Content uploaded by Pierre De Geest
Author content
All content in this area was uploaded by Pierre De Geest on Mar 16, 2018
Content may be subject to copyright.
This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
Author's personal copy
The climate of Socotra Island (Yemen): A first-time assessment of the timing
of the monsoon wind reversal and its influence on precipitation and
vegetation patterns
Paul Scholte
a
,
b
,
*
, Peter De Geest
c
a
Socotra Conservation and Development Program, Yemen
b
Kitabi College of Conservation and Environmental Management, Rwanda, c.o. Nieuwe Teertuinen 12C, 1013 LV Amsterdam, The Netherlands
c
Department of Geology, DGLG-WE, Pleinlaan 2, 1050 Brussels, Belgium
article info
Article history:
Received 18 February 2009
Received in revised form
10 January 2010
Accepted 20 May 2010
Available online 29 July 2010
Keywords:
Dragonblood
Drought
Fog
Indian Ocean Monsoon
Rainfall variability
Regeneration
abstract
The climate of Socotra, influenced by the Indian Ocean Monsoon, is poorly known, hampering under-
standing of its paleoclimate and (endemic) biodiversity. Mean annual rainfall and temperature, measured
in a network of meteorological stations from 2002 to 06, were 216 mm and 28.9
C. Combined with cloud
cover information from satellite images, this data provides clear ideas on inter- and intra-annual
variability. Precipitation derived from the northeast (NE) winter monsoon influences especially the NE
plateaus and windward side of the Haggeher Mountains because of orographic effects. The southwest
(SW) summer monsoon concentrates at the southern half of the island and generally produces less
rainfall. During the SW summer monsoon, clouds cover the highlands and plateaus south of the
Haggeher Mountains, creating fog. Preliminary measurements suggest that at higher altitudes, fog-
derived moisture may constitute up to two-thirds of total moisture, amounting up to 800 mm. The
predominant SW aspect of the enigmatic dragonblood tree underlines the importance of fog. Long-term
weather observations by Socotri put these short-term meteorological observations into a longer
perspective. Socotri informants also described the drought years when livestock populations crashed,
after which windows of opportunities for the regeneration of dragonblood and other grazing-sensitive
trees may have occurred.
Ó2010 Elsevier Ltd. All rights reserved.
1. Introduction
1.1. General
Socotra Island, situated in the north-western (NW) Indian Ocean
(Fig. 1) is known for its spectacular flora, including the dragonblood
tree (Dracaena cinnabari), frankincense (Boswellia spp.) and aloe
(Aloe spp.) that once dominated the world’s economy. The island’s
large number of endemic plant species (320, or 37% rate of specific
endemism) and numerous endemic animals, motivated its 2008
listing as natural World Heritage Site (UNESCO, 2008). The natural
diversity has its background in the weather pattern that is char-
acterised by large spatial, altitudinal, seasonal and inter-annual
variability. The climate is influenced by the seasonally reversing
monsoon wind system and oceaneatmospheric interactions such
as the Indian Ocean Dipole (IOD) (Saji et al., 1999; Webster et al.,
1999; Prasad and McClean, 2004) and the El Niño eSouthern
Oscillation (ENSO) (Neff et al., 2001; Abram et al., 2007). The
description of the island’s climate has, however, remained largely
anecdotal (Wellsted,1835; Popov, 1957; Mies and Beyhl, 1998), with
only preliminary site-specific assessments (Mies, 2001; Culek et al.,
2006).
Several paleoclimate reconstructions were inferred from
speleothems since Socotra’s location on the inter-tropical
convergence zone (ITCZ) is ideal to study the migration of climate
belts (Fleitmann et al., 2004, 2007; Shakun et al., 2007). The
application of climate-speleothem proxy relations should be
based on a detailed understanding of the present climate. Also
ecological studies, like the ones on the lack of regeneration of the
enigmatic dragonblood tree (Adolt and Pavlis, 2004; Attorre et al.,
2007) with a generation time of several centuries, can only be
interpreted based on a thorough understanding of climatic
variability.
This paper contributes to this understanding by assessing the
timing of the monsoon wind reversals and the relative importance
*Corresponding author at: Kitabi College of Conservation and Environmental
Management, Rwanda, c.o. Nieuwe Teertuinen 12C, 1013 LV Amsterdam,
The Netherlands.
E-mail address: PaulT.Scholte@gmail.com (P. Scholte).
Contents lists available at ScienceDirect
Journal of Arid Environments
journal homepage: www.elsevier.com/locate/jaridenv
0140-1963/$ esee front matter Ó2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jaridenv.2010.05.017
Journal of Arid Environments 74 (2010) 1507e1515
Author's personal copy
of the monsoon coupled rainfall periods originating from a north-
eastern (NE) direction versus a south-western (SW) direction. It
will assist in further understanding the unique botanic diversity of
Socotra, especially in relation with projected climate change
(Attorre et al., 2007). The presented Socotra’s climate dataset may
also contribute to the further understanding of the Indian Ocean
Monsoon system that affects the lives of almost half of the world’s
population, whereas its response to global change is not fully
understood (Leuschner and Sirocko, 2003).
1.2. The island’s geography
The Socotra Archipelago, consisting of four islands, is situated
between the Horn of Africa and the Arabian Peninsula (Fig. 1).
