1

Past environmental and climate changes in northern Tanzania

Vegetation and lake level variability in Empakaai Crater

Maria Ryner

Thesis contents

This doctoral thesis consists of a summary and four appended papers which will be referred to in the text by their roman numerals.

List of papers:

I Ryner, M.A., Bonnefi lle, R., Holmgren, K. and Muzuka A. 2004. Vegetation changes in Empakaai Crater, northern Tanzania, at 14,800 - 9300 cal yr BP. Review of Palaeobotany and Palynology 140, 163-174.

II Ryner, M., Gasse, F., Rumes, B and Verschuren, D. 2007. Climatic and hydrological instability in semi-arid equatorial eastern Africa during the late Glacial to Holocene transition: a multi-proxy reconstruction of aquatic ecosystem response in northern Tanzania. Palaeogeography Palaeoclimatology Palaeoecology, 248, XX-XX.

III Ryner, M., Holmgren, K. and Taylor D. A c.1200-year record of vegetation dynamics and lake level changes from Lake Emakat, northern Tanzania. Submitted to Quaternary International.

IV Westerberg, L-O., Holmgren, K., Börjeson, L., Håkansson, T., Laulumaa, V., Ryner, M. and Öberg, H. The development of the Engaruka irrigation system, Northern Tanzania. Physical and societal factors. Submitted to The Annals of Association of American Geographers.

The co-authorship of Papers I-IV:

The idea to work with lake sediments from the Empakaai Crater was fi rst proposed by Dr Alfred Muzuka and Pro-fessor Emeritus Wibjörn Karlén. I have planned and designed the study, initiated and carried out the fi eldwork, performed most of the analysis and led the paper writing in Paper I-III. In paper I Raymonde Bonnefi lle and Karin Holmgren contributed to the discussion and Alfred Muzuka participated in the fi eld work. In paper II Francoise Gasse contributed with the diatom analysis and interpretation, Bob Rumes with the ostracod and chironomid analysis and Dirk Verschuren with data interpretation and to the discussion. In paper III Karin Holmgren and David Taylor con-tributed to the interpretation and the discussion of the data. In paper IV I contributed with climate interpretation, data evaluation and participated in the overall discussion.

2

Maria Ryner

Introduction

The motivation behind this PhD project has been to con-tribute a piece in the puzzle of past climate variability in eastern Africa and, when possible, relate the results to human societal changes. The specifi c site, Empakaai Crater in northern Tanzania, is located in a highland region basically surrounded by semi-arid savannah, where rainfall variability is high and where supplies of water are scarce. An increased understanding about the amplitude of past climate changes, past climate vari-ability and the environment’s response to these changes in eastern Africa is needed in order to understand the impact and resilience of this ecosystem when subjected to external forces.

The climate system is a dynamic system that is natu-rally variable due to internal as well as external forces. This complexity requires many fi elds of expertise to un-tangle the diffi culties of interpreting past climate change. The use of indirect sources to reconstruct past climate changes demands knowledge within many disciplines in the fi eld of both human and natural sciences. The over-all aim is to translate the proxies to measures of e.g. temperature and/or rainfall, and to compare data from geographically different areas in order to understand the forcing factors behind the regional and temporal patterns in climate. The forces behind long term climate changes are fairly well understood, such as the ice age cycles, driven by changes in the astronomical cycles (Imbrie et al. 1984). For short term centennial-decadal changes or variability, the forcings are less well understood. A better understanding of natural climate changes on these scales is of interest, both in order to be able to distinguish an-thropogenically induced climate changes from naturally driven changes, and to provide information to societies for adaptation to climate change. An increasing number of reports demonstrate a strong anthropogenic compo-nent responsible for the recent increase in global mean temperature (IPCC 2007). Basing its report on several multiproxy temperature reconstructions the IPCC con-cludes that present day temperatures are likely the warm-est during the last 1300 years. The Southern Hemisphere and the tropics are however still largely underrepresented in the data base available and IPCC (2001, 2007) stresses the importance of additional information about past cli-mate change from the tropics and the Southern Hemi-sphere. In this thesis I have chosen, in most discussions, to limit the regional comparison to climate records and environmental changes found in eastern Africa.

Semi-arid regions, where rainfall variability is high, and where supplies of water are scarce, are often ex-pected to be vulnerable to changes in climate conditions (IPCC 2007, Hulme 1996). Millennia-scale large vari-ability in lake levels is well recorded in tropical Africa

(Gasse 2000) while the centennial-to-decadal-scale vari-ability is less well known. An improved understanding of climate variability in these regions on centennial to decadal time scale would not only increase the under-standing of natural climate variability on these time scales but would also facilitate more precise analysis of the interrelation between climate, environment and so-cieties. Understanding how early societies coped with and responded to environmental change, e.g. climate variability, may enable a better understanding of suitable adaptation strategies also for future management plans in these regions. In eastern Africa there are several stud-ies suggesting that a shifting climate has had an impact on past humans societies (Robertshaw et al. 2004, Ver-shuren 2004, Holmgren and Öberg 2006). Although we only have “snapshots” of the human history of eastern Africa, additional studies of climate variability in natu-ral archives as well as studies of human settlements and their coping mechanisms, will increase our possibilities to understand and untangle the reason for survival and failure of human societies.

Aim of study

The overall aim of this PhD project is to describe, from a palaeoecological and palaeohydrological point of view, an unexplored area in northern Tanzania, using lake sediments. The specifi c site, The Empakaai Crater, was chosen because it could be anticipated that the site was, owing to the environmental setting relatively un-dis-turbed by human impact, in comparison to other sites in eastern Africa, and that the lake within the caldera could contain continuous sediments spanning a long period of time. The caldera is also located at an, for eastern Africa, ecologically sensitive altitudinal range. For practical reasons the study focuses on the time periods for which suitable material in the sediment cores were found, i.e. between 14.8 and 9.3 ka (1000 cal radiocarbon years BP, 1950) (papers I and II) and the last 1200 years (papers III and IV).

Important objectives have been

• to document and interpret the composition of vari-ous climate and environmental proxy data in sedi-ment cores from lake Emakat

• to discern between local and regional factors gov-erning the composition of the proxy data studied

• to discern among climate and environmental factors and human infl uence governing the composition of the proxy data studied

3

Past environmental and climate changes in northern Tanzania

• to relate the detailed climate history obtained from Empakaai over the last 700 years to the societal de-velopment in near-by Engaruka.

Empakaai Crater and surroundings

The Empakaai Crater is a caldera belonging to the vol-cano Elanairobi (Fig. 1). The caldera is located in the Crater Highlands and origins from the orogeny prior to the development of the southern end of the eastern rift system. Elanairobi has not been active during the past 2 million years (Hay 1976). Two volcanoes on the east-ern fl ank of the highlands, the Kerimasi and Ol Doinyo Lengai, developed after the rift valley formation and the latter is still active. The lavas from Ol Doinyo Lengai are rich in alkali earth metal Na and K and are relatively rich in phosphorus (Dawson et al. 1990). The constituents of the ash and lavas (tephra) from Lengai dissolve quickly making the nearby surface and groundwater alkaline and carbonate rich (Dawson et al. 1990).

Lake Emakat, situated at the Empakaai Crater bot-

tom is a saline-alkaline 80 m deep lake. The depth of the lake, in combination with its relatively small size and restricted catchment, is rare in this part of eastern Africa. The lake is hydrologically closed and, while the alkalin-ity is caused by the composition of the eroded bedrock and the input of tephra from Ol Doinyio Lengai, it’s sa-linity is closely linked to precipitation and runoff:evapo-ration ratios (Cohen and Andrews 2003). The interannual lake level variation, caused by the seasonality in rainfall, is about 40 to 50 cm (Frame et al. 1975). The lake hy-drology and lake level fl uctuations over longer time span are discussed in papers II and III. Since there is no evi-dence of groundwater input to Empakaai we assume that lake level changes generally respond to relative changes between water input, e.g. precipitation and streamwater and evapotranspiration. During a series of years lake lev-el fall occurs when the evaporation greatly exceeds water input, as was reported in the late 1950s, when, following consecutive years of below average annual rainfall, lake Magadi in Ngorongoro Crater (20 km south of Empa-kaai) dried out completely and the lake level in Emakat was relatively low (Frame et al. 1975) (Fig. 2).

Climatically, the Crater Highlands is infl uenced by two major trade wind systems; the north-easterlies, originating from northeastern Africa and the Arabian Pe-ninsula, and the south-easterlies, which originate in the southwest regions of the Indian Ocean (McGregor and Nieuwolt 1998). In general, rainfall is associated with the seasonal movement of the ITCZ (intertropical con-vergence zone). The dominating easterly wind over the Crater Highlands creates a transect of decreasing rainfall from east to west with a mean annual rainfall of about 1000 mm on the eastern side to less than 200 mm/year on the plains to the west (Homewood and Rodgers 1991). Measurements from one year (March 1973-February 1974) at the eastern rim of Empakaai show the varia-bility in rainfall within one year (Fig. 1). The nutrient rich volcanic soils, most probably Andosols (Westerberg pers. comm. 2007), in combination with relatively high rainfall on the south eastern escarpment, support a forest along and above the eastern escarpment. These condi-tions are also suitable for agriculture, but a major part of the Crater Highlands is included in a protectorate, the Ngorongoro Conservation Area, NCA, which allows only restricted extensive human use.

