© 2002 International Center for Agricultural Research

in the Dry Areas (ICARDA)

All rights reserved.

ICARDA encourages the fair use of this material.

Proper citation is requested.

The author: Eddy De Pauw is Agroecologist at the International Center for

Agricultural Research in the Dry Areas (ICARDA).

Recommended Citation: De Pauw, E. 2002. An agroecological exploration of

the Arabian Peninsula. ICARDA, Aleppo, Syria, 77 pp.

ISBN: 92-9127-119-5

Key words: agroecology, desertification, GIS, Arabian Peninsula, arid zones

Front cover photograph

Landscape in Ahfara, Fujairah, United Arab Emirates. The relative lushness of
the vegetation is the result of above-normal rainfall. Pronounced rainfall
variability, both in space and time, is a key characteristic of climate in the
Arabian Peninsula, and responsible for tremendous variations in rangeland
productivity.

ICARDA
P.O.Box 5466, Aleppo, Syria
Tel.: (963-21) 2213433, 2225112, 2225012
Fax: (963-21) 2213490 / 2225105 / 2219380
E-Mail: ICARDA@CGIAR.ORG
Web site: http://www.icarda.cgiar.org
-3-
 Foreword
The Arabian Peninsula constitutes one of the largest contiguous arid zones in the
world. Within this context of aridity the region is surprisingly diverse in climatic
conditions, soil types, landscapes, and land use patterns. Agroecological niches occur
with edaphic conditions that can deviate substantially from those of surrounding areas
and often have a higher biomass or agricultural productivity. In the context of global
climate change this agroecological diversity is also an important source of abioticstress
resistance in plants against drought, high temperatures, and salinity.
However, the region is also ecologically fragile. Firstly, it is vulnerable to natural
processes, such as primary salinization, and wind and water erosion. In addition,
overgrazing, fuel-wood extraction, drought, and depletion of fossil water resources are
increasingly threatening the sustainability of the natural resource base, and may lead to
potentially irreversible desertification.
A rational approach to combating desertification requires in the first place
differentiating true degradation, as a result of over-exploitation, from processes and
conditions that are the natural outcome of the biophysical limits imposed by the harsh
climates that prevail in the Arabian Peninsula. Such an approach necessitates the
development of agroecological frameworks, which allow assessing the spatial and
temporal variations in the natural resource base and associated land use systems.
To combat desertification effectively, a good agroecological characterization is of vital
importance. Numerous thematic surveys in the form of soil survey reports, climatic
maps, and groundwater surveys already exist in the Arabian Peninsula. However, the
challenge is to develop integrated land and water resource information systems, based
on GIS-technology. This integration will allow linkage of multidisciplinary,
geographically referenced databases at different resolutions, and to develop decision-
support systems for more sustainable land use management and resource use
regulations.
By bringing together information sources from the international public domain
and the Arabian Peninsula itself, and processing them with state-of-the-art GIS
technology, this report aims to initiate this process of data integration at the regional
level. As such it will be of value for agricultural research planning, biodiversity
management, land use planning, and public awareness at the national and regional
level. We hope it will fill a major gap in our understanding of the agroecological
diversity, vulnerability, and agricultural productivity of one of the most important arid
regions in the world.

Prof. Dr Adel El-Beltagy

Director General, ICARDA

-iii-
 Acknowledgements
The author wishes to thank the following people and organizations for their
support and background efforts in the production of this technical bulletin.
• Dr John Peacock and Dr Ahmed Mustafa Tawfik, Regional Coordinators of
the Arabian Peninsula Program.
• The staff of the national agricultural research institutes in the region, who
have taken time and effort to provide information to the author. They are
too numerous to mention individually, but are listed in the Feasibility
Report.
• The staff of CODIS, at ICARDA, for polishing the manuscript.
• The donors of APRP, notably the Arab Fund for Economic and Social
Development, IFAD and OPEC, for their financial support.

-iv-
 Abstract
This report provides an overview of the agroecological characteristics of the
Arabian Peninsula.
The Arabian Peninsula is a vast plateau bounded by mountainous terrain. It
can be subdivided into 15 geomorphological regions. Its main characteristic is
aridity, due to low and erratic rainfall, and high temperatures. As a result,
productivity of rangelands and agriculture is variable and poor. The interaction
of temperature and precipitation gradients leads to a great diversity of climatic
conditions, which is evidenced by 22 distinct agroclimatic zones, of which eight
comprise 95% of the region.
The soils of the Arabian Peninsula reflect the general aridity of the climate.
Most are poorly developed, shallow, or are enriched in lime, gypsum, or salts.
Transported materials, such as sand dunes and sheets, cover large areas. That
said, there is no shortage of good agricultural soils. The obvious limitation to
put them into production is water availability. Where irrigation water is
available, standard fertility management practices are required, and, if
provided, allow maintenance and enhancement of soil quality.
Cropped areas are very limited in the Peninsula and most are irrigated,
although substantial rainfed areas exist in Yemen and in Saudi Arabia. Between
1980 and 1996, area under irrigation more than doubled, aided by modern
irrigation technology, such as center-pivot and drip irrigation. This use of fossil
groundwater, however, is not sustainable.
The Arabian Peninsula is perceived as having limited heterogeneity, poor
agricultural potential, and low population densities, and, therefore, it has
generated limited interest with regard to global biodiversity. This view is
oversimplified. The region has great agroecological diversity and much
potential as a source of genetic diversity and of abiotic stress resistance. To
achieve this goal there is a clear need to integrate existing thematic datasets into
agroecological frameworks for development. Specific methodologies, models,
and decision-support systems must be developed to achieve and make use of
this integration.
Priority should be given to the regional assessment of crop water
requirements with a view to enhancing water use efficiency, and agroecological
zoning for biodiversity conservation, rangeland rehabilitation, abiotic stress
identification, and development planning. Underpinning these research goals
should be a strengthening of climate monitoring networks.

-v-
 Contents
Foreword iii
Acknowledgements iv
Abstract v
1. Introduction 1
2. Human geography 3
3. Relief and geomorphology 5
4. Climate 11
4.1. General 11
4.2. Precipitation 12
4.2.1. Types and amounts 12
4.2.2. Seasonal patterns 12
4.2.3. Variability 14
4.3. Temperature 17
4.4. Evaporation and water deficit 22
4.4.1. Potential evapotranspiration 22
4.4.2. Aridity 22
4.5. Agroclimatic patterns 25
4.5.1. Agroclimatic zones 25
4.5.2. Similarity of climatic conditions 27
4.6. Climatic growing period 35
4.6.1. Types of growing period 35
4.6.2. Duration and onset of the growing period 38
4.7. Biomass productivity and climate 39
5. Soils of the Arabian Peninsula 44
5.1. General soil pattern 44
5.2. Soil management properties 49
5.2.1. Soil texture 49
5.2.2. Soil depth and stoniness 51
5.2.3. Soil fertility indicators 53
5.3. Conclusions 60
6. Agricultural production systems 60
7. Research priorities in agroecological characterization 65
7.1. Regional assessment of crop water requirements 65
7.2. Agroecological zoning 65

-vi-
 7.3. Improved climate monitoring 67
8. Methods and data sources 68
8.1. General maps 68
8.2. Relief and geomorphology 68
8.3. Climatic maps 68
8.4. Soil maps 75
8.5. Land use and cover maps 75
References 76

List of figures
Figure 1 Arabian Peninsula: General 4
Figure 2 Population density 6
Figure 3 Altitude 7
Figure 4 Elevation range and geomorphological regions 8
Figure 5 Wadi network in the northwest of the Arabian Peninsula
(Note: Wadis shown as blue lines. Elevation range is the
same as in Figure 3) 10
Figure 6 Mean annual precipitation (mm) 13
Figure 7a Seasonal distribution of precipitation. winter 15
Figure 7b Seasonal distribution of precipitation. spring 15
Figure 7c Seasonal distribution of precipitation. summer 15
Figure 7d Seasonal distribution of precipitation. autumn 15
Figure 8 Variability of annual rainfall, Muscat, Oman (1893-1978) 16
Figure 9 Probability distribution of annual rainfall, Muscat,
Oman (1893-1978) 16
Figure 10 Variability of monthly rainfall, Muscat, Oman (1893-1978) 17
Figure 11 Areas of the Arabian Peninsula with mean annual
temperature exceeding 30°C (in red), or below 20°C
(in magenta) 18
Figure 12 Mean annual temperature 19
Figure 13 Climate diagram for Salalah, Oman 20
Figure 14 Mean temperature of the coldest month 20
Figure 15 Mean temperature of the warmest month 20
Figure 16 Mean annual heat units 21
Figure 17 Mean annual potential evapotranspiration (mm) 23
Figure 18 Aridity index 24

-vii-
Figure 19 Annual precipitation deficit (mm) 24
Figure 20 Winter types (blue: cool; green: mild; yellow: warm) 26
Figure 21 Summer types (green: mild; yellow: warm; red: very warm) 27
Figure 22 Agroclimatic zones (classified according to UNESCO, 1979) 28
Figure 23a Station representative of agroclimatic zone A-M-W: Abha,
Saudi Arabia 29
Figure 23b Station representative of agroclimatic zone HA-M-VW:
Al Jouf, Saudi Arabia 29
Figure 23c Station representative of agroclimatic zone A-W-VW:
El Kod, Yemen 29
Figure 23d Station representative of agroclimatic zone SA-M-M:
Mabar, Yemen 29
Figure 23e Station representative of agroclimatic zone HA-W-VW:
Muscat, Oman 30
Figure 23f Station representative of agroclimatic zone A-M-VW:
Riyadh, Saudi Arabia 30
Figure 23g Station representative of agroclimatic zone HA-M-W:
Tabuk, Saudi Arabia 30
Figure 23h Station representative of agroclimatic zone SA-M-W: Taiz,
Yemen 30
Figure 24a Similarity in temperature and precipitation pattern
with Abha, Saudi Arabia 31
Figure 24b Similarity in temperature and precipitation pattern with
Al Jouf, Saudi Arabia 31
Figure 24c Similarity in temperature and precipitation pattern with
El Kod, Yemen 32
Figure 24d Similarity in temperature and precipitation pattern
with Mabar, Yemen 32
Figure 24e Similarity in temperature and precipitation pattern
with Muscat, Oman 33
Figure 24f Similarity in temperature and precipitation pattern
with Riyadh, Saudi Arabia 33
Figure 24g Similarity in temperature and precipitation pattern
with Tabuk, Saudi Arabia 34
Figure 24h Similarity in temperature and precipitation pattern
with Taiz, Yemen 34
Figure 25 Types of growing period 36

