Some factors affecting Ring Ouzels Turdus torquatus wintering in the Atlas Mountains of Morocco

2006·7.5 MB·

by  Colin Ryall

Some factors affecting foraging and habitat of Ring Ouzels
Turdus torquatus wintering in the Atlas Mountains of Morocco
Colin Ryalla andKevin Briggsb
Quelques facteurs affectant la recherche de nourriture et l’habitat du Merle a plastron Turdus
torquatus hivernant dans l’Atlas marocain. L’etude presentee avait pour objet d’examiner le role
des baiesdegenevriercommesourceprincipalede nourrituredes Merles aplastron Turdustorqua-
tus hivernant dans FAtlas marocain. Des deux especes de genevrier examinees, l’Oxycedre
Juniperus oxycedrus est la plus repandue, mais se trouve en faible densite dans d’autres milieux
boises. Dans les 43 sites etudies dans le Moyen et Haut-Atlas, les Merles a plastron ont ete
observes uniquement a l'interieur ou pres des bois de Genevrier de PhenicieJ.phoenicea. Malgre
le fait que le nombre de baies de genevrier mures dans de tels bois varie beaucoup (de 2,7 x 104
a 2,6 x 10(1 par ha), il apparait que le niveau des degats causes aux arbres par des coupes, et le
niveau general de perturbation, sont des facteurs plus determinants plus importants pour la
presence des Merles a plastron que le nombre de baies. Les bois de genevriers rencontres pendant
cette etude sont degrades et vieillissants, et il n’y a pas de regeneration, ce qui suggere un declin
a long terme avec des implications potentielles pour la future disponibilite de genevriers et la
survie des Merles a plastron.
Summary. This study aimed to shed light on the role ofjuniper berries as the principal food
source for Ring Ouzels Turdus torquatus wintering in Morocco’s Atlas Mountains. Of the two
juniper species surveyed, Prickly Juniper Juniperus oxycedrus was the most widespread, but
occurred at low densities in other types ofwoodland. Of43 sites surveyed in the Middle and
High Atlas, Ring Ouzels were onlyseen in orclose to PhoenicianJuniperJ.phoeniceawoodland.
Although the number ofripe juniper berries in such woodland ranged from 2.7 x 104to 2.6 x
1O'’ per ha, the degree ofdamage to the trees from cutting, indicative ofgeneral levels ofdistur-
bance, appeared to be a stronger determinant of Ring Ouzel presence than did the number of
berries. Juniper woodland encountered in this study was in a degraded and ageing state with no
recruitment by younger trees, suggesting a long-term decline with potential implications for
juniper availability and Ring Ouzel survival in the future.
R
ing Ouzel Turdus torquatus breeds in upland tion from Blackbirds T. merula, and problems on
areas ofEurope and Fennoscandia, and win- the migration routes or wintering grounds (Tyler
&
ters around the Mediterranean, North Africa and Green 1994, Murray et al. 1998, Stott et al.
the Middle East (Snow & Perrins 1998). 2002), but the exact reason remains unclear
Nominate T. t. torquatus, which breeds in Britain U(BKurfield 2002). Burfield (2002) suggested that as
and Fennoscandia, winters in southern Spain and birds appear to share wintering grounds with
north-west Africa, predominantly in the Atlas birds from stable continental populations, factors
mountains, from Morocco to Tunisia (Wernham causing this decline are most likely to be acting in
etal. 2002). the breeding grounds or migration routes.
Although populations in continental Europe Nevertheless, the ecology of Ring Ouzels during
( 7. t. alpestris) appear largelystable, there has been the 3-6 monthswhen theyare migratingandwin-
a decrease in numbers and a contraction ofrange tering outside their breeding area is poorly
in Spain and Britain (Heath etal. 2000, Wotton et understood.
al. 2002) and the species is now included on the The species’ winter food requirements are not
UK Red List (Gregory et al. 2002). The decline well understood, although they have been report-
has been attributed to a range offactors including ed mainly eating juniper berries in Morocco and
habitat change, disturbance, global climate Algeria (Heim de Balsac 1931, Heim de Balsac &
change, predation, pollution, increased competi- Mayaud 1962, Arthur et al. 2000). In the Sierra
60-BuiABCW13No 1(2006) RingOuzels winteringinMorocco:Ryall&Briggs
Nevada of Spain, Zamora (1990) found juniper 1,000 mto2,700 m, especiallyamongJ.phoenicea
berries, supplementedbyarthropods, to constitute andJ. thurifera at 1,800-2,200 m or in mixedJ.
the Ring Ouzel’s main dietinwinter. Zamoraalso oxycedrus / Quercus ilex woodland, often near
suggestedthatthis restricteddietreflectedthelim- rivers orwaterholes.
ited choice of food in this area, as Ring Ouzels The present study ofRing Ouzels overwinter-
wintering elsewhere in Spain ate other berry ing in the Atlas Mountains focused primarily on
species where available. their relationship with juniper and factors which
Four species of juniper occur in Morocco: may influence the availability of juniper berries,
Prickly Juniper Juniperus oxycedrus Phoenician and hence the survival ofRing Ouzels in winter.