The present climate supports an extensive Afromon-tane forest in Empakaai, but only 10% of the vegetation is composed of a mosaic of dry and moist forest (Rånge 2001). Bushland including thicket is common on both the inner and outer wall of the caldera (Frame et al. 1975, Rånge 2001), probably as a result of human induced fi res and agriculture (Fosbrook 1972, Frame et al. 1975, paper III). The vegetation in Empakaai Crater and surround-ings are discussed in more detail in papers I and III. The

Figure 2. Photographs from 1921 (Barns 1923) (top) and 1999 (bottom) showing Lake Emakat´s southeastern side. Note the tree remains in situ and the higher lake level in the 1921 pho-tograph.

4

Maria Ryner

modern vegetation in the Crater Highlands is described in Herlocker and Dirschl (1972) and the local fl ora in Empakaai is based on vegetation surveys conducted in 1974 (Frame et al. 1975) and 2000 (Rånge 2001).

Human history

The Empakaai Crater is situated in a region known for its long record of early human occupation (Leaky 1971). The number of people living in this region was however limited before the advent of farming and herding (Home-wood and Rodgers 1991, Ehret 1998). Around 500 BC the fi rst evidence of pastoralism in the Crater Highlands emerged (Fosbrook 1972) and from around 1000-1500 AD the Tatog (Datog) people, who today live immedi-ately to the south-southwest of the Highlands, occupied a large area of the north and central Tanzania, including the Crater Highlands (Homewood and Rodgers 1991).

The present mosaic of forest, scrubs and grassland, in the Crater Highlands and on the slopes of Empakaai Crater, has been suggested to be due to early occupation and limited cultivating activities (Herlocker and Dirschl 1972, Frame et al. 1975). The Masaai, now occupy-ing the Crater Highlands probably entered the northern Tanzania in the 1700s (Fosbrook 1948). However, an intensive use of the Crater Highlands by the Masaai is suggested to have occurred from 1850 AD and onward (Fosbrook 1972, Homewood and Rodgers 1991). East of the Crater Highlands, and below the escarpment, is one of the best preserved abandoned land use systems in eastern Africa, the ancient Engaruka irrigation system, dated to have been an active agricultural system between c. 1400-1820 AD (paper IV). This also witnesses of past human activities close to the caldera.

The late 1800s to early 1900s AD, contemporary with the European colonization of eastern Africa, was a time of famines and epidemics that had a tremendous

Figure 1. Map showing the location of Empakaai Crater in Ngorongoro Conservation Area (NCA), northern Tanzania and aclimate diagram from Empakaai (Mars 1973 to Feb. 1974) (Frame et al. 1975).

0 5 10 15 km2010

30 35 40

10

55

0

5 30 35 40

0

5

Crater

EMPAKAAI CRATER

The Crater H

ighlands

Mountain

∆

∆

∆

∆

∆∆

N

Tanzania

LakeVictoria

Kenya

Uganda Tanzania

NCA

0

20

40

60

80

100

120

140

160

180

200

Mar Apr May June July Aug Sept Oct Now Dec Jan Feb0

2

4

6

8

10

12

14

16

18Mar 1973 - Feb 1974

NgorongoroCrater

Northern HighlandsForest Reserve

Volcanic peak

Boundary of NCA

Rift

esc

arpm

ent

Engaruka

Mt Kermasi

Ol Doinyo Lengai

Elanairobi

Nainokanoka

Olmoti Crater

Serengeti Plain

Gol Mountains

Boundary of N. Highland Forest Res.

Village

5

Past environmental and climate changes in northern Tanzania

impact on the human and natural environment in Tanza-nia (Waller 1988) also affecting the people in the Crater Highlands (Homewood and Rodgers 1991). In the 20th century the Crater Highlands is recognized as an extraor-dinary beautiful area with high concentrations of wild animals and it therefore became a protectorate in 1959 (Fosbrook 1972). This protection limited the Masaai´s ability to use the Crater Highland and since then there has been a continuing debate between the Ngorongoro Conservation Area Authorities and the local people (Fos-brook 1972, Homewood and Rodgers 1991).

The presence of humans in the Empakaai and its surroundings has most probably impacted on the devel-opment of the landscape and also left an imprint in the sediments preserving the history of the past. To disentan-gle changes in sediments and its proxy data due to hu-man activities from changes caused by regional climate changes is one of many exciting challenges in paleoenvi-ronmental research.

The material and methods

Within this study several proxies from the lake sedi-ments have been used. Pollen (papers I and III), charcoal (paper III) and to some extent organic carbon content (papers II and III) represent catchment conditions, while diatoms, chironomids (paper II), inorganic carbon (pa-pers II and III) and sediment lithology (papers I and III) indicate changes in lake conditions. In addition to these proxies, dated tree remains in situ have been used as indi-

cators of dry conditions (papers III and IV). Photographs and aerial photographs from the 20th century have also contributed to the understanding of the Empakaai Crater catchment over the last century (papers III and IV).

In total twelve cores; six surface sediment gravity cores and six 1-2 m long piston cores were retrieved from Emakat (Fig. 3) of which three are used in this thesis. The original ambition, to sample 3 m long cores, proved to be impossible due to the presence of hard crus-tal-like layers.

All piston cores were run through a GEOTEK multi-sensor core logger to determine down-core variations in magnetic susceptibility, commonly used to correlate cores from a single lake (Verosub and Roberts 1995). The results however did not show any clear relation among the cores. The presence of hard layers and hiatuses (as later determined, see paper III and Muzuka et al. 2004) and observed differences in diatom composition among some of the cores (Helena Öberg pers. comm. 2007) sug-gest that large changes in lake level have occurred during the Holocene, to the extent that sampling sites, located at different distances from the present shore line, at some times in the past have been located above the lake shore line. This implies that the different cores may have hia-tuses in sediment accumulation which might explain the diffi culties to correlate the cores among each other.

Three of the piston cores, two from 20m water depth, E4 and E3, and one from 11 m water depth, E2b, were split length wise. These cores were chosen, because E4 and E3 supplemented each other and E2b showed

Figure 3. Map showing the location of the cores taken in Lake Emakat. Grey dot indicate surface cores and black dots indicate core retrieved with a modifi ed Livingstone corer.

3000

2800

2600

23002400

N

797060

504030

2010

Lake s

hore

1 km

E2b

E4

Emp2

Emp1

E3

1 km

6

Maria Ryner

some similarities in susceptibility with E3. However, the minerogenic material in E3 and E2b, with few biogenic remains and little organic material to date, made a cor-relation among the three very diffi cult and hence E4 was chosen as master core for further studies (papers I and III). Two of the surface sediment cores, from 25 m and 8 m water depth, labelled Emp1 and Emp2 respectively, were also chosen for further studies (paper III). These were chosen because they contained the longest sur-face sediment sequences and were located furthest apart among the short cores.

Pollen analysis was chosen as the main initial proxy in this thesis because i) the site has great potential for yielding relatively straightforward information about the history of the local vegetation, ii) the information de-rived from pollen may also be used in other studies in the region, e.g. archaeological and human geographical studies and iii) pollen is one of the most direct proxies of environmental change.

Charcoal is an amorphous carbon compound result-ing from incomplete combustion of plant tissue. Char-coal is well preserved in sediments and can therefore be used as a record of past fi res (Patterson et al. 1987). Charcoal fragments smaller than 25 µm is susceptible to wind transport and a large amount of these may record vast regional fi re activities, while larger fragments usu-ally refl ect local fi res (Patterson et al. 1987, Clark 1988). Natural fi res, caused by lightning, together with clear-ing and dry season burning, common in tropical arid and semiarid regions, are constantly increasing the fi re prone ecosystems (Bond et al. 2005). There are several meth-ods that may be used when extracting charcoal (Clark 1988). Counts of charcoal from samples processed for pollen analysis, as in this study, may produce underesti-mations of the actual content of charcoal (Turner et al. in press). However Pitkänen et al. (1999) found in sediment samples recording local fi res, that the most abundant charcoal fragment found in samples prepared for pollen analysis was the same as the majority of particles found in the sediments.

Both diatoms and invertebrates have successfully been used as indicators of lake level changes in lakes in eastern Africa (Gasse et al. 1997, Verschuren 2000, Barker et al. 2002). Diatoms are sensitive to lake-wa-ter chemistry changes, they commonly occur in great numbers in both modern lakes and lake sediments and a great number of diatom taxa are cosmopolitan or have a widespread geographical distribution (Gasse et al. 1995). Invertebrate fossil remains have also been extensively used as indicators of climatic change (De Deeker and Forester 1988). Ostracods are small bivalve crustacean that are common inhabitants of near-bottom freshwater environments, occupying both permanent and temporal water bodies. Their mainly benthic lifestyle and niche

specifi cation make them good indicators of littoral bi-otic response to changes in physical and chemical condi-tions (Forester 1991, Martens 1996, Park et al. 2003). The long persistence of larval head capsules in lake sediments has lead to widespread use of chironomids assemblage in palaeoecological studies (Walker 1995). In Africa the use of chironomids in palaeoenvironmental studies is relatively recent (Mees et al. 1991; reviewed in Verschuren and Eggermont 2006) and has proven to be a useful complement to diatom and ostracod studies.