-viii-
Figure 26a Example of an all-year-round dry period
(Al Jouf, Saudi Arabia) 37
Figure 26b Example of an intermediate growing period (Taiz, Yemen) 37
Figure 26c Example of a normal growing period (Aleppo, Syria) 38
Figure 27a Length of the first growing period, Yemen Highlands 38
Figure 27b Length of the second growing period, Yemen Highlands 39
Figure 27c Onset of the main growing period, Yemen Highlands 39
Figure 28 Temperature adaptability ranges for different crop groups 40
Figure 29a Biomass productivity index for crop group I,
Yemen Highlands 42
Figure 29b Biomass productivity index for crop group II,
Yemen Highlands 42
Figure 29c Biomass productivity index for crop group III,
Yemen Highlands 42
Figure 29d Biomass productivity index for crop group IV,
Yemen Highlands 42
Figure 30 Rangeland biomass productivity index 43
Figure 31 Soil associations 45
Figure 32 Soil associations: explanation of legend 47
Figure 33 Simplified soil map (dominant soils) 48
Figure 34 Legend of soil property distribution maps 50
Figure 35 USDA textural triangle and simplified textural classes 50
Figure 36 Distribution of coarse-textured soils 51
Figure 37 Distribution of medium-textured soils 52
Figure 38 Distribution of fine-textured soils 52
Figure 39 Dominant soil depth 53
Figure 40 Distribution of shallow soils 54
Figure 41 Distribution of gravelly and stony soils 54
Figure 42 Dominant organic carbon levels 55
Figure 43 Organic carbon pool 56
Figure 44 Dominant soil pH 57
Figure 45 Dominant CEC 57
Figure 46 Distribution of calcareous soils 58
Figure 47 Distribution of soils with hardened lime or gypsum 59
Figure 48 Distribution of soil salinity 59
Figure 49 Land use and land cover 61
Figure 50 Evolution of desert irrigation (1982-1993) 62

-ix-
Figure 51a Expansion of desert irrigation, observed from
AVHRR imagery(1983) 63
Figure 51b Expansion of desert irrigation, observed
from AVHRR imagery(1986) 63
Figure 51c Expansion of desert irrigation, observed from
AVHRR imagery(1990) 63
Figure 51d Expansion of desert irrigation, observed
from AVHRR imagery(1993) 63
Figure 52 Spatial pattern of desert irrigation 1982-1993 64
Figure 53 Mapping the distribution of plant communities or
species using landscape frameworks. 67

List of tables

Table 1 Population in the Arabian Peninsula 3

Table 2. Geomorphological regions of the Arabian Peninsula 9
Table 3. Areas under different moisture regimes 22
Table 4. Areas under different winter types 25
Table 5. Areas under different summer types 26
Table 6. Extent of agroclimatic zones of the Arabian Peninsula 29
Table 7. Adaptability ranges of different crop groups 40
Table 8. Main soil associations of the Arabian Peninsula 46
Table 9. Main soil types of the Arabian Peninsula 47
Table 10. Cropland in the Arabian Peninsula 60
Table 11. Conversion equations from PET(Hargreaves) to
PET(Penman-Monteith) 69
Table 12. Adaptability to temperature for different crop groups
(adapted from FAO, 1978) 74

-x-
1. Introduction
The Arabian Peninsula, also called Arabia, is a vast landmass, covering about
2,590,000 km2. It is bounded by the Red Sea on the west and southwest, the
Arabian Sea on the south, and the Gulf of Oman and the Persian Gulf on the
northeast. It is composed of seven countries (Bahrain, Kuwait, Oman, Qatar,
Saudi Arabia, the United Arab Emirates, and Yemen).
Arabia is one of the driest subcontinents in the world. With an average
precipitation of less than 100 mm per year it can be considered a desert region.
It is also one of the hottest, with daytime temperatures often exceeding 50°C in
summer. Yet, despite its general aridity, its ecosystems are surprisingly diverse.
The rich biodiversity of the Arabian Peninsula is important to agriculture. The
plants that are able to survive in this harsh environment might be carriers of
traits useful in developing new drought and heat tolerant crop varieties.
However, the biodiversity of the Arabian Peninsula is under threat. An
inherent fragility of the environment, combined with over-exploitation of the
vegetation resources, has severely reduced the plant cover and narrowed the
species pool. Huge parts of the Peninsula are now completely bare, not because
of agroecological constraints, but because of overgrazing and fuelwood
extraction. Given the long time required for biomass production in arid
environments, recovery under conditions that do not provide total plant
protection might be close to impossible. This is the true meaning of
‘desertification’ in a desert environment.
The Arabian Peninsula has enormous reserves of groundwater. In many
parts of the Peninsula this precious resource has been exhausted in order to
maintain agricultural production systems, such as irrigated field crops, which
are essentially not adapted to the over-riding climatic constraint of hyper-
aridity. Such systems are unsustainable because they consume huge amounts of
water, where the supply is virtually non-renewable on a human time scale.
Combating desertification in the Arabian Peninsula requires good

-1-
information on the different environments. To some extent this information is
already available. Many environmental studies have been undertaken, which
have produced inventories of climatic, soil, terrain, vegetation, and water
resources. Depending on the investments made by the governments of the
region, the level of detail, updating, and integration varies considerably
between countries. In addition, the access of the general public to this
information is not always easy. As a result, it is difficult to obtain a synthesis of
the agroecology at the level of the whole Peninsula.
Much information on the environments of the Arabian Peninsula exists also
in the international community, in the form of books, journal articles, and
international databases. Putting national and international data sources together
in a concise booklet and integrating them through a Geographical Information
System (GIS) is the main subject of this publication.
Given the size and diversity of this subcontinent, this publication is
restricted to the level of ‘exploration,’ hence the title. Nevertheless, it is hoped
that the ‘bird’s-eye view’ it provides will be of value for agricultural research
planning, biodiversity management, land use planning, and public awareness at
the national and regional level. In short, the publication is meant to fill an
important data gap and permit a better understanding of the resource diversity
and environmental problems of the Arabian Peninsula.
This booklet is organized in several sections. The first gives a brief
overview of the human geography of the Peninsula. The second describes the
characteristics of the natural environment in terms of relief, climate, soils, land
use, and cover. It also addresses the problem of land degradation assessment. A
third section looks into the current status of agroecological characterization in
the Arabian Peninsula, identifying knowledge gaps, thematic research
priorities, and follow-up studies at the national and regional level.
This booklet is richly illustrated with maps. These maps were derived,
through GIS techniques and methods of agroecological characterization, from
the various data sources to which the author had access. Section eight briefly
describes the methods used in generating the maps and lists the data sources.
The maps in this publication are also available on a separate CD as GIS files
(ARCView shape files and grids), and can be imported into compatible GIS
software.

-2-
2. Human Geography
The Arabian Peninsula comprises the countries of Bahrain, Kuwait, Oman,
Qatar, Saudi Arabia, the United Arab Emirates (UAE), and Yemen.
Notwithstanding the low aptitude of the subcontinent to support high
population densities (see section on ‘Climate’), these countries have witnessed
a tremendous population increase over the last 50 years (Table 1).

Table 1. Population in the Arabian Peninsula

Country Population Population change, % Density, persons/km2
1950 2000 1975-80 1995-2000 Overall Rural

Bahrain 116,000 617,000 4.9 2.0 778 5,789

Kuwait 152,000 1,972,000 6.2 3.1 82 1,017
Oman 456,000 2,542,000 5.0 3.3 9 3,759
Qatar 25,000 599,000 5.8 1.8 53 667
Saudi Arabia 3,201,000 21,607,000 5.6 3.4 8 90
UAE 70,000 2,441,000 14.0 2.0 27 968
Yemen 4,316,000 18,112,000 3.2 3.7 27 690
Source: World Resources Institute (2001) URL: http://earthtrends.wri.org/country_profiles

In all Gulf countries, population increased between 1950 and 2000, by a

factor of 5 to 30. The more spectacular growth rates (e.g., Kuwait, Qatar, UAE)
are mostly due to a large net immigration of guest workers. However, the
population data for Yemen, a country of net emigration, shows that increase in
the native population is also a major contributing factor to population growth in
the Peninsula.
The distribution of the population centers is shown in Figure 1. With the
exception of the agricultural hinterland of Riyadh, Saudi Arabia, most of the
population lives at the edges of the Peninsula, near coastal areas or in the
mountains bordering the sea. This is to be expected. Given the aridity of the
interior, people have historically concentrated in areas with higher rainfall (e.g.,
Yemen, Asir), or along trade routes. As the rainfall map shows, the density of
population centers is associated with the higher rainfall areas. Very few people,
with the exception of nomadic tribes, live in the desert interior.

-3-
-4-

Fig. 1: Arabian Peninsula: General

 The population density is shown in Figure 2. The subcontinent is
characterized by generally low population density and high urbanization, with
pockets of high rural population density. A high proportion of the country
population still makes a living in agriculture in Yemen, and to a lesser extent in
Saudi Arabia. Low cropland area (Table 1) and urbanization of the countryside
(e.g., Kuwait, Bahrain, Oman, Qatar, UAE) might artificially raise the rural
population density above levels that bear any relationship with the intensity of
rural land use.

3. Relief and Geomorphology

The Arabian Peninsula is a vast plateau, gently sloping northeastward from the
Red Sea to the eastern lowlands bordering the Persian Gulf. It is bounded on the
west, south and east by mountainous terrain. According to the digital elevation
model (DEM) GTOPO30 (Gesch and Larson, 1996), the altitude varies between
–37 m in the lowest point, in the Matfi salt flat south of Qatar, and 3660 m at
the Peninsula’s highest peak, Jebel An Nabi Shu’ayb. The elevation map,
derived from this DEM, is shown in Figure 3.
Elevation does not really show a landscape. In addition to its elevation, a
landscape is defined by its degree of dissection, or the range between high and
low points. Plains are defined by a very small elevation difference between
neighboring points, rolling topography by a higher difference, and mountains
by a very large difference. The map of Figure 4, derived from the GTOPO30
DEM, captures the ‘ruggedness’ of the landscapes of the Arabian Peninsula. It
should be noted that this map evidences errors in the DEM used. The
checkerboard pattern in the central Rub-al-Khali desert in Saudi Arabia is due
to inadequate coverage by detailed topographical maps with elevation
benchmarks and does not constitute a natural pattern.
The blue lines on Figure 4 delineate 15 major geomorphological regions
(modified after Guba and Glennie, 1998, and Barth, 1976). The superimposed
areas in red are salt flats. The geomorphological regions are briefly described in
Table 2.
Among the most common and important landscape elements of the Arabian
Peninsula are its drainage channels. These seasonal watercourses, or wadis,
drain wide catchment areas and high mountains through networks of well-
developed tributaries, ravines and runnels. An example of the drainage network

-5-
-6-

Fig. 2: Population density

-7-

Fig. 3: Altitude
-8-

Fig. 4: Elevation range and geomorphological regions

Table 2. Geomorphological regions of the Arabian Peninsula (adapted from Guba and Glennie, 998)
symbol Name Altitude range Description
(m)
1 Coastal plain 0-200 Includes two sub-regions, the Tihama bordering the Red Sea,
and the Batinah bordering the Gulf of Oman. Both plains are
mostly pediments, gently sloping upward from sea level to the
foothills of the mountains.
2 Western Escarpment Mountains
2a Midian 300-2000 Scarp mountains, very strongly dissected, with high peaks rising
Mountains up to 3000 m.
2b Hijaz 300-2000 Scarp mountains, rising less high, with interspersed high plateau
Mountains areas. Very complex lithology, including granitic, metamorphic,
volcanic, and sedimentary rocks
2c Asir 300-2500 Scarp mountains, very strongly dissected in the Asir, with high
Mountains peaks rising up to 3000 m. Towards the Yemen highlands the
and Yemen high plateau areas become more widespread. Granitic and
Highlands metamorphic rocks dominate the Asir, sedimentary rocks are
predominant in the northern Yemen Highlands, and extrusive
volcanic rocks in central and southern Yemen.
3 Arabian Shield
3a Sandstone 700-1000 High plateau of fairly uniform elevation covered with sandy
plateau soils.
3b Harrats 1000-1500 Area transitional towards the Najd plains at high elevation.
(western Structural slope from SW to NE. Complex terrain with salt flats,
Najd) pediments, and hills.
3c Central 800-1200 Mostly plains and plateaux with inselbergs and hill areas, built
plateau on the structural slope of the Arabian shield. Covered mostly by
granitic and metamorphic basement complex rocks. Includes
large areas with basaltic rocks.
3d Summan 250-500 Low-lying plateau with fairly uniform topography composed of
plateau flat-lying limestone with typical karst topography of sinkholes
and caves.
4 Central Arabian Cuesta
4a Dahna sands 500-800 Narrow belt of dunes and shifting sands extending over 1,300
and adjacent km and connecting the Rub Al-Khali with the Great Nafud.
areas
4b Tuwayq 500-1000 Cuesta region with 800-km-long escarpments composed of
mountain sedimentary rocks curving around the crystalline shield of the
systems Central Plateau. Elevation may locally rise to 1500 m.
5 Southern Arabian Deserts
5a Rub al-Khali 100-1000 The ‘Empty Quarter’ is the largest uninterrupted sand desert in
and adjacent the world. Contains both transverse and longitudinal dunes.
areas Individual dunes reach heights exceeding 200 m.
5b Wahiba Sands 0-300 Small sand sea formed by winds of the southwest monsoon, with
longitudinal dunes mainly.
6 Great Nafud 700-1000 Second largest sand desert of the Arabian Peninsula
7 Eastern Gulf 0-500 Coastal plain rising gently to inland plateau areas. Covered
Region mainly by unconsolidated beach sands, gravels, salt flats, and
aeolian sands.

