21

CHAPTER 2. WAYS FORWARD FOR CLIMATOLOGY

2.1 Although mostly arid or semi-arid, the Arab region encompasses a diversity of climates,

including temperate zones in the northern and higher elevations of the Maghreb and Mashreq,

tropical ocean climates in the Comoros, and varied coastal climates fronting the Mediterranean, Red

and Arabian Seas, the Gulf of Arabia and the Indian Ocean (Figures 2.3 and 2.4). Climate change is

expected to have different impacts in these different climate zones.

2.2 To date, the Arab region has not been addressed as a discrete region in climate change

research assessments, such as in the IPCC reports. Typically, information must be inferred from

analyses carried out in other regions. Recent literature from regional studies confirms the broad

conclusions of the IPCC Fourth Assessment Report (AR4) regarding increasing temperatures and

mostly reduced precipitation, but sometimes differs quite substantially regarding the details. This

problem is particularly severe for the Arabian Peninsula, where models disagree about whether

there will be more or less precipitation. The future climate in this region will depend in large part on

the position of the Inter-Tropical Convergence Zone (ITCZ) (see Box 2.1); all models project that it

will move further northward, but disagree on the precise amount and location of that shift. Much of

the region also falls in the transitional zone between areas with projected decreases in rainfall and

those with projected increases. Since models differ as to the precise location of that transition, it is

often difficult to project the exact magnitude of precipitation changes or even whether there will be

more or less rainfall.

2.3 While several efforts are underway to improve on the availability of downscaled climate

change information for the Arab region based on existing global and regional circulation models

and scenarios, the international climate change modeling community is moving ahead with new

global modeling approaches and climate change scenarios. This effort will include a new set of

improved and consistent modeling results that will be available for all regions, with the first results

appearing within the next one to two years.

2.4 Climate projections depend on having a good set of observational data to determine

current trends and to translate outputs from global circulation models to regional scales

(downscaling). Unfortunately, many observational and modeling gaps exist in the study of climate

change in the Arab region. In these circumstances, it is important not to fall into the trap of

extrapolating from global information and applying it to the region, or of jumping from

interpretations of long-term projections to statements about the near term. This chapter brings

together the various pieces of work that have been done for the Arab region, which provide an

insight into the region‘s future climate. The Arab region will remain predominately arid with some

areas becoming even drier and hotter, but rainfall patterns will change and the increase in flooding

events already being observed is likely to continue in the future.

22

Box ‎2.1 Some Basic Definitions22

Climate: Climate, in a narrow sense, is usually defined as the average of weather, or more rigorously, as the

statistical description in terms of the mean and variability of relevant quantities over a period of time ranging

from months to thousands or millions of years. The classical period for averaging these variables is 30 years,

as defined by the World Meteorological Organization. The relevant variables are most often surface

parameters, such as temperature, precipitation and wind. Climate, in a wider sense, is the state, including a

statistical description, of the climate system.

Climate change: Climate change refers to a long-term change in the state of the climate that can be

identified (e.g., by using statistical tests) by changes in the mean and/or the variability of its properties, and

that persists for an extended period, typically decades or longer. Climate change may be due to natural

internal processes or external forcings, or to persistent anthropogenic changes in the composition of the

atmosphere or in land use. Note that Article 1 of the UN Framework Convention on Climate Change

(UNFCCC) defines climate change as: ―a change of climate which is attributed directly or indirectly to

human activity that alters the composition of the global atmosphere and which is in addition to natural

climate variability observed over comparable time periods.‖ The UNFCCC thus makes a distinction between

climate change attributable to human activities altering the atmospheric composition, and climate variability

attributable to natural causes.

Climate scenario: A plausible and often simplified representation of the future climate, based on an

internally consistent set of climatological relationships that has been constructed for explicit use in

investigating the potential consequences of anthropogenic climate change, often serving as input to impact

models. Climate projections often serve as the raw material for constructing climate scenarios, but climate

scenarios usually require additional information, such as information about the observed current climate. A

climate change scenario is the difference between a climate scenario and the current climate.

Climate variability: Climate variability refers to variations in the mean state and other statistics (such as

standard deviations, the occurrence of extremes, etc.) of the climate on all spatial and temporal scales beyond

that of individual weather events. Variability may be due to natural internal processes within the climate

system (internal variability), or to variations in natural or anthropogenic external forcing (external

variability).

North Atlantic Oscillation (NAO): The North Atlantic Oscillation consists of opposing variations of

barometric pressure between areas near Iceland and near the Azores. It affects the strength and position of

the main westerly winds across the Atlantic into Europe and the Mediterranean. When the pressure

difference is high (NAO+), the westerlies are stronger and track more to the north leading to cool summers

and mild wet winters in Europe, but drier conditions in the Mediterranean. In the opposite phase, the

westerlies and the storms they bring track further south leading to cold winters in Europe, but more storms in

the Mediterranean and more rain in North Africa.

Inter-Tropical Convergence Zone (ITCZ) The Inter-Tropical Convergence Zone is an equatorial zonal belt

of low pressure near the equator where the northeast trade winds meet the southeast trade winds. As these

winds converge, moist air is forced upward, resulting in a band of heavy precipitation. This band moves

seasonally. In Africa, it reaches its northernmost position in summer and also interacts with the Indian

Monsoon, bringing rains to the southern part of the Arab Region (Southern Sahel), but the northward extent

varies from year to year making both conventional weather forecasting and climate modeling difficult.

Storm surge: Storm surge is a rise of the sea water above the normal level along a shore associated with a

low pressure weather system, typically tropical cyclones and strong extratropical cyclones. It is the result of

both the low pressure at the center of the storm raising the ocean surface, as well as the wind pushing the

22Largely these definitions are based on those provided within AR4.

23

water in the direction the storm is moving. The storm surge is responsible for most loss of life in tropical

cyclones worldwide.

Flash floods: Flash floods usually refer to rapid flooding that happens very suddenly, usually without

advance warning. They are different from regular floods, as they often last less than six hours. Flash floods,

with intense rainfall normally occur in association with the passage of a storm or tropical cyclone, especially

when the rain falls too quickly on saturated soil or dry soil that has poor absorption capacity. Flash floods

may also refer to a flooding situation, when barriers holding back water fail, such as the collapse of a natural

ice or debris dam, or a man-made dam.

Tropical cyclone: A storm system characterized by a large low-pressure center and numerous thunderstorms

that produce strong winds and heavy rain. Tropical cyclones strengthen when water evaporated from the

ocean is released as the saturated air rises, resulting in condensation of water vapor contained in the moist

air. They are fuelled by a different heat mechanism than other cyclonic windstorms such as nor'easters and

European windstorms. The characteristic that separates tropical cyclones from other cyclonic systems is that

at any height in the atmosphere, the center of a tropical cyclone will be warmer than its surrounds; a

phenomenon called ―warm core‖ storm systems. The term ―tropical‖ refers to both the geographic origin of

these systems, which form almost exclusively in tropical regions of the globe, and their formation in

maritime tropical air masses.

Heat wave: A prolonged period of excessively hot weather, which may be accompanied by high humidity.

There is no universal definition of a heat wave; the term is relative to the usual weather in the area.

Temperatures that people from a hotter climate consider normal can be termed a heat wave in a cooler area if

they are outside the normal climate pattern for that area. The term is applied both to routine weather

variations and to extraordinary spells of heat, which may occur only once a century. Severe heat waves have

caused catastrophic crop failures, thousands of deaths from hyperthermia and other severe damages.

DESPITE SPARSE OBSERVATIONAL DATA IT IS CLEAR THAT MOST OF THE ARAB

REGION IS BECOMING HOTTER AND DRIER

2.5 The climate in Arab countries ranges from Mediterranean, with warm and dry summers

and some wintertime precipitation, through subtropical zones, with variable amounts of summer

monsoon rains, to deserts with virtually no rain. During winter, variability in the North Atlantic

Oscillation (NAO) (see Box 2.1) influences the position of storm tracks, and annual variations in

precipitation in western and central North Africa (the Maghreb) are largely governed by this NAO

effect. The eastern part of the region (most of the Mashreq, Gulf and Center Regions), where it rains

mainly during the summer, is influenced by the Indian monsoon system, which is largely controlled

by the position of the ITCZ (see Box 2.1).

There is a scarcity of meteorological surface observation

2.6 With a few exceptions, the availability of climate-station data to establish baseline

climate across the Arab region is very limited compared to most other parts of the world (Figures

2.1 and 2.2). This hampers detection of climate change, as well as the interpretation of projected

changes, since changes must be compared to a verifiable current climate. The rescue of existing but

undigitized climate data and the establishment of well-chosen, permanent, high-quality

observational sites will be necessary to establish more rigorous models in the future.

2.7 The distribution of quality-controlled, long-term observational sites within the Arab

region is uneven. For historical reasons, a reasonable number of stations exist along the Nile and the

coast of the Mediterranean Sea, but in the desert regions, coverage is very sparse. Few data are

24

available for countries, such as Libya, Gaza and the West Bank, Saudi Arabia and Somalia. For

Algeria, Tunisia, Egypt, Iraq, Sudan and the Comoros, data is probably available, but not readily

accessible. Conflict in parts of the region disrupts both the collection and sharing of data. However,

it is likely that in many areas, additional data are being gathered by various agencies, but not

entered into more widely available meteorological databases. For example, research in Yemen

quickly identified a large number of additional meteorological observation sites held only by the

Ministry of Irrigation and Agriculture (Figure 2.2b, Wilby pers. comm., 2011).

Box ‎2.2 Spatial Distribution of Monthly Stations with a Least 10 Years of Data

Note: Available in the GPCC database (global number of stations in June: 64,471).

Source: Schneider et al. 2008.

Figure ‎2.1 Spatial Distribution of Monthly in-situ Stations with a Climatological Temperature Normal

Based on at Least 10 Years of Data. B. Additional Sites Found for Yemen After Extensive Discussions

with Local Authorities (Wilby, pers. comm. 2011)

Note: Available in the KNMI Climate Explorer Database. Source: http://climexp.knmi.nl

25

Figure ‎2.2 Yemen Observation Meteorological Sites

While aridity predominates, there is a wide range of climates within the Arab Region

2.8 As in most of the world, the critical climate variable for human settlement patterns in the

Arab region is rainfall. All desert regions (Mauritania, Algeria, Libya, Egypt, northern Sudan and

almost the entire Arabian Peninsula) have annual precipitation totals below 200 millimeters (Figure

2.3). The central parts of the Sahara have less than 50 millimeters. From southwestern Algeria to

western Egypt, no precipitation at all was observed during the twentieth century. Mediterranean

zones in the north and subtropical regions in the south typically get about 500 millimeters of rainfall

per year, whereas the annual rainfall in southern Sudan and the Comoros is more than 1,000

millimeters. As in other arid regions, precipitation varies greatly, with the coefficient of variation

(standard deviation of annual precipitation divided by the average annual precipitation) exceeding

100 percent in the deserts (see Appendix B.2). This means that there can be years with little or no

rainfall and years in which the rainfall greatly exceeds the average. It also means that it is difficult

to identify trends in the amount of rainfall.

26

Figure ‎2.3 Köppen-Geiger Climate Classification for the Arab World (1976–2000)

Source: Westphal 2010.

Figure ‎2.4 Mean Annual Precipitation in mm (left) and Inter-Annual Variability of Precipitation

Variability (coefficient of variation in %, right)

Note: Areas with mean annual precipitation of 0 are shown in white.

Source: Westphal 2010.

2.9 Mean annual temperature is between 20°C and 25°C in the desert regions, up to 28°C on

the Arabian Peninsula, between 15°C and 20°C in the Mediterranean and subtropical zones, and

close to 25°C in the tropical regions. In tropical zones, the annual amplitude is very small, whereas

variability increases further north. In the mountainous regions of Iraq and Lebanon, it is sometimes

even cold enough for occasional snowfall (see Annex 2).

2.10 Precipitation in most of the Arab world depends partly on the state of the NAO (see Box

2.1), which is the dominant source of climate variability in the Atlantic/European/Mediterranean

27

region. Its influence can be seen in weather patterns, streamflows and subsequent ecological and

agricultural effects. Cullen et al. (2002) identified two components of Middle Eastern streamflow

variability. The first principal component reflects rainfall-driven runoff and explains 80 percent of

the variability in river flows from December to March. This principal component is correlated, on

interannual to interdecadal timescales, to the NAO phase (in a positive NAO phase, the climate in

the Middle East is cooler and drier than average and vice versa when NAO is in a negative phase).

The second principal component (the so-called Khamsin) is related to spring snow melt in the

mountains and explains more than half of the streamflow variability from April to June. A

prevailing positive NAO phase, as in the 1990s and 2000s, can therefore result in drought

conditions in the region, including in the Euphrates-Tigris and Jordan Rivers Basins.

Arab countries have warmed and most have become drier

2.11 Despite some local deviations, the available evidence clearly indicates a warming trend

within the last 100 years or more. In the GHCN2 dataset (Schneider et al. 2008), there are 243

stations in the Arab region that have measured some temperatures since 1901. Most of them,

however, cover only short periods, and only 69 have sequences of more than 50 years of data,

mostly from Algeria, Egypt and Sudan. These 69 stations show an overall temperature increase of

0.2 to 0.3°C per decade, with slightly lower values along the coast of the Mediterranean Sea.

However, stations, which are often in larger cities, which for unknown reasons, showed a decrease

in temperature during the last century, notably in Bengasi (Libya), Damascus (Syria), and Dakhla

(Egypt) (Figure 2.5). The few stations with longer records suggest that precipitation has decreased

over the past century.

Figure ‎2.5 Temperature Trends for all Stations Available in the KNMI Climate Explorer Database

with More Than 120 Months of Data

Note: Red colour represents a positive trend, blue negative. The size of the circles indicates the size of the trend (see

legend). Significance level of a trend is indicated by a yellow dot (1-2 sigma) or square (above 2 sigma).

Source: KNMI.

28

A case study from the Maghreb: Morocco

2.12 Wilby (2007, 2008, 2010) has compiled climate change observations and projections on a

country by country basis for Morocco, Syria, Lebanon, Jordan and Yemen. In Morocco, the Atlas

Mountains divide the country into three climate zones (Figure 2.6): the Mediterranean zone in the

northeast; the Atlantic zone in the northwest; and the arid regions south of the Atlas Mountains. In

Morocco, precipitation is clearly connected to the NAO phase, with a positive NAO phase

associated with less than normal precipitation (Knippertz 2003, 2004).

