Evaluation and Projection of Vb-Cyclones and Associated North-Western Mediterranean Sea State in Regional Coupled Climate Simulations

Vb-cyclones propagating from the North-Western Mediterranean 9 Sea into Central Europe are often associated with extreme precipitation. This 10 study explores the state & process chains linking the North-Western Mediter11 ranean Sea and the Vb-event precipitation in Danube, Elbe, and Odra catch12 ments in regional coupled atmosphere-ocean climate simulations with COSMO13 CLM+NEMO. Two high-resolution simulations, an evaluation simulation (195114 2005) downscaling the centennial ERA-20C reanalysis and a continuous simu15 lation (historical 1951-2005 + RCP-8.5 future scenario 2006-2099) downscaling 16 the EC-EARTH global climate RCP-8.5 projection are used for this purpose. 17 Sea surface temperature (SST) validation with observations over the Mediter18 ranean Sea reveals a cold bias (≈ 1 – 1.5 K) in the historical & evaluation 19 simulations. There is a good agreement in mean annual Vb-cyclone frequency 20 between the evaluation (9.7 events/year) and the historical (10.1 events/year) 21 simulations. But, there are significant discrepancies in the seasonal cycle. The 22 mean cyclone intensity measured with minimum central pressure, track den23 sity, and precipitation rankings in the catchments also show good agreement. 24 A basin-average SST warming of ≈ 2.5 – 3 K, but insignificant changes in Vb25 frequency, mean intensity, and precipitation in the catchments are projected 26 by the end of the 21 century. The North-Western Mediterranean SST, evapo27 ration, and wind speed anomalies corresponding to the precipitation rankings 28 over the three catchments & associated process chains differ between the evalu29 ation & historical simulations. In the evaluation simulation, Vb-cyclone precip30 itation rankings correspond with SST, evaporation, and wind speed anomalies, 31 while in the historical & the future simulation no such correspondence is seen. 32 Especially the Adriatic & Ionian basins show no sensitivity to the Vb events 33 in the EC-EARTH driven simulation. The change in the processes might be 34 1 Institute for Atmospheric and Environmental Sciences, Goethe University Frankfurt am Main, Altenhöferallee 1, 60438 Frankfurt am Main, Germany. 2 now at Deutscher Wetterdienst, Offenbach, Germany 2 Praveen Kumar Pothapakula et al. because of the emergence of simulation biases inherited from the driving EC35 EARTH global simulation. 36

to infrastructure and life, and hence are in need of further investigations. 48 Over Central Europe, extra-tropical cyclones named Vb-cyclones are of-49 ten associated with extreme precipitation, especially in the summer season  river run-off model TRIP to make the regional system dynamically interactive 136 (Akhtar et al, 2018). The added value of such a coupled regional system was  In the current study, we apply the regional climate coupled model (COSMO- gion with high resolution coupled regional climate models (Ruti et al, 2016 2. Does the EC-EARTH driven coupled regional simulation produce Vb-183 events comparable to the ERA-20C reanalysis driven coupled regional simula-184 tion in the historical period. Do the Vb-cyclones change in the future?

217
The NEMOMED12 is the ocean component of the coupled regional climate    Furthermore, as our downscaled simulation with EC-EARTH was a part of Title Suppressed Due to Excessive Length 7 coordinated activity contributing to the Med-CRODEX phase II simulations, 254 we also used the SST data sets available from the Med-CORDEX phase-II 255 database to evaluate our simulations. The institutions and the models are 256 described in Table 1. CLMcom-GUF-CCLM5-0-9-NEMOMED12-3-6 EC-EARTH Observations and Reanalysis data sets The SST data sets from the models in Table 1 are available through the i.e., the Danube, Elbe, and Odra (shown as rectangles in black, orange, and 299 green colors respectively in Figure 3). The area-averaged precipitation sum 300 anomaly accumulated during the Vb-cyclone days over these catchments was 301 further ranked according to the intensity. The precipitation sum anomaly dur-302 ing the Vb-event was calculated by subtracting the day sum with the clima-303 tology day sum precipitation in the GUF evaluation, historical, and future 304 simulations respectively. In addition to the Vb-cyclone precipitation anomaly 305 rankings, we also showed the absolute precipitation amounts.

306
Corresponding to the Vb-cyclone precipitation anomaly rankings we an- Odra catchments. We adapted the methodology as in Krug  Vb-cyclones.

342
At the heart of the information theory lies the concept of Entropy (H). The

343
Entropy quantifies the uncertainty of a random variable X (Shannon, 1948) 344 and is defined as, where p(x) represents the probability of a state of the random variable X.

346
The summation goes through all the states of the random variable quantifying Where the p(x, y) represents the joint probability of a state corresponding 355 to the random variables X and Y . The M I is a symmetric quantity and thus 356 can not detect the direction of information exchange.

357
The T E builds upon the M I measure and is defined as mutual informa- Due to computational complexity in the estimation of joint probability 362 densities, the whole past of the source and target random variables are reduced 363 as follows, where τ and ω represents the time lags of the history of source and target

368
In this study, we chose the target variable to be the spatial averaged daily      Thereafter, we investigated the information exchange between the SST's 515 and the Vb-cyclone precipitation over the three catchments to diagnose the 516 process chains (Fig. 6). We noted differences in the information exchange spa-      circulation and lapse-rate changes for the future european summer climate.