Ashraf S. Zakey The Egyptian Meteorological Autority
A new inter-regional climate outlook forum for the Mediterranean Region. Development and dissemination of a consensus-based seasonal climate outlook. Periodic organization of activities aiming to strengthen the existing capabilities in seasonal forecasting.
In January 2012 in Algiers during the Scoping Workshop on Seasonal Climate Prediction for North Africa (PRESANORD 02), most of the participants expressed the potential benefits and their agreement to a possible extension of the existing RA I PRESANORD Outlook Forum to the Mediterranean Region involving RA I and RA VI. This proposal of a MedCOF, inheriting much of the work done by PRESANORD, may help to give momentum to regionally coordinate seasonal forecasting activities. In Egypt (January 2013) during the PRESANORD meeting clearly showed the determination of the involved countries to cooperate for the promotion of coordinated seasonal forecasting activities in the region.
Some potential participants in MedCOF are still in an incipient status of development of seasonal forecasting activities both in operations and research domains. MedCOF should facilitate the periodic organization of courses, workshops and other activities aiming to strength existing capabilities in seasonal forecasting and other working areas of MedCOF.
Way forward Joint LRF training (MedCOF, SEECOF, PRESANORD)? Once a year rotating N-S, E-W? Themes for the next LRF training session? Lectures/practical sessions?
Scoping Meeting for a Mediterranean Climate Outlook Forum. 12-14 junio 2013, Madrid. Ashraf Zakey First MEDITERRANEAN CLIMATE OUTLOOK FORUM (MedCOF-1) Belgrade, Serbia, November 13-19, 2013
Interesting Timescales Czaja and Frankignoul, 2002
Interesting Timescales #2 Days Weeks Month Season Decadal? Wind Storms Forest Fires Cold Air Outbreaks Dry/Wet Spells Drought Evolution Heat/Cold Waves Water Supply Energy Supply Normal Conditions (e.g. for Agri. Act. Planning)
Seasonal Forecast Overview Observed Temperature Matrix 1979 2010 1976 Temperature at location (x,y) Multivariate Linear Regression algorithm Predictors Matrix Forecasted Temperature at (x,y) for time (t) 9 expressed as probability of occurrence
Iteration over time and space Hindcasted Temperature at location (x,y) Hindcast Temperature Matrix 10 1979 2010 1976
Dominant tercile maps Towards a probabilistic view Multiregressive DRY NORM. WET Forecast Multiregressive Errors P forecast(d,n,w) PFcst PObs D,N,W s 2 ε : Forecast Error Variance SSE: Sum Square Errors n : Sample amount K : Predictors Prob. Dry Prob. Wet Norm.Ter.
Method Overview The multi-regressive method based on physical atmospheric indices and sea surface anomalies, at monthly time scale. Lead time choices are made on physical basis and on a maximization of the regression values between observed and forecasted field anomalies thus performing an adaptation for the best choice of predictors. 12
Winter Summer Predictions Summer variability could be associated to winter atmospheric variability (SV NAM, snow cover, etc.) and to Atlantic SST anomalies (Tripole and 1 st EOF Guinea).
Fall Winter Predictions Fall to Winter variability could be associated to Spring to SST anomalies through reemerging Atlantic SSTA mechanism (OUTLOOKS) plus the atmospheric fall winter variability (MONTLY). 14
Outlooks Temp850hPa Anom DJF April 2006 July 2006 November 2006 Reanalysis 1971-2000
there is still room for a cold period!
MedCOF-1 (Belgrade, Serbia, 13 19 Nov. 2013 ) Full report in: http://medcof.aemet.es/medcof/events/medcof1/medcof1.html Back-to-back with SEECOF-10 and LRF Training. First historical consensuated outlook for the whole Med. Region (winter 2013-14) Temperature Precipitation
MedCOF-2 (on-line) for JJA 2014 Web page (http://medcof.aemet.es ) and forum specially developed for online sessions First draft of procedures for verifying, diagnosing climate system current status and producing consensus outlooks Final products available in http://medcof.aemet.es/medcof/events/medcof2/medcof2.html JJA temperature
Evaluation of MedCOF products Evaluation of the most recent MedCOF seasonal forecasts is conducted at the start of each MedCOF meeting based on National Climate Reports for the previous season delivered by the participants and on the materials provided by RCCs on Climate Monitoring and Climate Data of both RAs. The reference climatological period is 1981-2010.
Potentially interested contributors for the COF International organisations : WMO, Regional level : RCCs (RCC RA VI, RCCs RA I including ACMAD), SEECOF, National level : NMHS of the SEECOF area, NMHS of North Africa, NMHS of South- Western Europe, Others : climate experts (if needed), communication advisors, Task Team of RAVI on COFs Working Group of RA VI on Climate and Hydrology
Results of MEDCOF -1
ANALYSIS AND VERIFICATION OF THE MEDCOF-1 CLIMATE OUTLOOK FOR THE 2013-14 WINTER SEASON FOR THE MEDITERRANEAN REGION (MED) 2 3 1 (A) (B) Graphical presentation of the climate outlook for the 2013-14 winter season for the Mediterranean region (a) Precipitation Outlook; (b) Temperature Outlook
Temperature As stated in the MedCOF-1 Consensus Statement for the Seasonal Climate Outlook for the 2013-14 winter season for the Mediterranean region, uncertainty for the temperature prediction is high in almost all the MedCOF region. A weak tendency is noticeable for the upper tercile in the western part of the domain, North African coasts and most of the Mediterranean Sea (regions 2 and 3 in figure 1 (B)). Most of the Balkan Peninsula, Turkey, South Caucasus region and Sahara show no clear signal and climatology is therefore assigned for the three categories (region 1 in figure 1 (B)).
