, E., M. Garcia-Diez, R. Vautard, S. Sobolowski, P. Zanis, G. Alexandri, R. M. Cardoso, A. Colette, J. Fernandez, A. Gobiet, K. Goergen, T. Karacostas, S. Knist, S. Mayer, P. M. M. Soares, I. Pytharoulis, I. Tegoulias, A. Tsikerdekis, and D. Jacob (2015), Regional climate hindcast simulations within EURO-CORDEX: evaluation of a WRF multi-physics ensemble, Geosci. Model Dev., 8(3), 603-618, doi: 10.5194/gmd-8-603-2015.
Page 6, line 14: Kotlarski et al. (2014) did not evaluate “forecasts”, but “hindcasts” or “evaluation runs”.
Page 9, line 11: It is not clear how PSS is calculated: is it calculated from frequency distributions at each grid cell and then averaged over the sub-domains or is it based on frequency distributions from sub-domain averaged variables? A clarification of this would increase the traceability of the paper and could be added to the “Varification data strategy”.
Page 9, line 15: the simulations are driven by analysis fields. Due to the high internal variability, it cannot be concluded that there is a PSS when the model is forecast mode.
Page 9, line 32-33: the high variability in HIRES may also be a consequence of the bilinear interpolation. When HIRES is conservatively remapped onto the ECMWF grid first, this source of variability can be avoided.
Page 10, line 21: typo “… the the …”
Page 13, line 16: which forecasts have been used (initial time and steps)? Information on this would also fit into the “Verification data strategy”.
Page 14, line 26-27: Does this mean, in HIRES Kain-Fritsch was active together with the Morrison scheme? Or is there a typo?
Page 15, line 8-13: The period from the July 26 to 28 was dominated by an extended trough in the eastern North Atlantic that developed from a large cyclone in the North Atlantic region from July 24 onwards. West and south to this cyclone a high pressure system developed (its centre lies approximately at 50°W/35°N). In total, this caused a drop of ~8.5 hPa during July 27 in the area averaged MSLP (see Figure 10). Means over large areas that average out these phenomena are not suitable to disentangle the boundary influences from the influences of SST and hence the sharp drop in MSLP as it is synchronously seen in all simulations as well as in the driving data cannot be convincingly be explained by the attempt on page 15 line 8-13, especially since Alexandru et al. (2007) demonstrated that internal variability depends on synoptic conditions independently from prescribed SST. A description of this issue can be easily used to demonstrate the difficulties that emerge when belt simulations are compared to observational based data (see General Comment).

Concerning Kida et al. (1991): yes, they describe an alternative way to nest LAMs into coarser resolution models. My intention to refer to Kida et al. (1991) was that it is one of the earliest studies that drastically demonstrate the effect of internal variability (at that time it was not called “internal variability”). In addition, it laid down the principals of spectral nudging based on the ideas of Anthes et al. (1989). Nowadays, spectral nudging is implemented in a broad range of models, including WRF.
Anthes, R.A., Y.H. Kuo, E.Y. Hsie, S. Low-Nam and T.W. Bettge (1989), Estimation of skill and uncertainty in regional numerical models. Quart. J. Roy. Meteor. Soc., 115, 763-806.
For the same reason I referred to Peagle et al. (1996).

Hide

RR by Anonymous Referee #2 (05 Apr 2017)

ED: Publish subject to minor revisions (Editor review) (10 Apr 2017) by Simon Unterstrasser

AR by Thomas Schwitalla on behalf of the Authors (20 Apr 2017) Author's response Manuscript

ED: Publish subject to technical corrections (28 Apr 2017) by Simon Unterstrasser

AR by Thomas Schwitalla on behalf of the Authors (28 Apr 2017) Manuscript

Short summary

Due to computational constraints, extended-range forecasts on the convection-permitting (CP) scale are often performed using a limited-area model. To overcome disturbances by lateral boundary conditions, a CP latitude belt simulation in the Northern Hemisphere was performed for July and August 2013. This approach allows for the study of resolution and parameterization impacts. The results demonstrate an improved representation of the general circulation and precipitation patterns.