-9-
Table 2. Continued
symbol Name Altitude range Description
(m)
8 Southern 300-1200 Includes the Hadramaut plateau and the raised plateau of
limestone Dhofar, which can locally reach an altitude of 2000 m, and dips
plateaux to the north. The Dhofar plateau is bounded southward by an
escarpment.
9 Hajar 500-2500 Steeply dissected narrow mountain range with heights up to
Mountains 3000 m, formed by sedimentary rocks.

density, covering the northwestern part of the Peninsula, is shown in Figure 5.

It is worth noting the absence of drainage channels in the eastern part, which is
occupied by sand dunes.
The wadis are common to all geomorphological units, with the exception of
sand dune areas, and support plant communities that are dependent on the water
regime. Along wadis the vegetation cover is usually denser, except under
conditions of overgrazing, fuelwood extraction, or aquifer over pumping.
However, the vegetation pattern is highly site-specific, determined by the
frequency of flooding, the stream velocity, type of sediments and coarse
materials deposited, and variability of rainfall in the catchment areas
((Kürschner, 1998).

Fig. 5: Wadi network in the northwest of the Arabian Peninsula(Note: Wadis shown as
blue lines. Elevation range is the same as in Fig.3)

-10-
4. Climate
4.1. General

The Arabian Peninsula is an ecoregion in which biomass productivity is

primarily limited by the availability of water. Although there are exceptions,
notably the Yemen Highlands and their extension into the Asir Mountains, the
region is essentially arid or even hyper-arid. In addition to generally low rainfall
amounts, rainfall distribution is usually unfavorable, coming in sudden and
erratic showers, and variability is high between years.
The weather in the Peninsula is controlled by four air masses. The main
reasons for the region’s aridity are its remoteness in relation to the major rain-
bearing weather systems, such as the North Atlantic depressions and the Indian
monsoon, and its exposure to air predominantly continental in origin.
With the exception of winter in the northern part of the Peninsula and high-
altitude locations, temperatures are high to very high, causing high evaporation,
but also high biomass productivity, if water is available.
The low and erratic rainfall causes large fluctuations in the productivity of
rangelands. It also enhances the importance of soils and landscape position in
capturing the little rainfall available. Their ability to generate, concentrate, or
receive runoff is the main reason for the ‘patchiness’ of vegetation cover in the
Peninsula.
Within the overall limitations imposed by aridity, there is an unexpected
diversity in climatic conditions. This diversity is usually related to differences
in temperature and moisture regimes as a result of different exposure to rain-
bearing systems, but also altitudinal gradients. The mountains at the edge of the
Peninsula generally act as ‘moisture traps.’ At certain times of the year the
influence of the mountains can be strong enough to generate their own weather
through erratic and intensive thunderstorm activity. This is certainly the case in
the Yemen and Asir highlands and the Hajar mountains in Oman. Topography
also influences climate by ‘guiding’ wind flows and rain along favored paths.
The Zagros mountains in western Iran, through this mechanism, play an
important role in generating precipitation over the extreme east of the Peninsula
(Fisher and Membery, 1998). In the same way, the western mountains influence
rainfall production along the Red Sea.

-11-
4.2. Precipitation
4.2.1. Types and amounts
Precipitation is mostly in the form of rainfall, although occasional snowfall has
been recorded in the Yemeni and Asir highlands.
Fisher and Membery (1998) report the regular occurrence of fog in the
western highlands, the Dhofar region, and the central desert of Oman. The
contribution of this ‘occult precipitation’ from fog, mist, low clouds, or dew, to
the regional waterbalance might be doubtful. However, these sources of hidden
precipitation can help significantly in creating, at a micro-scale, improved
conditions for more productive and diverse plant life, particularly grasslands
and woodlands.
The distribution of the mean annual precipitation is shown in Figure 6.
Generally speaking, precipitation levels are associated with elevation. The
highest rainfall occurs in the Yemeni highlands and Asir mountains, and to a
lesser extent in the mountains of northern Oman. The lowest precipitation is
recorded in the low-lying areas of the Rub al Khali, the Najd in the north of the
Peninsula, and the northern Red Sea coast.

4.2.2. Seasonal patterns

During winter the region is under the influence of polar continental air masses
that originate in Central Asia. The influx of these air masses is accompanied by
dry weather with generally clear skies and fairly low temperatures.
Occasionally the Peninsula is affected by polar maritime air coming from the
North Atlantic. These air movements are the remainder of the mid-latitude
depressions that have already traversed North Africa and the Mediterranean.
They are the main source of winter rainfall.
During summer the Peninsula is influenced by tropical continental air
masses, which bring hot and very dry air from Egypt and Sudan. These air
masses allow the region to become a stable high-pressure zone and source of
tropical continental air. Cloudless skies, low humidity, very high temperatures
(often >45°C), intensive surface heating, and dust characterize the weather
system during summer (Fisher and Membery, 1998). The Indian monsoon
system exercises some activity in summer, particularly in parts of Yemen,
southwestern Saudi Arabia, and coastal Oman. However, its influence is limited
by the strong tropical continental air mass, which prevails over the Peninsula at
the time.

-12-
-13-

Fig. 6: Mean annual precipitation (mm2)

 The seasonal distribution is shown in Figures 7a-7d as the percentage of the
annual rainfall that falls in each of the four seasons. This distribution is
explained by the prevalence of one or another weather system at different times
of the year.
Figure 7a shows the influence of the winter rainfall pattern, which is
common throughout Europe, North Africa, West Asia, and Central Asia. Figure
7b shows the importance of spring rainfall (also largely associated with the
westerly systems) for the central landmass of the Peninsula. Figure 7c shows
the area of influence of the Indian monsoon, with peaks in the Yemen highlands
and the Dhofar area in southern Oman. Autumn does not contribute much
rainfall, except in the northern Red Sea coast region (Figure 7d).

4.2.3. Variability
As in other arid parts of the world, high rainfall variability is the norm and the
impact of drought severe. Variability affects the amount and distribution of
rainfall at different time scales.
Figure 8 shows a typical example of inter-annual variations for Muscat,
time period 1893-1978, an unusually long dataset for the Arabian Peninsula. A
typical characteristic of rainfall in the arid zones is its negative skew. This
means that the probability of having rainfall below the mean is higher, but
compensated for by few high rainfall events, as represented by the peaks in
Figure 8. The probability distribution of annual precipitation in Muscat, as
approximated by a log-normal transform, is shown in figure 9.
This figure shows that the probability of exceeding the mean (105 mm) is
only 40%, illustrating the greater likelihood of smaller rainfall amounts.
At smaller time scales, variability increases even more. Figure 10 shows,
for each month of the year, the mean rainfall and the amounts that might not be
exceeded in one year out of four (2nd decile), and in four years out of five (8th
decile).
In the same figure it can be noted that between April and October, the 8th
decile is lower than the mean. This demonstrates that in areas (or times of the
year) with very low rainfall, the concept of an average rainfall pattern is a
statistical artifact. It is caused by the lumping together of a few high-rainfall
events with very low probability, with numerous low-rainfall events with high
probability.

-14-
 a b

-15-

c
d

Fig. 7: Seasonal distribution of precipitation. Clockwise from top left: winter (a), spring (b), summer (c), autumn (d)
 Fig. 8: Variability of annual rainfall, Muscat, Oman (period 1893-1978)

Fig. 9: Probability distribution of annual rainfall, Muscat, Oman (period 1893-1978))

-16-
 Fig. 10: Variability of monthly rainfall, Muscat, Oman (period 1893-1978)

The nature of individual rainstorms is such that they are often of limited
spatial extent with considerable gradients in intensity and amount. This implies
that in large areas in the middle of a ‘rainy season,’ the pattern in reality might
be one of intense rains separated by dry conditions or light falls (Jackson,
1977). This ‘spottiness’ of rainfall has also been suggested by Fisher and
Membery (1998).
The implication of increasing rainfall variability with decreasing time scale
is of fundamental importance to our understanding of vegetation growth,
biomass productivity, and climatic adaptation in the Arabian Peninsula. Since
temperature is usually not limiting, growth and flowering occurs whenever and
wherever water is available, irrespective of time of year and ‘statistical’ dry and
wet periods.

4.3. Temperature
The Arabian Peninsula is warm. More than 90% of its area has a mean annual
temperature of 20°C. A small area (shown in red) has a mean annual
temperature exceeding 30°C. The cooler areas are shown in magenta in figure
11. They correspond with the western Yemen highlands, the Asir mountains,
and the sandstone and limestone plateaux bordering Jordan.

-17-
Fig. 11: Areas of the Arabian Peninsula with mean annual temperature exceeding 30°C (in
red), or below 20°C (in magenta)
The major factors controlling temperature are elevation and latitude. From
south to north there is a clear cooling trend, owing to increased exposure to cold
continental air masses in winter. The map of mean annual temperature (Figure
12) illustrates these controls on temperature.
Temperature is strongly seasonal, with the lowest temperatures in the
period December-February and the highest in the period June-September. The
areas exposed to the Indian monsoon are an exception. These show a noticeable
temperature drop in July-August, as illustrated by the climate diagram for
Salalah (Figure 13).
Figures 14 and 15 show the temperature of the coldest and the warmest
month, respectively.
Temperature seasonality tends to increase from the southeast to the
northwest. The variability of temperature between years, in contrast with
rainfall variability, is very low.
Temperature patterns can also be represented as the distribution of available
atmospheric energy, which evaporates water or makes plants grow faster, for
example. This representation of temperature as a source of energy for plant
growth and biomass production can be done through the concept of
accumulated heat units or growing degree days, which sum the daily

-18-
-19-

Fig. 12: Mean annual temperature

 Fig. 13: Climate diagram for Salalah, Oman

Fig. 14: Mean temperature of the coldest Fig. 15: Mean temperature of the warmest
month month

temperatures above a threshold (e.g., 0 °C) for a specified period (e.g., one
year). The map of accumulated heat units in Figure 16 shows, unsurprisingly,
the same pattern as the map of mean annual temperature, only the units (°C
days) are different. It will be used later (see Section 4.7) to assess by proxy the
potential productivity of natural vegetation.