Socotra, focus of this study, is the largest island with a surface area
of 3625 km
2
(Cheung and DeVantier, 2006). Geological recon-
structions place Socotra close to present-day Dhofar (Oman),
(Fleitmann et al., 2004), whereas politically the Archipelago is part
of Yemen.
Socotra can be subdivided into three geomorphological zones:
predominantly alluvial coastal and inland plains, limestone
plateaus and the Haggeher Mountains. The island’s diagnostic
geographic feature and main watershed, the igneous Haggeher
mountain range, reaches 1540 m. The southern coastal plain,
measuring 70 by 5 km, is cut off by 300e400 m high escarpment
cliffs of the limestone plateaus. In the north the coastal plain is
narrower, interrupted by Wadi systems terminating in brackish
pools separated from the sea by spits and bars. The arid Zahr basin
dominates the western part of the island. The limestone plateaus,
characterised by karstic features (De Geest, 2006), cover more
than half of the island and are generally between 300 and 700 m
asl, reaching 800 m at Ma’alah in the west and 1000 m at Diksam
(Fig. 2). Wadi systems originating from the Haggeher Mountains
are deeply cut-in these plateaus in especially the southern part of
the island.
1.3. Monsoon climate
In winter, the atmospheric pressure gradient between the high-
pressure cell over the Eurasian continent and the low pressure ITCZ
over the southern Indian Ocean, results in moderate NE winds. In
spring the northern tropical and subtropical landmasses warm-up,
creating SW winds. In summer the ITCZ reaches its northernmost
position resulting in a maximum intensity of the Indian Ocean
Monsoon. In autumn the southward migration starts again and the
strength of the monsoon diminishes (Fleitmann et al., 2004)
(Fig. 3).
1.4. Limited meteorological data
The first measurements of temperature, wind directions and
humidity with casual remarks on rainfall were made in the early
19th ce ntu ry (Kerr and Edin, 1811e1824; Wellsted, 1835; De Gray
Birch, 1875; Forbes, 1903). During World War II, the British army
installed a weather station at Mouri Airport (northern coastal
plain) that functioned for three years (Popov, 1957). The timing
of the reversal in the monsoonal wind pattern and the related
rainfall periods drew most attention, as they were judged
Fig. 1. Location of the Socotra Archipelago in the NW Indian Ocean (adapted after
Kopp, 1999).
Fig. 2. Map of Socotra Island showing the location of the manual weather stations.
P. Scholte, P. De Geest / Journal of Arid Environments 74 (2010) 1507e15151508
Author's personal copy
differently to the extent of bringing precipitation by either the
SW (Mies and Beyhl, 1998) or NE monsoon (Wranik, 2003; Miller
and Morris, 2004). The complexity of Socotra’s climate was
further highlighted by Mies (2001) who, based on 24 h air
humidity measurements in the lower Haggeher (800 m),
postulated that half of the moisture in areas above 700 m may
originate from fog. More recently an automated weather station
was installed at Firmihin, at mid-elevation (440 m asl), providing
five years site-specific meteorological data for this limestone
plateau with dragonblood forest (Adolt and Pavlis, 2004)
exposed to the summer SW monsoon (Kral, 2005; Culek et al.,
2006).
2. Materials and methods
2.1. Weather station set-up and collection of meteorological data
In 2001, a network of 10 manual stations was set-up by the EPA
to which Homhil station was added in 2003 (Fig. 2). Its selection
was based on geometrical layout, geomorphological representation
and operational feasibility. Elevation was only partly covered with
no stations above 800 m. Each weather station was equipped with
a thermometer and a rainfall collector (diameter 19 cm, depth
10 cm) with a millimetre gradual level indicator, shielded from
direct sunlight and protected from harsh winds, humans and live-
stock. Since November 2001 and March 2002 respectively, rainfall
(accuracy of 1 mm) and temperatures (min. and max. with accuracy
of 0.1
C) were daily registered by EPA officers. Data was cross-
checked and monthly entered for analysis into Excel. An automatic
solar powered Cumulus weather station (508e100 series) was
placed at the Diksam plateau on 29 September 2003. Rainfall, air
temperature and relative humidity were measured every hour till
July 10, 2004. Wind direction and global radiation were registered
until January 15, 2005.
2.2. Satellite cloud cover images
A selection of 142 out of a total of 150 satellite images, freely
available on the internet as previews (USGS-Earthexplorer, USGS-
Glovis, NASA-Johnson Space Centre- The Gateway to Astronaut
Photography of the Earth, see web references), encompassing all
months of the year from 1962 to 2005, were used to assess cloud
formation and prevailing wind direction (De Flou, 2006). Images
were chronologically ordered, resulting in the representation of
6e17 satellite previews per month. By visually examining the cloud
cover on each image, indications were obtained on cloud formation
and the prevailing wind direction. Based on intra-monthly
Fig. 3. ITCZ migration pathway and the Indian Ocean Monsoon winds.
0
20
40
60
80
100
120
140
160
17
19
21
23
25
27
29
31
33
35
37
Jan-02
Mar-02
May-02
Jul-02
Sep-02
Nov-02
Jan-03
Mar-03
May-03
Jul-03
Sep-03
Nov-03
Jan-04
Mar-04
May-04
Jul-04
Sep-04
Nov-04
Jan-05
Mar-05
May-05
Jul-05
Sep-05
Nov-05
Jan-06
Mar-06
May-06
Jul-06
Sep-06
Nov-06
Rainfall (mm)
Temperature (°C)
Rainfall Av. Temp. Max. Temp. Min. Temp.