,
JuniperJ. phoenicea, Spanish JuniperJ. thurifera
and CommonJuniperJ. communis.The first three Methods
are common at moderate to high altitudes in During two visits to Morocco in the winters of
Morocco, but Common Juniper is rare, growing 1993 and 2000, a total of43 sites was examined
only on high mountains (Jahandiez & Maire along two transects through the Middle and High
1931). Atlas, respectively, plus an additional site (3.1) in
In Morocco, Ring Ouzels are common to the central HighAtlas (Fig. 1). Sites representative
locally abundant winter visitors, especially in the ofa range ofhabitats were assessed for the pres-
Central and Eastern High Atlas (Thevenot et al. ence of Ring Ouzel, juniper and other berry-
2003). They are rare and irregular in the western bearingspecies, potential food for Ring Ouzels, as
MiddleAtlas, butmore regularin theeasternpart. they were encountered along the transects. Food
Thevenot et al. (2003) state that they occur in availability and pressures on food sources were
open coniferous woodland on stone slopes from evaluated.
Figure 1. Map ofthe MiddleAtlasandHighAtlasMountainsofMoroccoshowingthepositionsofTransects 1 and2
andSite3.1.
CarteduMoyen etHaut-Atlas marocain, indicantlalocalisationdesTransects 1 et2 etduSite3.1.
RingOuzelswinteringinMorocco:Ryall&Briggs BullABCVol13No 1(2006)-61
Transects Juniper berryparameters
m
Transect 1 ran c.110 km, north to south from Ripe and unripe berries were counted per 0.25 2
Azrou, across the Middle Atlas, to Tounfite and of tree canopy surface on the north, south, east,
Midelt on the north-east slopes ofthe High Atlas westaspects and at the top often randomlyselect-
(Fig. 1).Thepreliminaryvisit, in December 1993, ed Phoenician and PricklyJunipertrees atselected
aimed to locate Ring Ouzels and potential berry sites. The number of ripe and unripe juniper
sources, focusing on juniper. Qualitative observa- berries per ha was extrapolated by calculating the
tions were made at nine sites (1.1-1.9). Juniper- surface area of juniper trees. The proportion of
rich sites were located with the aid ofinformation abortedor insect-parasitisedberrieswas calculated
from Eaux et Forets (Moroccan Water and at several sites and the berry diameter of
Forestry Department) and from Berber farmers PhoenicianJuniperandPricklyJuniperberrieswas
and shepherds. measuredwith a micrometer.
Transect2, inJanuary2000, ran primarilyeast
Casualobservationsandcoordinates
to northeastalongthe northern slopes ofthe High
Atlas from Demnate in the west to Midelt, c.230 Signs ofpressures on juniper tree density, survival
km to the east, and crossedTransect 1 atTounfite; and hence habitat quality were noted, as well as
so sites 1.7 and 2.19 are the same. On the second other factors that might effect survival of Ring
Ouzels, e.g. habitat degradation, disturbance and
transect, 33 locations (2.1-2.33) representing a
series ofdistinct habitats, at least 1 ha in extent, proximity of sites to water. Other berry-bearing
ranging from ploughed arable land to forest, were species were recorded alongwith topography, alti-
tude and compass orientation ofescarpments. For
assessed as they were encountered. However,
greater emphasis was given to areas containing Transect 1 bearings or altitudewere approximated
juniper and other berry-bearing species as these using maps, but for Transects 2 and site 3.1, a
Magellan GPS 2000XL global positioning system
were likely to be more important to Ring Ouzels.
An additional juniper-rich site (3.1) c.20 km was employed.
south ofTaddertwas surveyed (Appendix 3). Results
Ring Ouzeldensity In total, 43 siteswereinvestigatedin thetwophas-
At locations where Ring Ouzelswere observed, on es of this study over an altitude range of
m
Transect 2 and site 3.1, they were counted along 1,003-2,208 (Appendices 1-2), and 11 distinct
one 100 x 30 m transect through each site. habitat types were discerned (Table 1).
Surveys were carried out in the morning only,
Occurrence ofjuniper
between 09.00 and 12.00 hrs. Presence or absence
Three species of juniper were encountered:
ofother Turdusspp. and theirscats on andaround
Prickly, Phoenician and Spanish Juniper. Prickly
juniper trees was noted.
