Version classiqueVersion mobile

The Mediterranean region under climate change

 | 
Jean-Paul Moatti
, 
Stéphane Thiébault

Part 3. Adaptation, resilience, conservation of resources and prevention of risk

Sub-chapter 3.4.4. Seasonal forecast of droughts and water resources

Lauriane Batté, Fatima Driouech et Constantin Ardilouze

Texte intégral

Introduction

1Forecasting the upcoming season is a fundamentally different scientific issue from forecasting the weather over the next few days. Due to the intrinsically chaotic nature of the atmosphere, the predictability of temperature, winds, precipitation or pressure systems is typically limited to 10 days. Seasonal forecasts therefore set out to provide an outlook on statistical averages of climatic variables at ranges from one month to a year, and rely on large-scale, more slowly-evolving components of the climate system such as the ocean or land surface. The best known source of climate variability on a seasonal time scale is the El Nino Southern Oscillation phenomenon (ENSO), which occurs in the Tropical Pacific Ocean but has remote impacts on climate variability around the globe. Several approaches are commonly used to produce seasonal forecasts, namely dynamical climate modelling using global coupled models (GCMs), statistical/empirical approaches, or a combination of both.

2Dynamical seasonal forecasts are run as ensembles, starting from slightly different initial conditions and/or including in-run perturbations, to take into account initial condition and model uncertainties. These ensembles provide information on the probabilities of occurrence of departures from a mean model climate of the variables of interest. An example of such a forecast, presented as a synthesis plot for tercile probabilities, is shown in fig. 1. The mean climate is estimated by re-forecasting past seasons. These re-forecasts also provide necessary insights into the quality of the forecasting system, and enable the post-processing and calibration of model outputs to extract useful information for end-users.

Figure 1
Synthesis plot for July to September 2016 seasonal mean precipitation tercile probabilities based on Météo-France system 5 seasonal forecast initialized in June; the colors refer to the most likely tercile and its probability; the areas in white are where no preferred tercile is found. © Météo-France/DCSC

3A number of international efforts have focused on improving dynamical climate forecasts at the seasonal range. In Europe, several initiatives (e.g. the European Commission funded projects DEMETER, ENSEMBLES and SPECS) have developed multi-model forecasting strategies and studied predictability at seasonal or longer time scales. Seasonal forecasting systems based on coupled dynamical atmosphere and ocean models exhibit skill – with respect to common benchmarks as persistence or a past climatology – mainly over the Tropics, whereas limited skill is found over mid-latitude regions, including most of the Mediterranean basin (e.g. Weisheimer et al. 2011, Doblas-Reyes et al. 2013).

4Although the skill of seasonal forecasts is limited, their uptake by a wide end-user community has been encouraged by the Global Framework for Climate Services (GFCS). Over the Mediterranean region, several prototypes of end-to-end climate predictions and services have been developed and evaluated as part of the European Commission FP7 project EUPORIAS. As the region is prone to droughts and water vulnerability, one of the areas of focus for the Mediterranean is water management issues. This sub-chapter provides an overview of seasonal forecasting activities with a focus on these applications, by summarizing results from recent research, highlighting operational activities at Météo-France, the Direction de la Météorologie Nationale (DMN) in Morocco and the Mediterranean Climate Outlook Forum (MedCOF), and discusses future directions.

Predictability of droughts and water resources

5A first essential step in forecasting droughts and water resources on the seasonal time scale is to properly assess the predictability of relevant key climate variables, such as near-surface temperature and precipitation. As previously stated, skill for these variables is limited over the area of interest. However, recent results suggest that ENSO (Manzanas et al. 2014, Shaman 2014) as well as the North Atlantic Oscillation could have some impact on the Mediterranean region climate at on the seasonal rangelevel, therefore opening perspectives for future improvement of dynamical seasonal forecasting systems, and providing the basis of skill for statistical or combined approaches (such as those developed by Guérémy et al. 2011) in impact forecasting.

6For specific areas of interest, provided that a sufficient amount of data is available for model training, statistical models based on linear regression or maximum covariance analysis often exhibit considerable skill over a re-forecast period. However, these methods are highly dependent on statistical relationships which, in a changing climate, could be less robust in the upcoming decades.