Fig. 4. Overview of mean monthly rainfall and temperature on Socotra (2002e20 06) (mean annual temperature 272, 261, 569, 93, 224 mm respectively; mean annual temperature
29.2, 29.3, 28.2, 29.2, 28.8 C respectively).
P. Scholte, P. De Geest / Journal of Arid Environments 74 (2010) 1507e1515 1509
Author's personal copy
variations, the frequency of changes was visualised, allowing the
assessment of wind reversal within a month.
2.3. Fog collection
The Environmental Protection Authority (EPA), in collaboration
with CARE-Yemen, carried out an experiment to test if water needs
for isolated mountain communities could be met with fog collec-
tion. Four mist nets, ranging in size from 1 to 12 m
2
, were used to
collect fog in three locations on Diksam and one location on Momi
plateau from July to September 2004. Applying widely used net
types, these measurements allow comparisons with the well-
studied fog oasis of Dhofar, Oman (Hildebrandt and Eltahir, 2006;
Abdul-Wahab et al., 2007).
2.4. Oral history
We tapped Socotra’s long oral tradition, base of daily survival,
as a complementary source of climatic information through
informal interviews that we held amongst the EPA community
liaison officers. Indicated years should, however, be taken with
some caution as written records to triangulate this information are
lacking.
3. Results
3.1. Weather station records
Mean annual rainfall, measured at the 11 manual stations from
2002 to 2006, was 284 mm, or 216 mm when excluding the
abnormal Homhil records, with important inter-annual variability
(Fig. 4). Maximum rainfall was reached in November, with an
average of 120 mm or 42% of the mean average rainfall. In March,
July and August there was generally no rain.
The mean annual rainfall at individual stations showed
considerable variation (Fig. 5). At Mathre, Qariah, Hay as Salam,
Qedemeno and especially Homhil, the amount of rainfall at the end
of the year (winter rains) was higher than from April to June
(summer rains).
May was the hottest month with an average recorded temper-
ature of 31.2
C whereas a second pronounced warm period
occurred in SeptembereOctober with mean values of 29.1e29.0
C
respectively. This trend was observed in all stations, except
Ghahandad at the southern coast (Fig. 2:10) where the temperature
remained virtually constant throughout the year (Fig. 5). The
amplitude between the mean annual minima and maxima, repre-
senting differences between day and night temperatures, ranged
between 10.0
C and 2.6
C. The mean annual temperature was
28.9
C0.3
C, with 28.2
C in the wet year 2004 (Fig. 4). Popov
(1957) observed a nocturnal minimum temperature of 13.5
Cin
the Haggeher Mountains in January; we measured a minimum
temperature of 8.7
C during the early morning of January 15, 2006
at the summit. Herders told us it does get even colder, but no frosts
have ever been recorded in oral traditions during the period of
habitation of several millennia.
Precipitation and temperature data of the Diksam auto-
mated station can be correlated with data from nearby
Dihaher manual station. The relative humidity data showed
not only a high value (>95%) during periods of rainfall, but
also during cloudy months. Late September the mean relative
0
50
100
150
20
25
30
35
40
1
2
3
4
5
6
7
8
9
10
11
12
(mm)
(°C)
Months
(1) Qalansiya (3m asl)
Average rainfall: 67mm
Mean temperature: 30,0°C
0
50
100
150
20
25
30
35
40
1
2
3
4
5
6
7
8
9
10
11
12
(mm)
(°C)
Months
(2) Mathre (20m asl)
Average rainfall: 400mm
Mean temperature: 29,3°C
0
50
100
150
20
25
30
35
40
1
2
3
4
5
6
7
8
9
10
11
12
(mm)
(°C)
Months
(3) Qariah (30m asl)
Average rainfall: 174mm
Mean temperature: 29,7°C
0
50
100
150
20
25
30
35
40
1
2
3
4
5
6
7
8
9
10
11
12
(mm)
(°C)
Months
(4) Homhil (350m asl)
Average rainfall: 1012mm
Mean temperature: 26,2°C
0
50
100
150
20
25
30
35
40
1
2
3
4
5
6
7
8
9
10
11
12
(mm)
(°C)
Months
(5) Beytoh (150m asl)
Average rainfall: 249mm
Mean temperature: 27,4°C
0
50
100
150
20
25
30
35
40
1
2
3
4
5
6
7
8
9
10
11
12
(mm)
(°C)
Months
(6) Dirowah (120m asl)
Average rainfall: 187mm
Mean temperature: 29,3°C
0
50
100
150
20
25
30
35
40
1
2
3
4
5
6
7
8
9
10
11
12
(mm)
(°C)
Months
(8) Hay as Salam (180m asl)
Average rainfall: 358mm
Mean temperature: 28,2°C
0
50
100
150
20
25
30
35
40
1
2
3
4
5
6
7
8
9
10
11
12
(mm)
(°C)
Months
(9) Qedemeno (390m asl)
Annual rainfall: 191mm
Mean temperature: 27,2°C
0
50
100
150
20
25
30
35
40
1
2
3
4
5
6
7
8
9
10
11
12
(mm)
(°C)
Months
(10) Ghahandad (15 m asl)
Average rainfall: 141mm
Mean temperature: 29.7°C
0
50
100
150
20
25
30
35
40
1
2
3
4
5
6
7
8
9
10
11
12
(mm)
(°C)
(11) Bidollah (10m asl)
Average rainfall: 177mm
Mean temperature: 29,6°C
0
50
100
150
20
25
30
35
40
1
2
3
4
5
6
7
8
9
10
11
12
(mm)
(°C)
Months
Overall mean 2002-2006
Annual rainfall: 216 mm
Mean temperature: 28,9°C
0
50
100
150
20.0
25.0
30.0
35.0
40.0
1
2
3
4
5
6
7
8
9
10
11
12
(mm)
(°C)
Months
(7) Dihaher (760 asl)
Average rainfall: 168mm
Mean temperature: 27,6°C
Fig. 5. Overview of mean monthly rainfall and temperature (Max./Mean/Min.) for each manual weather station (2002e2006) (1e11) and the combined overall mean for Socotra.