-20-
-21-

Fig. 16: Mean annual heat units

4.4. Evapotranspiration and water deficit
4.4.1. Potential evapotranspiration
The evaporative demand of the atmosphere can be represented by the potential
evapotranspiration (PET), which is the evapotranspiration of a reference crop,
a grass cover. The PET can be calculated by the Penman-Monteith method from
elementary climatic parameters, such as temperature, air humidity, radiation,
and wind speed (Allen et al., 1998). The PET concept is the basis for the
assessment of crop water requirements and for scheduling irrigation.
The pattern of mean annual PET throughout the region (Figure 17) is very
similar to the pattern of mean annual temperature (Figure 12) and shows the
same trends, as governed by altitude and latitude. The PET is very high in the
interior of the Peninsula and decreases towards the edges.

4.4.2. Aridity
The Arabian Peninsula is characterized by a severe deficit of precipitation over
atmospheric water demand. The UNESCO (1979) ‘aridity index’ can be used to
quantify this deficit and map the severity of ‘dryness’ based on the ratio of
annual precipitation to annual potential evapotranspiration.
Figure 18 shows the distribution of the aridity index, corresponding with
the thresholds for the hyper-arid (<.0.03), arid (0.03-0.25), semi-arid (0.25-0.5),
and sub-humid climatic regimes (>0.5). The areas under each moisture regime
are summarized in Table 3. Nearly 99% of the Peninsula is either hyper-arid
(HA) or arid (A). Only part of the Yemen highlands is in the semi-arid (SA)
class. A very small area is sub-humid (SH). The relationships between moisture
regimes (as expressed by aridity index) and agriculture are discussed further in
section 4.5.1.
Table 3. Areas under different moisture regimes
Moisture regime Aridity index % km2
Hyper-arid (HA) < 0.03 32.02 1,020,107
Arid (A) 0.03 - 0.2 66.72 2,125,415
Semi-arid (SA) 0.2 - 0.5 1.22 38,788
Sub-humid (SH) > 0.5 0.04 1,199

These results reflect the huge negative balance between water supply from
precipitation and the evaporative demand of the atmosphere. The size of the
hydrological deficit on an annual basis is shown in figure 19.

-22-
-23-

Fig. 17: Mean annual potential evapotranspiration (mm)

 Fig. 18: Aridity index

Fig. 19: Annual precipitation deficit (mm)

-24-
4.5. Agroclimatic patterns
4.5.1. Agroclimatic zones
Climatic diversity in the Arabian Peninsula can be adequately represented by a
simple system of agroclimatic zones that takes into account the key
determinants of climate: moisture and temperature. UNESCO (1979) has
developed a simple system for differentiating agroclimatic zones based on three
major criteria:
• Moisture regime
• Winter type
• Summer type
The moisture regime is determined by the aridity index, as defined in
section 4.4.2. In the hyper-arid moisture regime there is almost no perennial
vegetation, with the exception of some bushes in riverbeds. In good years,
annual plants can grow, but agriculture and grazing are generally impossible
(UNESCO, 1979). In the arid moisture regime, scattered vegetation does grow,
which might include bushes, and small woody, succulent, thorny, or leafless
shrubs. Very light pastoral use is possible. Rainfed agriculture is only feasible
with some form of water harvesting and irrigation, and only where terrain
conditions are favorable or where there are local water resources. As a result,
agriculture, if any, is patchy. In the semi-arid moisture regime, vegetation is
denser and might include bushes, scrubs and even trees. Good grazing areas
might be found and rainfed agriculture is possible, albeit with great yield
fluctuations due to rainfall variability. Agriculture in either the arid or hyper-
arid classes requires irrigation (see section on land use). Even in the semi-arid
mountainous uplands of Yemen, agriculture is stabilized by terrace-based
supplemental irrigation.
The winter type is determined by the average mean temperature during the
winter months. Table 4 shows the winter type classes and the areas they occupy.
Figure 20 shows the spatial distribution of the winter types.

Table 4. Areas under different winter types

Winter type Temperature % km2
class (°C)
Cool (C) <10 2.81 89,622
Mild (M) 10-20 74.71 2,379,883
Warm (W) 20-30 22.48 716,005

-25-
 Fig. 20: Winter types (blue: cool; green: mild; yellow: warm)

In areas with cool winters, vegetation growth is limited by cold. If rainfall

is concentrated in winter, plants adapted to these conditions will be
characterized by rapid phenological development in spring and efficient soil
moisture extraction. In areas with mild and warm winters, vegetative growth
is possible in winter, and becomes more rapid with increasing temperature.
The summer type is determined by the average mean temperature during
summer months. Table 5 shows the summer type classes and the areas they
occupy. Figure 21 shows the spatial distribution of the summer types. The mild
summers are confined to the highest parts of the Yemen highlands.

Table 5. Areas under different summer types

Summer type Temperature class (°C) % km2
Mild (M) 10-20 0.48 15,417
Warm (W) 20-30 36.12 1,150,600
Very warm (VW) >30 63.40 2,019,492

-26-
 Fig. 21: Summer types (green: mild; yellow: warm; red: very warm)

The combinations between moisture regimes and winter and summer

types form individual climatic patterns, or agroclimatic zones. For example,
the pattern HA-M-VW represents a climate with a hyper-arid moisture
regime, mild winter type, and very warm summer type. In total, 22
agroclimatic zones have been differentiated in the Arabian Peninsula. Of
these, eight taken together comprise 95% of the region. The remaining 14 are
fairly small agroclimatic ‘islands’ with climatic conditions that are either
more humid or colder than surrounding areas.
The extent of the agroclimatic zones is shown, in order of importance, in
Table 6. The spatial distribution is shown in Figure 22.
Figures 23a-23h are climate diagrams of stations representing some
agroclimatic zones. Some zones cannot be represented due to their limited
extent and the lack of meteorological data (see further).

4.5.2. Similarity of climatic conditions

In all the maps in this chapter, the value of selected climatic parameters is
shown as classes with well-defined ranges, e.g., precipitation classes 0-10 mm,
10-20 mm, etc. These classes show similarity in a way that is independent of

-27-
-28-

Fig. 22: Agroclimatic zones (classified according to UNESCO, 1979)

Table 6. Extent of agroclimatic zones of the Arabian Peninsula
ACZ Representative % Km2 ACZ1 % km2
station
A-M-VW Riyadh 35.3 1,123,841 SA-M-M 0.4 11,522
A-M-W Abha 17.7 563,774 SA-W-W 0.1 3,481
HA-M-VW Al Jouf 15.5 495,144 A-M-M 0.1 1,925
HA-W-VW Muscat 7.2 229,484 SA-C-M 0.0 866
A-W-W 6.0 190,051 SH-M-M 0.0 828
HA-M-W Tabuk 5.0 160,135 HA-C-VW 0.0 663
A-W-VW El Kod 5.0 158,775 HA-C-W 0.0 469
HA-W-W 4.2 134,213 SA-C-W 0.0 426
A-C-W 2.4 75,341 SH-M-W 0.0 222
SA-M-W Taiz 0.7 22,492 SH-C-M 0.0 148
A-C-VW 0.4 11,580 A-C-M 0.0 127
1
All data from mabar station.

Fig. 23a: Station representative for Fig. 23b: Station representative for
agroclimatic zone A-M-W: Abha, Saudi agroclimatic zone HA-M-VW: Al Jouf, Saudi
Arabia Arabia

Fig. 23c: Station representative for Fig. 23d: Station representative for
agroclimatic zone A-W-VW: El Kod, Yemen agroclimatic zone SA-M-M: Mabar, Yemen

-29-
 Fig. 23f: Station representative for
Fig. 23e: Station representative for
agroclimatic zone A-M-VW: Riyadh, Saudi
agroclimatic zone HA-W-VW: Muscat, Oman
Arabia

Fig. 23g: Station representative for

Fig. 23h: Station representative for
agroclimatic zone HA-M-W: Tabuk, Saudi
agroclimatic zone SA-M-W: Taiz, Yemen
Arabia
the values in individual locations. There is, however, a very different approach
that consists of taking the value of a climatic parameter or index at one location
(the ‘match‘ location) and mapping similar locations (‘target’ locations). This
approach is valuable for assessing the likelihood of successful introduction of a
plant species in a different area, in the assumption that the more similar the
environments the more likely will be the adaptation.
The key is to be clear in the purpose and define similarity indices
accordingly. If the purpose is to assess adaptation to heat stress, a temperature-
based similarity index is needed. If the objective is to assess adaptation to
drought, a precipitation-based index is needed. In this publication the purpose
is to assess similarity both in moisture and temperature conditions, and for this
reason a combined similarity index has been developed (see Section 8.3.).
Figures 24a-24h show similarity between each part of the Arabian
Peninsula and a reference location. In this case the reference locations are the
stations that represent the main agroclimatic zones. In all cases, similarity is
shown on the same scale between zero and one, with zero indicating total

-30-
dissimilarity and one total similarity. These examples demonstrate that in some
cases the adaptability domain, as expressed by a high similarity index value, is
very widespread, and in other cases very limited.

Fig. 24a: Similarity in temperature and precipitation patterns with Abha, Saudi Arabia

Fig. 24b: Similarity in temperature and precipitation patterns with Al Jouf, Saudi Arabia

-31-
Fig. 24c: Similarity in temperature and precipitation patterns with El Kod, Yemen

Fig. 24d: Similarity in temperature and precipitation patterns with Mabar, Yemen

-32-
 Fig. 24e: Similarity in temperature and precipitation patterns with Muscat, Oman

Fig. 24f: Similarity in temperature and precipitation patterns with Riyadh, Saudi Arabia

-33-
Fig. 24g: Similarity in temperature and precipitation patterns with Tabuk, Saudi Arabia

Fig. 24h: Similarity in temperature and precipitation patterns with Taiz, Yemen

-34-
4.6. Climatic growing period
Growing period, as a climatic concept, is the time of year when neither moisture
nor temperature limit crop production. Developed about 30 years ago (e.g.
Cochemé and Franquin, 1967), it was subsequently adapted and applied by the
Food and Agriculture Organization of the United Nations (FAO, 1978-81) to
assess potential plant productivity and land suitability at the global, continental,
and regional scale.
The components of the climatic growing period (onset, duration, and end)
are determined by a waterbalance approach, which matches monthly rainfall to
monthly potential evapotranspiration. In technical terms the growing period is
the ‘period of the year during which the actual evapotranspiration exceeds a
critical threshold’ (De Pauw, 1983). This threshold is usually 50% of the
potential evapotranspiration.

4.6.1. Types of growing period

Three types of growing period are described for the Arabian Peninsula. Their
distribution is shown in figure 25.
The vast majority of the region is characterized by an all-year-round dry
period. There is no growing period because the critical ratios of actual to
potential evapotranspiration are not exceeded. In this moisture regime the
quality of the growing period is no longer determined by the rainfall but by the
level of potential evapotranspiration; the higher the latter, the higher the
irrigation requirements. An example of an all-year-round dry period is shown in
figure 26a.
In the Yemen highlands, northern Kuwait, and parts of the mountains of
northern Oman, a second type of growing period occurs. The intermediate
growing period lacks a humid sub-period: when rainfall exceeds potential
evapotranspiration soil moisture is not recharged. Crop production in such areas
is very risky and usually requires an additional source of water to stabilize
yields. However, this type of growing period has a reasonable productivity for
natural vegetation. An example, with two sub-periods, is shown in figure 26b.
In a very few areas of the Peninsula, located in the most rainy parts of the
Yemen highlands, the normal growing period occurs. It is characterized by a
humid sub-period, when soil moisture is recharged. This soil moisture can later
be released to crops when rainfall drops below the potential evapotranspiration,
thus buffering crops against drought stress. This type of growing period is of the

-35-
-36-

Fig. 25: Types of growing period

 Fig. 26a: Example of a year-round dry period (Al Jouf, Saudi Arabia)

Fig. 26b: Example of an intermediate growing period (Taiz, Yemen)

highest quality, ensuring optimal biomass productivity, and, if of sufficient

length, good crop yields. No representative station is available for the Arabian
Peninsula; therefore, the concept is illustrated with an example from Aleppo,
Syria (Figure 26c).