Figure ‎2.6 Winter Rainfall Regions for the Three Moroccan Precipitation Regimes: (I) Atlantic (ATL);

(II) Mediterranean (MED); and (III) South of the Atlas (SOA)

2.13 Annual and seasonal precipitation varies greatly in Morocco, with coefficients of

variation ranging from 25 percent in coastal areas to more than 100 percent in the Sahara (Knippertz

et al., 2003) (see also Figure 2.4). Nevertheless, precipitation has declined since the 1960s and, in

spring, rainfall has declined by 40 percent. The maximum length of dry spells has also increased by

more than two weeks over the same period (Wilby, 2007), and the Atlas Mountains have

experienced less rainfall (Chaponniere and Smakhtin, 2006).

A case study within the Mashreq

2.14 Recently, Wilby (2010) prepared climate change observations and projections for the

Mashreq countries of Jordan, Lebanon and Syria in the eastern Mediterranean. Here the data

availability is less favorable than in Morocco. The World Meteorological Organization (WMO) lists

11 stations for Jordan, 7 for Lebanon and 24 for Syria. Önol and Semazzi (2009) have gathered data

from a number of additional stations. However, regions without any data also exist, particularly in

northern Syria, eastern Jordan and western Iraq. The present-day climate in this region is quite

different from countries further west, with relatively cool and wet winters, and hot and dry

summers, generally, without any precipitation. Temperature and precipitation are strongly affected

by altitude and the distance from the sea (see e.g., Figure 2.7). In the Lebanon Mountains, up to

29

1400 millimeters/year of precipitation are observed (in winter often as snow), but the deserts of

southeast Syria and southern Jordan receive less than 100 millimeters/year. Temperatures above

50°C have been observed near the Dead Sea. To explore the time-space characteristics of

precipitation, daily rescaled data from the TRMM satellite observations (Kummerowet al., 2000;

Simpson et al., 1988) have been used.

2.15 A temperature rise since the 1970s has been observed in all three countries. Mahwed

(2008) considers meteorological records at 26 stations in Syria; Freiwan and Kadioglu (2008a,

2008b) examined monthly precipitation data from Jordan; and Shahin (2007) has looked at several

stations throughout the Middle East. The greatest warming has occurred for summer minimum

temperatures, which have risen at a rate of 0.4°C/decade. Consequently, a decrease in the diurnal

temperature range has been observed, which is consistent with earlier studies (Nasrallah and

Balling, 1993; Zhang et al., 2005). There is no clear indication whether precipitation has changed in

recent decades, but estimates from TRMM data (Figure 2.7) suggest a slight decrease in winter and

spring, probably related to shifts in cyclone tracks. However, the trends are small compared to the

interannual variability (Wilby, 2010). Figure 2.7 illustrates that a regional climate model (RCM) is

able to faithfully simulate precipitation climatology to a large extent.

30

Figure ‎2.7 Spatial Distribution of Rainfall; observed in right column and simulated by NNRP2-

RegCM (left) all Averaged Over the Period 1998–2009

Note: Rainfall (millimeters) obtained from the TRMM data averaged over 1998–2009 is shown in the right panels (b, d

and f).

Source: Almazroui (2011).

A case study from the Arabian Peninsula

2.16 Wilby (2008) also compiled an assessment of climate and climate change for Yemen,

despite the lack of reliable data. Obvious errors (Wilby, 2008) and missing data make it particularly

31

difficult to apply statistical downscaling procedures. In addition, a lack of obvious trends in rainfall

averages or extremes may be, in part, due to bad data. The only station in Yemen with a reasonably

long and reliable time series for precipitation is Aden, for which monthly means exist since 1880.

However, this station shows no significant trends in annual precipitation.

More extreme events are being observed

2.17 From a climate change point of view, changes in extremes are more interesting than

changes in average values. Unfortunately, researchers do not always use the same definitions of

extremes (Bonsal et al., 2001), making it difficult to make global or even regional comparisons.

Frich et al. (2002) tried to standardize definitions of extreme indices, but they focused on areas with

ample data, which meant they were difficult to apply in large parts of the world, including the Arab

region. To address these issues and to provide better input to the IPCC AR4, a WMO expert team

defined a number of indices recommended for use in all analyses of extremes. Numerous regional

workshops were held, one of them covering the Middle East (Sensoy et al., 2005), to collect and

analyze (including quality control and homogeneity testing) data for a region from Turkey to Iran

and from Georgia to the southern tip of the Arabian Peninsula, thus covering most of the eastern

part of the region of interest in this assessment. Fifty-two stations from 15 countries that passed all

quality checks were chosen (mapped in Figure 2.8 and listed in Table 2.1).

Figure ‎2.8 Location of Mideast Stations for which Extreme Climate Trends Have Been Calculated

Source: Zhang et al. (2005).

32

Figure ‎2.9 Percentage of Days with Maximum Temperature Above the 90 Percentile

Note: Time series anomalies, averaged over the stations shown in Figure 2.10, of the percentage of days with maximum

temperature above the 90th

percentile with respect to the period 1971-2000 and linear trend for the period 1950-2003.

Figure ‎2.10 Percentage of Days with Maximum Temperature Below the 10 Percentile

Source: Zhang et al. (2005).

2.18 Figures 2.9 and 2.10 show time series anomalies averaged over the region mentioned

above for the hottest days (i.e., days with a maximum temperature higher than 90% of those

observed over the baseline period 1971-2000) and the coolest days (i.e., maximum temperature is in

the lowest 10th percentile). A decrease in the number of the coolest days and a more recent abrupt

increase in the number of the hottest days are visible. Minimum daily temperatures show even

larger increases (not shown), which corroborates the finding that the diurnal temperature range is

decreasing. As an example, Figure 2.11 shows the geographical distribution of hottest day and

warmest night temperatures. Trends have become more coherent and more significant in recent

periods. Due to the large year-to-year temperature variations in desert regions, significance is

largest in the north, particularly during summer and autumn. For precipitation, the results are much

more variable, but if there is any trend at all, it is a small but insignificant increase.

33

Figure ‎2.11 Trend in the Annual Hottest Days (TX90p top) and Warmest Nights (TN90p bottom)

Series for the Periods 1950-2003 (left panels) and 1970-2003 (right panels)

Note: Upward triangles represent increasing trends, downward triangles, decreasing trends. Different triangle sizes

indicate different magnitudes of trends, expressed in percent per decade. Solid triangles represent trends significant at

the 5% level. Source: Zhang et al. (2005).

IPCC AR4 PROJECTS FURTHER WARMING AND ARIDITY

The global overview

2.19 For the IPCC AR4, a broad range of climate models were coordinated to perform a large

number of simulations for the known historical forcings (anthropogenic and natural) and various

future emission scenarios in order to assess climate change projections. Taken together with

information from observations, these coordinated simulations (referred to as a multi-model data set,

or MMD, in the IPCC AR4 and hereafter) provide a quantitative basis for assessing many aspects of

future climate change. All models assessed in the IPCC AR4 project increases in global mean

surface air temperature (SAT) continuing throughout the 21st century and driven by increases in

anthropogenic greenhouse gas concentrations, with warming proportional to the associated radiative

forcing. The best estimated projections indicate that decadal average warming by 2030 is insensitive

to the choice among the three non-mitigated scenarios from the IPCC Special Report on Emission

34

Scenarios (SRES), i.e., the B1, A1B and A2 (Figure 2.12), and is very likely to exceed the natural

variability observed during the 20th century (about 0.2°C per decade). Even if atmospheric

concentrations were held fixed at year 2000 levels, further warming of about 0.1°C per decade

would still be expected over the next two decades due to the fact that oceans respond slowly and

thermal expansion would still continue after stabilization. By the end of the 21st century (2090-

2099), projected global average SAT warming relative to 1980-1999, under the SRES emission

scenarios, will range from a best estimate of 1.8°C (likely range 1.1°C to 2.9°C) for the low

scenario (B1) to 4.0°C (likely range 2.4°C to 6.4°C) for the high scenario (A1F) to a best estimate

of 2.8°C (likely range 1.7°C to 4.4°C) for the moderate scenario (A1B) (IPCC AR4, 2007).

Figure ‎2.12 Projected Surface Temperature Changes for the Early and Late 21st Century Relative to

the Period 1980–1999

Note: The two panels show the AOGCM multi-model average projections for the B1 (top), A1B (middle) and A2

(bottom) SRES scenarios averaged over the decades 2020-2029 (left) and 2090-2099 (right). The dependency on the

scenario is insignificant for the near future, but will become increasingly important towards the end of the century.

Source: Meehl et al. (2007).

2.20 It is very likely that warming is greatest over land and at most of the high northern

latitudes, and least over the Southern Ocean and parts of the North Atlantic Ocean. Accompanying

projections using multi-model ensembles show increases of global average temperatures, mean

water vapor, evaporation and precipitation over the 21st

century. There is less certainty in the

regional patterns of precipitation changes compared with the high confidence in projected warming

patterns. In general, warming increases spatial variability of precipitation, with increases at higher

latitudes and in parts of the tropics and reductions in the subtropics.

AR4 projections for the Arab region

2.21 The Arab region lies within three neighboring sub-regions employed by the IPCC AR4:

South Europe and the Mediterranean (SEM); the Sahara (SAH); and East Africa (EAF). For the

purpose of this review, the African domains provide the most complete coverage of the areas of

Figure SPM.6

35

interest. Projected changes in climate are thus taken from global climate model (GCM) results for

the SAH and EAF sub-regions and the African domain (Figures 2.13 and 2.14).

Both the SAH and EAF regions are expected to warm by between 3°C and 4°C by the late 21st

Century under the SRES A1B emission scenario, which is roughly 1.5 times the global mean

response. Warming is evident in all seasons, with the greatest increase in summer (Figures 2.13

and 2.14, upper panel) (Christensen et al., 2007).

Annual rainfall is expected to decrease in much of Mediterranean Africa and the northern

Sahara, but increase in East Africa and the southern half of the Arabian Peninsula. A 20 percent

drying in the annual mean is typical along the African Mediterranean coast in the A1B scenario

by the late 21st century in nearly every MMD model, with drying extended into the northern

Sahara and down the west coast. The annual number of precipitation days is very likely to

decrease, and the risk of summer drought is likely to increase in the Mediterranean basin.

However, mainly due to a spread in the projected northward displacement of the ITCZ, there is

no consensus amongst the 21 MMD-GCMs about the projected precipitation changes over most

of the Arabian Peninsula and over the Sahel region. (Figure 2.16, middle and lower panels;

Christensen et al., 2007). Since these regions are mostly very dry, the implication is that little

change is expected, but a substantial increase in precipitation during the summer is possible if

the ITCZ moves further northwards as some models suggest it will.

Figure ‎2.13 Temperature Anomalies with Respect to the Period 1901 to 1950

Note: Temperature anomalies with respect to 1901 to 1950 for two ‗African‘ land regions (SAH, left, and EAF, right)

for 1906 to 2005 (black line) and as simulated (red envelope) by the IPCC AR4 models incorporating known forcings;

and as projected for 2001 to 2100 for the A1B scenario (orange envelope). The bars at the end of the orange envelope

represent the range of projected changes for 2091 to 2100 for the B1 scenario (blue), the A1B scenario (orange) and the

A2 scenario (red). The black line is dashed where observations are present for less than 50% of the area in the decade

concerned.

Source: Christensen et al. (2007).

36

Figure ‎2.14 Temperature and Precipitation Changes over Africa (and the Arabian Peninsula) based on

21 IPCC AR4 Models under the A1B IPCC Scenario

Note: Top row: Annual mean, NH winter (DJF), NH summer (JJA) temperature (in °C) changes between 1980-1999

and 2080-2099 averaged over 21 models. Middle row: same as top, but for relative change (in percentage) in

precipitation. Bottom row: number of models out of 21 that project increases in precipitation. Source: Christensen et al (2007).

2.22 In a climate change study undertaken in the Arab world, Westphal (2010) combined all

the MMD-GCM outputs of the A1B and B2 scenario projections for temperature, precipitation and

other relevant extreme indices for the Arab region. It is evident that, while all models agree on

strong warming, they agree on precipitation for only some parts of the region (cf. Figure 2.15). For

most countries on the Arabian Peninsula, as well as northern Sudan, southern Egypt, southern

Libyan and southern Algeria, consensus does not even exist regarding the sign of precipitation

change (white areas in Figure 2.16, lower panel). The wide disagreement is to be expected as

current precipitation is low and very variable and the models suggest that this will continue, with

sometimes a little less or a little more precipitation, but still very dry.

2.23 Even in regions where models generally agree on the sign of the change, there may still

be a large spread regarding the magnitude of the change. Figure 2.16 illustrates this by showing the

annual mean temperature change from 1980-1999 to 2080-2999 under the A1B scenario in Africa in

21 MMD-models used in the IPCC AR4. Although all models agree in a projected positive

temperature change, responses generated by individual models differ, ranging from rather weak

warming between 1.5°C and 3°C to very strong warming between 4°C and 7°C.

37

Figure ‎2.15 Zoom on the Arab Countries of Projected Climate Change. Compare with Figure 2.14

Figure ‎2.16 Zoom on the Arab Countries of Projected Climate Change. Compare with Figure 2.14

Note: Upper panel: mean annual temperature change (2080-2099 vs. 1980-1999) based on averaging the 24 IPCC AR4

GCMs. Lower panel, same as upper, but for precipitation based on averaging 23 GCMs. White areas indicate where

fewer than 2/3 of the models agree on the sign of the change. Compare with Figure 2.16

Source: Westphal (2010).

38

Figure ‎2.17 The Annual Mean Temperature Response in Africa in 21 IPCC AR4 Models

Note: Shown is the temperature change from the years 1980-1999 to 2080-2099 under the A1B scenario, averaging over

all available realizations for each model. The change averaged over all models is shown in the lower right hand corner.

Note the range of projected warming, e.g., models PCM and MIROC3.2

Source: Christensen et al. (2007).

39

2.24 Nevertheless, except for small areas in the south of the region, precipitation is likely to

decrease or change very little, and when increasing temperatures are taken into account, the region

will remain largely arid and even drier in some parts. Note that a large fraction of the area has been

masked by white as the models disagree on the sign of the change (see above for the comments to

Figure 2.14).