7Land-surface initialization and land-atmosphere coupling in dynamical systems could provide future improvements, particularly over regions around the Mediterranean Sea such as the Balkans (Ardilouze et al. 2016, Prodhomme et al. 2015). The land-surface variables are commonly initialized from model or reanalysis climatology due to the lack of timely gridded observations on the global scale for real-time forecasts. Inter-annual variability of soil conditions may provide a valuable source of predictability for both climate variables and impact indices over regions subject to drought.

8Due to the perception of low skill, the uptake of seasonal forecasts by end-users in Europe (and more generally over the Mediterranean region) has been limited up to now (Bruno-Soares and Dessai, 2016). Recent research efforts at Météo-France have focused on demonstrating the feasibility and added value of using dynamical seasonal forecast ensemble outputs as forcing for hydro-meteorological models to forecast river flows and soil wetness indices (see e.g. Céron et al. 2010; Singla et al. 2012; and the RIFF prototype in EUPORIAS, http://riff.euporias.eu/​).

Operational seasonal forecast activities

Météo-France

9Météo-France started real-time seasonal forecasting in the mid-90s and entered the EUROSIP consortium http://www.ecmwf.int/​en/​forecasts/​documentation-and-support/​long-range/​seasonal-forecast-documentation/​eurosip-user-guide/​multi-modelwith a coupled ocean-atmosphere system in 2005. As of 2016, Météo-France routinely provides dynamical seasonal forecasts each month as part of the EUROSIP consortium, and in a proof-of-concept phase of Copernicus Climate Change Services (C3S). These forecasts are based on the CNRM-CM GCM (Voldoire et al. 2013), using the ARPEGE-Climate v6 atmospheric component and the NEMO v3.2 ocean model.

10Briefings and synthesis maps can then be disseminated to a wide community through the WMO network, as Météo-France is a WMO Regional Climate Center node on long-range forecasting. These are based on information from the Météo-France system 5 but also other models contributing to the EUROSIP consortium, and balanced against the experience of past model performance. Monthly and daily model outputs can feed impact models for specific applications, including providing climatic indices.

Seasonal forecasts at DMN

11Seasonal forecasts have been developed at the Moroccan Meteorological Service (DMN) in order to provide decision makers with information that can help programming activities and works. Due to the importance of precipitation for agriculture and water resources and the different drought periods registered in the country (i.e. 1980-1984, 1990-1994), the need for information on the coming climate state in terms of seasonal precipitation has increased. Furthermore, in the context of climate change (more variability and climate extremes like drought and hot periods), seasonal forecasts are in increasing demand as they present an adaptation tool in the short term.

12During a period of about two decades, seasonal forecast developments and activities in Morocco have taken several steps forward, from a single deterministic forecast using uncoupled models to probabilistic forecasts issued from an operational chain including a coupled ocean-atmosphere model. Fig. 2 gives an illustration of the current chain. Each month, an ensemble of 27 potential climate evolutions for the following three months are developed using the French numerical climate model ARPEGE-Climate coupled with the Ocean Model NEMO3.2, atmospheric initial conditions issued from ECMWF (European Centre for Medium range Weather Forecast) and ocean initial conditions issued from MERCATOR-OCEAN (http://www.mercator-ocean.fr). This ensemble is used to generate probabilistic forecasts for tercile categories (above, below and near normal) comparatively to climatology for both precipitations and temperature. When considering meteorological drought, the relevance of precipitation is straightforward but temperatures are also important; a normal season with high temperatures can lead to a reduction in soil water and then a sort of drought (agricultural, hydrological). A dry season in terms of precipitation can have greater negative impacts when it combines high temperatures. Drought forecasts are also conducted by computing a drought index: the standardized precipitation index (SPI) of McKee (McKee et al. 1993, 1995). Fig. 3 shows an illustration of a past forecast for SPI.

Figure 2
Illustration of the operational seasonal forecast chain at DMN, composed of the ARPEGE atmospheric model, the NEMO ocean model and the TRIP river model, which exchange data using the OASIS coupler.