Juniper formed bushes or small trees averaging
m
Juniper treeparameters 2.8 in height and, being recorded at 21 ofthe
Quantitative data on juniper trees were collected 43 sites at 1,231-2,208 m, was the most wide-
in Transects 2 and site 3.1, i.e. canopy diameter, spread species. It grew mainly as a secondary
tree height, number oftrees per ha in 10 w 10 m2 species in stands ofPhoenician Juniper as 5-20%
quadrats chosen at random (see Table 2 and (mean 9%) ofthe trees present or in Holm Oak
Appendix 3). The degree of cutting damage to Quercus ilexwoodland as 5-30% (mean 21%). At
trees was scored on a scale of0-5 (0 = no visible two sites (1.9 and 2.2), Prickly Juniper was the
damage; 5 = only stump remaining). Stands of dominant tree species.
juniperwere classified as ‘undegraded’ ifthe mean Phoenician Juniper woodland was encoun-
scoredwas <2.5 and ‘degraded’ iftheyscored >2.6. tered mostly at the eastern end ofthe High Atlas,
At some sites (sites 2.1-2.19, 2.29 and 3.1) the around Tounfite and south ofMidelt, but also in
numberofjuniperstumps perhawascounted and the Central High Atlas, south of Marrakech at
atTounfite the diameter ofjuniper trunks at 0.25 1,918 m (3.1), at 1,003 m near Demnate (2.1)
m above ground was also recorded. and in lowernumbers in theMiddleAtlasat 1,800
m
and 2,000 (1.4 and 1.6). It grew as rounded
62-BullABCVol13No 1(2006) RingOuzelswinteringinMorocco:Ryall&Briggs
Table1.OccurrenceofRingOuzels Turdustorquatusandotherthrushes Turdusspp.in 11 differenthabitatssurveyed
in 1993and2001,andthe presenceofberrybearingspecies.
Tableau 1. Presencedu Merleaplastron Turdustorquatusetd’autresgrives Turdusspp.dans 11 habitatsdifferents exam-
inesen 1993et2001 etlapresenced’especesproduisantdes baies.
,
Vegetationtype Berry-bearingspecies Numberof Numberof Otherthrushespresent
andcondition* present** sites siteswith
surveyed RingOuzels
HolmOakwoodland PricklyJuniperJ.oxycedrus 7 0 Blackbird T.merula
UndegradedPhoenician PhoenicianJuniperJ.phoenicea 7 6 Blackbird T.merula
Juniperwoodland PricklyJuniperJ.oxycedrus MistleThrush T.viscivorus
(SpanishJuniperJ.thurifera) SongThrush T.philomelos
(Hawthorn Crataegusmonogyna) Redwing T.iliacus
(Mistletoe Viscumcruciatum)
(DogRose Rosacanina)
DegradedPhoenician PhoenicianJuniperJ.phoenicea 5 0 Blackbird T.merula
Juniperwoodland PricklyJuniperJ.oxycedrus MistleThrush T.viscivorus
(Hawthorn C.monogyna) Redwing T.iliacus
Undegraded/degraded PricklyJuniperJ.oxycedrus 2 0 Blackbird T.merula
PricklyJuniperwoodland (LentiscPistacialentiscus) MistleThrush T.viscivorus
Redwing T.iliacus
DegradedSpanish SpanishJuniperJ.thurifera 1 0 None
Juniperwoodland (PricklyJuniperJ.oxycedrus)
Scrubwoodland Hawthorn C.monogyna 6 1 Blackbird T.merula
Mistletoe V.cruciatum MistleThrush T.viscivorus
DogRoseRosacanina SongThrush T.philomelos
(PricklyJuniperJ.oxycedrus) Redwing T.iliacus
Cedarwoodland None 5 0 Blackbird T.merula
Plantation—pineorolive None 3 0 None
Lowherbandtussockgrassland None 2 0 None
Arableland None 4 0 None
Barestonyground None 1 0 None
*Conditionofvegetation:Undegraded=meanscorefortreedamageof0-2.5;Degraded=meanscorefortreedamageof2.6-5.0
**Scarcespeciesinparentheses
conical trees on bare stony ground, averaging Condition ofjuniper trees
m m
3.5 inheight, at 1,780-1,928 at 15 ofthe43 All juniper-rich sites had damaged trees, ranging
sites studied, at densities of 7.6-84 trees per ha in severity from removal ofa few lateral branches
and was invariably the dominant species. At to total destruction leaving splintered stumps
Tounfite, some individuals reached >8 m with (Figs. 3-4). On a scale of 0-5, all Phoenician
m
trunkdiameters up to 0.6 (Fig. 2).Table2 pres- Juniper sites contained damaged trees with nearly
entsparameters forjunipertrees atsevensites con- half rated at above 3 (moderate damage to only
taining PhoenicianJuniper, and further details are stump remaining) (Table 2). The situation was
provided inAppendix 3. similar for Holm Oak and Prickly Juniper
Spanish Juniper, albeit severely degraded by (Appendix 2). Local peoplewith donkeys carrying
cutting, was encountered at Inifif (1.5), near Col bundles of juniper wood were frequently
du Zad, in the MiddleAtlas. In the HighAtlas, it encountered. Fig. 5d shows that damage to
was found as an occasional secondary species to Phoenician Juniper correlated positively with the
Phoenician Juniper at sites 2.19 and 3.1 at alti- number ofjuniperstumps present (r= 0.788) and
tudes of c.1,920 m (Appendices 1-2). The trees negatively with presence ofRing Ouzels (Fig 5b:
were all c.3 m in height. r= -0.733), but not with number of ripe berries
(Fig 5a: r= -0.300) or altitude (Fig 5f: r= 0.047).