-37-
 Fig. 26c: Example of a normal growing period (Aleppo, Syria)

4.6.2. Duration and onset of the growing period

In most of the Arabian Peninsula, length and timing of the growing period are
highly variable as a result of pronounced rainfall variability (see section 4.2.3.).
The only areas where the growing period is sufficiently reliable to appear in
average data are the highlands of Yemen. This part of the Peninsula has two
rainy seasons, one in March-May, the other in July-September. The durations of
the resulting two growing periods are shown in figures 27a and 27b. The onset
date of the main growing period is shown in Figure 27c.

Fig. 27a: Length of the first growing period, Yemen Highlands

-38-
 Fig. 27b: Length of the second growing period, Yemen Highlands

Fig. 27c: Onset of the main growing period, Yemen Highlands

4.7. Biomass productivity and climate

Climate is the primary determinant of potential biomass productivity of plants

and crops. This is because assimilation – the capture by plants of carbon dioxide
from the atmosphere and its conversion into carbohydrates – is determined by
radiation energy and water availability. Biomass productivity should, therefore,
be related to climatic factors, in particular temperature (as proxy for the
radiation energy) and soil moisture. Apart from radiation and moisture regime,

-39-
the rate of assimilation and biomass production is strongly determined by crop
characteristics.
According to the response of assimilation rate to temperature, FAO (1978-
81) has proposed four crop groups (Table 7). Each crop group has a different
response function, or adaptability range, to temperature (Figure 28).
Table 7. Adaptability ranges of different crop groups
Crop Crop Optimal mean Examples
group types temperature
range
1 C3 15-20 Barley, bread wheat, chickpea, lentil,
olive, sunflower, cabbage, oats, rye,
grape, sugar beet; temperate grasses;
almost all trees
2 C3 adapted for 25-30 Cotton, groundnut, cowpea, soybean,
higher temperatures tobacco, sunflower, sesame, rice, fig,
grape, olive
3 C4 30-35 Maize, sorghum, sugarcane, all millets,
fonio rice; tropical grasses
4 C4 adapted for 20-30 Maize, sorghum, millets
lower temperatures

Fig. 28: Temperature adaptability ranges for different crop groups

Using this concept of crop adaptability groups, biomass productivity

indices have been developed for each crop group. For the exact definition of
each crop biomass productivity index (CBPI) refer to Section 8.3.

-40-
 The biomass productivity index was calculated for each crop group
(CBPI1, CBPI2, etc.) for each location in the Arabian Peninsula. The results
show that only in part of the Yemen highlands do the indices have non-zero
values. This is not surprising because the CBPI is strongly correlated with
growing period, which is absent in most of the Peninsula.
Figures 29a-29d focus on the Yemen highlands and show the values of the
CBPI for each crop group. Generally speaking, these figures show that the areas
are better adapted to crop groups 2, 3, and 4 than to crop Group 1.
To assess the potential productivity of rangelands, a different kind of index
is required that is less demanding in terms of moisture regime. The rangeland
biomass productivity index (RBPI) is the product of the aridity index (see
Section 4.4.2.) and the annual accumulated heat units (see Section 8.3.).
Distribution of the RBPI is shown in Figure 30.
The value of these biomass productivity indices is that they can be derived
from simple climatic data and allow extrapolation from site-specific
productivity measurements. It has to be realized that they provide a measure of
potential productivity, not current productivity, and, therefore, do not take into
account management factors, such as overgrazing, etc.

-41-
 Fig. 29a: Biomass productivity index for crop group I, Yemen Fig. 29b: Biomass productivity index for crop group II, Yemen
Highlands Highlands
-42-

Fig. 29c: Biomass productivity index for crop group III, Fig. 29d: Biomass productivity index for crop group IV, Yemen
Yemen Highlands Highlands
-43-

Fig. 30: Rangeland biomass productivity index

5. Soils of the Arabian Peninsula
5.1. General soil pattern
The soils of the Arabian Peninsula reflect the aridity of the climate. Most are
poorly developed, shallow, or are enriched in lime, gypsum, or salts. In
addition, transported materials, such as sand dunes and sheets, cover large
areas. The soils are mostly formed by the physical breakdown of geological
materials and their subsequent removal, sorting and deposition by wind and
water.
The distribution of the soils of the Arabian Peninsula is represented in the
Map of Soil Associations (Figure 31). Soil association maps show patterns of
soil occurrence, instead of the location of individual soils. Individual soils
cannot be located at the scale of a subcontinent. Soil associations are
characterized by the recurrence of a limited number of specific soil types within
particular landforms, but in different proportions. The main associations of the
Arabian Peninsula are listed in Table 8.
As shown in the example of Figure 32, the name of the soil association is
determined by the dominant soil type, the textural class, and the broad landform
type.
The association is further described by its full soil composition, which lists
each soil type and its proportion (in percent) within the soil association (Table
8).
The main soil types that occur in the Arabian Peninsula, classified
according to the FAO Soil Map of the World (FAO, 1974), are listed in Table 9.
Three textural classes are distinguished:
1: Coarse (predominantly sandy)
2: Medium (predominantly silty)
3: Fine (predominantly clayey)
These distinctions are further defined in the discussion on soil management
properties (Section 5.2.).
In addition, the soil associations do consider three broad landform classes:
a: flat topography; b: undulating topography; c: hilly topography
Looking only at the dominant soils within the different soil associations, the
soil association map can be simplified, as shown in the map of dominant soils
(Figure 33).

-44-
-45-

Fig. 31: Soil associations (FAO, 1995)

Table 8. Main soil associations of the Arabian Peninsula (FAO, 1995)
Soil Association % of the Soil Types and Proportions
Peninsula %1 %2 %3 %4 %5 %6

D/SS 5.80 DS 100 0 0 0 0 0

I-Y-bc 20.28 I 50 Y 50 0 0 0 0
I-Yh-Yk-1/2b 4.00 I 34 Yh 33 Yk 33 0 0 0
I-Yk 0.04 I 50 Yk 50 0 0 0 0
I-Yk-1/2a 0.26 I 50 Yk 50 0 0 0 0
I-Yk-2ab 2.41 I 50 Yk 50 0 0 0 0
Je61-2a 0.32 Je 90 Zo 10 0 0 0 0
Qa9-1a 0.48 Qa 100 0 0 0 0 0
Qc46-1/2ab 6.51 Qc 50 I 20 Y 20 R 10 0 0
Rc30-1ab 9.33 Rc 50 Qc 20 Yk 20 Z 10 0 0
Rc31-1/2ab 1.23 Rc 70 Z 30 0 0 0 0
Re1-1/2a 0.00 Re 100 0 0 0 0 0
SALT 0.05 ST 100 0 0 0 0 0
Yh22-1ab 4.96 Yh 30 Qc 20 Yk 20 Jc 10 Rc 10 Z 10
Yh3-1/2a 3.74 Yh 70 I 30 0 0 0 0
Yk25-1/2a 10.94 Yk 40 I 20 Yl 20 Jc 10 Z 10 0
Yk26-1ab 3.39 Yk 70 Qc 30 0 0 0 0
Yk27-2a 0.09 Yk 60 Z 30 I 10 0 0 0
Yk28-1a 6.40 Yk 60 Rc 30 I 10 0 0 0
Yk29-1/2a 0.25 Yk 60 Rc 30 Z 10 0 0 0
Yk32-a 8.12 Yk 60 I 20 Yl 20 0 0 0
Yl19-3ab 2.68 Yl 50 I 20 Yk 20 Rc 10
Yy10-2ab 3.79 Yy 60 I 20 Yk 20 0 0 0
Yy10-2ab 0.37 Yy 60 I 20 Yk 20 0 0 0
Yy12-a 0.54 Yy 60 I 30 Yk 10 0 0 0
Yy7-2/3a 0.05 Yy 50 Yk 20 Zo 20 Jc 10 0 0
Zg3-2/3a 1.18 Zg 70 Zo 30 0 0 0 0
Zo18-2ab 0.51 Zo 60 I 30 Yk 10 0 0 0
Zo19-1/2ac 1.35 Zo 50 I 30 Qc 10 Rc 10 0 0
Zo20-1/2a 0.19 Zo 60 Yh 30 Rc 10 0 0 0
Zo21-3a 0.03 Zo Composition unspecified
Zo22-2/3a 0.21 Zo
Zo27-3a 0.37 Zo
Zo28-3a 0.11 Zo 50 Zt 30 Jc 10 Zg 10 0 0

-46-
 Fig. 32: Soil associations: explanation of legend
Table 9. Main soil types of the Arabian Peninsula
Symbol Name Summary description

I Lithosols Undifferentiated very shallow soils; unsuitable for

agriculture
Je Eutric Fluvisols Alluvial soils with good fertility status; the best soils for
agriculture
Qa Albic Arenosols Strongly leached sandy soils, do not retain soil moisture;
unsuitable for agriculture
Qc Cambic Arenosols Slightly matured sandy soils, retain soil moisture better;
suitable for agriculture under sprinkler irrigation
Rc Calcaric Regosols Calcareous poorly developed soils, poor physical properties
for agriculture
Re Eutric Regosols Poorly developed soils with moderate fertility; poor
physical properties for agriculture
Yh Haplic Yermosols Undifferentiated very poorly developed soils of (semi-)
deserts; management properties vary considerably; full
irrigation is needed for all agricultural uses
Yk Calcic Yermosols Very poorly developed soils of (semi-) deserts with calcium-
enriched subsoil; unsuitable for agriculture
Yl Luvic Yermosols Very poorly developed soils of (semi-) deserts with clay-
enriched subsoil; can be made suitable for agriculture if full
irrigation is available
Yy Gypsic Yermosols Very poorly developed soils of (semi-) deserts with gypsum-
enriched subsoil; unsuitable for agriculture due to poor
physical properties and need for full irrigation
Zg Gleyic Solonchaks Saline soils with insufficient drainage; unsuitable for
agriculture
Zo Orthic Solonchaks Undifferentiated saline soils; unsuitable for agriculture

-47-
-48-

Fig. 33: Simplified soil map (dominant soils)

5.2. Soil management properties
Soil maps, particularly those at the regional or national level, are notoriously
difficult to interpret as a guide to agricultural management. In most cases, only
classification names are provided, but not the associated management
properties, which are assumed to be understood. This is rarely the case, since
most potential users of soil maps are not familiar with soil scientist jargon. For
this reason, maps are needed that show the spatial distribution of soil
management properties.
This is a particularly daunting task because soil management properties
tend to vary considerably, even within classified soils. The best that can be done
on a regional scale is to show the distribution of certain soil properties, which
can be associated with reasonable likelihood to the soil classification. For
example, Lithosols are by definition associated with shallow soil depth.
Arenosols are sandy soils, Fluvisols are likely to have optimal depth, water
holding capacity, and fertility status, Yermosols are associated with very low
organic carbon levels, and Solonchaks cover the wide spectrum of saline soils.
By looking at soil classification units as ‘indicators’ of soil management
properties, it is feasible to map the distribution of some properties, but not
others. It is possible to map soil texture, depth, stoniness, and some important
fertility indicators, such as organic carbon content, pH, and cation exchange
capacity (CEC). These are the more stable properties of soils, and are more
shaped by climate, geology, and landform, than by human intervention.
However, it is not possible to map physical properties, such as infiltration
capacity, aggregate stability, or nutrient availability (N, P, K), because these
properties are very site-specific and highly responsive to management.
The spatial distribution of some management properties is shown in the
following set of maps. Refer to the legend in Figure 34 to link the properties to
the dominant soil type and the other soil types of the soil association. The
yellow color indicates, for example, that the particular soil management
property is likely to occur in about 10-20% of the areas colored yellow.