13Forecast maps for precipitation and temperature are produced and included into the national monthly bulletin along with the SPI bulletin, the outputs of the statistical forecasts also developed at the DMN and other forecasts issued from WMO Global Producing Centers. A forecast statement is given for precipitation and temperature taking into account the different outputs; a way that gives generally more robust forecasts than when using a single model. Seasonal forecasts are also developed each month for the North African Regional Climate Centre (http://rccnara1.marocmeteo.ma/​) as the DMN is the node responsible for this activity.

Figure 3
November 2014 to January 2015 seasonal SPI forecast issued in October 2014, based on the ARPEGE-Climat–NEMO coupled system at DMN. The SPI is based on the probability of precipitation for any time scale. The probability of observed precipitation is then transformed into an index. It is being used in research or operational mode in more than 70 countries. Many drought planners appreciate the SPI’s versatility. It is also used by a variety of research institutions, universities, and National Meteorological and Hydrological Services across the world as part of drought monitoring and early warning efforts. (Standardized Precipitation Index User Guide)

14Beyond these bulletins, seasonal forecast activities include multisector works and studies that aim to develop and enhance appropriate use for the seasonal forecast by stakeholders, in order to allow them to take full benefit of climate information at a seasonal time scale. As an example, we can cite the recent initiative designed to include forecast meteorological parameters into a crop model and elaborate crop yield predictions. Such kinds of forecasts can help anticipating appropriate actions in case of drought and also be used for optimizing stock importation.

MedCOF

15Following in the footsteps of previously established Climate Outlook Fora, and in order to promote interactions and collaborations across the Mediterranean basin, the Mediterranean Climate Outlook Forum (MedCOF, http://medcof.aemet.es) was started in 2013 and holds outlook meetings twice annually, as well as training workshops.

16This forum acts as a platform for potential stakeholders over the region and provides consensus forecasts twice a year based on participating National Hydrometeorological Services (including seasonal forecast providers) and research institutes.

Future directions

17Future research, in the context of international collaborations, is set to focus on several aspects: improving the understanding of local and remote (via teleconnections) sources of predictability over the Mediterranean region; further assessing the skill of current forecast systems in a user-relevant setting; extracting the relevant signal from noise in ensemble forecasts; improving targeted applications by developing cutting-edge downscaling methods. Progress in these areas will rely on the continued integration of the different actors in the climate services chain, from seasonal forecast providers to end-users.

Bibliographie

References

Ardilouze C., Batté L., Bunzel F., Decremer D., Déqué M., Doblas-Reyes F.J., Douville H., Fereday D., Guemas V., Maclachlan C., Müller W., Prodhomme C., 2016
Multi-model assessment of the impact of soil moisture initialization on mid-latitude summer predictability. Submitted to Climate Dynamics.

Bruno Soares M., Dessai S., 2016
Barriers and enablers to the use of seasonal climate forecasts amongst organisations in Europe. Climatic Change, 137: 89–103, doi : 10.1007/s10584-016-1671-8

Céron J.-P., Tanguy G., Franchistéguy L., Martin E., Regimbeau F., Vidal J.-P., 2010
Hydrological seasonal forecast over France: feasibility and prospects. Atmos. Sci. Let., 11: 78–82, doi: 10.1002/asl.256

Doblas-reyes F. J., García-Serrano J., Lienert F., Biescas A. P., Rodrigues L. R. L., 2013
Seasonal climate predictability and forecasting: status and prospects. WIREs Clim Change, 4: 245–268. doi: 10.1002/wcc.217

EUROSIP
http://www.ecmwf.int/en/forecasts/documentation-and-support/long-range/seasonal-forecast-documentation/eurosip-user-guide/multi-model

Guérémy, J.-F., Laanaia N., Céron J.-P., 2012
Seasonal forecast of French Mediterranean heavy precipitating events linked to weather regimes. Nat. Hazards Earth Syst. Sci., 12:2389–2398, doi: 10.5194/nhess-12-2389-2012

Manzanas R., Frías M. D., Cofiño A. S., Gutiérrez J. M., 2014
Validation of 40 year multimodel seasonal precipitation forecasts: The role of ENSO on the global skill, J. Geophys. Res. Atmos., 119:1708-1719, doi: 10.1002/2013JD020680.