2.25 The large spread in model responses demonstrates the regional impact of the uncertainties

in climate projections. It is commonly recommended that multi-model ensembles, formed by GCMs

that have been driven by the same forcing scenarios, be used to generate quantitative measures of

uncertainty, particularly probabilistic information at a regional scale (Meehl et al., 2007;

Christensen et al., 2007). However, one must keep in mind that the regional probabilities generated

using these ensembles will not represent the full spread of possible regional changes, since even

multi-model ensembles explore only a limited amount of uncertainty (Meehl et al., 2007).

Projected changes in climate over the next few decades do not depend on the scenario chosen

2.26 It is important to realize that most of the uncertainty in climate projections for the end of

the century arises from the particular emission scenario (emission pathway) selected. However, in

the near term (until~2050), the scenario choice is not very important. This is clearly depicted in

Figure 2.12, which shows the projected mean surface temperature from an ensemble of climate

models available for the AR4. The left column represents the projected model mean change for the

near term, while the right column shows the projected change at the end of century, both for three

quite different emission scenarios. This figure shows how little influence the emission scenario has

on the projection in the coming decades, while the scenario is the dominant factor towards the latter

part of the century.

A NEW ROUND OF IMPROVED PROJECTIONS IS COMING

2.27 This chapter has focused on observation of recent trends in climate and on the modeling

of future climates produced for the IPCC AR4. These models are based on scenarios developed

over a decade ago and the compendium of models used in IPCC AR4 do not reflect some of the

advances in modeling made over the past decade. A new, coordinated modeling effort is underway.

The models are in advance of most of those in the existing compendium and all will be run at higher

resolutions, equivalent to regional climate models (RCMs), out to at least 2030s. There will be a

more focused effort on developing consistent downscaled outputs for all terrestrial regions of the

Earth and a major effort to improve the projection of changes in extreme events. The first results of

this effort will begin to appear within the next one to two years and can be expected to considerably

improve out understanding of future climates. This section of the paper describes in more detail

this new modeling effort.

Toward AR5 and beyond – a new framework for modeling climate change

A new generation of climate models

2.28 Since the IPCC AR4 was published, increasing efforts by the climate modeling

community have continued to: address outstanding scientific questions that arose as part of the

assessment process; improve understanding of the climate system; and provide estimates of future

40

climate change. A new set of coordinated climate model experiments, known as phase five of the

Coupled Model Intercomparison Project (CMIP5) (Taylor et al., 2011), has become a high priority

on the research agendas of most major climate modeling centers around the world. The results from

this new set of simulations, which will appear over the next few years, are expected to provide

valuable information and knowledge of particular relevance to the next international assessment of

climate science, e.g., the IPCC AR5.

2.29 Compared to the previous generation of models that contributed to the IPCC AR4, the

climate models participating in CMIP5 have been greatly improved through the adoption of new

findings about parameterizations of sub-grid scale physical processes, inclusion or further

development of aerosol schemes, carbon cycle models, variable vegetation cover, etc. In particular,

the new models more explicitly couple atmospheric and terrestrial carbon in a global carbon cycle

component (subsequently referred to as earth system models or ESMs). These coupled

carbon/climate model simulations provide a way of diagnosing the role of carbon-climate feedback

and quantifying the allowable emissions.

Figure ‎2.18 Representative Concentration Pathways (RCPs)

Note: (a) Changes in radiative forcing relative to pre-industrial conditions. Bold colored lines show the four RCPs; thin

lines show individual scenarios from approximately 30 candidate RCP scenarios that provide information on all key

factors affecting radiative forcing from the larger set analyzed by IPCC Working Group III during development of the

Fourth Assessment Report. (b) Energy and industry CO2 emissions for the RCP candidates. The range of emissions in

the post-SRES literature is presented for the maximum and minimum (thick dashed curve) and 10th to 90th percentile

(shaded area). The blue shaded area corresponds to mitigation scenarios; grey shaded area corresponds to reference

scenarios; the pink area represents the overlap between reference and mitigation scenarios.

Source: Mose et al., (2010).

A new way of making climate change scenarios

2.30 A set of new emission scenarios has been adopted in the CMIP5 framework (Moss et al.,

2008, 2010). Unlike the so-called SRES scenarios used during the past two decades, which explore

only emission pathways in the absence of climate policy, the new scenarios are defined by the

radiative forcing levels (i.e., the climate signal) of Representative Concentration Pathways (RCPs)

that are compatible with socioeconomic development, including adaptation and mitigation (van

41

Vuuren et al., 2010). Four RCPs are selected for CMIP5 experiments: one non-mitigated (RCP8.5),

and three taking into account various levels of mitigation (RCP6.0, RCP4.5 and RCP2.6), with

labels according to the approximate target radiative forcing at year~2100 (Figure 2.18). Each RCP

has been reviewed for internal consistency of the whole scenario, but only the information on

emissions, concentrations and land use are incorporated into the RCPs (Van Vuuren et al., 2010). In

comparison with the SRES, the RCPs provide more regionally detailed scenario information, such

as aerosol emissions, geographically explicit descriptions of land use and related emissions and

uptakes, and detailed specification of emissions by source type needed by new advances in climate

models. The RCPs establish a connecting and integrative thread that runs through the research

assessed by: the climate modeling community; the integrated assessment modeling community; and

the vulnerability, impacts and adaptation community.

A redefined focus on near-term changes

2.31 A new goal of the CMIP5 experiments aims specifically to provide decadal climate

predictions for the near term (out to about 2035). These experiments will be carried out with

atmosphere-ocean global climate models (AOGCMs) that are properly initialized for the ocean and

perhaps also sea ice and land surface, using either observations or initialization methods developed

recently. Some of the decadal simulations are expected to be performed using higher resolution in

order to better resolve regional climate and extremes. An enhanced resolution in such experiments

may enable a global horizontal grid scale as high as 50 km, which is equivalent to the scale of

regional climate models (RCMs), currently used in downscaling studies. These experiments will

also be able to better separate natural climate fluctuations from those that are anthropogenic in

origin.

The new scenarios to analyze mitigation outcomes

2.32 The CMIP5 experiments are intended to support the IPCC AR5, which is scheduled to be

published in 2013, although the work will continue beyond that date. Some early CMIP5 results

have already emerged. For example, in a recent study, Jones et al. (2011) reported results from

experiments employing ten AOGCMs and complex ESMs with an interactive carbon cycle from the

ENSEMBLES project (van der Linden and Mitchell, 2009) to simulate an ambitious mitigation

scenario of 21st century emissions, together with a contrasting medium-high emission scenario

from the SRES family (A1B). The study shows that the benefits of mitigation will not be realized in

temperature terms for several decades after emission reductions begin, and may vary considerably

between regions. The subset of the ESMs in the ensemble provides the allowable anthropogenic

carbon emissions under different scenarios as a direct model output.

Interpreting extreme events

2.33 A changing climate can lead to changes in the frequency, intensity or duration of an

extreme event, or, more generally to changes in the probability distribution function (PDF), and

eventually result in an unprecedented, previously unobserved extreme. From a meteorological point

of view, an extreme weather event is an event considered rare at a particular place and time of year.

An extreme event is commonly defined as one would normally be as rare or rarer than the 10th or

90th percentile of the observed probability density function for the meteorological phenomenon at

42

the location considered. When a pattern of extreme weather persists over time, such as during a

season, it may be classified as an extreme climate event, especially if it yields an average or total

that is in itself extreme (e.g., drought or heavy rainfall over a season). Extreme events usually

cannot be directly attributed to anthropogenic climate change, as there is always a finite chance the

event in question might have occurred naturally.

2.34 But extreme events are not interpreted purely from a meteorological point of view. A

weather or climate event, although not necessarily extreme in a statistical sense, may still have an

extreme impact, either by crossing a critical threshold in a social, ecological or physical system, or

because it occurs simultaneously with another event, which, in combination, leads to extreme

conditions or impacts. Conversely, not all extremes necessarily lead to serious impacts. The impact

of a tropical cyclone depends on where and when it makes landfall. Changes in phenomena, such as

monsoons, may affect the frequency and intensity of extremes in several regions simultaneously,

which indicates that the severity of an event may also depend on the overall geographical scale

being impacted. A critical or even intolerable threshold defined for a large region may be exceeded

before many, or any, of the smaller regions within it exceed local extreme definitions (e.g., local vs.

global drought).

Extremes occur even without climate change

2.35 Many weather and climate extremes are the result of natural climate variability (including

phenomena, such as El Niño), and natural decadal or multi-decadal variations in the climate provide

the backdrop for possible anthropogenic changes. Even if no anthropogenic changes in the climate

were to occur over the next century, a wide variety of natural weather and climate extremes would

still occur. Projections of changes in climate means are not always a good indicator of trends in

climate extremes. For example, observation and modeling show that precipitation intensity may

increase in some areas and seasons even as total precipitation decreases. Thus, an area might be

subject to both drier conditions and more flooding.

2.36 Tebaldi et al. (2006) illustrated this on a global scale (Figure 2.21). Based on a multi-

model analysis they found simulated increases in precipitation intensity for the end of the 21st

century (upper panel), along with a somewhat weaker and less clear trend of increasing dry periods

between rainfall events for the A1B scenario (lower panel). Precipitation intensity increases almost

everywhere, but particularly at middle and high latitudes where mean precipitation also increases

(Meehl et al., 2007). For the Arab region, the statistical signal is weak, but with an indication of

increasing risk of both increased precipitation intensity and increased length in dry day spells.

43

Figure ‎2.19 Changes in Extremes Based on Multi-Model Simulations from Nine Global Coupled

Climate Models

Note: Upper panel: Changes in spatial patterns of simulated precipitation intensity between two 20-year means (2080-

2099 minus 1980-1999) for the A1B scenario. Lower panel: Changes in spatial patterns of simulated dry days between

two 20-year means (2080-2099 minus 1980-1999) for the A1B scenario. Stippling denotes areas where at least five of

the nine models concur in determining that the change is statistically significant. Each model‘s time series was centered

on its 1980 to 1999 average and normalized (rescaled) by its standard deviation computed (after de-trending) over the

period 1960 to 2099. The models were then aggregated into an ensemble average at the grid-box level. Thus, changes

are given in units of standard deviations.

Source: Meehl et al. (2007).

Extreme impacts do not require extreme climate events

2.37 It is important to point out that events that may be perceived as extreme may actually be

due to the compound effect from: two or more extreme events; combinations of extreme events with

amplifying events or conditions; or combinations of events, which are not in themselves extreme,

Figure 10.18

44

but lead to an extreme event or impact when combined. The contributing events can be similar

(clustered multiple events) or very different. There are several varieties of clustered multiple events,

such as tropical cyclones generated a few days apart with the same path. Examples of compound

events resulting from events of different types are varied: for instance, high sea level coinciding

with tropical cyclone landfall, or a combined risk of flooding from sea level surges and

precipitation-induced high-river discharge (Van den Brink et al., 2005). Compound events can even

result from ―contrasting extremes,‖ for example, the projected near-simultaneous occurrence of

both droughts and heavy precipitation events mentioned above, or, more anecdotally, flash flooding

following bushfires due to fire-induced thunderstorms from pyrocumulus clouds (e.g., Tryhorn et al,

2008).

Extreme events as seen by climate models depend on resolution

2.38 Extreme events are often localized in both space and time (e.g., the track of an extreme

thunderstorm) so the coarse resolution climate models are not suited (nor designed) to capture the

many extreme events, which are considered as important from an impacts point if view. This has

been one of the major rationales for RCMs and downscaling activities. Christensen and Christensen

(2003; 2004), using Europe as an example, showed that by using a high resolution RCM nested in a

GCM, a more realistic pattern of precipitation change at the end of this century can be simulated.

Figure 2.20 illustrates this finding. For changes in the mean (left and middle panels), the increase in

precipitation in southern Europe is confined to the southern slopes of the resolved topography. In

the course resolution GCM, these are much wider than in the real world, while in the RCM, this

effect is confined to much smaller regions. Therefore, projected changes in extreme precipitation

(right panel) appear more credible when stemming from a high resolution model.

Figure 2.20 Example of the Benefit from High Resolution Modeling

Note: Left: simulated percentage change in precipitation at the end of the century with a coarse resolution GCM (250km

grid); Middle: same for RCM (50km grid); Right: percentage change in extreme precipitation for the same RCM. Blue

contours show model topography for every 500 meters according to model resolution.

Source: Christensen and Christensen (2003; 2004).

DOWNSCALED PROJECTIONS ARE AVAILABLE FOR THE ARAB REGION

2.39 Several assessments of future climate in the MENA region, building on new modeling

and downscaling exercises, have been made. Most of them, however, concentrate on the Middle

East, rather than on North Africa.

45

Box ‎2.3 Climate Models and Downscaling

The accuracy and representativeness of climate model data depends, among other factors, on the horizontal

resolution. It is therefore of importance to use a ―high enough‖ resolution to be able to represent features of

interest adequately, i.e. land/sea contrasts or fine scale topographical features

Figure ‎2.20 Index Map Color-Coding

Note: In this index map, color-coding is directly related to topographic height, with brown and yellow at the lower

elevations, rising through green, to white at the highest elevations. Blue areas on the map represent water. Image credit:

NASA/JPL-Caltech.

Source: http://photojournal.jpl.nasa.gov/catalog/PIA04965.

Figure ‎2.21 Model Topography at a Horizontal Resolution of 1° (about 100 km, top) and 0.2° (about 20

km., bottom)

46

Note: Compare to the map in Figure 2.21 a resolution on the order of 20 kilometers is necessary to resolve features like

the Iranian Plateau, the Anti-Atlas or the Al-Sarat in Yemen.

Source: DMI.

A straightforward way of adding spatial detail to GCM-based climate change scenarios could be so-called

perturbation experiments, where the GCM data is interpolated to a finer resolution and then these

interpolated changes are combined with observed high resolution climate data. In essence, however, this is

only an interpolation exercise, since it does not add any meteorological information beyond the GCM-based

changes; furthermore it implies that the spatial patterns of present-day climate are assumed to remain

constant in future.

Another option, dynamical downscaling, is more costly, namely to use a higher resolution limited-area model

(RCM) to generate climate change scenarios at a higher resolution. State-of-the-art RCMs have a typical

resolution of 25 km, although a few investigations have been conducted at higher resolution on the 5-10 km

scale). They are driven by boundary conditions taken and interpolated from the GCM. In a ―nudging‖ zone

of typically a few grid points in the RCM grid, the GCM values are relaxed towards the RCM grid. Inside

this transition zone, the meteorological fields are RCM-generated. Every few (typically 6) hours, new

boundary conditions are generated from the GCM data. In other words, the RCM can generate its own

structures inside the domain, but the large-scale circulation depends on the driving GCM. Many modelling

centres around the world follow this approach, which, however, requires substantial computer resources and

is comparatively slow.