RingOuzelswinteringinMorocco:Ryall&Briggs BullABCVol13No 1(2006)-63
0 0
Figure2. LargePhoenicianJuniperJuniperusphoenicea Figure4. Stumpofalargejuniperdestroyedfromfire-
treeofc.8 min near-perfectcondition, nearTounfite woodcollectionnearTounfite (Site2.16) intheeastern
(Site 2.19) in theeastern HighAtlas, Morocco. Denuded HighAtlas, Morocco (Colin Ryall)
groundfromovergrazingcan beseen (Colin Ryall) Souched’ungrandgenevrierdetruitparlacoupedebois
GrandGenevrierde PhenicieJuniperusphoenicead’envi- dechauffe, presdeTounfite (Site2.16) dansleHaut-
m
ron 8 en etatquasiparfait, pres deTounfite (Site2.19) Atlas marocainoriental (Colin Ryall)
dansle Haut-Atlas marocain oriental. Onpeutvoirlesol
denudeparlesurpaturage (Colin Ryall)
Tounfite, and even lone junipers in ploughed
fields, e.g. west ofBoumia and south ofAzrou.
Juniper berries
The number ofberries onjuniper and the ratio of
ripe to unripewere veryerratic, between sites and
between trees at each site (Table 2). Mean ripe
berrydensities forPhoenicianJunipertrees ranged
m
from 191-1,309 per 2ofcanopy and in Prickly
Juniper from 0-239 per m2 with many trees
*
devoid of berries. Where present Phoenician
Juniperalways contributed themajorityofberries.
Figure3. Largejunipertreein advancedstageofprogres- The number of ripe berries in stands of
sivedestruction from firewoodcollection near Phoenician Juniper ranged from 2.7 x 1 4—1.2 x
AguelmamedeSidiAli (Site 1.4) in theMiddleAtlas, 105per ha (mean = 0.7 x 1 5) for degraded sites
Morocco (Colin Ryall) and 1.1 x 105-2.6x 106perha (mean = 0.9x 106
)
Grandgenevrieren etatavancededestruction progressive for undegraded sites (Table 2). The mean ripe
parlacoupede bois dechauffe, presd’Aguelmamede berrycrop foralljuniperspecies atallsiteswas 6.9
Sidi Ali (Site 1.4) dansleMoyenAtlas marocain (Colin x 105per ha.
Ryall) At most sites, somejuniper berrieswere abort-
ed (shrivelled and lacking pulp) or parasitised by
Stumps of younger trees usually showed re- insects (a small exit hole, or white scale and lack-
sprouting from the base but this was absent in ing pulp). At site 2.2, PricklyJuniper dominated,
older stumps. 9.3% of berries were aborted or parasitised, for
All juniper woodland was heavily grazed by PhoenicianJuniper itwas 18% atsite 2.16, 6% at
sheep and browsed by goats, with the ground 2.17, 0.33% at 2.18 and 0.2% at 2.19 (overall
totally denuded of vegetation (Fig. 3) and with mean = 6.1%). The mean berry diameters for ten
copiousanimal droppings. Nojuniperseedlingsor trees atsite2.15 were 10.2 mm forPricklyJuniper
young trees were found. It was common to and 10 mm for Phoenician Juniper and in both
encounter juniper woodland in the process of species, ripe berries were red-brown and sweet
being cleared for agriculture, e.g. at Demnate and when ripe.