Lithological proxies have in this thesis been used to contribute to the general description of the sediments and are discussed in relation to the biological prox-ies (papers I-III). The lithological description is based on visual inspection and smear slides and carbon con-tent. Measuring the organic carbon content is a standard procedure in lake sediment studies. It contributes with information about relative organic matter content in the sediment, a proxy for lake productivity and/or infl ux of terrestrial organic debris. The control of sedimentary or-ganic matter may also be caused by dilution by minero-genic sediment component, usually associated with lake level change (Verschuren 1991). The inorganic carbon content e.g. carbonates usually represent the chemical precipitates and biotic remains (ostracods) deposited in the sediment.

Summary of papers

Paper I

Ryner, M.A., Bonnefi lle, R., Holmgren, K. and Muzuka A. 2004. Vegetation changes in Empakaai Crater, northern Tanzania, at 14,800 - 9300 cal yr BP. Review of Palaeobotany and Palynology 140, 163-174.

In this paper vegetation changes inferred from the pol-len record of a radiocarbon dated lake sediment core (E4) are presented and discussed. The record includes the time interval covering the late Pleistocene/Holocene transition. The pollen record is analysed at a resolution interval averaging 200 yr. The results show that, within the caldera, a Hagenia-forest starts to develop at 14.5 ka and is the dominating tree species until 13 ka. A change in vegetation after 13 ka, indicated by an increased propor-tion of Nuxia congesta in the forest and Artemisia in the afroalpine grassland corresponds in time to the northern hemisphere’s Younger Dryas cooling. Poaceae and Cy-peraceae increased at ~10.1 ka, indicating a signifi cant increase in local pollen, possibly attributed to lowered lake level, related to drier conditions. Although the Em-pakaai pollen record documents continuous forest condi-tions, from14.5 to 10.1 ka, the variation in the proportion of forest components seem to respond to environmental

7

Past environmental and climate changes in northern Tanzania

changes at the millennium scale.A comparison between four Hagenia pollen records

from Kenya, Uganda, Burundi and Tanzania (this study) covering the Pleistocene/Holocene transition reveals dif-ferent time of Hagenia increase at each site, while three of the records (Uganda, Burundi, Tanzania) all show a decrease in Hagenia c 13 ka (11-10.5 14C ka) (Fig 4.).

propose the local lake-system response to regional cli-matic and hydrological instability during the period of post-glacial warming. The aquatic biological proxy indi-cators suggest that the water level and chemistry of Lake Emakat evolved, fi rst from a shallow freshwater body at 14.8 ka to a deeper freshwater phase between ca.14.4 and 10.3 ka and then to a markedly shallower, alkaline-

Figure 4. Diagram indicating presence of Hagenia abyssinica during the Pleiostocene/ Holocene transition in: Rusaka, Burundi (Bonnefi lle et al. 1995); Muchoya, Uganda (Taylor 1990); Sacred lake, Mt Kenya (Coetzee 1967, Olago et al. 1999) and Lake Emakat, Tanzania (paper I). Dashed line indicates an approx. location of 12,000 14C yr BP. Doted line indicates when the Hagenia-expansion started at each site. Note: The dates are in radiocarbon years and the records are represented in different time scale due to uncertainty in age-depth models.

Paper II

Ryner, M., Gasse, F., Rumes, B and Verschuren, D. 2007. Climatic and hydrological instability in semi-arid equatorial eastern Africa during the late Glacial to Holocene transition: a multi-proxy reconstruction of aquatic ecosystem response in northern Tanzania. Palaeogeography Palaeoclimatology Palaeoecology, 248, XX-XX.

This paper reports new analysis of microfossil assem-blages (chironomids, diatoms and ostracods) on the same material as in paper I, ca. 14.8-9.3 ka. These data, in the context of available palaeoclimatic and palaeoeco-logical information from the region and together with previously published fossil pollen (Ryner et al. 2006, paper I) and carbon and nitrogen isotopic data (Muzuka et al. 2004) for the same sediment sequence, are used to

saline environment after ca.10.3 ka. The lake appears to have been deepest between 13.2 and 12.0 ka, at a time of climatic drying when moist montane forest vegetation within the lake’s catchment was being replaced by open wood- and scrubland (paper I). Some palaeohydrologi-cal changes reconstructed for Lake Emakat are in phase with lake evolution elsewhere in the region and thus possibly track broad-scale climate changes, but some are not. Collectively these multi-proxy paleolimnological data indicate a complex adjustment of the local aquatic ecosystem to temporal variations, both in total annual effective precipitation and its seasonal distribution. The lake’s hydrological response is also conditioned by local factors, notably its geological and topographic setting.

7,800

10,300

20 40

12,000 yr BP

Muchoya, Uganda(2,260 m a.s.l., 1º28 S, 29º80 E)

Lake Emakat, Tanzania(2,220 m a.s.l., 2º55 S, 35º50 E)

Rusaka, Burundi (2,070 m a.s.l., 3º45 S, 29º61 E)

Sacred Lake, Mt Kenya(2,350 m a.s.l., 0º05N, 37º32E)

11,100

13,900

20 40

20 40

14,050

10,560

9,700

12,200

20 40

8

Maria Ryner

Paper III

Ryner, M., Holmgren, K. and Taylor D. A c.1200-year record of vegetation dynamics and lake level changes from Lake Emakat, northern Tanzania. Submitted to Quaternary International.

Two ~30 cm-long surface sediment cores spanning the last 1200 years, as determined by radiocarbon dating, were analysed on their pollen and charcoal content. Together with information on periods of tree growth in the present lake area, a palaeoenvironmental synthesis for northern Tanzania is proposed. It is shown that the forested catchment located in the afromontane forest

Figure 5. Drawing showing the likely changes in lake level and the shore on the eastern side of Lake Emakat at three different periods in time and photographs from 1972 and 2000 of remains of one Nuxia tree standing in 3 m of water in 1972 (Frame et al. 1975) and by the lake shore line in 2000. Note that the depth and length are not drawn to scale.

1972

2000

Trees includingNuxia congesta

Poaceae Remains of Nuxia

Cyperaceae

-25

0

3

-6

-8

0 20 30

Emp2

-25

0

3

-6

-8

0 20 30

Emp2

c. 1600 AD

-25

0

3

-6

-8

0 20 30

Emp2

1972

2000

9

Past environmental and climate changes in northern Tanzania

zone, were subjected to frequent burning, with an period of increase in fi re at c. 1300-1600 AD, which coincides with an increase of human population in the region. In more recent times an increase in species typical found in human-induced environments are found. The near shore vegetation indicates lower lake levels, suggesting drier

for the history of this land use system. We now apply the new information on the climate history obtained from the Empakaai record (paper III), together with archae-ological and historical information, in order to analyse possible causes, environmental as well as societal, to the development and decline of the irrigated land use system

Figure 6. Periods of drier and wetter conditions in eastern Africa for the last 1200 years inferred from analysis of climate proxy data from Lake Edward (Russell and Johnson 2007), Lake Victoria (Stager et al. 2005), Lake Naivasha (Verschuren et al. 2000), Empakaai (Paper III), Lake Tanganyika (Alin and Cohen 2003) and Lake Malawi (Johnson et al. 2001).

Lake Emakat (1)

Lake Malawi (6)

Lake Naivasha (2)

Lake Tanganyika (5)

Lake Victoria (3)

800 1000 1200 1400 1600 1800 2000

800 1000 1200 1400 1600 1800 2000

Lake Edward (4)

Equator

0 800 Km

TANZANIA

UGANDA

KENYA

2..3

1.

. 4

. 5. 5. 5

.6

Calendar years AD

conditions, prior to c. 1200 AD at c 1450-1650 AD and in the late 18th, late 19th and early 20th century. Fig-ure 5 illustrates three different senarios, present situation (1999), higher lake level (1972) and lower lake level (c 1600 AD). The inferred evolution of lake level changes in Lake Emakat has similarites with many, but not with all, other palaeoenvironmental records in eastern Afri-ca (Fig. 6). The diffi culty with precise dating of events still limits the possibilities to discern the real timing and phasing of climate and environmental changes observed at different sites as well as to disentangle regional events from local ones.

Paper IV

Westerberg, L-O., Holmgren, K., Börjeson, L., Håkans-son, T., Laulumaa, V., Ryner, M. and Öberg, H. The development of the Engaruka irrigation system, Northern Tanzania. Physical and societal factors. Submitted to The Annals of Association of American Geographers.

The Engaruka system is an extensive fossil irrigation system located in northern Tanzania. Several hypotheses based on archaeological fi ndings have been suggested

(Fig. 7). Based on dated cultural layers we show that the irrigation system was initiated in the early 15th century, it was in it most extensive use between 1620 AD and 1720 AD and it declined in the early 19th century. We propose that, while the initiation of the society probably was stimulated both by good climatic conditions and an expanding long-distance caravan trade, the opposite situ-ation, i.e. a long period of drought and diminished long-distance trade, at ~1450-1650 AD, was not detrimental to the land use system. Rather, it may be so that the sur-vival and development of the system was forced by the harsh conditions at this time. In order to adapt to the dry conditions people managed the soils and practiced water conservation. The system continued to expand during the wetter conditions that followed and its maximum size most probably occurred at some time between 1680 AD and 1820 AD. The system was still active into the early 19th century when it fell into decline, probably due to a number of catastrophic events, such as imported diseases (smallpox), environmental deterioration due to popula-tion pressure and increased Masaai aggression. Later at the wake of these events some people took up irrigation at Engaruka again.