A third approach can be termed as statistical downscaling. Compared to the straight-forward interpolation

discussed above, additional meteorological knowledge can go into these models, for example, the height

dependence of precipitation. From the above figure, it is clear that the GCM always underestimates the

height or steep slopes of mountain ranges. Precipitation will therefore generally be underestimated in the

GCM compared to higher resolution approaches. Figures 2.21 and 2.22 also show that there generally is a

directional dependence of precipitation (as in the case of the Caucasus Range), i.e. that the increase in

precipitation with increasing resolution will not be uniformly distributed, but actually depend on wind

direction. At least three general methods of statistical downscaling exist: regression approaches, circulation

type schemes and stochastic weather generators, which are used to construct site-specific scenarios.

A weather generator is calibrated on an observed daily weather series over some appropriate period, usually

for a site, but possibly for a catchment or a small gridbox. It can then be used to generate any number of

series of daily weather for the respective spatial domain. Such a stochastic time series will in theory have the

correct (i.e., observed) climate statistics for that domain. The parameters of the weather generator can then

be perturbed using GCM output, allowing the generator to yield synthetic daily weather for the climate

change scenario. However, to derive the appropriate parameter perturbation from the GCM data, other

downscaling methods must be used. It is also important to ensure that low frequency variations (e.g. multi-

47

decadal variations) are adequately captured. Statistical downscaling approaches always require a very good

observational database to derive the statistical properties from. It is therefore not necessarily cheaper or

faster than running an RCM. It is also worth noting that downscaling methods generally cannot easily be

transported from one region to another. Just as with an RCM, the derived regional scenarios will depend on

the validity of the GCM data.

Eastern Mediterranean will become drier, especially in the rainy season

2.40 Figure 2.23 shows the change in the monthly mean rainfall over the eastern

Mediterranean. The most relevant finding is a significant decrease in precipitation (on the order of

40 percent) at the peak of the rainy season (December and January), not only for the West Bank and

Jordan, but also for the Eastern Mediterranean region as a whole (not shown). This is due to a

reduction in both the frequency and duration of rainy events. Before and after the rainy season, the

situation is less clear, with some areas projected to get wetter and others drier. These results are

broadly consistent with wider surveys of global models included in the IPCC AR4, which project a

decrease in annual total rainfall in the Middle East by the end of the 21st century (Lionello and

Giorgi, 2007; Kitoh et al., 2008; Evans 2009, Dai 2010) due to a reduction in the strength of the

Mediterranean storm track. The main mechanism is the northward displacement of this storm track

and, consequently, a reduction in the number of cyclones that cross the eastern Mediterranean basin

(Bengtsson et al., 2006).

48

Figure ‎2.22 Predicted Percentage Changes in Monthly Mean Rainfall for the Eastern Mediterranean

for October to March in the HadAM3P Regional Model Driven by HadCM3 GCM

Note: The horizontal resolution is 44 km. Scenario A2, (2071-2100) minus (1961-1990). Black dots indicate statistical

significance at the 95% level.

Source: Black (2009).

2.41 Evans (2009) compared 18 global climate models in the IPCC AR4. For these coarse

resolution models, a temperature increase of 1.4°C by mid-century and 4°C by the end of the

century was found. The southernmost region of the Mideast domain (Figure 2.24) is projected to

have a small increase in precipitation related to the northward displacement of the ITCZ. However,

precipitation sums are projected to remain below 200 millimeters (generally considered the

minimum for rainfed agriculture in this region).

49

Figure ‎2.23 Change in Precipitation for an Ensemble of Three Global Models, Scenario A2

Note: See double color bar: Hue shows the change in millimeter per year. Saturation intensity shows the change in

millimeters over the 2005 precipitation for 2050 (left) and 2095 (right). Dashed contours show significance at 0.9, 0.95

and 0.99 levels, respectively. Source: Evans (2009).

2.42 The rest of the domain, covering Turkey, the Eastern Mediterranean and Syria and

stretching across Northern Iraq into northwestern Iran is dominated by a decrease in precipitation.

These changes are consistent with decreasing storm track density and intensity (ECHAM5,

Bengtsson et al., 2006). The suggested consequences are a decrease of 170,000 km2 in rainfed

agriculture land, increases in the length of the dry season and changes in the timing of maximum

precipitation, which will affect the growing season.

Higher resolution modeling is needed to improve projections in the complex topography of the

Mashreq

2.43 In 2010, Evans used a regional model to repeat his 2009 assessment based on GCMs,

resulting in quite substantial differences in projections. Since only one GCM/RCM combination

was considered in the 2010 study, it is not clear whether the findings are also valid for other models.

Much of the difference arises from the better resolution of the topography of the RCMs. Figures

2.21 and 2.22 show the topography of the MENA region, as seen at three resolutions: (a) the digital

elevation model of the US Geological Survey (roughly 1 km.); (b) a typical state-of-the-art RCM

(~20 km.); and (c) a typical state-of-the-art GCM (roughly 100 km.). Three distinct features visible

in the RCM data, but not in the GCM data, are the Caucasus Range, the Iranian Plateau and the

double structure of the Zagros Mountains in western Iran. All affect the large-scale circulation, even

in regions far away, by causing upslope air mass movement and subsequent precipitation in the

RCM, whereas precipitation is dumped in wrong locations in the GCMs because they cannot

resolve these topographic structures.

2.44 In the RCM simulation, temperature changes to the end of the century (Figure 2.25) are

smallest in winter (around 2°C) and largest in summer (on the order of 5°C, but up to 10°C on the

Iranian Plateau).

50

Figure 2.25 MM5/CCSM Modeled Change [deg C] in Seasonal Mean Temperature (2095–2000)

Note: The 0.9 (0.99) significance levels is indicated by the thin (thick) dotted line. Scenario A2.

Source: Evans (2010).

2.45 Precipitation is projected to decrease over the Eastern Mediterranean and Turkey, but

increase along the Zagros Mountains in western Iran and the Saudi Desert in summer and autumn

(Figure 2.26). This reflects a change away from the direct dependence on storm tracks towards a

greater amount of precipitation triggered by the lifting of moist air along the mountains, a feature

totally absent in the coarse resolution simulations of Evans (2009). Further south, precipitation

increase goes along with a decrease in atmospheric stability. Seasonal distribution of rainfall is also

projected to change substantially (Figure 2.27).

Figure ‎2.24 for Seasonal Mean Precipitation (mm; 2095 –2000)

Source: Evans (2010).

51

Figure ‎2.25 MM5/CCSM Modeled Monthly Mean Precipitation for Early (solid line) and Late

Twenty-First Century, Scenario A2 (dashed line)

Source: Evans (2010).

2.46 For the Fertile Crescent, including northeast Syria, a minor increase of precipitation is

projected in late summer and early autumn. For this region, precipitation from storms will become

less important than for present-day climate, and upslope lifting processes will play a larger role. The

eastern Mediterranean, including Lebanon, Jordan, Gaza and the West Bank, and western Syria, is

projected to receive considerably less precipitation, mainly in winter and spring, due to changes in

the storm tracks. The Saudi Desert, on the other hand, will receive more precipitation in late

summer, when it is it projected that the ITCZ will be further north than in the present-day climate.

This will result in a decrease in vertical stability, leading to storms. An increase in autumn

precipitation is projected for southeastern Iraq and Kuwait (not shown), due to the advection of

moist air along the Zagros Mountains.

2.47 For the subtropical regions of Africa (such as Morocco and neighboring states), a

temperature increase of about 5°C is projected until the end of the century, with a maximum during

the summer (Christensen et al., 2007). This goes along with a decrease in soil moisture due to

decreasing rainfall and, in particular during the summer months, an enhanced risk of droughts

(Christensen et al., 2007). Less clear is the effect of NOA trends and the subsequent possibility of

enhanced droughts in a future climate, but a small tendency towards more positive NAO indices

and, therefore, less precipitation is backed up by two studies (Rauthe et al., 2004; Coppola et al.,

2005). Large differences exist between GCMs, with strong positive trends in ECHAM4 and no

discernible trends at all in HadCM2 (Wilby, 2007). According to Raible et al. (2010) and Krichak et

al. (2010), it could also be possible that increased polar intrusions following cyclone passages could

lead to no net winter precipitation change in the region.

Downscaled models project a hotter and drier Maghreb

2.48 In Morocco, the data density is high enough to allow for a statistical downscaling

approach. With HadCM3 as the driving GCM and using the A2 and B2 scenarios, climate

projections have been calculated for ten locations in Morocco (Wilby, 2007, Figure 2.6.) as shown

in Figure 2.28. 2.49

52

Figure ‎2.26 Mean Seasonal Temperature Changes for the 2020s (top panel), 2050s (mid panel) and

2080s (bottom panel) Downscaled from HadCM3 Under the A2 Emission Scenario

Source: Wilby (2007).

53

Figure ‎2.27 Calculated Budyko Dryness Index (AIB, Budyko, 1958)

Note: a) Reference simulation; b) A2 scenario simulation; c) B2 scenario simulation; d) Difference in A1B between the

A2 and reference simulations; e) Difference in A1B between the B2 and reference simulations; f) Same as d) but for 50

km simulation; and g) Same as e) but for 50 km (in stead of 25 km) simulation. Aridity regimes are defined as:

0<A1B≤1.1− humid (surplus moisture regime; steppe to forest vegetation), 1.1<A1B≤2.3− semi-humid (moderately

insufficient moisture; savanna), 2.3<AIB≤3.4− semi-arid (insufficient moisture; semi-desert), 3.4<A1B≤10−arid (very

insufficient moisture; desert), 10<A1B− hyper-arid (extremely insufficient moisture; desert).

Source: Gao & Giorgi (2008).

54

2.50 Projected temperature increases are smallest in Agadir (coastal station) and largest in

Ouarzazate in the Atlas Mountains, where summer temperatures are projected to increase by more

than 6°C by the 2080s. Along with this, the frequency and severity of heat waves will increase.

According to Wilby (2007), almost 50 days per year with a maximum above 35°C in Settat (near

the coast) and Beni Mellal (in the Atlas foothills) are projected by the end of the century. Except in

Agadir and Marrakech, less rainfall is projected, ranging from about a 25 percent decrease in the

south to approximately a 40 percent decrease in the agro-economic zone in the north.

2.51 Figure 2.29 shows an assessment of drought in the Mediterranean region (including

substantial parts of the Maghreb) for present-day climate and two scenarios, each at two different

horizontal resolutions (Gao and Giorgi, 2008). Both scenarios show the drought risk around the

Mediterranean Sea increasing from west to east, which is worst in the Levante, but notable also in

other regions, including southern Europe.

Figure ‎2.28 Seasonal Cycle of Rainfall Statistics for an Area Covering the Sinai, Jordan, Lebanon,

Palestine and Syria

Note: From top to bottom: total precipitation, wet days, mean precipitation on wet days, maximum precipitation per wet

day, probability of precipitation given no precipitation on the day before, and probability of precipitation given a wet

day the day before. Red lines represent (1961-1990), black lines represent (2071-2100) for the A2 SRES scenario. The

plots to the right show the difference between future scenario and baseline. Filled bars indicate significance at the 95%

level using a Student‘s t-Test. Source: Black (2009).

55

Agricultural lands are threatened by increasing aridity in the Mashreq

2.52 Figure 2.30 shows precipitation statistics for an area consisting of the Sinai, Jordan,

Lebanon, Palestine and Syria. Most prominent are a (statistically significant) decrease in the

number of rainy days, both following a dry or a wet day, and a general decrease in winter rainfall.

According to the GCMs in Christensen et al. (2007) and RCM experiments by Önol and Semazzi

(2009), temperatures in the region will increase on the order of 2°C in winter and up to 6°C in the

inland regions in summer. A reduction in winter precipitation on the order of 25 percent and an

increase of drought duration by up to 60 percent are expected based on the A1B scenario (Kim and

Byun, 2009). The authors also predict a northward expansion of the Arabian Desert and an increase

of autumn precipitation over the Fertile Crescent by up to 50 percent.

2.53 Although less favorable than for Morocco, the data density is sufficient to allow for

statistical downscaling by stations in Amman (Jordan), Kamishli (Syria) and Kfardane (Lebanon).

While for Kamishli, the present-day climate can be reconstructed quite well, results are less

convincing for Amman. This may be due to missing data and problems due to station relocation and

urban growth (Smadi and Zghoul, 2006). Depending on scenario and location, temperature

increases by 3-4°C (A2) and 2-3°C (B2), respectively, are obtained (Wilby, 2010), while rather

large decreases in precipitation are projected — up to 50 percent near the coast (Lattakia, Syria) and

around 15-20 percent inland (for example, in Palmyra, Syria). Note that this reduction in

precipitation takes place almost entirely during the winter, as generally no precipitation exists at all

during summer under present-day conditions.

2.54 As mentioned above, it has been suggested (Evans and Geerken, 2004) that the limit for

rainfed agriculture lies close to average annual precipitation of 200 millimeters per year. Taking

Amman as an example (average precipitation 1961-1990 was 260 millimeters), this suggests that

agriculture would no longer be possible in 2080. However, due to the large interannual variability

(coefficient of variation in Amman is 38 percent, which implies that there is a 1/3 chance for an

annual sum under 200 millimeters even in present-day conditions), agriculture could cease to be

economically viable considerably earlier. Based on HadCM3 data and the A2 scenario, the chance

of a dry year (<200 millimeters) would be 50 percent around 2020 rising to 80 percent around 2060.

This finding is consistent with previous studies that suggest that the 200 millimeters isohyet could

move on the order of 75 km northward by the end of the century (Evans, 2009). If precipitation is

too low, agriculture is traditionally replaced by grazing. The same scenario projects an increase in

the dry season length from 6 to 8 months by 2080 for Amman, which potentially poses a problem to

the herders.

2.55 Based on Christensen et al. (2007), GCMs suggest a temperature increase of 4°C in

winter and 6-7°C in summer. But unlike in the Maghreb and Levante regions, there is little

agreement between the models. No clear climate change sign is found for precipitation, with

indications for slightly wetter conditions in the south and drier conditions in the north, related to the

northward shift of the ITCZ.

2.56 Figure 2.31 shows results from another RCM simulation covering the Arabian Peninsula

and neighboring regions. Apart from the coastal plains in Yemen, Oman and Somalia, where the

salt marshes remain comparatively humid and therefore experience less warming, relatively uniform

warming is projected. For precipitation, changes are small in areas that are already arid under

56

present-day conditions, but major drying is projected for the Fertile Crescent and the regions further

west (Israel, Gaza and the West Bank, Jordan, Syria). Note this contradicts the finding by Kim and

Byun (2009).