5.2.1. Soil texture

Soil texture refers to the relative proportions of various particle size groups in
a mass of soil (Soil Survey Staff, 1951). Specifically, it refers to the proportions
of clay (<.002 mm), silt (.002-.05 mm), and sand (.05-2 mm). Soil scientists use
a system of 12 standard textural classes (Figure 35). For a regional assessment

-49-
of textural distribution these classes are reduced to the three mentioned in
section 5.1. The relationships between the standard and ‘regional’ textural
classes are shown in Figure 35.

Fig. 34: Legend of soil property distribution maps – % of area

Fig 35: USDA textural triangle and simplified textural classes

Soil texture is a key management property. It determines the amount of

water a soil can hold and make available to plants. In dry areas, soil texture is
a key determinant of irrigation management. Clayey soils are suitable for
different types of flood irrigation. They retain water very well, but are subject
to high evaporation losses. Silty soils can be irrigated fairly infrequently, but

-50-
have high percolation losses. Sandy soils demand very frequent irrigation,
usually by sprinkler or drip systems, or, if very coarse, cannot be irrigated. Soil
texture is also a determinant of soil structure and surface properties. Silty soils
tend to form a surface seal, which reduces infiltration capacity and promotes
runoff even on gentle slopes. In contrast, water infiltrates sandy soils and no
evaporation losses occur. As a result, higher biomass productivity of natural
vegetation is in the Arabian Peninsula often associated with sandy soils.
Figures 36-38 show the distribution of the coarse-textured, medium-
textured, and fine-textured soils, respectively, in the Arabian Peninsula.
These figures demonstrate that the prevailing perception of the Arabian
Peninsula as a sandy desert is incorrect. Most soils are medium-textured, with
silt as dominant soil component. Sandy soils are common, particularly in the
drier interior basins, but clayey soils are fairly uncommon.

5.2.2. Soil depth and stoniness

Soil depth and the related attribute, stoniness, are important management
properties, since they determine the feasibility of mechanization and the soil

Fig. 36: Distribution of coarse-textured soils(Note: for legend see Fig. 34)

-51-
Fig 37: Distribution of medium-textured soils(Note: for legend see Fig. 34)

Fig 38: Distribution of fine-textured soils(Note: for legend see Fig. 34)

-52-
moisture storage capacity. Very shallow or stony soils do not have sufficient soil
volume in which to store moisture, and therefore dry out more rapidly than deep
soils, or soils of similar depth without stones. For the same reason, these soils
have more difficulty absorbing rainfall during storms and thus generate much
runoff.
The distribution of soil depth and related attributes are shown in Figures 39-
41. The map of the dominant soil depth (Figure 39) is an oversimplification, but
shows a clear pattern. Mountain areas are invariably associated with high levels
of shallow (Figure 40) and stony soils (Figure 41). They occur mainly in the
western highlands of the Midian, Hejaz, and Asir, the escarpment of the Yemen
high plateau, and the Hajjarr and Mussandam mountains in Oman. In addition,
there are interspersed but fairly large areas of gravel plains throughout the
Peninsula. The most important ones occur in eastern Kuwait and south of the
Hajjar mountains.

5.2.3. Soil fertility indicators

As mentioned earlier, the fertility status of soils is, in general, so site-specific
and determined by management that any region-wide assessment is

Fig 39: Dominant soil depth

-53-
 Fig 40: Distribution of shallow soils(Note: for legend see Fig. 34)

Fig 41: Distribution of gravelly and stony soils(Note: for legend see Fig. 34)

-54-
meaningless, especially in a part of the world where moisture availability is the
prevailing constraint. However, certain soil characteristics, such as organic
carbon content, pH, CEC, and lime content, are indicative of the fertility status
and are more static and spatially invariant, being mostly determined by climate.
The organic matter content is the most important indicator of the general
fertility status of a soil. Although the composition of organic matter is also
important, soils with a high organic matter content are usually productive. Dry
and hot climates, which do not support a dense cover of vegetation, do not
promote organic matter decay and accumulation. As a result, the organic carbon
levels are very low in the Arabian Peninsula (Figure 42). Most of this organic
matter is concentrated in the topsoil. The subsoils are virtually devoid of
organic carbon, and, as a result, the total organic carbon pools are very low
(Figure 43). Given the considerable overgrazing and vegetation degradation
that has occurred during the last 30 years, it is suspected that even these low
levels of organic carbon quoted are optimistic estimates.
Soil pH controls availability and eventual deficiency or toxicity of certain
micronutrients, which are needed by plants in very small amounts but which are
essential for plant growth, such as zinc (Zn), iron (Fe), manganese (Mn), copper
(Cu), and boron (B). At high pH, Zn might be deficient in sensitive plants. B-

Fig. 42: Dominant organic carbon levels

-55-
 Fig. 43: Organic carbon pool
toxicity might occur in soils of volcanic origin, in other areas it might be
deficient (Ryan et al., 1997). The soil pH hovers in the Arabian Peninsula
around neutral or above (Figure 44). It is mainly controlled by the presence of
free calcium carbonate (see further), and is not limiting to plant growth, except
where it is the consequence of high soil salinity (pH >8.2).
Cation exchange capacity (measured in milliequivalents per 100 g soil)
quantifies the clay fraction’s capacity to retain nutrients, and is a measure of
resilience against nutrient depletion. In dry climates, the CEC is generally high
for most soils, with the exception of sandy soils, which do not have high enough
clay content. The CEC is therefore related to soil texture, particularly the
content of clay, sand, and coarse fragments. The patterns in Figure 45 reflect
broad textural groupings rather than particular soil characteristics, with high
CEC values associated with medium-and fine-textured soils.
The soils of the Arabian Peninsula are well supplied with free calcium
carbonate (CaCO3) as a result of inheritance from calcareous sediments and
rocks, but also due to the lack of leaching and weathering (Ryan et al., 1997).
Apart from its control on soil pH and availability of certain micronutrients,
CaCO3, within reasonable amounts, is not an impediment to crop productivity.
The distribution of calcareous soils is shown in Figure 46.

-56-
 The same qualification applies to gypsum (CaSO4.2H2O), which can be
present in substantive amounts, particularly in the Gypsic Yermosol subgroup.
Gypsum, like CaCO3, is another soluble mineral of sedimentary origin that is
very common in arid regions. Gypsiferous soils have special management

Fig. 44: Dominant soil pH

Fig. 45: Dominant cation exchange capacity

-57-
 Fig. 46: Distribution of calcareous soils(Note: for legend see Fig. 34)

properties. They can be irrigated with fairly saline water without causing
salinity build-up. However, inefficient irrigation with high percolation losses,
can cause the gypsum to dissolve, and the soil to collapse (possibly taking with
it irrigation canals).
In some soils, lime or gypsum have built up massively to form hard banks.
These ‘indurated’ soils are called ‘petrocalcic’ or ‘petrogypsic’ (petro, Greek
root for ‘rock’), depending on whether lime or gypsum is the cementing agent,
and are often confused with sedimentary rocks. They are obviously unsuitable
for agriculture. Their distribution is shown in Figure 47.
In contrast with most other soil types, saline soils show a great diversity in
their appearance. Unless salinity is very pronounced, chemical tests are needed
to recognize a saline soil. In order to be recognized as salt-affected, soils must
have a minimum concentration of salts at some time of the year within the root
zone. Soils are called saline when the total salt concentration, expressed in
electrical conductivity (ECe) of a saturated extract, is above 15 deciSiemens per
meter (dS m-1) at 25°C within 30 cm of the surface at some time of the year, or
more than 4 dS.m-1 if the pH (H2O 1:1) exceeds 8.5. An ECe of 15 dS m-1
corresponds with about 0.65% salt (Driessen and Dudal, 1991).
The distribution of soil salinity in the Arabian Peninsula is shown in Figure
48. Most of this salinity is due to primary salinization. This refers to the build-

-58-
up of salts as a result of lithological inheritance or topographical position, and
is a natural process within arid zones. This map does not provide a good
representation of secondary, or human-induced salinity. The latter has
significantly increased since the large resource surveys of the 1970s and 1980s,
due to the unregulated use of groundwater and large-scale irrigation
development, often without adequate drainage.

Fig. 47: Distribution of soils with hardened lime or gypsum

Fig. 48: Distribution of soil salinity(Note: for legend see Fig. 34)

-59-
5.3. Conclusions
Soil patterns in the Arabian Peninsula are controlled by the interaction of
climate, landforms, and geological parent materials. There is no shortage of
good agricultural soils; the obvious limitation to put them into production is
water availability. Where irrigation water is available, standard fertility
management practices are required, and, if provided, will allow maintenance or
enhancement of soil quality. Gypsiferous soils require careful irrigation
management, but they can be made productive.

6. Agricultural Production Systems

Notwithstanding extreme aridity and limited renewable water resources, the
Arabian Peninsula has developed indigenous agricultural production systems,
based on crop production under irrigation, and extensive livestock systems.
Rapid economic development in the latter half of the 20th Century has resulted
in significant changes in the traditional agricultural systems of the subcontinent.
Increased agricultural production has contributed to economic growth, but at
the price of degradation of natural resources, particularly the rangelands and the
non-renewable groundwater aquifers.
Country-level data indicate that the cropland areas are very limited in the
Peninsula (Table 10). With the exception of Saudi Arabia and Yemen, the

Table 10. Cropland in the Arabian Peninsula

Country Surface a Crop b Irrigation c
Bahrain 680 7.0 100
Kuwait 17,820 0.4 71
Oman 212,460 0.3 98
Qatar 11,000 1.5 76
S. Arabia 2,149,690 1.8 42
UAE 83,600 1.0 89
Yemen 527,970 2.9 31
Base year for statistics: 1997
a
Country area in km2
b
Crop: % of country area that is cropland (defined as sum of arable land and permanent
cropland)
c
Irrigation: irrigated land as a % of cropland area
Source: World Resources Institute, URL: http://earthtrends.wri.org/country_profiles

-60-
-61-

Fig. 49: Land use and land cover

majority of areas under crops are irrigated. Both statistics underscore the critical
limitation of water, since soil resources for agriculture are much less limiting
(see Section ‘Soils’).
The land use and land cover of the Arabian Peninsula are shown in Figure
49. It is clear that most of the subcontinent is either bare or under very sparse
vegetation. The lack of vegetation is probably as much the result of overgrazing
as aridity. Between 1980 and 1996 the livestock numbers nearly doubled, from
about 15 million to about 28 million sheep and goats, and from about 550,000
to 850,000 camels (FAO, 2001).
The best cover is found in the Yemen Highlands and Asir mountains. It
consists of open shrubland and woodland interspersed with rainfed agriculture.
(This is the only part of the Arabian Peninsula with a growing period adequate
for rainfed agriculture. See Section 4.6.2.)
Between 1980 and 1996, area under irrigation more than doubled, aided by
the use of modern irrigation technology, such as center-pivot and drip irrigation
(Figure 50). Some large areas in the deserts of Saudi Arabia are irrigated, as are
some valleys in Yemen (Figure 49). The spectacular growth in irrigated
agriculture in the center of the Arabian Peninsula between 1983 and 1993 is
shown in the four scenes from the 8-km resolution AVHRR (Advanced Very
High Resolution Radiometer) satellite (Figures 51a-51d). The areas with high
biomass productivity are shown in red or brownish colors. The rainfed areas of
the Yemen Highlands show up clearly, as do some coastal flats with halophyte
vegetation. All other inland areas in red or brown are irrigated. In 1983 (Figure
51a) there were barely any irrigated areas. Ten years later they reached their
maximum extent (Figure 51d).
The changes in biomass productivity as a result of irrigation in the desert
(and some crop area increase in the rainfed areas of Yemen and the Asir) are
shown in Figure 52. It should be noted that the small spots scattered across the
image are probably artifacts due to errors in processing the satellite signal.
Since most of the irrigated agriculture is fed by fossil aquifers, which are
barely recharged, this type of agriculture is obviously not sustainable.