10

Maria Ryner

Discussion

The pollen record

In the fossil pollen record there are a few pollen taxa that dominate. The most common tree taxa found in the fossil pollen record are Hagenia abyssinica and Nuxia conges-ta (Fig. 8). In this thesis Hagenia is used as an indicator of more moist conditions favouring open forest canopy, while presence of Nuxia indicates dry Afromontane conditions. Hagenia abyssinica is found on most of the higher mountains between Ethiopia and the Nyika Pla-

teau in northern Malawi, at altitudes from 1800 to 3600 m a.s.l. (White 1983). Characteristically Hagenia form pure stands in narrow zones (often interrupted) between taller types of forest and shrubland of the Ericaceous belt (White 1983). In Empakaai Crater, Hagenia trees are found in the moist parts of the caldera (Frame et al. 1975, Rånge 2001). In Uganda Hagenia is found above the moist bamboo zone (Hamilton 1982) and on Mt Kili-manjaro in the cloud forest (Hemp 2005). Hagenia is also abundant on Mt Kenya up to the opened moorland at c 3600 m a.s.l. (Hedberg 1951, Noad and Birnie 1992). The palaeoenvironmental study from Empakaai covering the period 14.8 to 9.3 ka suggest moister conditions at the same time as a Hagenia forest is expanding (papers I and II), which is in agreement with other palaeoenviron-

Drier conditions

Wetter conditions

Periods of trade

Datings of ruinswith confidence interval

1300 1400 1500 1600 1700 1800 1900

Fires Fires

S

NN

N

N

NS

S

SS

S

SS

C

S

SS

CC

SS

S

S

Intermediate conditions

AD

SCN Location of datings: northern, centraland southern part of settlements

Figure 7. Climate conditions at Empakaai 1300-2000 AD. Empakaai Crater is located c. 15 km from Engaruka. Periods of long-distance trade and datings of settlements (with confi dence intervals) are indicated.

11

Past environmental and climate changes in northern Tanzania

mental studies from Burundi (Bonnefi lle et al. 1995) and Uganda (Taylor 1990). In the pollen record covering the last c. 1200 the Hagenia record more diffi cult to inter-pret and (paper III) and is here also discussed in terms of pollen dispersal, pollen preservation and pollen counts. In this record the importance of an open landscape and the occurrence of fi res seemed to have had a impact on the presence of Hagenia and the light demanding char-acter of Hagenia may be of greater important when the requirements for moisture is suffi cient for its growth. Ac-cording to White (1983), Hagenia is not closely related to climate conditions and in Ethiopia Hagenia have been assigned to dry afromontane forest (Friis 1992).

Nuxia congesta is a shrub or tree varying in height from 2 to 20 m. It is occurring from coastal to medium altitude evergreen forest but also in rocky gorges, and in dry areas among boulders (Palgrave 1983). It is com-monly associated with dry afromontane forest found in highland regions in sub-Saharan Africa (White 1983). There are three Nuxia species that are common in east-ern Africa but the Nuxia congesta pollen is relatively easy to distinguish from the other types (Fig. 8) (pers. comm. David Taylor 2006). Nuxia is found in palae-oecological studies in highland eastern Africa (Bon-nefi lle et al. 1995, Marchant and Taylor 1998, Vincens et al. 2005) however only in small numbers. Frame et al. (1975) suggest that dry evergreen forest with Nuxia congesta was once extensive in Empakaai Crater, but only small remnants have survived repeated dry season burning and tree-cutting.

Other tree pollen found in notable numbers in the lake sediments are pollen from Podocarpus, Olea and Junipe-rus (Fig 8), all characteristic genera of some of the driest types of montane forest (Hamilton 1987). However Po-docarpus and Olea have species, Podocarpus latifolius and Olea capensis respectively that prefer relatively wet conditions (Lind and Morrison 1974, Noad and Birnie 1989) and these species may therefore conceal changes in available moisture (Hamilton 1982, Lamb et al. 2003). In the Lake Emakat record Podocarpus and Olea pollen occur throughout all sediments studied and without dras-tic changes in relative abundance. Two species of Olea, O. capensis and O. europea ssp africana are found in the caldera today. Podocarpus is not present in Empakaai to-day (Frame et al. 1975, Rånge 2000) nor recorded in the vegetation survey over Ngorongoro Conservation Area (Herlocker and Dirschl 1972). Podocarpus absence in the present day fl ora may indicate that the conditions today

Figure 8. Microscopic photographs (Zeiss Axiophot) of pollen from the Lake Emakat sediments showing a) Hagenia abys-sinica, b) Nuxia congesta, c) Olea capensis d) Podocarpus e) Juniperus (*1000 magnifi cation).

a) Hagenia abyssinica

b) Nuxia congesta

c) Olea capensis

d) Podocarpus

e) Juniperus

12

Maria Ryner

are less favourable for its survival than in the past, or that its suitability as timber has led to its complete removal (Dharani 2002) or that the fossil pollen found in the sedi-ments represent long distance transported pollen (Ham-ilton 1987). Juniperus procera occurs most commonly at altitudes between 1500-3000 m (Noad and Birne 1989). Juniperus is mostly a constituent of open evergreen scle-rophyllous mountain forest and above 1300 mm annual rainfall Juniperus is usually absent (Farjon 1992) and has light demanding characteristics (Lind and Morrison 1974, Fetene and Feleke 2001).

The climate record from Lake Emakat, Empakaai Crater

Even though the two time-slices studied in this thesis, one 5500 and the other 1200-years-long, represent sig-nifi cantly different episodes in the African environmen-tal history, the same pollen taxa are to a large extent present in both records. Of the 78 taxa identifi ed in the pollen record from the late Pleistocene/early Holocene transition, 82% are present also in the surface sediments. The large abundance of montane forest taxa in the pollen records from the transition and the last 1200 years AD may indicate that the montane forest was more extensive in both periods than it is today even though it can not be fi rmly concluded. The different dispersal of different pollen taxa may infl uence the abundance of pollen grains found in each records. Forest trees usually have more easily dispersed pollen than most bush species (Hamil-ton 1976) and the most common tree types in the pollen records may thus be over-represented in comparison to bush/schrub vegetation. Therefore, in this thesis, the rel-ative changes in the proportion between these pollen taxa have been used for the interpretation of relative changes in vegetation cover, rather than attempting to document the true composition of the complete vegetation.

The overall vegetation response to changing condi-tions is small in comparison to the lake response. While the pollen fl ora has been fairly similar over time, the lake conditions were markedly different in early Holocene compared with the last 1000 years (papers II and III). Although the lake is alkaline in both the record from the late Pleistocene/early Holocene transition and the last ~1200 years AD the diatom study shows that the early material record a lake, at its most saline stage, much less saline (~6000 μS cm−1), than the present lake (27,000 μS cm−1) (paper II and Muzuka et al. 2004). The few and highly corroded diatoms (own observations) and chi-ronomids (Eggermont pers. comm. 2005) in the younger material prevent us from drawing any conclusions about the salinity during the last 1000 years. The diatoms that were present however were mainly Nitzschia (pers. comm. Helena Öberg 2005) which are the same diatom

species found in the present lake (Kilham 1971) while in the early material Nitzschia only occurred in small amounts (paper II).

A conclusion that can be drawn from the above dis-cussion on pollen and lake proxies is that the lake seems to respond more quickly to changing conditions than does the catchment vegetation. Since the lake is closed and there are no evidence of tectonic activity the only water source is direct precipitation and runoff from with-in the caldera. There are several factors that can cause an increase in lake level; i) an increase in rainfall and/or ii) a decrease in temperature leading to less evaporation from the lake surface or iii) increased runoff from the catch-ment due to heavy rainstorms or less vegetation cover. The record of short-lived changes in effective precipita-tion, however signifi cant their temporary effect on lake water balance are, may be absent in the pollen record caused by a muted vegetation response. The vegetation and lake responses may therefore partly be driven by multiple factors at different spatial and temporal resolu-tion. A combination of the two has in this thesis been used to minimize the different possible reasons behind observed changes.

Climate change in eastern Africa from 15ka to the last 1000 years.

Climate change in eastern Africa over the last 15 ka has been described in several reviews (e.g. Street and Grove 1976, Street-Perrott and Perrott 1993, Olago et al. 1999, Gasse 2000, Barker et al. 2004). At 22 ka inso-lation in northern hemispheric tropics rose from a mini-mum to maximum at 12 ka (Berger and Loutre 1991) and soon after 22 ka Africa experienced large changes in climate (Gasse 2000) when the Pleistocene/Holocene transition commenced (Blunier et al. 1998). Below is a brief review of millennial-scale changes recorded from eastern Africa over the last 15,000 years and in-corporating the results from this thesis. In fi gure 9 lake records from eastern Africa, covering extensive parts of the last 15,000 years are illustrated together with the Lake Emakat record. The records are based on vari-ous proxies (diatoms, isotopes, organic matter, pollen) and interpreted in terms of wind (Johnson et al. 1996, Talbot and Laerdal 2000), lake level changes (Talbot and Laerdal 2000, Gasse et al. 2002, Barker et al. 2002, 2003, Chalié and Gasse 2002, paper II) and catchment vegetation change (Vincens et al. 2005, paper I). In the fi gure black bars indicate wetter, dark grey intermediate and light grey bars indicate drier conditions.