Figure ‎2.29 Left: Temperature Projections with Respect to the 1990s: 2020s (top), 2040s (middle),

2070s (bottom) in °C. Right: Same but for change in precipitation as % change from the 1990s

Source: Hemming D, Betts R, & Ryall D. 2007. Here from Tolba and Saab, 2009.

SEA LEVELS WILL CONTINUE TO RISE

2.57 Although sea level rise is not a meteorological phenomenon, it is discussed here because it is

indirectly caused by human activities and it is one of the external drivers influencing many of the issues

discussed later in this volume.

57

Global mean sea level has been observed to have steadily increased since 1870

2.58 The sea level of the sea at the shoreline is determined by many factors that operate over a

great range of temporal scales: hours to days (tides and weather), years to millennia (climate) and

longer. The land itself can rise and fall and such regional land movements must be accounted for

when using tide gauge measurements for evaluating the effect of oceanic climate change on coastal

sea levels. Coastal tide gauges indicate that the global average sea level rose during the 20th

century. Since the early 1990s, sea level has also been continuously observed by satellites, with

near-global coverage. Satellite and tide gauge data agree at a wide range of spatial scales and show

that global average sea level has continued to rise during this period. Figure 2.32 illustrates the

observed steady increase in global sea level since 1870 to the present. In total, the global increase

has been approximately 18 centimeters. Sea level changes show geographical variation because of

several factors, including the distributions of changes in ocean temperature, salinity, winds and

ocean circulation. Regional sea level is affected by climate variability on shorter time scales, for

instance, with El Niño and the NAO, leading to regional interannual variations, which can be either

greater or weaker than the global trend.

Figure ‎2.30 Annual Averages of the Global Mean Sea Level Based on Reconstructed Sea Level Fields

Since 1870 (red), Tide Gauge Measurements Since 1950 (blue) and Satellite Altimetry Since 1992

(black)

Note: Units are in millimeters relative to the average for 1961 to 1990. Error bars are 90% confidence intervals.

Source: IPCC, 2007.

2.59 The IPCC AR4 states that over the period 1961 to 2003, global ocean temperature has

risen by 0.10°C from the surface to a depth of 700 m. This temperature increase results in a thermal

expansion of seawater that has contributed substantially to sea level rise in recent decades. Over the

period with data available, warming was particularly large between 1993 and 2003, but somewhat

smaller afterwards. The thermal expansion of the oceans, however, is not sufficient to explain the

observed sea level rise, which also occurs from cryospheric changes, including the increased

melting of glaciers and ice caps, and the thawing and calving of the Greenland inland ice.

Figure TS.18

58

2.60 Large-scale, coherent trends of salinity have also been observed over the last few

decades. All available data point toward a freshening in subpolar latitudes and a salinification in

parts of the subtropical and tropical oceans. These trends are consistent with changes in

precipitation and evaporation and probably an increased freshwater flow into the Arctic from river

runoff and thawing land ice. Furthermore, oceans have become more acidic and, as a consequence,

the amount of emitted carbon dioxide taken up by the oceans has decreased.

2.61 Recent expert meetings (e.g., from the climate conference held in Copenhagen in

December 2009) discussed whether the IPCC AR4 had underestimated the amount of sea level rise

and that ocean warming is about 50 percent greater than the IPCC had previously reported. Material

presented at the above conference suggested that the rate of sea level rise increased during the

period from 1993 to the present, mainly due to the growing contribution of ice loss from Greenland

and Antarctica. Observations show that the area of the Greenland ice sheet at freezing point or

above for at least one day during the summer period increased by 50 percent during the period 1979

to 2008, particularly during the extremely warm summer of 2007 (Steffen and Huff, 2009; Mote,

2007).

2.62 Ice sheets may also lose mass through ice discharge, which is also sensitive to regional

temperature. The best estimate is that the Greenland ice sheet has been losing mass at a rate of 179

Gt/year since 2003, corresponding to a contribution to global mean sea level rise of 0.5

millimeters/year (AMAP, 2009). This is approximately twice the amount estimated by the IPCC in

2007.

Climate models project sea level will continue to rise in the 21st century

2.63 Climate models are consistent with the ocean observations and indicate that thermal

expansion is expected to continue to contribute significantly to sea level rise over the next 100

years. New estimates suggest a sea level rise of around one meter or more by 2100 (IOP, 2009).

Since deep ocean temperatures change slowly, thermal expansion would continue for many

centuries even if atmospheric greenhouse gas concentrations were stabilized today.

2.64 In a warmer climate, models suggest that the ice sheets could accumulate more snowfall,

which would lower the sea level. However, in recent years, any such tendency has probably been

outweighed by accelerated ice flow and greater discharge. The processes of accelerated ice flow are

not yet completely understood, but are likely to continue to result in overall net sea level rise from

the two large ice sheets of Greenland and Antarctica.

2.65 The greatest climate- and weather-related impacts of sea level are due to extremes

associated with tropical cyclones and mid-latitude storms, on time scales of days and hours. Low

atmospheric pressure and high winds produce large local sea level excursions called ―storm surges,‖

which are especially serious when they coincide with high tide. Changes in the frequency of the

occurrence of these extreme sea levels are affected both by changes in mean sea level and in the

meteorological phenomena causing the extremes.

59

FROM CLIMATE MODELS TO IMPACT ANALYSES

More effort is needed to develop good impact models to translate climate to potential impacts

2.66 The previous sections of this chapter have attempted to explain what is understood about

observed and projected climate change, and have emphasized the strengths and limitations of both

observational data and the climate models used to make projections. We know that the Earth is

warming and that the atmosphere is becoming more unstable as a result. Great uncertainty remains

about the precise magnitude of changes in climate and weather patterns at any given location and

time. Some of this uncertainty derives from the inherently chaotic nature of the weather and climate

and from limitations in the models, but much is derived from uncertainty about future human

actions that release greenhouse gases and change land cover. Despite these uncertainties, decisions

still have to be made about water management, agricultural practices, land use and infrastructure,

where the consequences of the decisions extend well into what we know will be significantly

changed climates. These decisions are often guided by ―impact models‖ that, among many other

factors, take into account climate and climate variability. These were typically guided by climate

observations from the recent past (e.g., 1961 to 1990 weather data), but the past is no longer a good

prophesy of the future. The challenge is to find ways of representing both the understanding of

future climate and the uncertainties in models of water use, decisions about land use and

infrastructure design.

2.67 A number of steps must be taken when deciding how to incorporate future climate change

and variability into impact models. The first is to assess how the impact model considers current

climate variability. Many models, e.g., many economic models, treat climate as a constant and

factor climate variability into a general error term or safety factor, as in bridge and building design.

In some cases, the design of these decision processes will have to be reassessed. Can the safety

margins simply be extended? Will these still lead to effective and efficient decisions? In some

cases, new approaches to modeling impacts will be needed. Other decisions are based on detailed

models of stream flow or agricultural productivity, which are driven by detailed inputs of

meteorological data. In these cases, the challenge is to replace inputs from the recent past with

inputs from well selected models of the future. But many of these models require data at a much

finer resolution than can be produced by GCMs and even RCMs and at time scales (e.g., hourly or

daily data) that are often not available from modeling runs.

The challenge of output resolution –in both space and time scales

2.68 The horizontal resolution of GCMs is typically 100-200 km, which is usually too coarse

for direct use by impact models. Climate projections with higher spatial resolution can be obtained

by dynamical downscaling, using high-resolution RCMs driven by initial and boundary conditions

provided by GCMs (e.g., Rummukainen, 2010). This improved resolution makes it possible to

include subgrid variability, like small-scale topography, and to resolve sub-grid processes to

improve the modeling of regional precipitation, for example. Indeed, a number of studies show that

the results from impact models using downscaled climate information can be quite different from

results directly based on GCM output (e.g., Mearns et al., 2001; Olesen et al., 2007).

2.69 Different GCMs and RCMs use different physical formulations to describe atmospheric

processes, interactions between the atmosphere and the land and oceans, etc., and, thus, provide

60

different projections of future climate. The spread of results from these different models represents

an uncertainty regarding future climates that must be accounted for by impact modelers. Too often

impact models have been driven by the results of a single GCM or RCM. However, the

uncertainties are now being explored by ensemble simulations applying different combinations of

GCMs and RCMs (e.g., Christensen et al., 2007b; Kjellström et al., 2010). One way of directly

including this uncertainty is to calculate the probability distribution of important outputs (e.g., daily

maximum temperatures) based on the information from the full set of model simulations (Déqué

and Somot, 2010). In recent studies, the contribution of individual models to the probability

distribution has been weighted according to their performance with respect to a given set of metrics

(Christensen et al., 2010). The weighting procedure itself adds an extra source of uncertainty and

the method is still being explored. For example, the ability of a model to predict present-day climate

could be seen as an indication that it will also be able to predict future climate. However, feedback

mechanisms that were important in the past could be less important in a future climate with a

different forcing, so the weightings themselves may not be appropriate (Christensen et al., 2008;

Reifen and Toumi, 2009). In addition, models may perform differently for different regions. The

Coordinated Regional climate Downscaling Experiment (CORDEX) initiative will study this and

offer new regional projections for a number of regions across the world, including the Arab region,

although not as a single entity, but rather distributed across several proposed regional domains

(Giorgi and Asrar, 2009).

2.70 The probability distributions discussed in the previous paragraph usually apply to a point

in time. However, for impact studies, time series are generally needed, making the probabilistic

approach more difficult. Large-scale community efforts have addressed how to pick the most

relevant climate information from an ensemble of model simulations (Christensen et al., 2007b;

Christensen et al., 2009). As for the weighting procedure, the selection of individual models is in

itself a source of uncertainty and should be done carefully to avoid biasing the resulting analyses.

2.71 Most downscaling methods combine climate model output with observed data. In this

way, local features not captured by the climate model can be incorporated in the scenario data,

allowing the impact model to be used with data in the form that it was developed and tailored to.

The most appropriate method of preparing climate scenario data depends on the local region and on

the specific needs of the impact models.

2.72 Any requirement to use time series in order to calibrate and validate an impact model

prevents the direct use of GCM or RCM data, as the climate models do not reproduce sequences of

real weather events. The models are not designed nor meant to be able to do this. Instead, a climate

model is constructed with the aim of being able to reproduce climate, which, in a narrow sense, is

usually defined as the average weather or, more rigorously, as the statistical description in terms of

the mean and variability of relevant quantities over a period of time, ranging from months to

thousands or millions of years. The classical period for averaging these variables is 30 years, as

defined by the WMO. The relevant quantities are most often surface variables, such as temperature,

precipitation and wind. Climate in a wider sense is the state of the climate system, including a

statistical description. Even downscaled climate change information is still not akin to a weather

prediction. Many climate variables in the model output are systematically offset and need to be bias

corrected, which requires observational data.

2.73 Methods to transform climate model output into the impact application are therefore

called for. The climate change scenarios can be further downscaled to a regional or local level by

61

applying a weather generator (WG) approach. For example, a stochastic WG based on the series

approach (Racsko et al., 1991) has been proven to give a realistic representation of the duration of

wet and dry days and, thus, the duration of droughts, while the frequently applied Markov chain

models (Semenov et al., 1998) is less successful in achieving this. Such an approach has the benefit

of allowing for a good representation of current climatic conditions by calibrating the WG to

observed data, and allowing not only the inclusion of changes in mean climate from the RCMs (or

GCMs), but also changes in climatic variability. This approach also has the flexibility to generate

longer and multiple time series of synthetic weather data for use in impact models, thus allowing for

better quantifying the variability in response to climate change.

2.74 Even if suitable time series of weather data are generated to drive an impact model, the

question remains as to what extent is the impact model itself calibrated and tested for weather

conditions beyond those experienced in observed conditions. This should be clarified when the

model is used under climate change conditions, in particular when addressing values well outside

past experiences or when extreme events are being assessed.

From weather forecasting to seasonal-to-decadal prediction

2.75 A climate model is designed to be able to capture the statistical properties of the

geographically and temporarily varying weather that characterizes climate; it is not designed to

predict the actual weather. A weather prediction or forecast, on the other hand, is typically

formulated as a categorical statement about the state of the weather within a certain time frame,

typically up to 10 or 15 days (medium-range forecasts). More recently, monthly and seasonal time

scale forecasts are also provided by many centers. These forecasts typically provide qualitative

assessments of the most likely weather development, primarily based on long-term variations, e.g.,

expectation of droughts or moist conditions. A commonality between all forecasts is the

dependency on a well-defined initial state that must have its origin in an observed state of the

atmospheric (or more broadly the climate) system. This typically involves some kind of data

assimilation system by which the forecast model is constrained towards the observed evolution of

the weather and/or oceanic state, moist or dry land surfaces, etc. The techniques to adequately

combine modeled information, typically represented on a grid, with point measurements at which

most observations are made, comprise a large research field of their own.

2.76 In climate modeling, the model typically begins from an idealized initial state, which

aims to represent conditions ―typical‖ for the period it is intended to represent, often pre-industrial

conditions where the levels of greenhouse gases and anthropogenic aerosol loads in the atmosphere

were low. Then known or estimated external forcings from the emission of greenhouse gases and

aerosols, varying solar insolation, volcanism and other drivers (e.g., land use) are introduced into

the freely running model. In this way, a modeled representation of the evolution of the climate since

the industrial revolution is made. But the state of the climate in such a model for a particular

decade, say 2001-2010, is not meant to be compared with the real world for that decade. The model

is designed to capture the changing characteristics of major phenomena such El Niños decadal

fluctuations in monsoons, the NAO, etc. but not forecast particular events.

2.77 This is a problem when the challenge is to address the near future climate development

because natural fluctuations on the decadal time scale are major determinants. In essence, the

probable evolution of climate over the next three to four decades is more difficult to depict than the

62

climate at the end of the century under a prescribed emission scenario. This is very often forgotten

when impact assessments are called for.

2.78 Some predictability in the atmosphere on seasonal, interannual and decadal time scales

can arise from internally generated natural climate variability—often connected to oceanic

variability—and certain types of external forcing (solar and volcanic eruptions). Internal variability

that results in, e.g., extensive, long-lasting, upper ocean temperature anomalies, has the potential to

provide seasonal, interannual and even decadal predictability in the overlying atmosphere, both

locally and remotely through atmospheric ―teleconnections.‖ These teleconnections can guide

predictions for the next few seasons. The question of whether or not skill extends to forecasts for

the following decade and beyond is a scientific ―hot topic‖ and currently a central issue in the

upcoming AR5.