P. Scholte, P. De Geest / Journal of Arid Environments 74 (2010) 1507e15151510
Author's personal copy
humidity dropped sharply to 50%, and rose again during
(rainy) November. December and January were relative humid
months with cold precipitation. February to April were dry
months, until it started raining at the end of April or begin-
ning of May. Humidity stayed high during the month of June
(Fig. 6). The daily amplitude in global radiation was filtered out
by taking the daily mean temperature, with a clear inverse
relationship between the latter and relative humidity. The
wind reversal periods were marked by the wind direction data
(Fig. 7); at the end of April the wind shifted from a predomi-
nant NE direction towards a constant S or SE direction, lasting
till late September.
3.2. Relation cloud cover ewind direction
Late June till late September was characterised by a well-
developed SW-cloud cover type with orographic cloud formations
over the southern cliffs, the elevated southern limestone plateaus
and southern Haggeher Mountains, which all experienced fog
formation, resulting in daily drizzles. The western part of the island
was less covered by clouds, with the exceptions of Ma’alah and
a small plateau further to the SW. The northern coastal regions
remained cloudless (Fig. 8:fei). The SW monsoon cloud cover was
very stable, without exception over the years. In October both SW-
and NE-cloud cover types were found (Fig. 9). All SW-cloud cover
Fig. 6. Rainfall, Temperature and Relative Humidity at Diksam station (Sept. 2003eJuly 2004).
Fig. 7. Wind direction and Global Radiation, at Diksam station (Sept. 20 03eJan. 2005).
P. Scholte, P. De Geest / Journal of Arid Environments 74 (2010) 1507e1515 1511
Author's personal copy
types occurred before October 16; three out of four NE-cloud cover
types occurred after October 26, indicating a quick wind reversal
during October. November and December showed a clear NE cloud
cover type with orographic cloud formations over the NE limestone
plateaus (Momi), the area north of the Haggeher Mountains and the
western part of the island at the elevated coastal cliffs and north of
Ma’alah. There was also an important cloud intrusion following the
NE-SW limestone range west of Shibehon penetrating the island to
the northern parts of the southern limestone cliffs (Fig. 8: k and l).
The NE cloud cover type was stable during November and
December in respectively 75% and 85% of the images (Fig. 9).
January and especially February showed less cloud cover. In
January, most of the time clouds covered the northern flanks of the
Haggeher which was less consistent in February when the first
clouds at the southern coastal cliffs were observed (Fig. 8: a and b).
Images with a clear NE cloud cover type were limited to the first
half of February, which suggests the beginning of the spring-inter
monsoon by the end of February (Fig. 9). In March and April, there
was limited cloud cover with no clear wind direction component,
characteristic for the inter monsoon (Fig. 8:aed). From March till
April, SW-cloud cover types increased from 20% to almost 50% of
the images (Fig. 9). Yet in April, the SW-pattern was very variable
and sometimes clouds only formed along the NE-SW limestone
range west of Shibehon and along the southern coastal cliffs
(Fig. 8: d). About half of the April-images were without pattern,
indicating that April starts with the end of the inter-monsoon and
ends with the beginning of the SW monsoon. In May the SW-cloud
cover component developed further, with the building-up of the
characteristic SW-cloud cover type in June (Figs. 8:e, f and 9).
3.3. Quantities of fog harvested with nets
Fog harvested with mist nets during the summer monsoon
reached up to the equivalent of 10 mm of rain per day. The variation
Fig. 8. Cloud cover of Socotra, as indicated by a monthly selection of the most characteristic previews.
Fig. 9. Classification of cloud patterns on 142 satellite images of Socotra (see text).
P. Scholte, P. De Geest / Journal of Arid Environments 74 (2010) 1507e15151512
Author's personal copy
in collected quantities underlines the importance of location and
aspect.