From c 15 ka rising lake levels are recorded in many lakes in the region (Roberts et al. 1993, Talbot and Laer-dal 2000, Gasse 2000, Barker et al. 2002; 2003, paper II). At the same time humid forest species become more

13

Past environmental and climate changes in northern Tanzania

common in many highland regions (Taylor 1990, Bon-nefi lle et al. 1995, Jolly et al. 1997 and 1998, paper I). Humid conditions are also observed at Lake Emakat from 14.8 ka as deduced from the pollen record (Fig. 9) even though lake levels decrease between 14.4 and 13.2 ka (paper II). The lake level record at Emakat is in agreement with similar conditions at Lake Massoko in southern Tanzania (Barker et al. 2000), while other records only show a short regressive phase, such as the records from Lake Victoria (Talbot and Laerdal 2000) and Lake Abiyata (Chalié and Gasse 2002), or no change at all (Barker et al. 2003, Filippi and Talbot 2005). A dry interval within c 13-11 ka is found in many lakes (Gillespie et al. 1983, Roberts et al. 1993, Johnson et al. 2002, Stager et al. 2002, Talbot et al. 2007) and in one peat bog record (Bonnefi lle et al. 1995) and is sug-gested to coincide with the Younger Dryas chronozone (13.2-11.8 ka). Lakes from the southern part of the re-gion show a less consistent signal. Rukwa (Barker et al. 2002) records no dry interval between 13-11 ka, while Malawi does (Johnson et al. 2002, Gasse et al. 2002, Filippi and Talbot 2005). The lake level in Massoko was already reduced in 13 ka and increase from 12 ka (Barker et al. 2003). Lake Emakat indicates no lake level lower-ing (paper II), although there is slightly drier conditions

c. 13.2 ka (paper I), as indicated in the vegetation. Most lake sequences recording a lake level lowering return to wetter conditions (Chalié and Gasse 2002, Barker et al. 2003, Filippi and Talbot 2005). Wet conditions are also inferred from palynological data from Lake Edward fol-lowed by a more variable climate from c.9 ka (Beuning and Russell 2004). Lake records from Mt Kenya also in-fer increased precipitation from c. 11.1 to 8.6 ka (Barker et al. 2001). These wetter conditions are not observed in the Lake Emakat record, instead the record show a marked change towards drier conditions c 10.3 ka (paper I), not evident from elsewhere in eastern Africa.

Around 8 ka a period of low lake levels is found in Lake Victoria (Talbot and Laerdal 2000, Stager et al. 2003) and Lake Malawi (Johnson et al. 2002, Filippi and Talbot 2005) (Fig. 2). The period coincides with the brief but extensive cooling observed in northern high latitudes at 8.2 ka (e.g. Alley et al. 1997). Again the signal from eastern Africa is not consistent. Bonnefi lle and Chalie (2000) propose an increase in rainfall in highland eastern Africa. Lake Massoko and Lake Rukwa (Thevenon et al. 2002, Barker et al. 2003, 2002) report continuing wet conditions and so does L Abyiata further north (Chalié and Gasse 2002).

After 5.7 ka, numerous sites record a change towards

Figure 9. Periods of drier and wetter conditions in eastern Africa from Lake Abyiata (Chalié and Gasse 2002), Lake Victoria (Tal-bot and Laerdal 2000, Stager et al. 2003), Lake Emakat (paper I, II and III), Lake Rukwa (Barker et al. 2002, Vincens et al. 2005), Lake Massoko (Barker et al. 2000; 2003), Lake Malawi (Gasse et al. 1996, Johnson et al. 1996, Filippi and Talbot 2005). Black bars indicate wetter, dark grey intermediate and light grey drier conditions.

0100020003000400050006000700080009000100001100012000130001400015000

Equator

0 800 Km

TANZANIA

UGANDA

KENYA

3.

. 1

.5 . 4. 6

2.

? ? ?

?

?

L. Abyiata (1)

L Victoria (2)

L Emakat (3)

L Massoko (5)

L Rukwa (4)

L Malawi (6)

Calibrated years BP

14

Maria Ryner

drier conditions. Lake Malawi, Lake Rukwa, Lake Vic-toria and Lake Abiyata record lake level lows (Chalié and Gasse 2002, Barker et al. 2003, Stager et al. 2003, Powers et al. 2005) ending the “African humid period”. In the last 5000 years drier conditions are dominating in most records although several lake records report in-tervals with wetter conditions (Talbot and Laerdal 2000, Barker et al. 2000, Barker et al. 2001, Chalie and Gasse 2002, Stager et al. 2003).

A fi nal dry interval on a millennia timescale occurred c 1000 years ago. Several lake sediment records show a dry period of about 300 years length (Verschuren et al. 2000, Lamb et al. 2003, Stager et al. 2005, Russell and Johnson 2005, Filippi and Talbot 2005, paper III).

This brief summary of the paleoclimate record from eastern Africa reveals several regionally felt intervals, but with the length and character differing from one site to another. Differences between the records may illu-minate the range of variability that is found in the lake proxy records and/or in the large uncertainties inherited in radiocarbon age models and/or in the ambiguity in the interpretation of the proxy data. Differences between sites are to be expected since local climate and catch-ment conditions at each site infl uence the record. The use of multiproxy lake records, where single proxies respond to different processes, increase not only the complexity but also the possibility to evaluate if changes found orig-inates from a local or regional source. A more prominent diffi culty still is the limitations in the radiocarbon chro-nologies and age calibrations performed for the differ-ent sites. There are a good number of sites and records from eastern Africa, but there is still no consensus about the regional climate evolution. High resolution continu-ous records are still scarce but are warranted to verify or falsify the increasing number of suggested scenarios for climatic changes in eastern Africa. The Empakaai record also suffers from relatively low resolution and hiatuses. However the sequences presented are relatively well dat-ed and the changes better constrained than in many of the other records from the region.

Human and environmental proxy records in eastern Africa

Since the second millennia AD, human infl uence on lake and peat bog catchments has been recognised from several locations in eastern Africa. Paleoenvironmental records from Uganda (Marchant and Taylor 1998, Taylor et al. 2000, 2005, Lejju et al. 2005), Ethiopia (Bonnefi lle and Mohammed 1994, Legesse et al. 2002), Kenya (Lamb et al. 2003), Tanzania (Vincens et al. 2003, Cohen et al. 2005, paper III) contemporary with archaeological fi nd-ings (Sutton 1993, Robertshaw and Taylor 2000, Taylor et al. 2005, paper IV) and linguistic sources (Ehret 1998,

2002) all suggest concentration of human settlements and an expansion of agricultural activities and farming in the region. Whether the fl uctuations in human activities are related to economical, political factors or a response to fl uctuations in climate is not always clear. Both social and natural factors are probably involved (paper IV). The great regional variability in the present climate of eastern Africa (Nicholson 2000) may well have been as variable within the last 1000 years, as indicated by differences in climate between sites (Verschuren 2000, Holmgren and Öberg 2006, paper III). Comparing palaeoenvironmental records at different sites with each other as well as with archaeological fi ndings and linguistic, oral and written history all incorporates some uncertainty of imprecise dating and gaps in the records preventing an in depth analysis of causes and effects. However an increasing number of sources are valuable for palaeoenvironmental and human historical reconstruction and a combination of them can deepen the discussion about factors contrib-uting to the establishment and decline of human centres (Hulmes 1996, Holmgren and Öberg 2006 and paper IV).

Conclusions

The major fi ndings of this thesis are:

• The record from Lake Emakat in Empakaai Cra-ter, presented in this thesis, contributes palaeoen-vironmental information from two time-slices in the African environmental history; 14.8 to 9.3 ka and the last 1200 years.

• A complete history covering the whole of Holocene has not yet been possible due to sedimentation hia-tuses and/or disturbed and reworked sections in the recovered cores.

• The 5500 year long sequence, 14.8-9.3 ka, records a complex system with contradicting results between catchment and lake conditions that might be ex-plained by internal response to external forces.

• A Hagenia-forest developed at 14.5 ka and was the dominating tree species until 13 ka. A change in veg-etation after 13 ka corresponds in time to the north-ern hemisphere’s Younger Dryas cooling. Lower lake levels and drier conditions occurred at ~10.1 ka.

• The Lake Emakat record is partly in agreement with other eastern African palaeoclimatological records both in the older and the younger sequence. However there is especially one period, 10.3-9.3 ka, that has

15

Past environmental and climate changes in northern Tanzania

little resemblance with other eastern African records suggesting either a locally induced change that infl u-enced the whole catchment, or that the chronology is not reliable over this period and that this dry event are present in other records but at another time. A third alternative is that this record responds to a re-gional change that has not yet been observed in other palaeoenvironmental records from the region.

• The last 1200 years is characterized by recurrent lake level fl uctuations with a long period of low lake level between ~ 1450 and 1650 AD. The vegetation evolution is inferred to some extent related to in-creased human activities recorded from an increase in fi re frequency, between c. 1300 and 1600 AD.

• The initiation of an irrigation system in Engaruka, 15 km from Empakaai Crater, occurred during rela-tively wet conditions. The systems continuing devel-opment is not entirely dependent on good climate conditions even though the most extensive phase of the system between 1680 and 1820 occurred during signifi cantly wetter conditions than at present.

• The vegetation in the caldera seems to respond to high amplitude, low frequency changes while the lake responds both to the low frequency changes and to low amplitude, high frequency changes.