64-BullABCVol13No 1(2006) RingOuzelswinteringinMorocco:Ryall&Briggs
Table2.Junipertreeparametersandfruitcropatseven PhoenicianJuniperJuniperusphoenicearich sitesonTransect2
andsite3.1, andoccurrenceofRingOuzels Turdustorquatus
Tableau2. Parametresdesgenevriersetproductiondebaiesdansseptlocalites richesen Genevrierde PhenicieJuniperus
phoeniceale longduTransect2etau Site3.1 etpresencede Merlesaplastron Turdustorquatus
,
SiteNo. Treespp. Mean Meannumber Meantree Juniper Meanberrycount Berriesperha RingOuzels
present* canopy ofeach condition stumps /m2canopy(SE) per
diameter Junipersp. rating perha 100'50m
(m)(S£) perha(SE) (0-5)** (SE) transect(SE)
ripe unripe ripe unripe
2.15 Jp 3.75 33.9 1.5 0 436 712 3.3x 105 5.3x 105 14
(0.34) (4.3) (106) (180) (4.04)
Jo 2.17 1.5 2 0 0 0 0
(0.30) (0.52)
2.16 Jp 3.14 7.6 3.5 9.2 1033 158 1.2x 105 1.9x 104 0
(0.22) (0.76) (0.55) (189) (63)
Jo 2.08 2.2 3.5 0 0 0 0
(0.24) (0.73)
2.17 Jp 3.39 18.4 2.4 7 448 68 1.5x 105 2.2x 104 1
(0.30) (1.9) (1.8) (137) (14) (0.70)
Jo 2.4 12.8 3.7 239 0 2.8x 104 0
(0.86) (4.2) (276)
2.18 Jp 3.82 83.6 2.5 11.9 759 58 1.5x 106 1.5x 105 3.4
(0.43) (10.8) (3.17) (168) (23) (0.45)
Jo 2.64 3.6 3 0 0 0 0
(0.31) (3.5)
2.19 Jp 5.62 40 2 5.2 1309 92 2.6x 106 1.8x 105 7.8-
(0.37) (2.23) (1.82) (224) (80) (3.45)
Jo 2.9 16 2 0 0.8 0 0
(0.78) (4.66) (0.9)
Jt 3.15 0.4 1 - - -
(0.10) (0.45)
2.29 Jp 2.12 20 4 30.8 191 157 2.7x 104 2.2x 104 0
(0.55) (3.16) (5.13) (191) (157)
3.1 Jp 3.63 24.8 2.5 2 214 63 1.1 x 105 3.2x 104 3.4
(0.30) (2.41) (1.22) (77) (45) (0.97)
Jo 3.63 19.2 2.5 170 0 6.7x 104 0
(0.36) (2.79) (162)
Jt 3.1 0.8 2.5 319 480 3.9x 103 5.8x 103
(0.33) (0.55) (20) (76)
'Jp=J.phoenicea;Jo=J.oxycedrus,Jt=J.thurifera
"Conditionofvegetation:Undegraded=meanscorefortreedamageof0-2.5;Degraded=meanscorefortreedamageof2.6-5.0
(seeMethods,Junipertreeparametersforfurtherdetails)
Otherpotentialfoodsources Presence ofRing Ouzels
Table 1 shows that other berry species, namely RingOuzels, seen atclosequarters, appearedto be
Hawthorn Crataegus spp., Dog Rose Rosa canina, nominate T.t. torquatus.Althoughotherspeciesof
Red-berried Mistletoe Viscum cruciatum and thrush were recorded in all six types ofberry rich
Lentisc Pistacia lendscus, were present in damper habitat (Table 1), Ring Ouzels were onlyencoun-
sites, such as valley bottoms, northern slopes and tered in undegraded Phoenician Juniper wood-
where overgrazingwas less severe. land, except once where one was seen feeding on
RingOuzelswinteringinMorocco:Ryall&Briggs BullABCVol13No 1(2006)-65
Figure5a PhoenicianJunipertreecondition
Figure5dPhoenicianJunipertreecondition
vs. logberrybount{r=4).3fl©} vs. numbertreestumps(r=0.703}
II 86 i- 4300 *
I 20
B 10 \
* 0
0 1 2 3 4 5
Treeconditionscore Treeconditionscore
Figure5b RingOuzelsvs. PhoenicianJuniper
condition(r=-0.733}
Treeconditionscore
Figure5cRingOuzelsvs.altitude Figure5fPhoenicianJunipertreecondition
(r=-0.085) vs.altitude(r=0.047)
n 15
JZ
iq
1 10
©
N
o 5
1,800 1,850 1,900 1,950 1,820 1,840 1,860 1,880 1,900 1,920 1,940
Altitude(m) Altitude(m)
Figures 5a-f. Correlationcoefficients (r) foraseriesofparameters relatingto RingOuzels Turdustorquatusandjuniper
in the HighAtlas.
