Evaluating regional climate models over complex topography

Ivan Güttlerivan.guettler@cirus.dhz.hr

Meteorological and Hydrological Service of CroatiaResearch and Development Division

Workshop on advances in meso- and micro-meteorology3–4 November 2014, Jezerčica Thermae,Donja Stubica, Croatia

Many thanks to collaborators: Č. Branković, M. Gajić-Čapka, M. Patarčić, L. Srnec, I. Stepanov, C. Jones, G. Nikulin

Funding: FP7-CLIMRUN, MZOS project, DHMZ, HRZZ-CARE

Outline

1.Introduction

2. RCM T2m systematic errors and T2m projections

3. RCM total precipitation errors and the impact of horizontal

resolution

4. Linking T2m, total cloud cover and surface radiation errors

5. Summary

1/15Workshop on advances in meso- and micro-meteorology

3–4 November 2014, Jezerčica Thermae,Donja Stubica, Croatia

2/15

50 km 25 km

12.5 km 6.25 km

FP6-ENSEMBLES regional climate models (RCMs): 25 km horizontal resolution

DHMZ CORDEX experiments: 50 km and 12.5 kmSMHI RCA3 experiments: 50 km, 25 km, 12.5 km, 6.25 km

In this talk:

Introduction

3/15

Branković et al. (2013) Clim. Dyn.

RCM T2m systematic errors and T2m projections

4/15RCM T2m systematic errors and T2m projections: RCM+ERA40

5/15

RCM+reanalysis (e.g. ERA40)Total Error ≈ RCM Error

RCM+GCM (e.g. ECHAM5/MPI-OM)Total Error ≈ RCM Error + GCM Error

RCM T2m systematic errors and T2m projections: RCM+ECHAM5/MPI-OM

6/15RCM T2m systematic errors and T2m projections

7/15RCM total precipitation errors and the impact of horizontal resolution: topography

Güttler et al., submitted

RCA3+(ERA40, ECMWF analysis)

Period: 1987-2008

RCA3 vs. MeteoSwissRhiresD overSwitzerlandandMETNO KLIMAGRID over Norway

8/15

Mean wintertotal precipitation

95th percentile ofthe total precipitation

RCM total precipitation errors and the impact of horizontal resolution: Switzerland

9/15RCM total precipitation errors and the impact of horizontal resolution: Switzerland

10/15RCM total precipitation errors and the impact of horizontal resolution: Norway

Mean wintertotal precipitation

95th percentile ofthe total precipitation

Hydrostatic core; 23 σ-levels

Emanuel (1991) convection scheme

Hotlslag et al. (1993) PBL scheme

CCM3 radiative transfer

BATS LSM

RCM: RegCM4 (Giorgi et al. 2012)

ERA-Interim boundaryconditions & SST

Period: 1989-2008

EURO-CORDEX domain

Horizontal resolution: 50 km and 12.5 km

R1 R2

R3

R4

11/15Linking T2m, total cloud cover and surface radiation errors

50 km

12.5km 12.5km

12/15Linking T2m, total cloud cover and surface radiation errors

12.5km

13/15Linking T2m, total cloud cover and surface radiation errors

14/15

SNS: surface net SW flux; SHF: sensible heat fluxSNL: surface net LW flux; LHF: latent heat flux

Linking T2m, total cloud cover and surface radiation errors: surface energy budget

15/15Workshop on advances in meso- and micro-meteorology

3–4 November 2014, Jezerčica Thermae,Donja Stubica, Croatia

Summary

1. The amplitude of the systematic errors on the local scales can be

substaintal. However, in some cases the climate change signal is not strongly

sensitive to the systematic errors.

2. The increase in the horizontal resolution is not sufficient for the error

reduction. Total precipitation is expected to benefit the most from the

horizontal resolution increase.

3.Many satellite-based climatology datasets are emerging. Often not fit for

the trend detection but very usefull model evaluation.

Thank you for attention!