-62-
 a

Fig 51a, b, c, d: Expansion of desert

irrigation, observed from AVHRR imagery.
From top left: situation in 1983 (a), 1986
(b), 1990 (c) and 1993 (d).

-63-
-64-

Fig. 52: Evolution of desert irrigation 1982-1993

7. Research Priorities in Agroecological
Characterization
The Arabian Peninsula suffers from a public perception that it is a subcontinent
of limited heterogeneity, poor agricultural potential, and low population
densities. This view is oversimplified, and the Peninsula is far from monolithic
in its agroecological characteristics. In fact, some agroecological niches,
admittedly small, have high agricultural potential. The subcontinent also has
important potential as a source of genetic diversity and abiotic stress resistance.
In order to realize this potential, there is a clear need for better agroecological
characterization of the Peninsula. Numerous thematic surveys in the form of
soil survey reports, climatic maps, and groundwater surveys exist, therefore, the
need for new resource
inventories is probably
limited. The real need is to
integrate this information in
the form of agroecological
frameworks for development.
Specific methodologies,
models, and decision-support
systems must be developed to
achieve and make use of this
Fig. 50: Evolution of irrigated areas in the Arabian
integration. Peninsula (1980-1996)
The number of synoptic
and climatological stations for a subcontinent of this size is inadequate to map
the variations in agroclimatic conditions, particularly in very dry, hot, and
mountainous areas. With the exception of a few areas, time series of climatic
data are generally short and often interrupted. As well, the temperature,
radiation, moisture, humidity, and wind regimes in the Peninsula’s
agroecological niches will need to be properly characterized. This will require
a network of well-monitored, integrated research sites.
The following research priorities in agroecological characterization have been
identified for the Arabian Peninsula (De Pauw, 1998):
• Regional assessment of crop water requirements for enhancing water use
efficiency
• Agroecological zoning for biodiversity conservation, rangeland
rehabilitation, abiotic stress identification, and development planning
• Improved climate monitoring

-65-
7.1. Regional assessment of crop water requirements
A basic principle behind efficient water use is irrigation to meet the crop water
requirement. Crop water requirements can be calculated from climatic and crop
data. The main challenge in estimating crop water requirements is extrapolation
of climatic and crop coefficient data, obtained by measurement at specific sites,
to areas without data. As mentioned earlier, the climatic station network in the
Arabian Peninsula is sparse and records are recent. In addition, apart from
literature data on crop coefficients, relatively little work has been done on the
reassessment of crop coefficients for local crop cultivars under the specific
climatic conditions of the subcontinent. As a result, assessment of crop water
requirements is usually very site-specific and difficult to generalize to
agroecological zones.
Figure 17, which shows mean annual potential evapotranspiration, is a
rough basis for estimating regional water requirements. This work needs to be
improved, however, by using climatic data on a shorter time scale, and linking
this with crop coefficients for specific crops and cultivars, calibrated in
representative sites.

7.2. Agroecological zoning

There is considerable data available on climate, soils, and water resources in the
Arabian Peninsula. These information sources range from fair to excellent,
depending on the country and theme. With the exception of Yemen, little work
has been done in the region to integrate these data for targeted agricultural
research, ecosystem and biodiversity studies, and development planning. As a
result, data that are potentially highly valuable are under-analyzed, under-
utilized, and under-valued.
Area profiles (agroecological zones) are required to identify different
environments as a basis for a holistic approach to development planning and
resource conservation. Modeling, state-of-the-art interpolation techniques, GIS,
and remote sensing could be used to integrate climatic, soils, and land cover
data into agroecological zones. Agroecological frameworks can form a rational
basis for agricultural research priority setting, promote transferability and
compatibility across countries of research results, and provide an integrated
spatial view of resource availability, quality, use, and degradation risk.
The concept of this research theme is illustrated in Figure 53. A macro-
scale agroecological characterization allows assessment of edaphic diversity at
the level of broad landscapes. This edaphic diversity is nearly always associated

-66-
 Fig. 53: Mapping the distribution of plant communities or species using landscape
frameworks.

with different plant communities. The plants of the Arabian Peninsula have
been described in many studies, but they have not been mapped in a consistent
framework that allows extrapolation across the region. An agroecological
characterization at macro-scale would address the issue of biodiversity
assessment by linking landscapes (with their climatic, landscape and soil
patterns) to plant communities and species.

7.3. Improved climate monitoring

Techniques of data spatialization and remote sensing, such as those used in this
publication, are valuable tools for extracting maximum information content
from data sparse areas. However, they cannot compensate fully for the
tremendous climatic data gap in the Arabian Peninsula. Meteorological services
in the region should make an effort to service the most needy areas, using
agroecological frameworks as a basis for siting weather stations in
representative locations. Automatic stations, with data loggers and data
transmission by telephone or satellite, can overcome the remoteness of such
areas.
In addition, different meteorological data networks, including those
independent of the national meteorological services, should form partnerships,
assisting each other with data collection, pooling, and sharing.

-67-
8. Methods and Data Sources
8.1. General maps
Roads and population centers
Obtained from the Digital Chart of the World (ESRI, 1993).
Annual precipitation background: see ‘Climatic maps.’

Population density
Obtained from Columbia University’s Center for International Environmental
Science Information Network (CIESIN) Gridded Population Database of the
World (GPW version 2). This database contains estimates of population density
(in square kilometers) of the world in 1995.
[On-line document. URL:
http://sedac.ciesin.org/plue/gpw/index.html?main.html&2 ]

8.2. Relief and geomorphology

Altitude
Derived from 1-km resolution global DEM GTOPO30 (USGS, 1996).
Elevation range
Derived from 1-km resolution global DEM GTOPO30 (USGS, 1996) by
applying a range filter on the 8 neighboring cells of each grid cell.
Geomorphological regions
Polyline on-screen digitizing of boundaries from Guba and Glennie (1998) on
the elevation range background.
Salt flats: polygon digitizing based on FAO (1995), with corrections based
on elevation range background. (Afterwards added as a mask.)
Wadi network: obtained from the Digital Chart of the World (ESRI, 1993).

8.3. Climatic maps

All climatic maps were prepared by converting point data into grid datasets
(with 1 km resolution) through spatial interpolation methods. The point data
were obtained from an international climatic database, FAOCLIM (FAO,

-68-
1995b), supplemented with data obtained from meteorological records of the
region. (Particularly for Oman, Qatar, the United Arab Emirates, and, to a lesser
extent, Yemen, stations could be added to the FAOCLIM database.)
The basic climatic variables processed were mean maximum and minimum
monthly temperature (Tmax and Tmin), mean monthly precipitation (Prec), and
mean monthly potential evapotranspiration (PET). The latter parameter was
pre-processed for the stations of the FAOCLIM database using the Penman-
Monteith method (Allen et al., 1998). For stations with only temperature data
available, but not the other variables required by the Penman-Monteith formula
(sunshine/radiation, wind, humidity), a two-step approach was followed to
estimate PET according to the Penman-Monteith method by regression:
1. Calculate PET according to the Hargreaves method (Choisnel et al., 1992):

PET HG= .0023 * Ra * (Tmean + 17.8) * √ (Tmax – Tmin)

With:
Ra: extraterrestrial radiation (calculated from latitude and time of year);
Tmean: mean temperature
Tmin: minimum temperature
Tmax: maximum temperature

2. Estimate Penman PET from Hargreaves PET using regression equations for
climatically homogeneous regions. The regions were obtained from the
Köppen climatic classification and the regression equations used are shown
in Table 11.
Table 11. Conversion equations from PET (Hargreaves) to PET (Penman-Monteith)
Köppen Description Equation r2
region
BSs Semi-arid climate with summer drought PETPM = 1.1058 PETHG –14.909 .90
BSw Semi-arid climate with winter drought PETPM = 0.1478 PETHG1.3689 .85
BW Arid climate PETPM = 1.1594 PETHG – 7.3988 .81
PETPM: PET calculated according to the method of Penman-Monteith (mm)
PETHG: PET calculated according to the method of Hargreaves (mm)
TP: mean monthly temperature (°C)

From these four basic climatic parameters, the following derived climatic
parameters were calculated: mean average monthly temperature (Temp), and
the mean average temperature in summer (Tsum) and winter (Twin).
The interpolation technique used was a thin plate smoothing spline using
the method of Hutchinson (1995) and the software package ANUSPLIN. This

-69-
method is essentially a radial basis interpolation function of the type:
B(h) = (h2 + R2) log (h2 + R2)
With:
B: weight at the grid node
H: anisotropically rescaled, relative distance from the point to the node
R2 smoothing factor specified by the user

In the approach of Hutchinson the smoothing factor, or inversely, the degree

of complexity of the created ‘climate surface,’ is determined automatically from
the database by minimizing a measure of predictive error of the fitted surface
given by the generalized cross validation (GCV). In the surface fitting
procedure three independent spline variables were used, longitude, latitude, and
elevation above sea level. They were considered the most appropriate for fitting
surfaces related to temperature or precipitation parameters.
Twelve monthly climate surfaces were created for each of the basic climatic
parameters (Tmax, Tmin, Temp, Prec, and PET), and a surface for Tsum and
Twin. These elementary climate surfaces were combined into various derived
climate surfaces using formulas and models, which will be explained in the
following sections.
Aridity index (AI)
12 12
AI = Σ
i=1
prec/
Σi=1 pet
with i: month number
prec: total precipitation during month I
pet: total potential evapotranspiration (Penman-Monteith) during month I

Precipitation deficit (PD)

12
PD = Σ (prec – pet )
i=1
i i

Annual heat units (AHU)

12
AHU= Σ (Temp x NumDays )
i=1
i i Temp>Threshold

-70-
with: Temp: mean monthly temperature (°C) during month i
NumDays: number of days in month i
Threshold: temperature below which no accumulation is done (in this study: 0°C)

Climatic growing period (GP)

GP_ON = (Date)aet/pet>Threshold
GP_END = (Date)aet/pet<Threshold
LGP = GP_END – GP_ON

with: GP_ON: growing period onset date

GP_END: growing period end date
LGP: length of growing period

In the Arabian Peninsula low temperature is not a significant limiting

factor, therefore only the moisture limitation is considered to determine the
growing period.
The criterion used for the definition of a moisture-limited growing period
is whether the ratio of actual evapotranspiration (AET) to potential
evapotranspiration (PET) for any particular month is higher than a user-defined
threshold. If it is, that month is part of a growing period, if it is not, that month
is not part of the growing period. The start date of the growing period is
obtained from linear interpolation of the AET/PET ratios between the last
month that is part of the growing period, and the first month that is not part of
the growing period. The end date, inversely, is obtained by linear interpolation
of the AET/PET ratios between the last month that is part of the growing period,
and the first one that is not part of the growing period.