The environmental and climate information derived from this research in Empakaai Crater endorse the initial hypothesis that this caldera contributes to the overall cli-matic history of eastern Africa. The palaeoenvironmen-tal community together with other fi elds of research is facing the fact that the more we learn about processes in nature the more we increase our understanding of its complexity. One fundamental factor of importance for mapping leads and lags in palaeoclimate is chronological control; therefore the timescale has always been and still is one of the great challenges.

Acknowledgment

There are a great number of people involved in this thesis to which I am very grateful.

First of all I like to thank my supervisor Karin Holmgren for her support, patience, encouragement and optimism. You have done a tremendous job get-ting me started after the three maternity leaves, al-ways been there with your support when I have been in doubt and kept me going until the very end. My co-supervisors Lars-Ove Westerberg, Peter Kuhry (the

last 3 years) and Stig Jonsson (until the Licentiate de-gree) have given valuable comments and given criti-cal input on my manuscripts. Wibjörn Karlén and Eve Arnold, thank you for your enthusiasm and genuine interest in science which encouraged me to continue with graduate studies.

I would like to thank Alfred Muzuka at the Insti-tute of Marine science, University of Dar es Salaam; this project would not have been possible without your cooperation. The Tanzania Commission for Science and Technology (COSTECH) and Tanzania Wildlife Research Institute (TAWIRI) kindly author-ised research permission and fieldwork. Particular thanks to Victor Runyoro, a conservation authority officer at Ngorongoro. I would like to thank Kent Rånge, the personnel at Scantan tours and the Masaai for invaluable assistance during field work. Many thanks to Neduvoto Mollel, Tropical Pesticides Re-search Institute, for conducting the vegetation survey in Empakaai Crater.

I thank Guillaume Buchet, Centre Européen de Recherché et d’ Enseignment des Geosciences de l’Environnement, for introducing me to the world of palynology. I also would like to thank my co-authors Raymonde Bonnefi lle, Françoise Gasse, Bob Rumes, Dirk Verschuren, David Taylor. I am truly grateful to the PLATINA (People Land and Time in Africa) members, sharing the exciting work with the Engaruka project.

I want to thank all the colleagues at the Depart-ment of Physical Geography and Quaternary Geology, for joyful discussions over a cup of coffee. I would like to acknowledge the administrative and logistic staff at the department for their great help and support. Thank you all fellow PhD students, many of you who are now doctors, for always sharing a laugh. Thank you, Elin my roommate, for good discussions and for your patience with my outburst over computers and peripherals. Hanna you have been of great support during the final process of the thesis, thank you.

The financial support to this work was provided by the Swedish Society for Anthropology and Ge-ography (SSAG), Lillemor and W:son Ahlmann, Axel Lagrelius, Carl Mannerfelt and Margit Ahl-tins foundations, and the Swedish Foundation for International Cooperation in Research and Higher Education (STINT) and also the Swedish Research Council (VR) and the Granholm Foundation award-ed to Karin Holmgren.

Last but not the least I want to thank my parents, Berit and Gunnar, for always being supportive and that you for numerous of evenings have been taking care of the three most important persons in my life, Liam, Sean and Jo-Linn. Thank you Tom for being there for them when “mamma jobbar”.

16

References

Alin, S.R. and Cohen, A.S., 2003: Lake-level history of Lake Tanganyika, East Africa, for the past 2500 years based on ostracode-inferred water-depth reconstruction. Palaeogeog-raphy Palaeoclimatology Palaeoecology, 199: 31-49.

Alley, R.B., Mayewski, P.A., Sowers, T., Stuiver, M., Taylor, K.C. and Clark, P.U., 1997: Holocene climatic instability: A prominent, widespread event 8200 yr ago. Geology, 25: 483-486.

Barker, P.A., Street-Perrott, F.A., Leng, M.J., Greenwood, P.B., Swain, D.L., Perrott, R.A., Telford, R.J. and Ficken, K.J., 2001: A 14,000-year oxygen isotope record from diatom silica in two alpine lakes on Mt. Kenya. Science, 292: 2307-2310.

Barker, P., Telford, R., Gasse, F. and Thevenon, F., 2002: Late Pleistocene and Holocene palaeohydrology of Lake Rukwa, Tanzania, inferred from diatom analysis. Palaeogeography Palaeoclimatology Palaeoecology, 187: 295-305.

Barker, P., Williamson, D., Gasse, F. and Gibert, E., 2003: Cli-matic and volcanic forcing revealed in a 50,000-year diatom record from Lake Massoko, Tanzania. Quaternary Research, 60: 368-376.

Barker, P., Telford, R., Merdaci, O., Williamson, D., Taieb, M., Vincens, A. and Gibert, E., 2000: The sensitivity of a Tanza-nian crater lake to catastrophic tephra input and four millen-nia of climate change. Holocene, 10: 303-310.

Barker, P., Talbot, M.R., Street-Perrott, F.A., Marret, F., Scourrse, J. and Odada E., 2004: Late Quaternary climatic variability in intertropical Africa. In Battarbee R., Gasse, F. and Stickely, C. (Eds), Past climate variability through Eu-rope and Africa, Amsterdam, Kluwer, pp: 117-138.

Barker, P.A., Street-Perrott, F.A., Leng, M.J., Greenwood, P.B., Swain, D.L., Perrott, R.A., Telford, R.J. and Ficken, K.J., 2001: A 14,000-year oxygen isotope record from diatom silica in two alpine lakes on Mt. Kenya. Science, 292: 2307-2310.

Barns, T.A., 1923. Across the Great Craterland to the Congo: a sequel to ”The Wonderland of the Eastern Congo”. Ernest Benn, London. 276 pp.

Berger, A. and Loutre, M.F., 1991: Insolation Values for the Climate of the Last 10,000,000 Years. Quaternary Science Reviews, 10: 297-317.

Beuning, K.R.M. and Russell, J.M., 2004: Vegetation and sedi-mentation in the Lake Edward Basin, Uganda-Congo during the late Pleistocene and early Holocene. Journal of Paleolim-nology, 32: 1-18.

Blunier, T., Chappellaz, J., Schwander, J., Dallenbach, A., Stauffer, B., Stocker, T.F., Raynaud, D., Jouzel, J., Clausen, H.B., Hammer, C.U. and Johnsen, S.J., 1998: Asynchrony of Antarctic and Greenland climate change during the last glacial period. Nature, 394: 739-743.

Bond, J.B. and Keeley, J.E., 2005: Fire as a global “herbiv-ore”. The ecology and evolution of fl ammable ecosystems.

TRENDS in Ecology and Evolution, 20: 387 394.Bonnefi lle, R. and Mohammed, U., 1994: Pollen-Inferred Cli-

matic Fluctuations in Ethiopia During the Last 3000 Years. Palaeogeography Palaeoclimatology Palaeoecology, 109: 331-343.

Bonnefi lle, R. and Chalie, F., 2000: Pollen-inferred precipita-tion time-series from equatorial mountains, Africa, the last 40 kyr BP. Global and Planetary Change, 26: 25-50.

Bonnefi lle, R., Riollet, G., Buchet, G., Icole, M., Lafont, R. and Arnold, M., 1995: Glacial/interglacial record from intertrop-ical Africa, high resolution pollen and carbon data at Rusaka, Burundi. Quaternary Science Reviews, 14: 917-936.

Chalie, F. and Gasse, F., 2002: Late Glacial-Holocene diatom record of water chemistry and lake level change from the tropical East African Rift Lake Abiyata (Ethiopia). Palaeo-geography Palaeoclimatology Palaeoecology, 187: 259-283.

Clark, J.S., 1988: Particle Motion and the Theory of Charcoal Analysis - Source Area, Transport, Deposition, and Sam-pling. Quaternary Research, 30: 67-80.

Coetzee, J.A., 1967: Pollen analytical studies in East and Southern Africa. Palaeoecology of Africa 3.

Cohen, A.S. and Andrew S.A. (Eds), 2003: The chemical en-vironments of lakes. The history and evolution of Lake sys-tems. Oxford University Press, NC, USA, pp: 69-89.

Cohen, A.S., Palacios-Fest, M.R., Msaky, E.S., Alin, S.R., Mc-Kee, B., O’Reilly, C.M., Dettman, D.L., Nkotagu, H. and Lezzar, K.E., 2005: Paleolimnological investigations of anthropogenic environmental change in Lake Tanganyika: IX. Summary of paleorecords of environmental change and catchment deforestation at Lake Tanganyika and impacts on the Lake Tanganyika ecosystem. Journal of Paleolimnology, 34: 125-145.

Dawson, J.B, Pinkerton, H, Norton, G.E. and Pyle, D.M., 1990: Physicochemical properties of alkali carbonatite lavas: Data from the 1988 eruption of Oldoinyo Lengai, Tanzania. Geol-ogy 18, 260-263.

De Deckker, P., Forester, R.M., 1988: The use of ostracods to reconstruct continental palaeoenvironmental records. In P. De Deckker, J.-P. Colin, J.-P. Peypouqet, (Eds), Ostracods in the Earth sciences. Elsevier, pp: 175-199.

Dharani, N., 2002: Field guide to common Trees and Shrubs of East Africa. Struik publisher, Cape town, South Africa. 320 pp.

Ehret, C., 1998: An African Classical Age: Eastern and South-ern Africa in World history 1000BC to AD 400. James Cur-rey. Oxford. 354 pp.

Ehret, C., 2002: The civilization of Africa: A history to 1800. James Currey. Oxford. 480 pp.