3.4. Local climatic knowledge
We reviewed oral knowledge, passed on through genera-
tions (Morris, 2002), for a long-term understanding of the
Socotra climate (Socotri word in between [ ]). The NE
Monsoon from late October to February is called [serb], during
which a cooling, much appreciated N wind [serbihi]brings
rains. The [serb] winter rains are plentiful and prolonged,
especially at the high NE and central plateaus and the interior
valleys followed by the Central Highlands. The northern parts
of the Haggeher Mountains and Momi Plateau can be covered
incloudandmistandexperiencesomedrizzleinterspersed
with rainfall. Mid-February, a transition period begins, known
as the short summer [qeyat] lasting until April. In the
lowlands, it is generally uncomfortably hot. The [qeyat]rains
are unpredictable in time and space. The short period of
intense heat and stillness in April is called [minqeyat], towards
the end of which clouds begin to gather. The transitional pre-
monsoon season [doti], running from the second half of April
to the end of May or early June, is a period with possible rain
again. In years when the [doti] rains are light or fail, the SW
monsoon season [horf] from June to late September, will be
very hard. When [doti] rains fail it is said that the next [serb]
rains have a tendency to fail too. Normally though during the
[horf] the mountains, and to some extent the high limestone
plateaus, are covered in mist, with a constant drizzle. [Zer-
ebhen] is the transition period in late SeptembereOctober, as
the SW Monsoon winds decline and shift to the NE Monsoon.
The first day is marked by a short period with no winds, after
which the winds start to shift and clouds pile up and disperse
repeatedly, accompanied in many areas by heavy dewfall.
Linked with the survival of livestock, the island’s interior
economic base, droughts are well remembered and named.
Most droughts are the consequence of failing [doti] summer
rains, followed by poor [serb] winter rains. Especially 1844
[Difareten]:‘hearing the nails clapping’, 1942e1943 [Dimindah]:
‘when the vultures where eating corpses’and 1954 were
reported as severe droughts and the island was subsequently
de-populated (Morris, 2002). Droughts also occurred in 1972,
1976 and 1981 when summer rains failed. Even the winter rains
failed in 1978e1979 and 1980e1981. The summer rains of
1984e
1985 and 1987 were very light, causing considerable
livestock losses. Summer rains generally failed in 1993 and were
completely absent in 1994. In 1999 with failing summer rains,
cattle in the Haggeher Mountains died, and the number of goats
and sheep dropped with 40% (Scholte et al., 2008). Even more
disastrous than these droughts are severe winter floods after
failing summer rainfall, as weakened cattle die from cold and
houses are destroyed. Socotri on the central limestone plateaus
thus remember 1999 [enoh di mewaat]: ‘year of the dead’,as
subsequent floods were followed by a measles epidemic
(Morris, 2002).
4. Discussion and conclusions
4.1. General weather pattern
The annual weather pattern on Socotra consists of the winter
and summer monsoon, separated by autumn and spring transition
periods. The autumn transition period lasts from early till late
October. Wind changes from an SW towards NE direction,
although with high variability. Relative humidity decreases and
temperature rises towards the end of the period, when the first
rains may fall. During the winter monsoon, from late October to
early February, winds blow from NE direction bringing the largest
annual rainfall in November. It affects the entire island, although
the northern regions are more influenced due to orographic
rainfall. The subsequent spring transition period, from mid-
February until the first half of April, is generally dry, hot, rather
cloudless while the general wind direction changes from the NE
towards the SW with less variability. The summer monsoon starts
with increasing wind speed, bringing rainfall in the second half of
April, sometimes lasting till early June, influencing the southern
regions of the island and only sporadically reaching the northern
regions (Culek et al., 2006). From July until mid-August harsh
winds generally blow from a western direction creating cloud
cover above the southern coastal and especially higher altitude
plateaus. Late August eearly September, the SW wind generates
high humidity.
Rainfall peaks around November, sometimes starting at the
end of September and generally lasting till early February (NE
derived winter rains) (Fig. 5). The second period of rainfall with
(less) precipitation is AprileMay, which sometimes start in
March and may last until June (SW derived summer rains). Both
rainy seasons are alternated by dry periods, occurring in March
and JulyeAugust. Comparing both rainy seasons from 2002 to
2006, we notice that they did not develop during the same
periods and were not always equal in duration (1e5 months). At
the manual stations of Qalansiya, Beytoh, Dirowah, Dihaher,
Ghahandad and Bidollah, the differences were less pronounced,
although also in these places winter precipitation dominated
(Fig. 5). The annual rainfall registered at Homhil, with excep-
tional annual maxima well over 1000 mm in both 2003 and 2004
(Fig. 5), suggests errors in registration. Excluding Homhil, the
mean annual rainfall registered on Socotra in 2002e2006 was
216 mm.
The NE continental winds of the winter monsoon take up
moisture while passing over the warm Arabian Sea, explaining the
wet winter monsoon on Socotra. The 400e600 m high limestone
cliffs at the northern and southern coast and elevated plateaus
around the Haggeher Mountains cause orographic uplifts. Conse-
quently, two cloud cover types can be distinguished: an SW
monsoon type, with cloud formations especially in the southern
parts of the island and an NE monsoon cloud cover type with cloud
formations especially above the northern parts of the island. Our
analysis of 142 preview satellite images showed the inter-monsoon
wind reversal periods to be rather constant in timing over the last
40 years.
Table 1
Fog measurements during the SW monsoon, Socotra island (JulyeSept. 2004), see Fig. 2 for areas.
Location Area Altitude
(m asl)
Collector
Size (m
2
)
Trial duration
(days)
Days with fog
and no rain
Total water
collected (l)
Water collected
(l m
2
day
1
)
Shibhan Diksam 500 1 63 60 639 10.14
Dimenhen Diksam 500 2 66 48 1187 8.99
Difre-ahten Diksam 700 12 89 80 4135 3.87
Dihof Momi 500 6 74 19 80 0.18
P. Scholte, P. De Geest / Journal of Arid Environments 74 (2010) 1507e1515 1513
Author's personal copy
4.2. Importance of fog
Vegetation, in particular tree species such as dragonblood, is
able to capture cloud water (‘fog’) by their canopies, producing
so-called horizontal precipitation (Hildebrandt and Eltahir, 2006).