Coefficientsde correlation (r) pouruneseriedeparametres concernantleMerleaplastron Turdustorquatusetle
genevrierdans le Haut-Atlas marocain.
small rosehips in scrub close to sparse Phoenician ings and copious droppings, on and around the
Juniper (site 1.8). It is noteworthy that all sites trees. Suchdensescatdepositswereabsentinother
where RingOuzelswereseen orindicatedbyplen- types ofwoodland. Ring Ouzels were not evident
tiful scats were less than 500 m from a source of at sites containing Prickly Juniper without
water. Ring Ouzels were mainly in fast-moving PhoenicianJuniper, whether PricklyJuniper dom-
flocks ofc.4-30 birds accompanied bysmall num- inated (sites 1.9 and 2.2 only) orwas secondaryto
bers of other Turdus spp. (Blackbird T. merula Holm Oak (e.g. sites 2.3, 2.4 and2.14).As shown
,
Mistle Thrush 77 viscivorus, Song Thrush 77 in Table 2, no Ring Ouzels were recorded where
philomelos and Redwing 77 iliacus). There was no Phoenician Juniper was seriously degraded (mean
correlation between altitude and number ofRing tree damage score >2.6), and Fig. 5b shows a
Ouzels seen (Fig. 5c). strong negative correlation between condition of
Near Tounfite, Ring Ouzels’ preference for Phoenician Juniper trees and number of Ring
PhoenicianJuniperwas attested by frequentsight- Ouzels present (r= -0.733). However, the number
66-BullABCVol13No 1(2006) RingOuzelswinteringinMorocco:Ryall&Briggs
of ripe berries in Phoenician Juniper correlated (c.50%) can do so. The zero berry counts in
weakly with number of Ring Ouzels seen (r = Prickly Juniper at several sites may be due to a
0.355, Fig. 5e). concentration of male plants or to a local berry
No Ring Ouzelswere seen in the MiddleAtlas failureinfemaletrees.Jordano (1991) pointedout
duringthe 1993 visit (Transect 1) although atA'in thatthe monoecious state in PhoenicianJuniperis
Nokra, Berber shepherds said theysometimes saw variable but is more than 90% in Morocco. Thus
them in small groups inwinter. Prickly Juniper must usually contribute a small
proportion of the total berry crop where both
Discussion species occur.
This study, albeit small in scale, is a first attempt As wide-ranging, opportunistic feeders, Ring
to quantifyfactors that mayinfluence the survival Ouzels may be attracted primarily to extensive
ofRing Ouzels wintering in the Atlas Mountains areas of juniper, e.g. Phoenician Juniper wood-
and indicates areas for further elucidation. land, with high berry densities and low levels of
disturbance, thus maximising foraging success,
Juniperas afoodsourceforRing Ouzels rather than to isolated patches or individual fruit-
All Ring Ouzels were seen in or close to ing trees, e.g. Prickly Juniper, outnumbered
PhoenicianJuniper, often accompanied bysmaller amongst non-berry bearing species, e.g. by Holm
numbers ofother Turdus spp., closely paralleling Oak, though these mayserve as stop-offpoints for
the observations ofSnow (1952), Heim de Balsac migrants.
(1931) and de Juana & Santos (1981). Heim de The condition ofPhoenician Juniper trees, as
Balsac (1931), Blondel (1962) andThevenot etal. well as number ofberries, seems to be akeydeter-
(2003) considered both Prickly Juniper and minant of the presence of Ring Ouzel. We only
Phoenician Juniper berries a major part of their found Ring Ouzel at five sites where cutting dam-
winter diet. In the Spanish Sierra Nevada too, age to Phoenician Juniper was low, but not in
Ring Ouzels feed almost exclusively on juniper degraded stands, even with good berry crops. At
during the winter (Zamora 1990; pers. obs.) and an intermediate level of cutting, where trees still
their occurrence is strongly correlated with berry produce a good berry crop, disturbance from the
availability (Jordano 1993). This apparent speci- frequent visits by small-scale wood collectors and
ficity for juniper may apply only during midwin- livestockmaykeep RingOuzels awaymuch ofthe
—
ter because, as reported by other workers, Ring time afactorwhichalsooperatesinpartsoftheir
Ouzels occur in varied habitats during migration breeding range (Burfield 2002).
andthenexploitotherspecies ofberry, in addition However, not all berries on a juniper tree are
to juniper. edible.Theytaketwoyearsto ripenso, inautumn,
In this study, RingOuzelswereonlyseenfeed- trees contain both ripe and unripe berries, the
ing in Prickly Juniper where it was with ratio being variable (Table 2). In addition, a vari-
PhoenicianJuniper, but not where it grew among able proportion of berries are either aborted or
Holm Oak. This may reflect overall berry avail- parasitised by insects, resulting in berries lacking
ability. The berries ofthe two species are so simi- pulp and with reduced nutritional value (Ionesco
lar, physical andto taste, thatselection ofone over & Sauvage 1969). Garcia etal. (1999) found that
theotherseemsunlikely.Themeanripeberrycrop Ring Ouzels rejected aborted and parasitised
for all juniper sites (both juniper species) of6.9 x Common Juniper berries in the Sierra Nevada,
105 per ha closely matches that for Common thus necessitatinggreaterforagingeffort.Thepro-
Juniper (7 x 105per ha) recorded in the Sierra portion ofunpalatable berries can be substantial;
Nevada (Garcia et al. 1996) but, like Jordano we found levels of aborted or parasitised ranged
(1993), we found that the crop varied widely from 0.2-18% (mean 6.1%), butTraveset & Sans
between areas. As a monoecious species (bearing (1994) recorded moth infestation levels to range
male and female flowers on the same plant), all from 3-50% ofthe crop in PhoenicianJuniper in
Phoenician Juniper trees can potentially bear the Balearic Islands.
