Font Size

Abstract: The prevailed atmospheric blocking over Eastern Europe and Western Russia during July and August of 2010 led in the development of the devastating Russian heat wave. Therefore the question whether the event was predictable or not is highly important. The principal aim of this study is to examine the predictability of this high-impact atmospheric event on a seasonal time scale. To this end, a set of dynamical seasonal simulations have been carried out using an Atmospheric Global Circulation Model (AGCM). The impact of various model initializations on the predictability of this large scale event and its sensitivity to the initial conditions has been also investigated. The ensemble seasonal simulations are based on a modified version of the lagged average forecast method using different lead-time initializations of the model. The results indicated that only a few individual members reproduced the main features of the blocking system 3 months ahead. Most members missed the phase space and the propagation of the system setting limitations in the predictability of the event.

Discussion: The predictability of the Russian heat wave on a seasonal time scale has been investigated in this study. The dynamical seasonal simulations have been carried out using the state-of-the-art CAM3 AGCM. The impact of various model initializations on the predictability of the event has been also investigated because such comprehensive prognostic systems are sensitive to the initial conditions due to the chaotic nature of the atmosphere. According to the synoptic analysis, the Russian heat wave provoked by a strong omega blocking system persisted over Eastern Europe and driving warm air from Africa and Arabic peninsula to Western Russia. The vertical temperature profile over Moscow reveals an intense inversion layer coexisting with a dry air mass in the lower troposphere resulting to amplification of the anticyclone. During the blocking period the orientation of the anticyclone favored a cold northerly airflow towards the Indian Ocean which interacts with low-level warm and humid air and triggered heavy rainfall across Northern Pakistan.

Seasonal simulations of the event were based on a modified version of LAF method constructing 61 independent ensemble members initialized on January and April 2010. Each ensemble member has been integrated for 8 and 5 months ahead respectively and in this way, for the period of JJA were produced 31 members on a 5-8 months lead time and 30 members on a 2-5 months lead time.

As far as the predictability is concerned, only a few individual members in April reproduced the main features of the blocking system almost 3 months before the event. For both set of simulations the ensemble spread is relatively limited over Eastern Europe while the areas of high uncertainty are mainly located over central Russia. Most members displaced the basic characteristics of the phase space and the velocity of the system shifting the center eastward and predicting a short-lived blocking pattern. Despite the fact of the long lead period, both January and April members provided similar confidence of the forecast reliability. Thus, almost the entire members initialized on April 2010 and having 2-5 months lead time did not provide any further predictability improvement. Thus the predictability seems to be independent to the forecast horizon varying from seasonal to intra-annual time scales.

The results of this study underline the main difficulties and limitations in the seasonal simulation of such high-impact weather event. Many studies confirm that the seasonal scale predictability may be feasible but further work is required to properly assess these findings (Palmer and Anderson 1994; Hastenrath, 1995; Rowell, 1998; Lee et al., 2011). However, since the LAF method is operationally feasible, due to the fact that the LAF ensemble members can be produced during the normal operational cycle, it is of great importance to investigate furthermore the performance of such ensemble forecasting system. To this end, more high-impact weather events should be considered in order to evaluate the forecast skill and assess the effectiveness of the seasonal prediction.

Quantitative verification statistics of WRF predictions over the Mediterranean region

The Weather Research and Forecasting (WRF) limited area model with the embedded Non-hydrostatic Mesoscale Model (NMM) dynamical core has been installed and appropriately configured in the parallel computing infrastructure of the Department of Geography at Harokopio University of Athens in order to provide accurate regional forecasts for the entire Mediterranean basin and the Black Sea. Despite the production of conventional weather predictions, the model forecasts support many others operational and research activities such as driving local hydrological models for flash floods predictions, especially over small catchments, producing fire weather indexes and fire risk assessments during summer and providing estimations of the maximum wind power for areas with dense wind farms installations. The quality of the forecasts is related to the model capability in providing reliable precipitation, temperature, and wind estimations at resolutions that tend to meet the needs of the abovementioned specific applications. Many studies were based on models objective error statistics and verified both upper air and surface predictions of various numerical weather predictions models (Colle et al., 1999; Gallus, 1999; Mass et al., 2002). However, much attention has been put on the WRF model predictions in surface winds and temperature fields which can be greatly influenced by synoptic to mesoscale phenomena and local terrain heterogeneities. Recent studies revealed a rather systematic tendency of the WRF model to overestimate the 10-m wind speed and to underestimate the 2-m daytime temperature (Roux et al., 2008, Gozzini et al. 2008). The comparison with the Eta model (Cheng et al., 2005) revealed that WRF produces larger 2-m temperature and dewpoint mean absolute and bias error than the Eta model, and overpredicts the 10-m wind speed.

In this article the performance of the WRF weather forecasts has been assessed using as reference the surface measurements available from the World Meteorological Organization (WMO) network. Surface observations from more than 900 conventional stations were used to verify and compare categorical model forecasts of the 10-m wind field, 2-m air temperature and sea level pressure every 3 hours and the accumulated 6-h precipitation for two consecutive years (2009 and 2010). In this end, a verification procedure has been developed based on the estimation of traditional objective verification techniques such as bias, root mean square error (RMSE) and threat scores for the continuous and discrete predictants.

On the basis of traditional objective verification techniques (like bias, RMSE, threat scores) preliminary results showed that the mean sea level pressure is systematically underestimated in a range of 0.1-0.5 hPa. The systematic error of the near surface air temperature indicated a diurnal signal, in which the moderate cold bias of almost -0.7°C on evening hours turned to +0.5°C warm bias during daytime. The seasonal distribution of the statistical scores revealed a cold, up to 1°C, bias of the minimum and maximum temperatures for the transient and summer seasons while the maximum temperatures were overestimated up to 0.5°C during winter. Moreover, the minimum temperatures indicated a systematic cold bias up to 0.5°C at the stations located under the 250m. This may be attributed to the model domain inadequate representation of terrain characteristics. The maximum temperatures were also underestimated up to 0.5°C for the elevations exceeding the 750m during the transient and summer seasons.

Concerning the wind speed forecasts, the WRF model systematically overestimated the wind speed up to 1.5 ms-1 during evening hours while the forecast error remains almost constant (2.5-3 ms-1) for the entire forecast time. However the seasonal variability of the bias score suggested that autumn and winter are the seasons that mostly contributed to the systematic overestimation. The definition of the forecast error local maxima during the evening hours and over the cold period of the year suggests a rather unrealistic description of the near surface heat and momentum fluxes. Additionally, the stations located at low and moderate altitudes mostly contribute to the wind speed overestimation. This may in part be related to the discrepancy between the elevation represented in the model domain and the actual elevation at which observations were made.

The 6-h accumulated precipitation bias score indicated a systematic overestimation of the light-to-moderate (0.5-6 mm) thresholds, which was more prominent at the day time, while the medium-to-high thresholds (6-24 mm) were underestimated for the entire forecast period. The equitable threat score was continuously decreased with increased precipitation categories. However, domain-average precipitation statistics reveal only part of the model performance, since it has been found substantial spatial variations in precipitation forecast accuracy (Mass et al., 2002).
Petros Katsafados, 12th Annual WRF Users’ Event, 20-24 June 2011, Boulder CO, USA.


You are here: Home articles