Observed quantities of horizontal precipitation are comparable
with measurements in the fog oasis of Dhofar (Oman) where, with
similar green shade filters, 11.5 l m
2
day
1
was collected in
JulyeSeptember 2005 (Abdul-Wahab et al., 2007). The Diksam fog
measurements, close to the Dihaher weather station (Fig. 5:7) with
an annual average rainfall of 168 mm, suggest that during the SW
monsoon, locally quantities of moisture received through fog are in
the range of 357e567 mm, i.e. 68e77% of total moisture. This is in-
line with measurements in Dhofar where Hildebrandt and Eltahir
(2006) found total moisture three times as high as rainfall. This
results in a strikingly well-wooded vegetation surrounded by
shrublands characteristic of much drier areas (Miller and Morris,
2004; Hildebrandt and Eltahir, 2006). The lack of fog at Momi
(Table 1) showed its localised nature, making it difficult to
extrapolate results over larger areas. The area south of the
Haggeher generally experiences a dense cloud cover, while Momi
and the northern shores are situated in the ‘fog shadow’of the SW
monsoon (Fig. 8).
The importance of fog as a source of moisture on Socotra has
been neglected, and has, for example, not been integrated in
palaeoclimatic models of the Indian Ocean Monsoon System
(Fleitmann pers.comm. 2007). Little is known about its evolution in
times of increased aridification.
4.3. Climate and vegetation
The climate of the Socotra lowlands is comparable with
surrounding Arabian and African continents, with mean annual
temperatures approaching 30
C and rainfall generally confined to
the winter (Le Houerou, 2003). Yet with a locally higher mean
annual rainfall (up to 200 mm), lowland vegetation on Socotra is
more diverse than its continental counterparts. Halophytes are
confined to the driest sea-ward oriented areas in Socotra whereas
elsewhere Croton socotranus shrubland dominates (Mies, 2001).
The midlands of Socotra with their various exposures to either
the SW or NE monsoon harbour a diversified vegetation dominated
by succulent shrubland (Mies, 2001; Miller and Morris, 2004). The
spectacular appearance of the bottle trees desert rose Adenium
obsesum and cucumber tree Dendrosycios socotranum contests Le
Houerou (2003)’sstatement that midlands in the Eritreo-Arabian
Domain ‘lack originality’. Striking in the midlands are also the cliffs
with fog catching microhabitats where frankincense trees, also
being sheltered from grazing, regenerate. Here they may find an
evolutionary refuge in times of aridification. The predominant
winter precipitation makes the midlands more comparable to their
counterparts on the African continent (Eritrea, Djibouti, Somali-
land) than on the Arabian Peninsula as illustrated by the dominant
Buxanthus hildebrandtii evergreen shrub that Socotra’s midlands
share with these areas. The differences with the seasonal cloud
forest vegetation in Dhofar and Hawf (Oman, Yemen) may be linked
to the occurrence of fog during the summer (Miller and Morris,
1988; Hildebrandt and Eltahir, 2006), whereas on Socotra fog
bridges the dry season and winter rainy season.
Kral (2005) and Culek et al. (2006) showed that the montane
vegetation south of Haggeher is influenced by moisture arriving
with the SW monsoon in the form of a fine drizzle or fog. This
corresponds to the predominant SW aspect of seven out of nine
main areas where the island’sflagship species, dragonblood trees,
can be found (Attorre et al., 2007). The remaining two areas
comprise Serahon (W-aspect) where dragonblood is of poor
condition and Skand (SE) where, at the summit of the Haggeher
Mountains and higher than the other areas, dragonblood has the
greatest potential to survive periods of increased aridity (Attorre
et al., 2007).
The lack of regeneration of dragonblood and frankincense trees
has been the subject of much discussion if ‘the goat or climate is to
be blamed’(Mies, 2001; Adolt and Pavlis, 2004; Miller and Morris,
2004; Attorre et al., 2007; Scholte et al., 2008; Habrova et al., 2009).
Increasing insight leads us to hypothesise that it is not the present
(absolute) goat grazing pressure that causes the lack of regenera-
tion, but increasingly reduced spatial and temporal grazing
dynamics (Scholte et al., 2008). Reduced mobility is triggered by an
increasingly sedentary pastoral lifestyle (Morris, 2002). Goat
populations have historically fluctuated dramatically, mainly as
a function of rainfall. The introduction of new management
practises such as water provision, veterinary care, supplementary
fodder and transport of animals by truck to other areas have,
however, increased their survival during droughts (Scholte et al.,
2008). Dragonblood and frankincense stands on Socotra are
remarkably homogeneous (e.g. Habrova et al., 2009), and may have
regenerated en masse in the aftermath of droughts when grazing
pressure has remained low whereas rainfall re-established. There is
an analogy to the homogenous Acacia tortilis woodlands in East
Africa that can be traced back to the late 1880s when rinderpest
ravaged cattle populations (Prins and Van der Jeugd, 1993). Follow-
up research, assisted by the dragonblood age-assessments of Adolt
and Pavlis (2004) and Habrova et al. (2009) may test this hypoth-
esis, by investigating if tree stands can be traced to the above
indicated drought years when goat populations crashed.