berries, whereas in Prickly Juniper and Spanish A further key factor in determining Ring
Juniper, being dioecious (bearingmale and female Ouzels’ choice offeeding site appears to be prox-
flowers on separate plants), only female trees imity to water. In this study, locations where the
RingOuzelswinteringinMorocco:Ryall&Briggs BullABCVol13No 1(2006)-67
birds were found were all within a few hundred erative ability with age was recognised by
&
metres of a water source. Ring Ouzels need to Emberger (1938) and Metro Sauvage (1955).
drink regularly whilst feeding on juniper berries. This habitat degradation is further exacerbated by
Heim de Balsac (1931), Arthur etal. (2000) and the lack ofrecruitment ofyoung trees due proba-
Thevenot etal. (2003) have commonly observed bly to a combination ofovergrazing and drought,
them drinking at rivers and waterholes in seen also in the Sierra Nevada (R. Zamora pers.
&
Morocco, as they do also in the Sierra Nevada comm.), and Quezel Barbero (1981) noted a
when feeding on juniper berries (pers. obs.). This completelackofjuniperregenerationintheregion
issue clearly needs further investigation. oftheAtlas Mountains.
Despite a statement to the contrary byArthur
Occurrence ofjuniper etal. (2000), juniper is no longer being used sus-
Prickly Juniper was widespread, and like Quezel tainably in the Atlas. Auclair (1996) pointed out
(1980), we found it associated with either Holm that traditional controls used to workwell to pre-
Oak or Phoenician Juniper woodland as a second- serve mountain forest but population increase in
aryorsubdominantspecies. PhoenicianJuniperwas parts ofthe High Atlas is resulting in permanent
less widespread but where present was the domi- forest loss; indeed, wood removal is twice the rate
nant tree, forming open woodland interspersed of production and stocking levels twice the sus-
with smaller numbers ofPricklyJuniperand Holm tainable level.
m
Oak.Trees averaged3.3 in height, althoughnear Lone junipers in vast areas ofploughed arable
Tounfite some reached 8 m, the maximum for this land on the lower mountain slopes and plains tes-
species (Maire etal. 1932), with some trunk diam- tify to large-scale clearance ofjuniper woodland.
etersof0.6 m, indicatingtrees more than 500years Wearelateinalong-termprocessofdeforestation.
old, based on annular rings of stumps. In In Roman times more than halfofNorth Africa
Emberger’s (1938) day, PhoenicianJuniperwasvery was densely forested (Blondel & Aronson 1999),
widespread in both the Middle and High Atlas up whereas 17% remains in Morocco, including oak,
to 2,200 m, but we found it to be frequent if cedar and juniper forest. Conacher & Sala (1998)
m
patchily distributed at 1,780-2,208 in the High observed that agricultural clearance and deforesta-
Atlas, and scarce in the Middle Atlas. Spanish tion in the mountains ofNorth Africa intensified
m
Juniper has an altitudinal range of 1,800-3,150 from the late-18th centurydue to excessivewood-
(Emberger 1938). Its apparent rarity in our study cutting and overgrazing by sheep and goats. This
reflects the limits ofaltitudesvisited. Its presence at scenario involves thelong-termfragmentation and
m
2,000 near Inififin the Middle Atlas (1.5) was destruction ofa key resource for Ring Ouzels. As
also noted bySauvage (1956). Zamora (1990) and Jordano (1993) found in
Spain, the Ring Ouzel is most probably the main
Juniperdamage anddecline dispersion vector for Juniper species in North
Most ofour sites had trees showing moderate to Africa.
&
severe damage, which concurs with Quezel
Barbero (1981), who described stands of Conclusions
Phoenician Juniper as very degraded by human We focused on the occurrence ofRing Ouzels in
influence and livestock. We commonly saw local relation to species of juniper, berry crop and
Berberswith donkeys, and even trucks, laden with degree ofdamage to trees. It must be recognised
juniper and Holm Oak wood. Ofcourse, damage that thiswas asmall-scale study, with RingOuzels
may be less severe at sites more remote than those only being detected at seven sites, and so the fol-
in our study. All three juniper species are used for lowing conclusions must therefore be considered
burning for cooking, heating and construction by provisional.
the Berbers (Auclair 1996), and Phoenician and The well-established link between wintering
Spanish Junipers are used for livestock feed in Ring Ouzels and juniper is confirmed, but this
&
droughts (Ionesco Sauvage 1969). link is primarily with Phoenician Juniper, which
We noted thatstumpsofsmallerjunipersoften contributes far more berries than Prickly Juniper,
re-sprouted but, in older trees, some more than and therefore offers the most productive foraging
500 years old, this was absent. This loss ofregen- option.