Precipitation Uncertainty in precipitation predictions is large. There is no preference for any climate defined categories in most of the MedCOF domain (region 3 in figure 1 (a)). There is a slightly higher probability (40%) that the amount of precipitation will be exceeded in the central Mediterranean region (region 2 in figure 1(a)). The southern part of the Iberian Peninsula and the Atlantic Facade of the African region show some slight tendency for the dry tercile (region 1 in figure 1(a)).
Analysis of the 2013-14 winter season Precipitation percentage of the 1951-2000 average for winter 2013/14. Source: GPCC temperature anomalies (1981-2010 reference) for winter 2013/14. Source: KNMI, data from NOAA NCDC
Analysis of the winter season temperature and precipitation anomalies are based on seasonal bulletins on climate in the WMO region I and VI for the winter of 2013-14 (WMO RA I RCC Node on Climate Monitoring: http://www.meteo.tn/htmlen/donnees/winter_2014.pdf ; WMO RA VI RCC Offenbach Node on Climate Monitoring: http://www.dwd.de/rcc-cm) and national reports from MedCOF participants.
North Africa (RA I) [ Temperature] The winter of 2013-2014 was warmer than normal in most regions of North Africa. Winter season mean temperature ranged between 8 C in Djefla in the north east of Algeria and 21 C in Asswan in the southeast of Egypt. In most of the North African region, anomalies ranged between +1 C and +2 C. Seasonal temperature was above normal in the north of Algeria, the north of Tunisia, the north of Libya, the east of Morocco and the north of Egypt. Anomalies were even higher with values up to +3 C above normal. This was registered in the center of Algeria, the south of Tunisia, and the south of Egypt. The exception occurred in the southern region of Libya with an anomaly of 0 C below normal 1981-2010.
North Africa (RA I) [Precipitation ] The winter of 2013-2014 was drier than normal in most regions of north Africa especially in the south of Algeria, Libya and Egypt. Libya reported drought events in January and February. The extreme north of Africa had normal winter precipitation totals. Exceptions occurred in the extreme north of Algeria and the center of Tunisia which are considered wetter than normal with a maximum percentage of 130% of the normal in the center of Tunisia.
Verification of the MedCOF-1 climate outlook for the 2013-14 winter season North Africa (RA I) [Temperature] The seasonal temperature over coasts of North Africa would be above normal with low probability. It was not possible to predict the winter season temperature in the south regions, i.e. the south of Algeria, the south of Morocco and most regions of Libya and Egypt due to the equal probabilities for below-, near-, or above-normal conditions. This indicates that the climate outlook for the winter season air temperature was not able to predict temperature anomalies registered for most north African regions.
Verification of the MedCOF-1 climate outlook for the 2013-14 winter season North Africa (RA I) [Precipitation] According to MedCOF-1, it was predicted that the winter season precipitation totals would be near or below average in the west of North Africa covering all of Morocco. This prediction turned out to be quite accurate. In the rest of the North African domain it was not possible to predict the summer season precipitation totals due to dry season masking
Results of MEDCOF -2
ANALYSIS AND VERIFICATION OF THE MEDCOF-2 CLIMATE OUTLOOK FOR THE 2013-14 Summer SEASON FOR THE MEDITERRANEAN REGION (MED) 2 3 1 (A) (B) Graphical presentation of the climate outlook for the 2013-14 summer season for the Mediterranean region (a) Temperature Outlook; (b) Precipitation Outlook;
Summer temperature outlook Although for most of the MedCOF domain the uncertainty for the temperature prediction is high, there is a weak tendency for the upper tercile in the Eastern and Central part of the domain and most of the Mediterranean Sea (regions 1 and 2). The Atlantic facade of the domain, most of the Balkan Peninsula and Turkey show no signal and climatology is therefore assigned for all three categories(region 3).
Summer Precipitation outlook Precipitation in the Central and Eastern of the MedCOF domain shows a slight preference for the above normal tercile (region 1). The Atlantic regions of the Iberian Peninsula and France and Northern Balkans show no signal different from climatology (region 2). The Southern part of the domain is masked as dry season and therefore no categories are assigned (region D). Note that it is necessary to express seasonal forecasts in terms of probability due to inherent uncertainty. Any further advice on the forecast signals, smaller scales, shorter-range updates and warnings will be available throughout the summer from the National Meteorological Services, along with details on the methodology and skill of long-range predictions.
Conclusion get on board (the bit has left the station) Any Questions Thank You