4.4. Perspectives
Socotra’s plateaus and mountains should be considered as water
towers of the island, providing fresh water that allows diverse life
in an otherwise (semi-) arid environment. With the rapidly
expanding human population and its demands, amongst other
things expressed through increased (irrigated) agriculture, this
dependence will become further pronounced. Climate change that
has been projected to cause considerable aridification on Socotra
(Attorre et al., 2007) is expected to put further pressure on the
system. The maintenance and, where possible increase, of tree
cover that intercepts fog will be of crucial importance for the
hydrological balance of the island.
The increased insights offered by this study notwithstanding,
a more quantitative understanding of the inter-annual variability of
Socotra’s weather, especially in relation to the expected climate
change in the 21st century, can only be reached by continued data
collection through an enlarged meteorological station network that
includes Socotra’s highlands. Of particular interest is the further
understanding of the spatial importance of fog and its contribution
to the hydrological balance of the island. Continuing the meteoro-
logical monitoring and linking it with the ecological monitoring
and subsequent development of management guidelines should be
a priority for the Environmental Protection Authority and its
research and management partners.
Socotra’s climate with its large spatial and temporal variation
has shaped the evolution of its flora and fauna, resulting in its
present outstanding biodiversity. Its continued survival in times of
climate change and rapidly increasing human pressure will once
again depend on it.
Acknowledgements
We would like thank our EPA-SCDP colleagues for the stimu-
lating working environment and the possibility to use the
P. Scholte, P. De Geest / Journal of Arid Environments 74 (2010) 1507e15151514
Author's personal copy
meteorological database. Ahmed Saeed Suleyman, Mohammed
Nageb (EPA) and Mohammed Saad (CARE International Yemen)
supervised the data collection. Ismail Mohamed (EPA) was of great
help in tapping Socotra’s oral history. Mathias De Flou, Morgan De
Dapper (Dept. Geography, University of Ghent) and Eddy Keppens
(Dept. Geology, University of Brussels) assisted with an earlier
version of this paper. The participation of PdG was made possible
through a scholarship of the Institute for the Promotion and Inno-
vation through Science and Technology (IWT-SB/23527/De Geest).
We would further like to thank Catherine Cheung, Lyndon DeVantier
and Stephany Kersten for critically commenting the manuscript.
References
Abdul-Wahab, S.A., Al-Hinai, H., Al-Najar, K.A., Al-Kalbani, M.S., 2007. Fog water
harvesting: quality of fog water collected for domestic and agricultural use.
Environmental Engineering Science 24, 446.
Abram, N.J., Gagan, M.K., Liu, Z., Hantoro, W.S., McCulloch, M.T., Suwargadi, B.W.,
2007. Seasonal characteristics of the Indian Ocean dipole during the Holocene
epoch. Nature 445, 299e302.
Adolt, R., Pavlis, J., 2004. Age structure and growth of Dracaena cinnabari pop-
ulations on Socotra. Trees 18, 43e53.
Attorre, F., Francesconi, F., Taleb, N., Scholte, P., Saed, A., Alfò, M., Bruno, F., 2007. Will
dragonblood survive the next period of climate change? Current and future
potential distribution of Dracaena cinnabari (Socotra, Yemen). Biological
Conservation 138, 430e439.
Cheung, C., DeVantier, L. (Eds.), 2006. Socotra eA Natural History of the Islands and
Their People. Odyssey Books and Guides, Airphoto International Ltd., Hong
Kong.
Culek, M., Kral, K., Habrova, H., Adolt, R., Pavli, J., Madera, P., 2006. Socotra’s annual
weather pattern. In: Cheung, C., DeVantier, L. (Eds.), Socotra eA Natural History
of the Islands and Their People. Odyssey Books and Guides, Airphoto Interna-
tional Ltd., Hong Kong, pp. 42e45.
De Flou, M., 2006. Study of the Physical Environment of Socotra island: Climate and
Geomorphology (in Dutch). Unpublished Master Thesis, Gend University,
Belgium.
De Geest, P. (Ed.), 2006, Soqotra Karst Project (Yemen), 2000e2004. Berliner
Höhlenkundliche Berichte, Band 20.
De Gray Birch, W., 1875. The Commentaries of the Great Afonso Dalboquerque
Second Viceroy of India. Translated from the Portuguese Edition of 1774.
Hakluyt Society, London.
Fleitmann, D., Matter, A., Burns, S.J., Al-Subbary, A., Al-Aowah, M.A., 2004. Geology
and Quaternary climate history of Socotra. Fauna of Arabia 20, 27e44.
Fleitmann, D., Burns, S.J., Mangini, A., Mudelseed, M., Kramers, J., Villa, I., Neffe, U.,
Al-Subbary, A.A., Buettner, A., Hippler, D., Matter, A., 2007. Holocene ITCZ and
Indian monsoon dynamics recorded in stalagmites from Oman and Yemen
(Socotra). Quaternary Science Reviews 26, 170e188.
Forbes, H.O., 1903. The Natural History of Sokotra and Abd-el-Kuri. Special Bulletin
of the Liverpool Museums. Henry Young & Sons, London.
Habrova, H., Cermak, Z., Pavlis, J., 2009. Dragon’s blood tree-Threatened by
overmaturity, not by extinction: dynamics of a Dracaena cinnabari woodland in
the mountains of Soqotra. Biological Conservation 142, 772e778.