Thre – R0
GP_ON = M_Start+NDays
R1 – R0

Thre – Rn-1
GP_END = M_End+NDays2
Rn – Rn-1

with: M_Start: the number of days from 1 January up to the end of the last month that is not
part of the growing period

-71-
 M_End: the number of days from 1 January up to the end of the month preceding the last
month of the growing period
NDays: number of days in the first month of the growing period
NDays2: number of days in the last month of the growing period
Thre: |AET/PET threshold for defining a growing period (user-defined; for this study set
to .5)
R0: AET/PET ratio for the month preceding the first month of the growing period;
R1: AET/PET ratio for the first month of the growing period;
Rn-1: AET/PET ratio for the month preceding the last month of the growing period;
Rn: AET/PET ratio for the last month of the growing period.

If more than one growing period occurs, a distinction is made between the
first and the second growing period, but the calculation procedure is the same.

Similarity index
A combined temperature-precipitation based similarity index is calculated as
follows1:
1. For each grid cell the 12 monthly mean temperature (Temp) and precipitation
values (Prec) are taken;
2. The square deviations with the match locations are summed:

12
[(
Tr = ∑ 10 Tempi − Ti
2
)]
i =1
and
12
(
Pr = ∑ Pr eci − Pi
2
)
i =1
3. The deviations are sorted and ranked into arrays [Tr]n and [Pr]m.
4. The similarity index for temperature in a grid cell j is then calculated as:

(
 1 − rank Tr j , Tr 
Ts j = 100 1 − 
)
 N −1 

1
Procedure developed by Dr. F. Pertziger, SANIGMII, Tashkent, Uzbekistan.

-72-
and similarity in precipitation as:
(
 1 − rank Pr j , Pr 
Ps j = 100 1 − ,
)
 M −1 

( )
in which rank b, A is a ranking number of b in array A .

5. The combined temperature-precipitation similarity is calculated as:

(T S WT )2 + (PS W P )2
S = 100 ,
WT2 + W P2

where the WT and WP are the weights assigned to temperature and precipitation, respectively. In
this study, equal weights have been used for WT and WP.

Biomass productivity indices

A distinction is made between biomass productivity indices for natural
vegetation/rangeland and for crops.
One rangeland biomass productivity index (RBPI) is defined as follows:

RBPI = AHU x AI

with: AHU: annual heat units (°C days)

AI: aridity index

The RBPI can, therefore, be considered as the atmospheric energy available

for biomass production, as expressed by accumulated temperature, adjusted for
the moisture regime.
Biomass productivity indices for crops can be developed using the same
principle, except that temperature outside the time bounds of the moisture-
limited growing period are not considered.
The first step is to calculate a daily adjusted thermal increment (ATI) as a
function of the adaptability range for each crop group.

-73-
Table 12. Adaptability to temperature for different crop groups (adapted from FAO,
1978)

Crop group T0 Topt1 Topt2 Tx

I 5 15 20 33
II 10 25 30 45
III 15 25 35 50
IV 10 20 30 45
Adaptability expressed in relation to four cardinal temperature points (all in °C):
T0 : the daytime temperature below which no assimilation takes place (cold-limited);
Topt1 : the lower daytime temperature threshold above which maximum assimilation takes
place;
Topt2 : the higher daytime temperature threshold above which assimilation rate declines;
Tx : the day-time temperature above which no assimilation takes place (heat-limited)

For a particular daytime temperature Tday (as assimilation takes place

during the day), ATI can be defined as follows:

ATI =0 [Tday <= T0 or Tday >= Tx]

ATI = Tday – T0 [Tday > T0 and Tday < Topt1]
ATI = (Topt1 + Topt2)/2 - T0 [Tday >= Topt1 and Tday <= Topt2]
ATI = Tx – Tday [Tday > Topt2 and Tday < Tx]

The concept of ATI is thus related to the concept of heat units, but the
accumulation is weighted according to the distance of the real daytime
temperature from the optimal daytime temperature for each crop group.
The daytime temperature is estimated from the minimum and maximum
temperature as:
Tmax - Tmin
Tday = Tmean +
π

The ATI values are summed for each crop group on a daily basis between
the onset and end dates of the growing period. Since daily data are needed for
operational reasons, and the interpretation derived from this exercise does not
require high precision, the daily values can be interpreted from the monthly
temperature values through linear interpolation.

-74-
The biomass productivity index for each crop group can then be defined as:
GP _ END
CBPI j = ∑ ( ATI )
i = GP _ ON
i, j

with: j: crop group

I: day number
ATI: adjusted thermal increment (°C)
GP_ON: growing period onset date
GP_END: growing period end date

8.4. Soil maps

All soil information was obtained, directly or indirectly, from the FAO Soil Map
of the World (SMW).

Soil associations
Obtained directly from the digital version of the SMW (FAO, 1995).

Derived soil properties

The original resolution of the derived soil property maps on the Digital SMW
is 5 arc-minutes (about 10 x 10 km). To allow mapping of the derived soil
properties at 1 km resolution, the original Soil Associations vector file was
converted to a 30 arc-second grid. The QuickBasic code of the viewing program
IMAGES.BAS was then modified to allow operation on a 30 arc-second grid
and to export the generated maps as ASCII grids.

8.5. Land use and cover maps

Land use/cover

Map clipped from CWANA map in Celis and De Pauw (2001).

Irrigated areas
NDVI (Normalized Difference Vegetation Index) from AVHRR 8 x 8 km
downloaded from Goddard DAAC as monthly datasets for period 1981-94.
(URL:

-75-
http://daac.gsfc.nasa.gov/CAMPAIGN_DOCS/LAND_BIO/GLBDST_Data.ht
ml)
Threshold of NDVI > 0.46 used to superimpose masks for years 1982,
1985, 1987, 1990, and 1993 representing expansion of irrigated areas.

References
Allen, R.G., Pereira, L.S., Raes, D. and Smith M. 1998. Crop evapotranspiration: Guidelines for
computing crop water requirements. FAO Irrigation and Drainage Paper 56. FAO, Rome.
300 pp.
Barth, H.K. 1976. Natürraumliche Gliederung und Einzellandschaften. Pages 51-90 in Saudi
Arabien. Natur, Geschichte, Mensch und Wirtschaft (Blume, H.H. ed.) H. Erdmann Verlag
Celis, D. and De Pauw, E. 2001. Assessment of land cover/land use and growing period
characteristics in CWANA region using AVHRR imagery. ICARDA Research Report,
ICARDA, Aleppo, Syria.
Choisnel, E., de Villele, O. and Lacroze, F. 1992. Une approche uniformisée du calcul de
l'évapotranspiration potentielle pour l'ensemble des pays de la Communauté Européenne.
Centre Commun de Recherche, Commission des Communautés Européennes, EUR 14223.
178 pp.
Cochemé, J. and Franquin, P. 1967. An agroclimatology survey of a semi-arid area in Africa
south of the Sahara. World Meteorological Organization, 86, No. 210, TP 110.
De Pauw, E. 1983. Agricultural potential and constraints in relation to growing periods and their
variability in Tanzania. Pédologie 33:2: 147?170.
De Pauw, E. 1998. Status of agroecological characterization in the Arabian Peninsula. Report
prepared for the Regional Technical Coordination Meeting held in Manama, Bahrain, 24-26
May 1998. Arabian Peninsula Regional Program, ICARDA, Aleppo, Syria.
Driessen, P.M. and Dudal, R. (eds.) 1991. The major soils of the world. Lecture notes on their
geography, formation, properties and use. Agricultural University of Wageningen and
Katholieke Universiteit Leuven. 310 pp. ISBN 90-800725-1-6.
ESRI (Environmental Systems Research Institute). 1993. Digital chart of the world. ESRI,
Redlands, California, USA
FAO (Food and Agriculture Organization of the United Nations). 1974. FAO-UNESCO soil
map of the world. Vol.I: Legend. UNESCO, Paris.
FAO (Food and Agriculture Organization of the United Nations). 1978?81. Reports of the
agro?ecological zones project. World Soil Resources Report No. 48, Vol.1 Africa. FAO,
Rome.
FAO (Food and Agriculture Organization of the United Nations). 1988. FAO-UNESCO soil
map of the world: revised legend. World Soil Resources Report No. 60. FAO, Rome, Italy.
FAO. (Food and Agriculture Organization of the United Nations). 1995. Digital soil map of the
world and derived soil properties. CD-ROM, version 3.5, November 1995.
FAO. (Food and Agriculture Organization of the United Nations). 1995b. FAOCLIM 1.2. A
CD-ROM with a compilation of monthly worldwide agroclimatic data. FAO
Agrometeorology Working Paper Series No.11. FAO, Rome, Italy. 68 pp. (and one CD-
ROM)

-76-
FAO. (Food and Agriculture Organization of the United Nations). 2001. FAOSTAT Agriculture
Data. FAO Statistical Databases. http://apps.fao.org/page/collections?subset=agriculture
Fisher, M. and Membery, D.A. 1998. Climate. Pages 5-38 in Vegetation of the Arabian
Peninsula. Geobotany 25 (Ghazanfar, S.A. and Fisher, M. eds.). Kluwer Academic
Publishers. Dordrecht, the Netherlands, 362 pp.
Gesch, D.B. and Larson, K.S. 1996. Techniques for development of global 1-kilometer digital
elevation models. in Pecora Thirteen, Human interactions with the environment.
(And on-line document. URL: http://edcdaac.usgs.gov/gtopo30/README.html)
Guba, I. and Glennie, K. 1998. Geology and geomorphology. Pages 39-62 in Vegetation of the
Arabian Peninsula. Geobotany 25 (Ghazanfar, S.A. and Fisher, M. eds.). Kluwer Academic
Publishers , Dordrecht, the Netherlands, 362 pp.
Hutchinson, M.F. 1995. Interpolating mean rainfall using thin plate smoothing splines.
International Journal of Geographical Information Systems, 9: 385-403.
Kürschner, H. 1998. Biogeography and introduction to vegetation. Pages 63-98 in Vegetation of
the Arabian Peninsula. Geobotany 25 (Ghazanfar, S.A. and Fisher, M. eds.). Kluwer
Academic Publishers, Dordrecht, the Netherlands, 362 pp.
Ryan, J., Masri, S., Garabet, S., Diekmann, J. and Habib, H. 1997. Soils of ICARDA's
agricultural experiment stations and sites. ICARDA, Aleppo, Syria, 107 pp.
USDA (United States Department of Agriculture) 1951. Soil survey manual. USDA,
Agriculture handbook No. 18. USDA, Washington, D.C., USA.
UNESCO (United Nations Educational, Scientific and Cultural Organization) 1979. Map of the
world distribution of arid regions.
Map at scale 1:25,000,000 with explanatory note. UNESCO, Paris. 54 pp.
World Resources Institute. 2001. EarthTrends.
On-line database. URL: http://earthtrends.wri.org/country_profiles

-77-
 An agroecological exploration of the
Arabian Peninsula

by

Eddy De Pauw

International Center for Agricultural Research in the Dry Areas

-78-