2.79 The magnitude of the changes in the atmosphere associated with any decadal

predictability at a specific location is extremely small compared with the size of the day-to-day

variability linked to changes in weather at that same location. Thus, the decadal signal is difficult to

recognize amongst the much larger variability arising from unrelated changes in weather. So while

the decadal signal can underpin predictability in changes to decadal averages in climate variables,

the same signal provides little if any enhancement to our ability to forecast weather.

2.80 The level of predictability and apparent predictive skill arising from both internal and

external forcing can vary markedly from place-to-place and from variable-to-variable. Recent

research (see e.g., Meehl et al. (2009), Murphy et al. (2010), Latif et al. (2010) and Hurrell et al.

(2010)) indicates that predictability is absent or very limited in some variables in most locations

over the surface of the earth. So we cannot accurately predict all aspects of the climate over the

coming decades even if we could perfectly resolve all the technical issues confronted when

developing and conducting predictions. The nature of the climate system precludes the possibility of

reliable decadal predictions of some climate variables in some locations. In such cases, the best

estimate that can be provided for the quantities and regions in question for future decades is the

information contained in historical climate records.

The practice of impact modeling

2.81 The best choices for integrating climate modeling projections with impact modeling

depends on location, time period and application, and is best made through the cooperation of

climate modeling and impact specialists. It is not possible to discuss the myriad of options in this

chapter, but in Annex 1, we provide some guidelines as to how a team responsible for developing

impact models to guide decisions that are likely to be affected by climate change might step through

the process of developing advice guided by the best available understanding of future climates.

63

APPENDIX B.1 DEALING WITH CLIMATE RISKS – A CHECKLIST

This checklist is designed to guide the leadership of a program or large sectoral or cross-sectoral project that is assessed as being subject to climate

risks. Examples include the construction of a new dam, a major rural road upgrade, the design of an agricultural irrigation program, or a major

coastal tourism development. The checklist is designed to assist in incorporating climate risk, including climate change, in the program design. It

is essentially a top-down approach to adaptation that unfortunately so often brings adaptation considerations rather late to the decision process

(Wilby & Dessai, 2010). This is not the only, nor necessarily, the most effective, approach to adaptation planning (see World Bank World

Development Report 2010; Lempert et al., 2004; Wilby & Dessai, 2010; Hallegate et al., 2011 for a wider discussion of adaptation in the context

of wider development goals). However, the circumstances described above are common in development planning and especially that which

focuses on infrastructure.

Note: The order of the checklist is such that items nearer the top are most likely to precede those below, but this is NOT a purely sequential

process.

1. Assess where you stand in the decision process

a. If the decision process is in its early stages, then

follow a process to identify goals and needs of

stakeholders, and the range of adaptation options

and criteria by which they may be evaluated. This

is the most desirable situation, but this checklist

cannot elaborate this open-ended and case-

specific process any further.

Clearly, this is the best stage to begin considering adaptation actions. It allows the

full range of climate-resilient options to be considered. See the references in the

introductory paragraph for guidance.

b. Are the core features of the design in place? If so,

what information relating to future climate might

challenge these decisions?

Here, the assumption is that adaptation to climate risk and climate change is a

component of a wider development objective, i.e., this is not a ―stand-alone‖

adaptation project.

c. What are the feasible options for the modification

or fine-tuning of the existing design?

These options need to be identified so that appropriate decisions can be made

about how much assessment of the climate risks is needed. At this point, simple

64

―rules-of-thumb‖ about climate change may need to be applied. For example,

using expert judgment to assess the effect of a warmer climate with more variable

precipitation on the proposed project.

d. What is the time horizon relevant to the decision

process?

The time horizon is not necessarily limited to the longevity of a piece of

infrastructure. For example, implementing improved flood resistance of a planned

road system may be sufficient for the normal design horizon for roads, but it could

also lead to denser settlement in inherently flood-prone areas, leading to

unacceptable outcomes in the climates of the more distant future.

2. What does the latest IPCC Report and

subsequent commentaries and updates say about

your project location and sector(s)?

The IPCC Reports will contain a summary of the current climate, observed

changes, climate change projections and the possible impacts affecting your

region.

www.ipcc.ch

3. Are more recent (since the most recent IPCC

Report) or more nationally or regionally explicit

assessments available?

a. Many countries have their own national

assessments of climate change and its impacts

Check National Communications to the UNFCCC at

http://unfccc.int and search for National Reports

b. Assess the knowledge base and validity of all

assessments

Many national and sub-national reports are summaries from the IPCC with little

additional information

c. Identify major discrepancies (if any) between the

regional/explicit assessments of the IPCC and

seek to establish an explanation

Discrepancies may arise because of newer information, more regionally specific

information, different sources of information, or errors and misinterpretations.

4. Establish a climate base line The climate baseline is a description of the current climate (often a 30 year period,

such as 1961-1990), including averages, variability and trends. It is used to both

establish the current conditions and most modeled projections (next section) are

interpreted as changes to this baseline.

a. Seek cooperation and input from the national

meteorological office and similar authorities, e.g.,

Baseline climate information may exist in a form that is not publically available.

E.g. data may not exist in digitized form or if digitized, basic quality control may

65

for sea level monitoring still be lacking.

b. Decide upon the variables that may affect the

outcome and design of the project

They are not only temperature and precipitation. They are more likely to be

complex variables, such as run-off, dry spells, rainfall intensity, wind storms and

dust storms. Some of these variables are available from GCM modeling, while

others will have to be derived through secondary modeling.

c. Explore the existing public data bases for both

station and gridded data

See references in this chapter for coverage of the Arab region.

d. Calculate averages, trends and measures of

variability for the variables of interest

See the World Bank Climate Portal where much of this information is available

(http://sdwebx.worldbank.org/climateportal/)

e. Share the base line data with appropriate experts

(sectoral and climate) and review – revisit

Section 1 above if necessary

These analyses may give insights to additional options in the project design

5. Establish the relevant range of climate

projections

a. Currently, 22 IPCC AR4 models form a core set

and are available from the sources listed opposite.

List of sources for 22 GCMs

b. Assess whether particular scenarios should be

chosen (i.e., A1B, A2, B1, B2). Note that for

projections in the near term (next few decades),

the choice scenario makes little difference.

The A1B is closest to our current trajectory. A common pairing is B1 (most

effective mitigation) with A1B (in-effective mitigation).

c. Do not reject any GCM unless data is missing, or

an authoritative climatological reason indicates it

is inappropriate.

Some users seek a single ―best‖ GCM for their region. This is unwise as the

criteria for choosing the best are unclear; one GCM might project current

precipitation better in your region, while another matches temperatures better.

Also, the ability to project current climate does not necessarily indicate the ability

to project future scenarios.

d. Check the performance of each GCM selected for

a broad match with the major observed

meteorological phenomena, such as rainy

Poor performance of a GCM for your region in projecting major patterns of

seasonality, etc., such as monsoon patterns or rainfall seasonality is an a priori

rationale for rejecting that GCM. But cases may exist where all models appear to

66

seasons, timing of monsoons, etc. If major

discrepancies are found, seek further technical

help before using the model.

be disqualified. In such cases, seek further technical assistance.

e. Check climate change projections for each of the

selected models to see if any appear to be

outliers, i.e., producing results very different

from the other models, or strong discontinuities

between observed and near future projections. If

major discrepancies are found, seek further

technical help before using the model.

A very different climate projection for a single model may indicate that the model

should be disregarded. However, it may also be due to internal variability, which

simply indicates that the climate change signal for your region is not a robust

feature in that model.

f. As a minimum requirement, explore a plausible

range by using GCMs with a low, medium or

high global mean temperature or precipitation

response to a particular scenario.

The use of a bracketing set of GCMs is a common approach when running the

impact modeling/assessment is time consuming and/or expensive.

6. Assess options for evaluating impacts An enormous range of approaches is available for evaluating the impacts of

climate change and variability, ranging from well-established damage formulas,

watershed models, agricultural yield and production models and so on. They

cannot be dealt with in detail here. Instead, some generic advice, applicable in

most situations, is provided.

a. How was, or will, climate be incorporated into

the assessment of the design adequacy and

projected outcomes of the project?

Was climate treated as a constant in the original design? If so, what performance

criteria that might affect decision making are likely to be affected by different

projected climates? If standard hazard models (e.g., 1:100 year flood levels) are

used, can these be updated for the different climate scenarios?

b. If qualitative methods were used (e.g., expert

judgment, Delphi techniques, etc.) go to Section

8 below.

c. If quantitative impact models are to be used,

which climate variables will be used as inputs?

These are usually well prescribed for quantitative models.

67

i. Has the climate response of the impact

model been sufficiently tested (validated)

against observational data? If not, what was

the justification for the impact model‘s

selection?

ii. Does this validation still apply to a changed

climate?

For example, would a changed climate go beyond the range/domain of validation,

and especially for the impact of extreme events? Or might climate change

introduce new phenomena not considered in the original model, such as

salinization of cropland or even flooding in a location currently safe from

flooding?

iii. If not, you may proceed, but with cautious

interpretation of the modeling results. Also,

seek additional ways of validating the

models under the changed conditions.

iv. How sensitive are the final outcomes of the

models to the climate parameters?

The sensitivity of the model to variation in the input parameters should be a part

of model development and testing. It may be that the important outputs from the

impact model are sensitive to only some of the climate inputs. This will help to

focus on best describing the important climate variables. For example, a crop

growth model may not be sensitive to seasonal changes in mean rainfall, but very

sensitive to the timing of the first rains of the wet season. Are your climate models

able to provide the information that matters?

v. Do the models treat climate as a fixed (e.g.,

use climatic means) or a stochastic (e.g., use

historical weather data) input?

If the model treats climate as fixed, you will need to design a series of model runs

covering the range of climate projections being considered. Usually, this means

running the impact model under current climate, and for example, a +1ºC

scenario, a +2ºC scenario, etc. If the model uses stochastic weather patterns, then

you will need to consider how to generate realistic weather sequences for the

future climate projections; for example, via a ―weather generator‖ which generates

random weather sequences within prescribed bound of means, variability and

cross correlation between variables, such as temperature and precipitation.

68

vi. Might the modeling results under a changed

climate affect other components of the

modeling or the decision making process?

For example, the current planning may have implicit or explicit assumptions about

the availability of water for irrigation, constant trends in yields or constant

commodity prices. These may not be valid under a changed climate. If so, the

impact modeling may need to be modified or extended to deal with these

additional factors.

d. Are the climate projections gathered in Section 5

above sufficient to rerun the models under a

changed climate?

i. If not, can the necessary data be provided

from the raw data repositories of the model

runs?

The GCMs provide extremely detailed descriptions of the Earth‘s atmosphere and

land surface for every few minutes of their run. Most of these data are not saved,

but far more data are saved than appear in most compendiums and may be

available from the modeling groups who conducted the experiments.

ii. Can proxies for the missing data be found?

Can the missing data be assumed constant, or

can the models be run with a set of estimates

of the missing data that bracket the expected

range?

Some aspects of climate change are known to scale more or less linearly with

global temperature. To assess whether this may apply in a specific context,

technical assistance is required. Some guidance is also available in IPCC reports

(e.g. Chapter 11 in the contribution from WG-I to AR4).

In some cases, you can exchange place for time, e.g., if data for a particular

application in a region does not exist, analogue current climates may be identified

and pseudo data from such a place may be used as a proxy. In any case, further

technical advice should be sought.

e. How important are other variables (e.g.,

population trends, commodity prices, increasing

demand, yield improvements) in determining the

outcomes of the modeling and/or in guiding the

decision? In other words, what sort of climate

change might cause you to change or revise the

project design – does climate matter in your

decision process?

This is essentially a revisit of Section 1 above to ensure that you are still on track.

69

f. If the answers to the above questions suggest that

climate change may affect the project design, and

that the impact models can incorporate climate

change effectively, then proceed to run the

impact models under a representative range of

climate scenarios determined in Section 5 above.

7. Consider how to apply the changed climate to the

impact model

a. Direct use of GCM output is rarely applicable Most GCMs have well documented ‗biases,‘ such as a tendency of being clearly

too wet or too dry in a particular region, exhibiting too weak an annual cycle for

certain climate parameters, or systematically underrepresenting extreme values

(e.g., heavy precipitation). Such biases must be corrected for before using their

output for driving impact models.

b. The most common approach is to use changes

(often called deltas).

Future climate projections are calculated as the difference between the model

results for the time period of interest (e.g., 2030) and the model results for

baseline period (e.g. 1961 to 1990) – i.e., the delta. This is added to the observed

climate data for the same baseline period.

c. If the decisions are sensitive to the impact

model‘s treatment of climate variability, seek

expert advice on ‗bias correction.‘

Delta change in its most simple form will preserve the higher order climate

statistics of the current period (i.e., its variability, longer-term cycles). This need

not be the case in the future, nor even as depicted by the climate models.

8. Reassessing qualitative treatments of the impacts

of climate, climate variability and climate change

on the decision process.

a. In most complex programs or projects, some

elements of the design and decision process will

be based on qualitative assessments. These

should still be subject for scrutiny for climate

impacts.

For example, the task may be a program for building schools in remote areas.

Detailed climate modeling is not necessary, but it would be useful to consider the

environment in which these schools will operate over their lifespan, e.g., higher

temperatures, higher flood risk, etc.

70

b. Involve climate experts in ‗expert judgment‘

exercises

c. Ensure that stakeholders, civil society

representatives, etc., are adequately briefed on

climate risks and have the opportunity to apply

this information to their inputs

d. Look at recent trends in extreme weather events

and disasters

The CRED database provides comprehensive disaster information

(http://www.cred.be/) and a review of other sources in Tschoegl et al., 2006.

e. Apply sensible rules-of-thumb, e.g.:

i. It will be a warmer world (the IPCC will

give estimates for your location and future

date), but some places will also experience

more extreme cold events due to changes in

weather patterns

Most impact specialists will be able to make an assessment as to the nature and

extent of such changes in climate. They may advise that no changes in design are

needed, be able to recommend adjustments (e.g., in the school example, further set

back from rivers, upgrading of water supply), or conclude that further analysis and

possibly modeling is needed.

ii. Precipitation will marginally increase in

most regions (but check IPCC Reports), but

most places will be effectively drier due to

longer dry spells, higher temperatures and

increased evaporation

iii. Precipitation probably will occur more

erratically, with both extreme dry (drought)

and extreme wet (flood) periods becoming

more common

iv. Coastal regions will be subject to rising sea

levels, but probably also to increased storm

surges and wind damage, which are likely to

be more important in the near future

9. Cross-check the results of the impact modeling by

71

whatever means available

a. Do the results scale sensibly with different levels

of climate change?