Mckee T.B., Doesken N.J., Kleist, J., 1993
The relationship of drought frequency and duration to time scale. In: Proceedings of the Eighth Conference on Applied Climatology, Anaheim, California, 17–22 January 1993. Boston, American Meteorological Society, 179–184.

Mckee T.B., Doesken N.J., Kleist J., 1995
Drought monitoring with multiple timescales. In: Proceedings of the Ninth Conference on Applied Climatology, Dallas, Texas, 15–20 January 1995. Boston American Meteorological Society, 233–236

Prodhomme C., Doblas-Reyes F. J., Bellprat O., Dutra E., 2015
Impact of land-surface initialization on sub-seasonal to seasonal forecasts over Europe, Climate Dynamics, 1-17, doi: 10.1007/s00382-015-2879-4

Shaman J., 2014
The Seasonal Effects of ENSO on European Precipitation: Observational Analysis. Journal of Climate, 27: 6423-6438. doi: 10.1175/JCLI-D-14-00008.1

Singla S., Céron J.-P., Martin E., Regimbeau F., Déqué M., Habets F., Vidal J.-P., 2012
Predictability of soil moisture and river flows over France for the spring season, Hydrol. Earth Syst. Sci., 16:201-216, doi: 10.5194/hess-16-201-2012

Voldoire A., Sanchez-gomez E., Salas Y Mélia D., Decharme B., Cassou C., Sénési S., Valcke S., Beau I., Alias A., Chevallier M., Déqué M., Deshayes J., Douville H., Fernandez E., Madec G., Maisonnave E., Moine M.-P., Planton S., Saint-Martin D., Szopa S., Tyteca S., Alkama R., Belamari S., Braun A., Coquart L., Chauvin F., 2013
The CNRM-CM5.1 global climate model: description and basic evaluation. Climate Dynamics, Special Issue: IPSL-CM5 and CNRM-CM5, 40(9-10): 2091-2121, doi: 10.1007/s00382-011-1259-y

Weisheimer A., Palmer T.N., Doblas-Reyes F. J., 2011
Assessment of representations of model uncertainty in monthly and seasonal forecast ensembles, Geophys. Res. Lett., 38: L16703, doi: 10.1029/2011GL048123.

Table des illustrations

Légende Figure 1Synthesis plot for July to September 2016 seasonal mean precipitation tercile probabilities based on Météo-France system 5 seasonal forecast initialized in June; the colors refer to the most likely tercile and its probability; the areas in white are where no preferred tercile is found. © Météo-France/DCSC
URL http://books.openedition.org/irdeditions/docannexe/image/23973/img-1.jpg
Fichier image/jpeg, 367k
Légende Figure 2Illustration of the operational seasonal forecast chain at DMN, composed of the ARPEGE atmospheric model, the NEMO ocean model and the TRIP river model, which exchange data using the OASIS coupler.
URL http://books.openedition.org/irdeditions/docannexe/image/23973/img-2.jpg
Fichier image/jpeg, 204k
Légende Figure 3November 2014 to January 2015 seasonal SPI forecast issued in October 2014, based on the ARPEGE-Climat–NEMO coupled system at DMN. The SPI is based on the probability of precipitation for any time scale. The probability of observed precipitation is then transformed into an index. It is being used in research or operational mode in more than 70 countries. Many drought planners appreciate the SPI’s versatility. It is also used by a variety of research institutions, universities, and National Meteorological and Hydrological Services across the world as part of drought monitoring and early warning efforts. (Standardized Precipitation Index User Guide)
URL http://books.openedition.org/irdeditions/docannexe/image/23973/img-3.jpg
Fichier image/jpeg, 168k

Auteurs

Météo-France, France
Climatologist, Centre National de Recherches Météorologiques, Météo-France and CNRS, France
lauriane.batte@meteo.fr

Météo-France, France
Climatologist, Centre National de Recherches Météorologiques, Météo-France and CNRS, France
constantin.ardilouze@meteo.fr

Le texte et les autres éléments (illustrations, fichiers annexes importés) sont sous Licence OpenEdition Books, sauf mention contraire.

Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search