Hildebrandt, A., Eltahir, E.A.B., 2006. Forest on the edge: seasonal cloud forest in
Oman creates its own ecological niche. Geophysical Research Letters 33, 1e4.
Image Science and Analysis Laboratory, NASA-Johnson Space Center. The Gateway
to Astronaut Photography of Earth. http://eol.jsc.nasa.gov/sseop/clickmap/
(accessed March 2006 and December 2009).
Kerr, R.S., Edin, F.A.S., 1811e1824. General History and Collection of Voyages and
Travels, Arranged in Systematic Order: Forming a Complete History of the
Origin and Progress of Navigation, Discovery, and Commerce, by Sea and Land,
from the Earliest Ages to the Present Time. Ramsay, Edinburgh.
Kopp, H., 1999. Abiotische Geofaktoren. In: Wranik, W. (Ed.), Sokotra. Mensch und
Natur. Jemen-Studien 14 Wiesbaden, pp. 3e22.
Kral, K., 2005. Assessment and Mapping of Forest and Shrub Biogeocenoses by
Geoinformation Methods, Part 2 eSocotra Island, [unpublished Ph.D. Thesis].
Mendel University of Agriculture and Forestry in Brno, Brno.
Leuschner, D.C., Sirocko, F., 2003. Orbital insolation forcing of the Indian Monsoon-
a motor for global climate changes? Palaeogeography, Palaeoclimatology,
Palaeoecology 197, 83e95.
Le Houerou, H.N., 2003. Bioclimatology and phytogeography of the red sea and aden
gulf basins: a monograph. Arid Land Research and Management 17, 177e256.
Mies, B., 2001. Flora und Vegetationsokologie der Insel Soqotra. Essener Okologi-
sche Schriften 15. Westarp Wissenschaften, Essen, Germany.
Mies, B.A., Beyhl, F.E., 1998. The vegetation ecology of Soqotra. In: Dumont, H.J. (Ed.),
Soqotra eProceedings of the First International Symposium on Soqotra Island:
Present and Future. United Nations Development Programme, New York,
pp. 35e81.
Miller, A.G., Morris, M., 1988. Plants of Dhofar-the Southern Region of Oman,
Traditional, Economic and Medicinal Uses. Office for Conservation of the
Environment, Oman.
Miller, A.G., Morris, M., 2004. Ethnoflora of the Soqotra Archipelago. The Royal
Botanic Garden, Edinburgh.
Morris, M., 2002. Manual of Traditional Land Use Practices in the Soqotra Archi-
pelago. GEF YEM/96/632 Report. Unpublished Report The Royal Botanic Garden,
Edinburgh, UK.
Neff, U., Burns, S.J., Mangini, A., Mudelsee, M., Fleitmann, D., Matter, A., 2001. Strong
coherence between solar variability and the monsoon in Oman between 9 and
6 kyr ago. Nature 411, 290e293.
Popov, G.B., 1957. The vegetation of Socotra. Journal of the Linnean Society (Botany)
55, 706e720.
Prasad, T.G., McClean, J.L., 2004. Mechanisms for anomalous warming in the
western Indian Ocean during dipole mode events. Journal of Geophysical
Research 109, 123.
Prins, H.H.T., Van der Jeugd, H.P.V., 1993. Herbivore population crashes and
woodland structure in East Africa. Journal of Ecology 81, 305e314.
Saji, N.H., Goswami, B.N., Vinayachandran, P.N., Yamagata, T., 1999. A dipole mode in
the tropical Indian Ocean. Nature 401, 360e363.
Shakun, J.D., Burns, S.J., Fleitmann, D., Kramers, J., Matter, A., Al-Subary, A., 2007. A
high-resolution, absolute-dated deglacial speleothem record of Indian Ocean
climate from Socotra Island, Yemen. Earth and Planetary Science Letters 259,
442e456.
Scholte, P., Miller, A., Shamsan, A.R., Suleiman, A.S., Taleb, N., Milroy, T., Attorre, F.,
Porter, R., Carugati, C., Pella, F., 2008. Goats: Part of the Problem or the Solution
to Biodiversity Conservation on Socotra? Report to UNESCO-IUCN to Support
Socotra’s Listing as World Heritage Site.
UNESCO, 2008. Socotra Archipelago. http://whc.unesco.org/en/list/1263 (accessed
16.02.09).
U.S. Department of the Interior eU.S. Geological Survey United States. Earth-
explorer. http://edcsns17.cr.usgs.gov/EarthExplorer (accessed March 2006 and
December 2009).
U.S. Department of the Interior eU.S. Geological Survey United States. USGS Global
Visualisation Viewer. http://glovis.usgs.gov/ (accessed March 2006 and
December 2009).
Webster, P.J., Moore, A.M., Loschnigg, J.P., Leben, R.R., 1999. Coupled ocean-atmo-
sphere dynamics in the Indian Ocean during 1997e98. Nature 401, 356e360.
Wellsted, J.R., 1835. Memoir on the island of Socotra. Journal of the Royal
Geographical Society, London V, 129e219.
Wranik, W., 2003. Fauna of the Socotra Archipelago eField Guide. Universität
Rostock, Germany.
P. Scholte, P. De Geest / Journal of Arid Environments 74 (2010) 1507e1515 1515