For example, if you used GCMs representing low, medium and high outcomes,

and/or scenarios representing low and high mitigation effectiveness, do the results

of the impact modeling follow these trends in the way expected. If not, can

plausible explanations be found and can these explanations be further tested?

b. How do the results compare with previous

studies?

c. If a range of climate projections has been

considered in the impact modeling, do any of the

projections lead to unacceptable or undesirable

outcomes?

Here you need to consider whether the option chosen is robust to the feasible

range of climate outcomes.

d. Have you left out any factors that you thought

might be important but did not have enough

information to include? If so, document the

reasons why and warn users of the omission.

Could you make estimates that might bracket the

scale and/or impact of these omitted factors?

For example, you may have good estimates of sea level rise but no reliable

estimates of changes in the height and frequency of storms surges, so they were

omitted. In terms of flooding impact, the storm surges are possibly far more severe

and will affect you much earlier than sea level rise.

e. Have you overlooked some other factor that

would ‗swamp‘ any climate signal?

For example, studies of crop yields under climate change often predict decreases

of 10% to 20% by 2050. But the overall impact model may have a built in

assumption that technological improvement will increase crop yields by 1% to 2%

per year, as it has over past decades. If the latter estimate is wrong by only half a

percent or so, and technological improvement is declining in many regions, this

trend will swamp any climate change effects. Beware of ‗crackpot rigour‘ – i.e., a

very detailed analysis of the wrong problem.

f. Are you prepared to recommend changes in plans

based on the results?

g. Would you risk your own money or livelihood or

life on your advice? You might be!

72

73

APPENDIX B.2 MEAN ANNUAL AND MEAN MONTHLY PRECIPITATION (MM) AND INTER-ANNUAL PRECIPITATION

VARIABILITY FOR THE ARAB WORLD

Country January February March April May June July August September October November December Mean

Annual

Inter-

annual

Variability

(CV - %)

Algeria 10.1 8.5 9.2 7.6 6.5 3.1 2.0 4.0 5.9 8.1 12.1 10.4 87.6 75.0

Bahrain 21.7 13.5 13.3 7.3 0.0 0.0 0.0 0.0 0.0 0.0 5.2 17.6 78.7 51.3

Comoros 320.0 237.9 203.7 226.5 134.8 132.6 106.5 81.5 32.7 50.8 72.1 171.4 1770.2 17.9

Djibouti 9.0 14.1 28.1 22.7 10.6 2.4 27.1 54.1 25.0 7.4 10.6 8.0 219.0 50.3

Egypt 4.9 3.6 3.7 2.4 3.5 1.9 2.5 2.6 1.4 2.7 2.9 4.9 37.1 56.4

Iraq 34.7 31.0 30.8 28.2 11.1 0.7 0.3 0.3 0.4 7.0 23.0 29.9 197.4 23.0

Jordan 23.7 22.7 17.2 8.3 2.7 0.2 0.3 0.4 0.1 3.4 11.2 19.3 109.6 35.9

Kuwait 21.6 14.4 16.7 15.7 4.9 0.3 0.6 0.6 0.3 1.6 15.6 21.9 114.2 31.0

Lebanon 159.6 129.7 87.9 48.8 15.2 3.0 0.0 0.1 2.5 30.8 84.9 139.5 701.9 25.0

Libya 8.7 5.5 5.2 2.8 2.1 0.5 0.4 1.2 1.6 4.5 6.2 9.1 47.8 49.5

Mauritania 0.6 0.7 0.3 0.6 1.2 6.8 21.1 41.2 25.8 4.7 1.3 1.2 105.4 48.7

Morocco 39.0 41.4 44.5 36.2 21.4 8.4 2.4 4.3 14.1 29.0 42.9 47.5 331.0 33.0

Oman 7.4 11.4 12.4 12.7 6.0 7.5 11.1 9.2 3.0 3.9 4.5 7.5 96.7 49.7

74

Qatar 7.0 19.7 17.0 7.0 1.9 1.1 2.0 1.9 0.6 0.8 2.1 9.1 70.2 46.6

Saudi

Arabia 6.1 5.9 12.2 16.2 9.1 1.9 3.9 3.6 1.0 2.2 7.2 6.5 75.8 24.8

Somalia 2.8 3.5 11.4 53.6 53.7 16.3 14.7 10.4 13.4 42.5 31.2 15.7 269.4 34.9

Sudan 1.0 2.2 7.5 21.0 43.5 56.1 91.3 109.4 65.8 34.5 7.8 2.3 442.4 42.6

Syria 55.1 47.2 40.6 31.7 13.6 2.3 0.4 0.3 1.5 15.6 33.7 50.2 292.1 26.2

Tunisia 33.7 28.9 28.9 23.1 15.5 8.0 2.6 5.7 21.1 30.5 31.5 32.8 262.3 33.7

United

Arab

Emirates

12.1 14.0 15.7 7.4 2.1 0.2 0.2 0.2 0.0 0.3 3.9 12.5 68.6 52.9

West Bank

and Gaza 129.1 99.1 60.2 19.7 3.2 0.0 0.0 0.0 0.0 14.2 59.5 113.0 498.0 29.3

Yemen 7.2 8.3 16.4 24.0 17.5 9.9 27.2 28.4 11.9 6.0 8.8 9.2 174.8 28.3

Source: Westphal (2010).

APPENDIX B.3 MEAN MONTHLY TEMPERATURE (°C) FOR THE ARAB WORLD

Country January February March April May June July August September October November December Mean

Annual

Algeria 12.0 14.6 18.0 22.0 26.4 30.8 32.6 31.9 28.7 23.5 17.5 13.0 22.6

Bahrain 15.9 17.9 21.8 26.9 32.3 34.9 36.6 36.1 33.3 28.7 23.0 18.0 27.1

Comoros 26.0 25.9 26.2 25.9 25.0 23.5 22.8 23.1 23.9 25.1 26.0 26.1 25.0

Djibouti 23.5 24.4 26.0 27.7 29.4 31.6 32.4 31.7 30.2 27.6 25.4 23.8 27.8

Egypt 13.2 14.8 17.9 22.4 26.5 28.8 29.3 29.4 27.5 24.6 19.3 14.7 22.4

Iraq 9.0 11.1 15.0 20.7 26.5 30.8 33.3 33.1 29.5 23.9 16.3 10.5 21.7

Jordan 8.4 9.9 13.0 17.6 22.1 25.0 26.8 27.1 25.0 21.1 15.1 10.1 18.4

Kuwait 12.4 14.5 19.0 24.2 30.5 34.5 36.0 35.8 32.7 27.1 20.0 14.1 25.1

Lebanon 6.5 7.0 9.6 13.4 17.2 20.5 22.7 23.3 21.4 18.1 12.8 8.0 15.0

Libya 12.5 14.7 17.9 22.1 26.4 29.1 29.4 29.2 27.4 23.5 18.2 13.8 22.0

Mauritania 20.2 22.2 25.0 28.0 31.3 33.4 33.3 32.2 31.2 29.4 25.0 20.6 27.7

Morocco 9.4 10.8 13.1 15.3 18.8 22.7 26.0 26.1 22.6 18.2 13.7 10.3 17.3

Oman 19.9 20.7 23.3 26.6 29.4 29.9 28.5 27.4 27.3 26.1 23.7 21.4 25.3

Qatar 18.4 19.3 22.4 26.5 30.6 32.4 33.7 33.5 31.6 28.2 24.0 20.3 26.8

Saudi 15.4 17.1 20.4 24.5 28.9 31.5 32.0 32.1 30.2 25.8 20.9 16.8 24.6

76

Arabia

Somalia 25.1 25.9 27.3 28.1 28.3 27.9 27.1 27.1 27.7 26.8 25.8 25.1 26.9

Sudan 22.1 23.7 26.4 28.9 30.2 30.1 28.9 28.4 28.5 28.0 25.2 22.6 26.9

Syria 6.1 7.8 11.3 16.3 21.6 26.0 28.8 28.7 25.3 20.0 13.0 7.7 17.7

Tunisia 9.9 11.4 14.3 17.4 21.5 25.9 28.6 28.7 25.7 20.8 15.3 11.3 19.2

United

Arab

Emirates

17.1 18.3 21.6 26.2 31.2 33.9 35.0 34.7 32.5 28.9 24.1 19.3 26.9

West Bank

and Gaza 10.8 11.6 13.7 17.3 21.0 23.5 25.4 25.9 24.5 21.9 17.1 12.6 18.8

Yemen 18.4 19.3 20.9 23.6 26.0 27.9 27.0 26.8 26.1 23.1 20.4 19.2 23.2

Source: Westphal (2010).

REFERENCES

Almazroui, M: Sensitivity of a regional climate model on the simulation of high intensity rainfall

events over the Arabian Peninsula and around Jeddah (Saudi Arabia) Theor Appl Climatol

(2011) 104:261–276, DOI: 10.1007/s00704-010-0387-3

Alderwish, A. and El-Eryani, M., 1999: An approach for assessing the vulnerability of the water

resources of Yemen to climate change. Clim. Res. 12, 85-89.

AMAP (2009) The Greenland Ice Sheet in a Changing Climate: Snow, water, ice and permafrost

in the Arctic (SWIPA) Edited by Dahl-Jensen D, Bamber J, Bøggild CE, Buch E, Christensen

JH, Dethloff K, Fahnestock M, Marshall S, Rosing M, Steffen K, Thomas R, Truffer M, van den

Broeke CJ. Arctic Monitoring and Assesment Programme (AMAP), Oslo, 115 pp.

Black, E., 2009: The impact of climate change on daily precipitation statistics in Jordan and

Israel. Atm. Sci. Lett. 10, 192-200, doi: 10.1002/asl.233

Black, E., 2010: Water and society in Jordan and Israel today: an introductory overview. Phil.

Trans. R. Soc. A 368,5111-5116, doi: 10.1098/rsta.2010.0217.

Bonsal, B.R., X. Zhang, L.A. Vincent and W.D. Hogg, 2001: Characteristics of daily and

extreme temperatures over Canada. J. Climate 14, 1959-1976.

Chaponniere, A. and V. Smakhtin, 2006: A review of climate change scenarios and preliminary

rainfall trend analysis in the Oum er Rbia Basin, Morocco. Working Paper 110 (Drought Series:

Paper 8), International Water Management Institute (IWMI), Colombo, Sri Lanka, 16 pp.

Christensen, J.H. and O.B. Christensen, 2003: Severe Summer Flooding in Europe, Nature, 421,

805-806.

Christensen, J.H., B. Hewitson, A. Busuioc, A. Chen, X. Gao, I. Held, R. Jones, R.K. Kolli, W.-

T. Kwon, R. Laprise, V. Magaña Rueda, L. Mearns, C.G. Menéndez, J. Räisänen, A. Rinke, A.

Sarr and P. Whetton, 2007: Regional Climate Projections, Chapter 11 in: Climate Change 2007:

The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report

of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen,

M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.)]. Cambridge University Press,

Cambridge, United Kingdom and New York, NY, USA.

Christensen JH, Carter TR, Rummukainen M, Amanatidis G. 2007b. Evaluating the performance

and utility of regional climate models: the PRUDENCE project. Climatic Change 81(Suppl 1):1-

6. DOI:10.1007/s10584-006-9211-6.

Christensen JH, Boberg F, Christensen OB, Lucas-Picher P. 2008: On the need for bias

correction of regional climate change projections of temperature and precipitation. Geophys.

Res. Lett., 35, L20709, doi:10.1029/2008GL035694.

Christensen JH, Rummukainen M. 2009. Choosing ENSEMBLES regional Climate Model

(RCM) simulations and projections for regional climate change and impact analyses. Appendix

for ENSEMBLES deliverable D3.2.2. Available at http://ensembles-eu.metoffice.com/

Christensen JH, Kjellström E, Giorgi F, Lenderink G, Rummukainen M. 2010. Clim. Res. 44,

179-194. doi:10.3354/cr00916.

78

Christensen, O.B. and J.H. Christensen, 2004: Intensification of extreme European summer

precipitation in a warmer climate. Global and Planetary Change, 44, 107-117.

Coppola, E., F. Kucharski, F. Giorgi and F. Molteni, 2005: Bimodality of the North Atlantic

Oscillation in simulations with greenhouse gas forcing. Geophys. Res. Lett. 32, L23709.

Cullen, H.M., A. Kaplan, P.A. Arkin and P.B. de Menocal, 2002: Impact of the NAO on Middle

Eastern climate and streamflow. Clim. Change 55, 315-338.

Dai, A., 2010: Drought under global warming: a review. WIREs Clim. Change 2, 45-65, dooi:

10.1002/wcc81.

Evans, J.P., 2009: 21st century climate change in the Middle East. Clim. Change 92, 417-432,

doi: 10.1007/s10584-008-9438-5.

Evans, J.P., 2010: Global warming impact on the dominant precipitation processes in the Middle

East. Theor. Appl. Climatol. 99, 389-402, doi: 10.1007/s00704-009-0151-8.

Evans, J. and R. Geerken, 2004: Discrimination between climate and human-induced dryland

degradation , J. Arid Environm. 535-554.

Freiwan, M. and M. Kadioglu, 2008a: Climate variability in Jordan. Internat. J. Climatol. 28, 68-

89.

Freiwan, M. and M. Kadioglu, 2008b: Spatial and temporal analysis of climatological data in

Jordan. Internat. J. Climatol. 28, 521-535.

Frich, P., et al., 2002: Observed coherent changes in climatic extremes during the second half of

the twentieth century. Clim. Res., 19, 193–212.

Gao, X.J andF. Giorgi 2008 Increased aridity in the Mediterranean region under greenhouse gas

forcing estimated from high resolution simulations with a regional climate model Global and

Planetary Change 62 3-4 195-209

Giorgi, F., C. Jones, and G. R. Asrar, 2009: Addressing climate information needs at the regional

level: The CORDEX framework. WMO Bull., 58, 175–183.