Eggermont, H. and Verschuren, D., 2004: Sub-fossil Chirono-midae from East Africa. 1. Tanypodinae and Orthocladiinae. Journal of Paleolimnology, 32: 383-412.

Eggermont, H. and Verschuren, D., 2004: Sub-fossil Chirono-midae from East Africa. 2. Chironominae (Chironomini and Tanytarsini). Journal of Paleolimnology, 32, 413-455.

17

Past environmental and climate changes in northern Tanzania

Farjon, A., 1992: The taxonomy of multiseed junipers (Ju-niperus Sect. Sabina) in southwest Asia and east Africa (Taxonomic notes on Cupressaceae I). Edinburgh Journal of Botany 49: 251-283.

Fetene, M. and Feleke, Y., 2001: Growth and photosynthesis of seedlings of four tree species from a dry tropical afromon-tane forest. Journal of Tropical Ecology, 17: 269-283.

Filippi, M.L. and Talbot, M.R., 2005: The palaeolimnology of northern Lake Malawi over the last 25 ka based upon the elemental and stable isotopic composition of sedimentary or-ganic matter. Quaternary Science Reviews, 24: 1303-1328.

Forester, R.M., 1991: Ostracode assemblages from springs in the western United States: implications for paleohydrology. Memoirs,-Entomological-Society-of-Canada, 155: 181-201.

Fosbrooke, H.A. 1948: An Administrative survey of the Masai Social System. Tanganyika Notes and Records 26: 1-49.

Fosbrooke, H., 1972: Ngorongoro: The Eight Wonder. Andre Deutsch Ltd, London. 240 pp.

Frame, G., Frame, L.H. and Spillett, J.J., 1976: An ecologi-cal survey and development plan for the Empakaai Crater ecosystem (Ngorongoro Conservation Area). Serengeti Re-search Contribution 212. 472 pp.

Friis, I., 1992: Forest and forest trees of northeast tropical Afri-ca. Royal botanical Gardens, Kew Bulletin Additional Series XV, London. 396 pp.

Gasse, F., 2000: Hydrological changes in the African tropics since the Last Glacial Maximum. Quaternary Science Re-views, 19: 189-211.

Gasse, F., Juggins, S. and Khelifa, L.B.; 1995 Diatom-Based Transfer-Functions for Inferring Past Hydrochemical Char-acteristics of African Lakes. Palaeogeography Palaeoclima-tology Palaeoecology, 117: 31-54.

Gasse, F., Barker, P., Gell, P.A., Fritz, S.C. and Chalie, F., 1997: Diatom-inferred salinity in palaeolakes: An indirect tracer of climate change. Quaternary Science Reviews, 16: 547-563.

Gasse, F., Barker, P. and Johnson, C.J., 2002: A 24,000 yr dia-tom record from the northern basin of Lake Malawi. In E. Odada and D.O. Olago (Eds), East African Great Lakes: Limnology, Palaeolimnology and Biodiversity. Kluwer aca-demic publisher, Amserdam, 393-414.

Gillespie, R., Streetperrott, F.A. and Switsur, R., 1983: Post-Glacial Arid Episodes in Ethiopia Have Implications for Cli-mate Prediction. Nature, 306: 680-683.

Hamilton, A.C., 1972: The interpretation of pollen diagrams from highland Uganda. Palaeoecology of Africa 7: 45 149.

Hamilton, A.C., 1982: Environmental history of East Africa – a study of the Quaternary. Academic Press Ltd, London. 328 pp.

Hamilton, A.C., 1987: Vegetation and climate of Mt Elgon dur-ing the late Pleistocene and Holocene Palaeoecology of Af-rica and the Surrounding Islands, 18: 283-304.

Hay, L.R. (Ed), 1976: Geology of the Olduvai Gorge – a study of sedimentation in a semiarid basin. University of Califor-nia Press, Berkley. 203 pp.

Hedberg, O., 1954: A pollen-analytical reconnaissance in tropi-cal East Africa. Oikos, 5: 137-165.

Hemp, A., 2005; Climate change-driven forest fi res margin-alize the impact of ice cap wasting on Kilimanjaro. Global Change Biology 11, 1013-1023.

Herlocker, D.J. and Dirschl, H.J., 1972: Vegetation of the Ngorongoro Conservation Area, Tanzania. Canadian Wild-life Service Report Series 19. 39 pp.

Homewood, K.M. and Rodgers, W.A., 1991: Masaailand Ecol-ogy – pastoralist development and wildlife in Ngorongoro, Tanzania. Cambridge University Press, Cambridge. 298 pp.

Holmgren K. and Öberg, H., 2006: Climate change in the Southern and East Africa during the past millennium and its implications for societal development. Environment, devel-opment and society, 8: 185-195.

Hulme, M. (Ed) 1996: Climate change and Southern Africa: an exploration of some potential impacts and implications for the SADC region. Norwich and Gland: Climatic Research Unit and WWF International. 104 pp.

Imbrie, J., Hays, J.D., Martinson, D.G., McIntyre, A., Mix, A.C., Morley, J.J., Pisias, N.G., Prell, W.L. and Shackel-ton, N.J., 1984: The orbital theory of Pleistocene climate: support from a revised chronology of the marine delta-O18 record. In: A. Berger, J. Imbrie, J. Hays, G. Kukla and B. Saltzman (Eds), Milankovich and Climate. Reidel Dor-drecht, pp: 269-306.

IPCC, 2001: Climate change 2001: synthesis report. IPCC, Ge-neva. 184 pp.

IPCC, 2007: Climate change 2007: The physical science basis-Summary for policymakers IPCC, Geneva. 18 pp.

Johnson, T.C., 1996: Sedimentary porcesse and signals of past climate change in the large lakes of Eastern African Rift Val-ley. In T.C. Johnson and E.O.Odada (Eds), The Limnology , Climatology and Paleoclimatology of the East African Great Lake. Gorndon and Breach, Amsterdam, pp: 267-312.

Johnson, T.C., Scholz, C.A., Talbot, M.R., Kelts, K., Ricketts, R.D., Ngobi, G., Beuning, K., Ssemmanda, I. and McGill, J.W. 1996. Late pleistocene desiccation of Lake Victoria and rapid evolution of cichlid fi shes. Science, 273: 1091-1093.

Johnson, T.C., Barry, S.L., Chan, Y. and Wilkinson, P., 2001: Decadal record of climate variability spanning the past 700 yr in the Southern Tropics of East Africa. Geology, 29: 83-86.

Johnson, T.C., Brown, E.T., McManus, J., Barry, S., Barker, P. and Gasse, F. 2002. A high-resolution paleoclimate record spanning the past 25,000 years in southern East Africa. Sci-ence, 296: 113-H.

Jolly, D., Harrison, S.P., Damnati, B. and Bonnefi lle, R., 1998: Simulated climate and biomes of Africa during the late qua-ternary: Comparison with pollen and lake status data. Qua-ternary Science Reviews, 17: 629-657.

Jolly, D., Taylor, D., Marchant, R., Hamilton, A., Bonnefi lle, R., Buchet, G. and Riollet, G., 1997. Vegetation dynamics in central Africa since 18,000 yr BP: pollen records from the

18

Maria Ryner

interlacustrine highlands of Burundi, Rwanda and western Uganda. Journal of Biogeography, 24: 495-512.

Kilham, P., 1971: Biogeochemistry of African lakes and rivers. Ph.D. thesis, Duke University, Durham, NC. 199 pp.

Lamb, H., Darbyshire, L. and Verschuren, D., 2003: Vegetation response to rainfall variation and human impact in central Kenya during the past 1100 years. Holocene, 13: 285-292.

Legesse, D., Gasse, F., Radakovitch, O., Vallet-Coulomb, C., Bonnefi lle, R., Verschuren, D., Gibert, E. and Barker, P., 2002: Environmental changes in a tropical lake (Lake Abiy-ata, Ethiopia) during recent centuries. Palaeogeography Pal-aeoclimatology Palaeoecology, 187: 233-258.

Le Turdu, C., Tiercelin, J.J., Gibert, E., Travi, Y., Lezzar, K.E., Richert, J.P., Massault, M., Gasse, F., Bonnefi lle, R., Deco-bert, M., Gensous, B., Jeudy, V., Tamrat, E., Mohammed, M.U., Martens, K., Atnafu, B., Chernet, T., Williamson, D. and Taieb, M., 1999: The Ziway-Shala lake basin system, Main Ethiopian Rift: Infl uence of volcanism, tectonics, and climatic forcing on basin formation and sedimentation. Pal-aeogeography Palaeoclimatology Palaeoecology, 150: 135-177.

Leakey, M.D., 1971: Olduvai Gorge: Excavations in beds I & II 1960- 1963, Cambridge University Press, Cambridge; Disclosing the Past, Doubleday & Co., New York. 306 pp.

Lejju, B.J., Taylor, D. and Robertshaw, P., 2005: Late-Holocene environmental variability at Munsa archaeological site, Uganda: a multicore, multiproxy approach. Holocene, 15: 1044-1061.

Marchant, R. and Taylor, D., 1998: Dynamics of montane for-est in central Africa during the late Holocene: a pollen-based record from western Uganda. Holocene, 8: 375-381.

Martens, K., Davies, B.R., Baxter, A.J. and Meadows, M.E., 1996: A contribution to the taxonomy and ecology of the os-tracoda (Crustacea) from Verlorenvlei (Western Cape, South Africa). South African Journal of Zoology, 31: 23-36.