68-BullABCVol13No1(2006) RingOuzelswinteringinMorocco:Ryall&Briggs
The presence of Ring Ouzels correlated with • the status of other berry species and their
the condition ofPhoenician Juniper. Where trees importance for migrating and wintering Ring
wereseverelydamaged, indicativeofahighlevelof Ouzels
humandisturbance, therewas noevidenceofRing • the importance and availability ofwater as a
Ouzelvisits, evenwhereripeberrieswereplentiful. factor limiting the Ring Ouzel’s ability to
Several factors are resulting in a long-term exploit available berrysupplies.
process ofjuniperwoodland decline:
Acknowledgements
• unsustainable harvesting for firewood and
forage WethankFarnborough Collegeforfinancialsupport
for this work throughout, along with the British
• loss ofregenerative ability in the ageing stock
TrustforOrnithologyandBirdExplorationFundfor
ofjuniper trees
their assistance with the 1993 phase of the study.
• an ageing population ofjuniper due to lack of Thanks also to Prof. Jacques Franchimont, Zin
recruitment ofyoung trees from overgrazing, LabaadinAfhaz and Mick Green for their assistance
agricultural clearance and drought. on the 1993 research trip, toJeanne Capey for help
Burfield (2002) points out that, if UK birds with translation and to the late Rae Vernon for his
share theirwinteringgroundswith thosefrom sta- helpful information. Patrick Bergier commented on
ble continental populations, the factors causing adraft ofthis paper.
their decline are most likely to be acting in the
References
abtreperdesienngt,grliotutlnediss kornomwingroaftitohne erocuotleosg.yHoorwmeovveer,- Arthur, D. S. C., Ellis, P. R., Lawrie, R. & Nicoll, M.
2000. Observations ofwintering Ring Ouzel and
ments of the two races, T t. torquatus and T. t. their habitat in the High Atlas Mountains,
alpestris, during the 5-6 months they are migrat- Morocco. ScottishBirds21: 109-115.
ing and wintering in the mountains of North
Auclair, L. 1996. L’Appropiation communautaire des
Africa. Neither is it known how much juniper is forets dans le Haut Atlas marocain. Cahiers des
needed to support this Ring Ouzel population, SciencesHumaines32: 177-194.
particularlyinayearwithapoorberrycropand/or Blondel,J. 1962. MigrationprenuptialedanslesMonts
water is scarce near berry sources. desKsours (Saharaseptentrional).Alauda30: 1-29.
The total acreage of juniper woodland is Blondel,J. &Aronson,J. 1999. BiologyandWildlifeof
unknown, as is its condition and the rates offrag- the Mediterranean Region. Oxford: Oxford
mentation and loss. Juniper berryavailability may UniversityPress.
not yet be a limiting factor for the species but as Burfield, I. J. 2002. The breeding ecology and conser-
destruction continues this point must eventually vation of the Ring Ouzel Turdus torquatus in
come. Poor foraging in winter, for whatever rea- Britain. Ph.D. thesis. Queens’ College, University
son, means poorer condition for migration and ofCambridge.
breeding. Conacher,A.J. & Sala, M. 1998. LandDegradation in
Our study covered a small number of more Mediterranean Environments ofthe World: Nature
accessible sites, during a short part ofthe winter andExt&ent, Causes andSolutions. Chichester: John
period and did not include Spanish Juniper sites, Wiley Sons.
which occur at higher altitudes. Nevertheless, our Emberger, L. 1938. LesArbresduMarocetCommentles
findings indicate aspects that future, more exten- Reconnaitre. Paris: Lacrosse. &
sive studies could focus on: Garcia, D., Gomez, J. M., Hodar, J. A. Zamora, R.
1996. Ecologia reproductiva del enebro en Sierra
• how much juniper is required to support the Nevada: factores que determinan la regeneration
current Ring Ouzel population, in viewofthe natural de las poblaciones. In Chacon,J. & Rosua,
variability ofberryproduction J. L. (eds.) First Conf. Lnt. Sierra Nevada:
• the extent and status ofjuniper-rich habitat in Conservacion y Desarrollo Sostenible, Univ. of
Algeria and less accessible parts of the Granada,vol. 2: 441-453.
MoroccanAtlas Garcia, D., Zamora, R., Gomez,J. M. & Hodar, J. A.
• the condition and rate of loss of the juniper 1999. Bird rejection ofunhealthy fruits reinforces
woodland that remains the mutualism between juniper and its avian dis-
persers. Oikos85: 536-544.
RingOuzelswinteringinMorocco:Ryall&Briggs BullABCVol13No1(2006)-69