Haylock, M.R., N. Hofstra, A.M.G. Klein Tank, E.J. Klok, P.D: Jones and M. New, 2008: A

European daily high-resolution gridded data set of surface temperature and precipitation for

1950-2006. J. Geoph. Res. 113, D20119

Hurrell, J., T. Delworth, G. Danabasoglu, H. Drange,K. Drinkwater, S. Griffies, N.Holbrook, B.

Kirtman, N. Keenlyside,M. Latif, J. Marotzke, J. Murphy,G. Meehl, T. Palmer, H. Pohlmann, T.

Rosati, R. Seager, D. Smith, R. Sutton, A. Timmermann, K. Trenberth,J. Tribbia, and M.

Visbeck, 2010: Decadal Climate Prediction: Opportunities and Challenges in Proceedings of

OceanObs‘09: Sustained Ocean Observations and Information for Society (Vol. 2), Venice, Italy,

21-25 September 2009, Hall, J., Harrison, D.E. & Stammer, D., Eds., ESA Publication WPP-306,

doi:10.5270/OceanObs09.cwp.45.

IPCC, 2007: Summary for Poicymakers. In: Climate Change 2007: The Physical Science Basis.

Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental

Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B.

79

Averyt, M.Tignor and H.L. Miller (eds.)]. Cambridge University Press, Cambridge, United

Kingdom and New York, NY, USA.

IOP, 2009: Climate Change: Global Risks, Challenges and Decisions, Copenhagen 10.-12.

March 2009. All sessions. IOP Conference Series: Earth and Environmental Series. Available

online at: http://www.iop.org/EJ/volume/1755-1315/6.

Johns, T.C, J.-F. Royer, I. Ho¨schel, H. Huebener, E. Roeckner, E. Manzini, W. May, J.-L.

Dufresne, O. H. Ottera°, D. P. van Vuuren, D. Salas y Melia, M. A. Giorgetta, S. Denvil, S.

Yang, P. G. Fogli, J. Körper, J. F. Tjiputra, E. Stehfest, C. D. Hewitt, 2011: Climate change

under aggressive mitigation: the ENSEMBLES multi-model experiment. Clim Dyn., DOI

10.1007/s00382-011-1005-5.

Kim, D.-W. and H.R. Byun, 2009: Future patterns of Asian drought under global warming

scenario. Theor. Appl. Climatol. 98, 138-150.

Kitoh, A., A. Yatagai and P. Alpert, 2008: First super-high resolution model projection that the

ancient ―Fertile Crescent‖ will disappear in this century. Hydrol. Res. Lett. 2, 1-4.

Kjellström E, Giorgi F. 2010. Introduction. Clim Res 44:117-119.

Knippertz, P., 2003: Tropical-extratropical interactions causing precipitation in Northwest

Africa: Statistical analysis and seasonal variations. Mon. Wea. Rev. 131, 3069-2076

Knippertz, P., 2004: A simple identification scheme for upper-level troughs and its application to

winter precipitation variability in Northwest Africa. J. Climate 17, 1411-1418

Knippertz, P., M. Christoph and P. Speth, 2003: Long-term precipitation variability in Morocco

and the link to the large-scale circulation in recent and future climates. Meteorol. Atmosph. Phys.

83, 67-88.

Krichak, S.O., P. Alpert and P. Kunin, 2010: Numerical simulation of seasonal distribution of

precipitation over the eastern Mediterranean with a RCM. Clim. Dyn. 34, 47-59.

Kummerow, C., et al., 2000. The status of the Tropical Rainfall Measuring Mission (TRMM)

after two years in orbit. J. Appl.Meteorol. 39, 1965–1982.

Latif, M., T. Delworth, D. Dommenget, H. Drange, W. Hazeleger, J. Hurrell, N. Keenlyside, G.

Meehl and R. Sutton, 2010: Dynamics of Decadal Climate Variability and Implications for its

Prediction in Proceedings of OceanObs‘09: Sustained Ocean Observations and Information for

Society (Vol. 2), Venice, Italy, 21-25 September 2009, Hall, J., Harrison, D.E. & Stammer, D.,

Eds., ESA Publication WPP-306, doi:10.5270/OceanObs09.cwp.53

Lionello, P. and F. Georgi, 2007: Winter precipitation and cyclones in the Mediterranean region:

future climate scenarios in a regional simulation. Adv. Geoscience 12, 153-158.

Mahwed, K., 2008: Vulnerability assessment of the climate sector. Enabling activities for

preparation of Syria‘s initial communication to the UNFCCC (Project no. 00045323), 47 pp.

Mearns, LO, Hulme M, Carter TR., Leemans R, Lal M, Whetton P, Hay L, Jones RN, Katz R,

Kittel T, Smith J, Wilby R, Mata, LJ, Zillman J. 2001: Climate Scenario Development. In

Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third

Assessment Report of the Intergovernmental Panel on Climate Change [Houghton JT, Ding Y,

80

Griggs DJ, Noguer M, van der Linden PJ, Dai X, Maskell K, Johnson CA (eds.)]. Cambridge

University Press, Cambridge, United Kingdom and New York, NY, USA, 881 pp.

Meehl, G.A., T.F. Stocker, W.D. Collins, P. Friedlingstein, A.T. Gaye, J.M. Gregory, A. Kitoh,

R. Knutti, J.M. Murphy, A. Noda, S.C.B. Raper,I.G. Watterson, A.J. Weaver and Z.-C. Zhao,

2007: Global Climate Projections. In: Climate Change 2007: The Physical Science Basis.

Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental

Panel on Climate Change [Solomon, S.,D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt,

M. Tignor and H.L. Miller (eds.)]. Cambridge University Press, Cambridge, United Kingdom

and New York, NY, USA.

Meehl, G. A., L. Goddard, J. Murphy, R. J. Stouffer, G. Boer, G. Danabasoglu, K. Dixon, M. A.

Giorgetta, A. Greene, E. Hawkins, G. Hegerl, D. Karoly, N. Keenlyside, M. Kimoto, B. Kirtman,

A. Navarra, R. Pulwarty, D. Smith, D. Stammer, and T. Stockdale (2009): Decadal prediction:

Can it be skillful? Bull. Amer. Meteorol. Soc., DOI:15 10.1175/2009BAMS2778.1.

Moss RH, Babiker M, Brinkman S, Calvo E, Carter T, Edmonds J, Elgizouli I, Emori S, Erda L,

Hibbard K, Jones R, Kainuma M, Kelleher J, Lamarque JF, Manning M, Matthews B, Meehl G,

Meyer L, Mitchell J, Nakicenovic N, O‘Neill B, Pichs T, Riahi K, Rose S, Runci P, Stouffer R,

van Vuuren D, Weyant JW, T., van Ypersele JP, Zurek M, 2008: Towards New Scenarios for

Analysis of Emissions, Climate Change, Impacts, and Response Strategies. Intergovernmental

Panel on Climate Change, Geneva.

Moss RH, Edmonds JA, Hibbard KA, Manning MR, Rose SK, van Vuuren DP, Carter TR,

Emori S, Kainuma M, Kram T, Meehl GA, Mitchell JFB, Nakicenovic N, Riahi K, Smith SJ,

Stouffer RJ, Thomson AM, Weyant JP, Wilbanks TJ, 2010: The next generation of scenarios for

climate change research and assessment. Nature 463:747-756.

Mote, T.L.,2007: Greenland surface melt trends 1973 – 2007: Evidence of a large increase in

2007, Geophys. Res. Lett., 34, L22507, doi: 10.1029/2007GL031976.

Murphy, J., Kattsov, V., Keenlyside, N., Kimoto, M., Meehl, G., Mehta, V., Pohlmann, H.,

Scaife, A., and Smith, D., 2010: Towards Prediction of Decadal Climate Variability and Change.

Procedia Environmental Sciences, 1, 287-56 304, doi:10.1016/j.proenv.2010.09.018.

Nasrallah, H.A. and R.C. Balling, 1993: Spatial and temporal analysis of Middle-Eastern

temperature changes. Clim. Change 25, 153-161.

Olesen JE, Carter TR, Diaz-Ambrona CH, Fronzek S, Heidmann T, Hickler T , Holt T, Minguez

MI, Morales P, Palutikof J, Quemada M, Ruiz-Ramos M, Rubæk G, Sau F, Smith B, Sykes M.

2007. Uncertainties in projected impacts of climate change on European agriculture and

ecosystems based on scenarios from regional climate models. Climatic Change 81, Suppl. 1,

123-143.

Önol, B. and F.H.M. Semazzi, 2009: Regionalization of climate change simulations over the

Eastern Mediterranean. J. Climate 22, 1944-1961

Oroud, I.M., 2008: The impacts of climate change on waterresources in Jordan. In: Zereini, F.

and H. Hotzl (eds.): Climatic changes and water resources in the Middle East and North Africa,

Springer, New York.

81

Raible, C.C., B. Ziv, H. Saaroni and M. Wild, 2010: Winter synoptic-scale variability over the

Mediterranean Basin under future climate conditions as simulated by the ECHAM5. Clim. Dyn.

35, 473-488.

Racsko P, Szeidl L, Semenov M. 1991. A serial approach to local stochastic weather models.E

col. Model, 57:27-41.

Rauthe, M., A. Hense and H. Paeth, 2004: A model intercomparison study of climate change

signals in extratropical circulation. Global and Environmental Change – Human and Policy

Dimensions 9, 45-58.

Reifen C, Toumi R. 2009. Geophys. Res. Letters 36, L13704, 10 doi:10.1029/2009GL038082.

Rummukainen M. 2010. State-of-the-art with regional climate models. Wiley Interdisciplinary

Reviews: Climate Change, 1:82–96. doi:10.1002/wcc.8.

Schneider U, Fuchs T, Meyer-Christoffer A, Rudolf B (2008) Global precipitation analysis

products of the GPCC. Global Precipitation Climatology Centre (GPCC), DWD, Internet

Publikation, 1–12.

Semenov MA, Brooks RJ, Barrow EM, Richardson CW. 1998. Comparison of the WGEN and

LARS-WG stochastic weather generators for diverse climates. Climate Research 10, 95-107

Sensoy, S., T.C. Peterson, L.V. Alexander and Z. Xuebin, 2007: Enhancing Middle East climate

change monitoring and indexes. Bull. Amer. Met. Soc. 88, 1249-1254.

Shahin, M., 2007: Water resources and hydrometeorology of the Arab region. Water Science and

Technology Library, vol. 59, Springer, The Netherlands, 586 pp.

Simpson, J., Adler, R.F., North, G.R., 1988. A proposed Tropical Rainfall Measuring Mission

(TRMM) satellite. Bull. Am. Meteorol. Soc. 69, 278–295.

Smadi, M.M. and A. Zghoul, 2006: A sudden change in rainfall characteristics in Amman,

Jordan during the mid-1950s. Amer. J. Environm. Changes 2, 84-91.

Sowers, J., A. Vengosh and E. Weinthal, 2011: Climate change, water resources and the politics

of adaptation in the Middle East and North Africa. Clim. Change 104, 599-627, doi:

10.1007/s10584-010-9835-4.

Steffen, K., and Huff, R.,2009: University of Colorado at Boulder, personal communication.

Taylor, K.E.,R.J. Stouffer, G.A. Meehl, 2011: A Summary of the CMIP5 Experiment Design.

(Available from http://cmip-pcmdi.llnl.gov/cmip5/).

Tebaldi, C., K. Hayhoe, J.M. Arblaster, and G.A. Meehl, 2006: Going to the extremes: An

intercomparison of model-simulated historical and future changes in extreme events. Clim.

Change, 79, 185-211

Tryhorn, L., A. Lynch, and R. Abramson, 2008: On the meteorological mechanisms driving

postfire flash floods: A case study. Monthly Weather Review, 136, 1778-1791.

Tschoegl, L., R. Below, Guhar-Sapir, D. 2006: An Analytical Review of Selected Data Sets on

Natural Disasters and Impacts. Centre for Research on the Epidemiology of Disasters (CRED),

Université catholique de Louvain, Belgium. (Available at http://www.em-

dat.net/documents/Publication/TschoeglDataSetsReview.pdf).

82

Van den Brink, H.W., G.P. Konnen, J.D. Opsteegh, G.J. Van Oldenborgh, and G. Burgers, 2005:

Estimating return periods of extreme events from ECMWF seasonal forecast ensembles.

International Journal of Climatology, 25(10), 1345-1354.

van der Linden P, Mitchell JFB (eds) (2009) ENSEMBLES: climate change and its impacts.

Summary of research and results from the ENSEMBLES project. Met Office Hadley Centre,

Exeter.

van Vuuren, D., K. Riahi, R.H. Moss, J. Edmonds, A. Thomson, N. Nakicenovic, T. Kram, F.

Berkhout, R. Swart, A. Janetos, S. Rose and N. Arnell, 2010: Developing new scenarios as a

common thread for future climate research. (Available from http://www.ipcc-

wg3.de/meetings/expert-meetings-and-workshops/files/Vuuren-et-al-2010-Developing-New-

Scenarios-2010-10-20.pdf/view?searchterm=New%20Scenarios).

Westphal, MI (2010) The Arab World historical climate and future climate projections, WB

background paper.

Wilby, R.L., 2007: Climate change scenarios for Morocco. Final Technical Report, WB

background paper.

Wilby, R.L., 2008: Climate change scenarios for the Republic of Yemen. Revised Technical

Report, WB background paper.

Wilby, R.L., 2010: Climate change projections and downscaling for Jordan, Lebanon and Syria.

Draft Synthesis Report, WB background paper.

Wilby, R.L., J. Troni, Y. Biot, L. Tedd, B.C. Hewitson, D.M. Smith and R.T.Sutton, 2008: A

review of climate risk information for adaptation and development planning. Int. J. Climatology

29, 1193-1215.

Yatagai, A., P. Xie and P. Alpert (2008): Development of a daily gridded precipitation data set

for the Middle East, Adv.in Geosci. 12, 165-170.

Zhang, X., E. Aguilar, S. Sensoy, H. Melkonyan, U. Tagiyeva, N. Ahmed, N. Kutaladze, F.

Rahimzadeh, A. Taghipour, T.H. Hantosh, P. Alpert, M. Semawi, M.K. Ali, M.H.S. Al-Shabibi,

Z. Al-Oulan, T Zatari, I. Al Dean Khelet, S. Hamoud, R. Sagir, M. Demircan, M. Eken, M.

Adiguzel, L. Alexander, T.C. Peterson and T. Wallis, 2005: Trends in Middle East climate

extreme indices from 1950 to 2003. J. Geophys. Res. 110, D22104, doi:10.1029/2005JD006181.