McGregor, G.R. and Nieuwolt, S., 1998: Tropical Climatology – an introduction to the Climates of theLow Latitudes. John Wiley, Chichester. 339 pp.

Muzuka, A.N.N., Ryner, M. and Holmgren, K., 2004: 12,000-Year, preliminary results of the stable nitrogen and carbon isotope record from the Empakai Crater lake sediments, Northern Tanzania. Journal of African Earth Sciences, 40: 293-303.

Nicholson, S.E., 2000: The nature of rainfall variability over Africa on time scales of decades to millenia. Global and Planetary Change, 26: 137-158.

Noad, T.C. and Birnie, A., 1989: Trees of Kenya. Nairobi, Kenya. 308 pp.

Olago, D.O., Street-Perrott, F.A., Perrott, R.A., Ivanovich, M. and Harkness, D.D., 1999: Late Quaternary glacial-intergla-cial cycle of climatic and environmental change on Mount Kenya, Kenya. Journal of African Earth Sciences, 29: 593-618.

Palgrave Coates K., 1983: Trees of Southern Africa. Struik

Publisher, Cape Town. 959 pp.Park, L.E., Cohen, A.S., Martens, K. and Bralek, R., 2003:

The impact of taphonomic processes on interpreting pale-oecologic changes in large lake ecosystems: ostracodes in Lakes Tanganyika and Malawi. Journal of Paleolimnology, 30: 127-138.

Patterson, W.A., Edwards, K.J. and Maguire, D.J., 1987: Mi-croscopic Charcoal as a Fossil Indicator of Fire. Quaternary Science Reviews, 6: 3-23.

Pitkanen, A., Lehtonen, H. and Huttunen, P., 1999: Compari-son of sedimentary microscopic charcoal particle records in a small lake with dendrochronological data: evidence for the local origin of microscopic charcoal produced by forest fi res of low intensity in eastern Finland. Holocene, 9: 559-567.

Roberts, N., Taieb, M., Barker, P., Damnati, B., Icole, M. and Williamson, D., 1993: Timing of the Younger Dryas Event in East-Africa from Lake-Level Changes. Nature, 366: 146-148.

Robertshaw, P. and Taylor, D., 2000: Climate change and the rise of political complexity in western Uganda. Journal of African History, 41:1-28.

Robershaw, P., Taylor, D., Doyle, S. and Marchant, R., 2004: Famine, Climate and Crises in western Uganda. In R Bat-tarbee., F Gasse,. and C Stickely, (Eds), Past climate vari-ability through Europe and Africa, Amsterdam, Kluwer, 535-549.

Russell, J.M. and Johnson, T.C., 2005: A high-resolution geo-chemical record from Lake Edward, Uganda Congo and the timing and causes of tropical African drought during the late Holocene. Quaternary Science Reviews, 24: 1375-1389.

Russell, J.M. and Johnson, T.C., 2007: Little Ice Age drought in equatorial Africa: Intertropical Convergence Zone migra-tions and El Nino-Southern Oscillation variability. Geology, 35: 1-24.

Ryner, M.A., Bonnefi lle, R., Holmgren, K. and Muzuka, A., 2006: Vegetation changes in Empakaai Crater, Northern Tanzania, at 14,800-9300 cal yr BP. Review of Palaeobotany and Palynology, 140: 163-174.

Ryner, M., Gasse, F., Rumes, B and Verschuren, D., 2007: Cli-matic and hydrological instability in semi-arid equatorial East Africa during the late Glacial to Holocene transition: A multi-proxy reconstruction of aquatic ecosystem response in northern Tanzania. Palaeogeography, palaeoclimatology and palaeoecology, 248: XX-XX.

Rånge K., 2001: The Empakaai Crater – sammanställning övervegetationsförhållandena I Empakaaikratern, norra Tan-zania, under augusti 2000. MSc student report, Dep. of Phys.Geogr. and Quat. Geol., Stockholm Univ. 25 pp.

Ssemmanda, I., Ryves, D.B., Bennike, O. and Appleby, P.G., 2005: Vegetation history in western Uganda during the last 1200 years: a sediment-based reconstruction from two crater lakes. Holocene, 15: 119-132.

Stager, J.C., Cumming, B.F. and Meeker, L.D., 2003: A 10,000-year high-resolution diatom record from Pilkington

19

Past environmental and climate changes in northern Tanzania

Bay, Lake Victoria, East Africa. Quaternary Research, 59: 172-181.

Stager, J.C., Westwood, J., Grzesik, D. and Cumming, B.F. 2005. A 5500-year environmental history of Lake Nabug-abo, Uganda. Palaeogeography Palaeoclimatology Palae-oecology, 218: 347-354.

Street, F.A. and Grove, A.T., 1976: Environmental and climatic implications of late Quaternary lake-level fl uctuations in Af-rica. Nature, 261: 385-390.

Street-Perrott, F.A. and Perrott, R.A. 1993. Holocene vegeta-tion, lake levels and climate of Africa. In: H.E. Wright, jr., J.E. Kutzbach, T.I. Webb, F.A. Street-Perrott and B. P.J. (Eds), Global climate since the Last Glacial Maximum. Uni-versity of Minnesota Press, Minnesota, pp: 318-356.

Sutton, J.E.G., 1993: The Antecedents of the interlacustrine kingdoms. Journal of African History 34: 33-64

Talbot, M.R. and Laerdal, T., 2000: The Late Pleistocene-Holocene palaeolimnology of Lake Victoria, East Africa, based upon elemental and isotopic analyses of sedimentary organic matter. Journal of Paleolimnology, 23: 141-164.

Talbot, M.R., Filippi, M.L., Jensen, N.B. and Tiercelin, J.J., 2007: An abrupt change in the African monsoon at the end of the Younger Dryas. Geochemistry Geophysics Geosystems, 8

Taylor, D. M., 1990: Late Quaternary pollen records from two Ugandan mires, evidence for environmental changes in the Rukiga Highlands of southwest Uganda. Palaeogeography. Palaeoclimatology. Palaeoecology, 80: 283-300.

Taylor, D., Lane, P.J., Muiruri, V., Ruttledge, A., McKeever, R.G., Nolan, T., Kenny, P. and Goodhue, R., 2005: Mid- to late-holocene vegetation dynamics on the Laikipia Plateau, Kenya. Holocene, 15: 837-846.

Taylor, D., Robertshaw, P. and Marchant, R.A., 2000: Environ-mental change and political-economic upheaval in precolo-nial western Uganda. Holocene, 10: 527-536.

Thevenon, F., Williamson, D. and Taieb, M., 2002: A 22 kyr BP sedimentological record of Lake Rukwa (8 degrees S, SW Tanzania): environmental, chronostratigraphic and climatic implications. Palaeogeography Palaeoclimatology Palae-oecology, 187: 285-294.

Turner R, Kelly, A. and Roberts, N. in press: A critical assess-ment and experimental comparison of microscopic charcoal extraction method. In: G. Fiorentino and D. Magri (Eds) Charcoal from the past: cultural and palaeoenvironmental implications. Proceedings of the Third International Meet-ing of Anthracology, Cavallino (Lecce), June 2004. BAR International Series, Archaeopress, Oxford, UK.

Walker, I.R., 1995: Chironomids as indicators of past climatic change: In P. Armitage,, P. Cranston and L. Pinder (Eds), The Chironomidae. The biology and ecology of non-biting midges, Chapman and Hall, London, pp. 405-422.

Waller, R.D., 1988: Emutai: crises and response in Masaailand 1883-1902. In D. Johnson and D. Anderson (Eds), The Ecol-ogy of survival: case studies from Northeast African History. Lester Crook Academic Publishing, pp:73-114.

Verosub, K.L. and Roberts, A.P., 1995: Environmental Mag-netism - Past, Present, and Future. Journal of Geophysical Research-Solid Earth, 100: 2175-2192.

Verschuren, D., 1991: Taxonomy and ecology of subfossil Chironomidae (Insecta, Diptera) from Rift Valley lakes in central Kenya. Archiv für Hydrobiologie Supplement, 107: 467-512.

Verschuren, D. 2004. The last 2000 years in tropical Africa. In Battarbee R., Gasse, F. and Stickely, C. (Eds) Past climate variability through Europe and Africa, Amsterdam, Kluwer, pp: 139-158.

Verschuren, D. and Eggermont, H., 2006: Quaternary pale-oecology of aquatic Diptera in tropical and Southern Hemi-sphere regions, with special reference to the Chironomidae. Quaternary Science Reviews, 25: 1926-1947.

Verschuren, D., Laird, K.R. and Cumming, B.F., 2000: Rainfall and drought in equatorial east Africa during the past 1,100 years. Nature, 403: 410-414.

White, F., 1983: The vegetation of Africa. UNESCO, Paris. 356 pp

Vincens, A., Williamson, D., Thevenon, F., Taieb, M., Buchet, G., Decobert, M. and Thouveny, N. 2003. Pollen-based veg-etation changes in southern Tanzania during the last 4200 years: climate change and/or human impact. Palaeogeogra-phy Palaeoclimatology Palaeoecology, 198: 321-334.

Vincens, A., Buchet, G., Williamson, D. and Taieb, M., 2005: A 23,000 yr pollen record from Lake Rukwa (8 degrees S, SW Tanzania): New data on vegetation dynamics and climate in Central Eastern Africa. Review of Palaeobotany and Paly-nology, 137: 147-162.